diff -u --recursive --new-file v2.5.1/linux/Documentation/DocBook/kernel-hacking.tmpl linux/Documentation/DocBook/kernel-hacking.tmpl --- v2.5.1/linux/Documentation/DocBook/kernel-hacking.tmpl Fri Oct 5 12:06:51 2001 +++ linux/Documentation/DocBook/kernel-hacking.tmpl Thu Dec 20 08:54:23 2001 @@ -18,8 +18,8 @@ - 2000 - Paul Russell + 2001 + Rusty Russell @@ -651,6 +651,29 @@ + + <function>cpu_to_be32()</function>/<function>be32_to_cpu()</function>/<function>cpu_to_le32()</function>/<function>le32_to_cpu()</function> + <filename class=headerfile>include/asm/byteorder.h</filename> + + + + The cpu_to_be32() family (where the "32" can + be replaced by 64 or 16, and the "be" can be replaced by "le") are + the general way to do endian conversions in the kernel: they + return the converted value. All variations supply the reverse as + well: be32_to_cpu(), etc. + + + + There are two major variations of these functions: the pointer + variation, such as cpu_to_be32p(), which take + a pointer to the given type, and return the converted value. The + other variation is the "in-situ" family, such as + cpu_to_be32s(), which convert value referred + to by the pointer, and return void. + + + <function>local_irq_save()</function>/<function>local_irq_restore()</function> <filename class=headerfile>include/asm/system.h</filename> diff -u --recursive --new-file v2.5.1/linux/MAINTAINERS linux/MAINTAINERS --- v2.5.1/linux/MAINTAINERS Tue Dec 18 14:56:34 2001 +++ linux/MAINTAINERS Mon Dec 17 10:49:31 2001 @@ -1322,8 +1322,10 @@ S: Maintained SCSI SUBSYSTEM +P: Jens Axboe +M: axboe@suse.de L: linux-scsi@vger.kernel.org -S: Unmaintained +S: Maintained SCSI TAPE DRIVER P: Kai Mäkisara diff -u --recursive --new-file v2.5.1/linux/Makefile linux/Makefile --- v2.5.1/linux/Makefile Tue Dec 18 14:56:34 2001 +++ linux/Makefile Tue Dec 25 15:45:00 2001 @@ -1,7 +1,7 @@ VERSION = 2 PATCHLEVEL = 5 -SUBLEVEL = 1 -EXTRAVERSION = +SUBLEVEL = 2 +EXTRAVERSION =-pre2 KERNELRELEASE=$(VERSION).$(PATCHLEVEL).$(SUBLEVEL)$(EXTRAVERSION) diff -u --recursive --new-file v2.5.1/linux/arch/alpha/kernel/setup.c linux/arch/alpha/kernel/setup.c --- v2.5.1/linux/arch/alpha/kernel/setup.c Fri Nov 16 18:38:39 2001 +++ linux/arch/alpha/kernel/setup.c Tue Dec 25 15:39:20 2001 @@ -440,12 +440,6 @@ return MKDEV(TTY_MAJOR, 64 + c->index); } -static int srm_console_wait_key(struct console *co) -{ - /* Huh? */ - return 1; -} - static int __init srm_console_setup(struct console *co, char *options) { return 1; @@ -455,7 +449,6 @@ name: "srm0", write: srm_console_write, device: srm_console_device, - wait_key: srm_console_wait_key, setup: srm_console_setup, flags: CON_PRINTBUFFER | CON_ENABLED, /* fake it out */ index: -1, diff -u --recursive --new-file v2.5.1/linux/arch/i386/kernel/signal.c linux/arch/i386/kernel/signal.c --- v2.5.1/linux/arch/i386/kernel/signal.c Fri Sep 14 14:15:40 2001 +++ linux/arch/i386/kernel/signal.c Wed Dec 19 15:27:12 2001 @@ -28,7 +28,7 @@ #define _BLOCKABLE (~(sigmask(SIGKILL) | sigmask(SIGSTOP))) -asmlinkage int FASTCALL(do_signal(struct pt_regs *regs, sigset_t *oldset)); +int FASTCALL(do_signal(struct pt_regs *regs, sigset_t *oldset)); int copy_siginfo_to_user(siginfo_t *to, siginfo_t *from) { diff -u --recursive --new-file v2.5.1/linux/arch/i386/kernel/vm86.c linux/arch/i386/kernel/vm86.c --- v2.5.1/linux/arch/i386/kernel/vm86.c Fri Jul 6 17:05:07 2001 +++ linux/arch/i386/kernel/vm86.c Wed Dec 19 15:27:12 2001 @@ -62,7 +62,7 @@ ( (unsigned)( & (((struct kernel_vm86_regs *)0)->VM86_REGS_PART2) ) ) #define VM86_REGS_SIZE2 (sizeof(struct kernel_vm86_regs) - VM86_REGS_SIZE1) -asmlinkage struct pt_regs * FASTCALL(save_v86_state(struct kernel_vm86_regs * regs)); +struct pt_regs * FASTCALL(save_v86_state(struct kernel_vm86_regs * regs)); struct pt_regs * save_v86_state(struct kernel_vm86_regs * regs) { struct tss_struct *tss; diff -u --recursive --new-file v2.5.1/linux/arch/ia64/hp/hpsim_console.c linux/arch/ia64/hp/hpsim_console.c --- v2.5.1/linux/arch/ia64/hp/hpsim_console.c Thu Oct 12 14:20:48 2000 +++ linux/arch/ia64/hp/hpsim_console.c Tue Dec 25 15:39:20 2001 @@ -24,14 +24,12 @@ static int simcons_init (struct console *, char *); static void simcons_write (struct console *, const char *, unsigned); -static int simcons_wait_key (struct console *); static kdev_t simcons_console_device (struct console *); struct console hpsim_cons = { name: "simcons", write: simcons_write, device: simcons_console_device, - wait_key: simcons_wait_key, setup: simcons_init, flags: CON_PRINTBUFFER, index: -1, @@ -54,17 +52,6 @@ if (ch == '\n') ia64_ssc('\r', 0, 0, 0, SSC_PUTCHAR); } -} - -static int -simcons_wait_key (struct console *cons) -{ - char ch; - - do { - ch = ia64_ssc(0, 0, 0, 0, SSC_GETCHAR); - } while (ch == '\0'); - return ch; } static kdev_t diff -u --recursive --new-file v2.5.1/linux/arch/m68k/amiga/config.c linux/arch/m68k/amiga/config.c --- v2.5.1/linux/arch/m68k/amiga/config.c Tue Dec 18 14:56:34 2001 +++ linux/arch/m68k/amiga/config.c Tue Dec 25 15:39:20 2001 @@ -98,7 +98,6 @@ extern void amiga_floppy_setup(char *, int *); #endif static void amiga_reset (void); -static int amiga_wait_key (struct console *co); extern void amiga_init_sound(void); static void amiga_savekmsg_init(void); static void amiga_mem_console_write(struct console *co, const char *b, @@ -112,7 +111,6 @@ static struct console amiga_console_driver = { name: "debug", - wait_key: amiga_wait_key, flags: CON_PRINTBUFFER, index: -1, }; @@ -740,33 +738,6 @@ } return 0; -} - -static int amiga_wait_key (struct console *co) -{ - int i; - - while (1) { - while (ciaa.pra & 0x40); - - /* debounce */ - for (i = 0; i < 1000; i++); - - if (!(ciaa.pra & 0x40)) - break; - } - - /* wait for button up */ - while (1) { - while (!(ciaa.pra & 0x40)); - - /* debounce */ - for (i = 0; i < 1000; i++); - - if (ciaa.pra & 0x40) - break; - } - return 0; } static NORET_TYPE void amiga_reset( void ) diff -u --recursive --new-file v2.5.1/linux/arch/m68k/mac/debug.c linux/arch/m68k/mac/debug.c --- v2.5.1/linux/arch/m68k/mac/debug.c Mon Jun 11 19:15:27 2001 +++ linux/arch/m68k/mac/debug.c Tue Dec 25 15:39:20 2001 @@ -248,17 +248,6 @@ return( scc.cha_b_data ); } -int mac_scca_console_wait_key(struct console *co) -{ - int i; - do { - for( i = uSEC; i > 0; --i ) - barrier(); - } while( !(scc.cha_a_ctrl & 0x01) ); /* wait for rx buf filled */ - for( i = uSEC; i > 0; --i ) - barrier(); - return( scc.cha_a_data ); -} #endif /* The following two functions do a quick'n'dirty initialization of the MFP or @@ -395,9 +384,6 @@ /* Mac modem port */ mac_init_scc_port( B9600|CS8, 0 ); mac_console_driver.write = mac_scca_console_write; -#ifdef CONFIG_SERIAL_CONSOLE - mac_console_driver.wait_key = mac_scca_console_wait_key; -#endif scc_port = 0; } else if (!strcmp( m68k_debug_device, "ser2" )) { diff -u --recursive --new-file v2.5.1/linux/arch/m68k/mvme147/config.c linux/arch/m68k/mvme147/config.c --- v2.5.1/linux/arch/m68k/mvme147/config.c Tue Dec 18 14:56:34 2001 +++ linux/arch/m68k/mvme147/config.c Tue Dec 25 15:39:20 2001 @@ -241,31 +241,7 @@ restore_flags(flags); } - -static int m147_scc_wait_key (struct console *co) -{ - volatile unsigned char *p = (volatile char *)M147_SCC_A_ADDR; - unsigned long flags; - int c; - - /* wait for rx buf filled */ - while ((*p & 0x01) == 0) - ; - - save_flags(flags); - cli(); - - *p = 8; - scc_delay(); - c = *p; - - restore_flags(flags); - return c; -} - - void mvme147_init_console_port (struct console *co, int cflag) { co->write = m147_scc_write; - co->wait_key = m147_scc_wait_key; } diff -u --recursive --new-file v2.5.1/linux/arch/m68k/q40/config.c linux/arch/m68k/q40/config.c --- v2.5.1/linux/arch/m68k/q40/config.c Tue Dec 18 14:56:34 2001 +++ linux/arch/m68k/q40/config.c Tue Dec 25 15:39:20 2001 @@ -72,10 +72,8 @@ extern int ql_ticks; -static int q40_wait_key(struct console *co){return 0;} static struct console q40_console_driver = { name: "debug", - wait_key: q40_wait_key, flags: CON_PRINTBUFFER, index: -1, }; diff -u --recursive --new-file v2.5.1/linux/arch/m68k/sun3x/prom.c linux/arch/m68k/sun3x/prom.c --- v2.5.1/linux/arch/m68k/sun3x/prom.c Mon Jun 11 19:15:27 2001 +++ linux/arch/m68k/sun3x/prom.c Tue Dec 25 15:39:20 2001 @@ -96,7 +96,6 @@ sun3x_prom_write, /* write */ NULL, /* read */ NULL, /* device */ - NULL, /* wait_key */ NULL, /* unblank */ NULL, /* setup */ CON_PRINTBUFFER, diff -u --recursive --new-file v2.5.1/linux/arch/mips/au1000/common/serial.c linux/arch/mips/au1000/common/serial.c --- v2.5.1/linux/arch/mips/au1000/common/serial.c Fri Oct 5 12:06:51 2001 +++ linux/arch/mips/au1000/common/serial.c Tue Dec 25 15:39:20 2001 @@ -2922,35 +2922,6 @@ serial_out(info, UART_IER, ier); } -/* - * Receive character from the serial port - */ -static int serial_console_wait_key(struct console *co) -{ - static struct async_struct *info; - int ier, c; - - info = &async_sercons; - - /* - * First save the IER then disable the interrupts so - * that the real driver for the port does not get the - * character. - */ - ier = serial_in(info, UART_IER); - serial_out(info, UART_IER, 0x00); - - while ((serial_in(info, UART_LSR) & UART_LSR_DR) == 0); - c = serial_in(info, UART_RX); - - /* - * Restore the interrupts - */ - serial_out(info, UART_IER, ier); - - return c; -} - static kdev_t serial_console_device(struct console *c) { return MKDEV(TTY_MAJOR, 64 + c->index); @@ -3075,7 +3046,6 @@ name: "ttyS", write: serial_console_write, device: serial_console_device, - wait_key: serial_console_wait_key, setup: serial_console_setup, flags: CON_PRINTBUFFER, index: -1, diff -u --recursive --new-file v2.5.1/linux/arch/mips/baget/vacserial.c linux/arch/mips/baget/vacserial.c --- v2.5.1/linux/arch/mips/baget/vacserial.c Sun Sep 9 10:43:01 2001 +++ linux/arch/mips/baget/vacserial.c Tue Dec 25 15:39:20 2001 @@ -2622,43 +2622,6 @@ serial_outp(&scr_info, VAC_UART_INT_MASK, ier); } -/* - * Receive character from the serial port - */ -static int serial_console_wait_key(struct console *co) -{ - struct serial_state *ser; - int ier; - int lsr; - int c; - struct async_struct scr_info; /* serial_{in,out} because HUB6 */ - - ser = rs_table + co->index; - scr_info.magic = SERIAL_MAGIC; - scr_info.port = ser->port; - scr_info.flags = ser->flags; - - /* - * First save the IER then disable the interrupts so - * that the real driver for the port does not get the - * character. - */ - ier = serial_inp(&scr_info, VAC_UART_INT_MASK); - serial_outp(&scr_info, VAC_UART_INT_MASK, 0x00); - - do { - lsr = serial_inp(&scr_info, VAC_UART_INT_STATUS); - } while (!(lsr & VAC_UART_STATUS_RX_READY)); - c = serial_inp(&scr_info, VAC_UART_RX); - - /* - * Restore the interrupts - */ - serial_outp(&scr_info, VAC_UART_INT_MASK, ier); - - return c; -} - static kdev_t serial_console_device(struct console *c) { return MKDEV(TTY_MAJOR, 64 + c->index); @@ -2812,7 +2775,6 @@ name: "ttyS", write: serial_console_write, device: serial_console_device, - wait_key: serial_console_wait_key, setup: serial_console_setup, flags: CON_PRINTBUFFER, index: -1, diff -u --recursive --new-file v2.5.1/linux/arch/mips/dec/promcon.c linux/arch/mips/dec/promcon.c --- v2.5.1/linux/arch/mips/dec/promcon.c Thu Oct 12 14:20:48 2000 +++ linux/arch/mips/dec/promcon.c Tue Dec 25 15:39:20 2001 @@ -30,11 +30,6 @@ } } -static int prom_console_wait_key(struct console *co) -{ - return prom_getchar(); -} - static int __init prom_console_setup(struct console *co, char *options) { return 0; @@ -50,7 +45,6 @@ name: "ttyS", write: prom_console_write, device: prom_console_device, - wait_key: prom_console_wait_key, setup: prom_console_setup, flags: CON_PRINTBUFFER, index: -1, diff -u --recursive --new-file v2.5.1/linux/arch/parisc/kernel/pdc_cons.c linux/arch/parisc/kernel/pdc_cons.c --- v2.5.1/linux/arch/parisc/kernel/pdc_cons.c Sun Sep 16 21:23:15 2001 +++ linux/arch/parisc/kernel/pdc_cons.c Tue Dec 25 15:39:20 2001 @@ -105,10 +105,6 @@ static struct console pdc_cons = { name: "ttyB", write: pdc_console_write, - read: NULL, - device: NULL, - wait_key: pdc_console_wait_key, - unblank: NULL, setup: pdc_console_setup, flags: CON_PRINTBUFFER|CON_ENABLED, // |CON_CONSDEV, index: -1, diff -u --recursive --new-file v2.5.1/linux/arch/ppc/8260_io/uart.c linux/arch/ppc/8260_io/uart.c --- v2.5.1/linux/arch/ppc/8260_io/uart.c Mon May 21 17:04:46 2001 +++ linux/arch/ppc/8260_io/uart.c Tue Dec 25 15:39:20 2001 @@ -2216,54 +2216,6 @@ info->tx_cur = (cbd_t *)bdp; } -/* - * Receive character from the serial port. This only works well - * before the port is initialize for real use. - */ -static int serial_console_wait_key(struct console *co) -{ - struct serial_state *ser; - u_char c, *cp; - ser_info_t *info; - volatile cbd_t *bdp; - volatile smc_uart_t *up; - - ser = rs_table + co->index; - - /* Pointer to UART in parameter ram. - */ - up = (smc_uart_t *)&immr->im_dprambase[ser->port]; - - /* Get the address of the host memory buffer. - * If the port has been initialized for general use, we must - * use information from the port structure. - */ - if ((info = (ser_info_t *)ser->info)) - bdp = info->rx_cur; - else - bdp = (cbd_t *)&immr->im_dprambase[up->smc_rbase]; - - /* - * We need to gracefully shut down the receiver, disable - * interrupts, then read the input. - */ - while (bdp->cbd_sc & BD_SC_EMPTY); /* Wait for a character */ - cp = __va(bdp->cbd_bufaddr); - - if (info) { - if (bdp->cbd_sc & BD_SC_WRAP) { - bdp = info->rx_bd_base; - } - else { - bdp++; - } - info->rx_cur = (cbd_t *)bdp; - } - - c = *cp; - return((int)c); -} - static kdev_t serial_console_device(struct console *c) { return MKDEV(TTYAUX_MAJOR, 64 + c->index); @@ -2274,7 +2226,6 @@ name: "ttyS", write: serial_console_write, device: serial_console_device, - wait_key: serial_console_wait_key, setup: serial_console_setup, flags: CON_PRINTBUFFER, index: CONFIG_SERIAL_CONSOLE_PORT, diff -u --recursive --new-file v2.5.1/linux/arch/ppc/8xx_io/uart.c linux/arch/ppc/8xx_io/uart.c --- v2.5.1/linux/arch/ppc/8xx_io/uart.c Fri Nov 2 17:43:54 2001 +++ linux/arch/ppc/8xx_io/uart.c Tue Dec 25 15:39:20 2001 @@ -97,7 +97,6 @@ static void serial_console_write(struct console *c, const char *s, unsigned count); static kdev_t serial_console_device(struct console *c); -static int serial_console_wait_key(struct console *co); #if defined(CONFIG_SERIAL_CONSOLE) && defined(CONFIG_MAGIC_SYSRQ) static unsigned long break_pressed; /* break, really ... */ @@ -218,7 +217,6 @@ name: "ttyS", write: serial_console_write, device: serial_console_device, - wait_key: serial_console_wait_key, setup: serial_console_setup, flags: CON_PRINTBUFFER, index: CONFIG_SERIAL_CONSOLE_PORT, @@ -2400,11 +2398,6 @@ } return((int)c); -} - -static int serial_console_wait_key(struct console *co) -{ - return(my_console_wait_key(co->index, 0, NULL)); } #ifdef CONFIG_XMON diff -u --recursive --new-file v2.5.1/linux/arch/ppc/amiga/config.c linux/arch/ppc/amiga/config.c --- v2.5.1/linux/arch/ppc/amiga/config.c Tue Dec 18 14:56:34 2001 +++ linux/arch/ppc/amiga/config.c Tue Dec 25 15:39:20 2001 @@ -104,7 +104,6 @@ extern void amiga_floppy_setup(char *, int *); #endif static void amiga_reset (void); -static int amiga_wait_key (struct console *co); extern void amiga_init_sound(void); static void amiga_savekmsg_init(void); static void amiga_mem_console_write(struct console *co, const char *b, @@ -118,7 +117,6 @@ static struct console amiga_console_driver = { name: "debug", - wait_key: amiga_wait_key, flags: CON_PRINTBUFFER, index: -1, }; @@ -734,33 +732,6 @@ } return 0; -} - -static int amiga_wait_key (struct console *co) -{ - int i; - - while (1) { - while (ciaa.pra & 0x40); - - /* debounce */ - for (i = 0; i < 1000; i++); - - if (!(ciaa.pra & 0x40)) - break; - } - - /* wait for button up */ - while (1) { - while (!(ciaa.pra & 0x40)); - - /* debounce */ - for (i = 0; i < 1000; i++); - - if (ciaa.pra & 0x40) - break; - } - return 0; } static NORET_TYPE void amiga_reset( void ) diff -u --recursive --new-file v2.5.1/linux/arch/ppc/kernel/m8xx_setup.c linux/arch/ppc/kernel/m8xx_setup.c --- v2.5.1/linux/arch/ppc/kernel/m8xx_setup.c Fri Nov 16 10:10:08 2001 +++ linux/arch/ppc/kernel/m8xx_setup.c Mon Dec 17 11:12:05 2001 @@ -55,12 +55,6 @@ extern void m8xx_ide_init(void); -#ifdef CONFIG_BLK_DEV_RAM -extern int rd_doload; /* 1 = load ramdisk, 0 = don't load */ -extern int rd_prompt; /* 1 = prompt for ramdisk, 0 = don't prompt */ -extern int rd_image_start; /* starting block # of image */ -#endif - extern unsigned long find_available_memory(void); extern void m8xx_cpm_reset(uint); diff -u --recursive --new-file v2.5.1/linux/arch/ppc/kernel/prep_setup.c linux/arch/ppc/kernel/prep_setup.c --- v2.5.1/linux/arch/ppc/kernel/prep_setup.c Fri Nov 16 10:10:08 2001 +++ linux/arch/ppc/kernel/prep_setup.c Mon Dec 17 11:12:05 2001 @@ -111,12 +111,6 @@ extern int probingmem; extern unsigned long loops_per_jiffy; -#ifdef CONFIG_BLK_DEV_RAM -extern int rd_doload; /* 1 = load ramdisk, 0 = don't load */ -extern int rd_prompt; /* 1 = prompt for ramdisk, 0 = don't prompt */ -extern int rd_image_start; /* starting block # of image */ -#endif - #ifdef CONFIG_SOUND_MODULE EXPORT_SYMBOL(ppc_cs4232_dma); EXPORT_SYMBOL(ppc_cs4232_dma2); diff -u --recursive --new-file v2.5.1/linux/arch/sh/kernel/setup.c linux/arch/sh/kernel/setup.c --- v2.5.1/linux/arch/sh/kernel/setup.c Fri Nov 16 18:38:39 2001 +++ linux/arch/sh/kernel/setup.c Tue Dec 25 15:39:20 2001 @@ -140,15 +140,6 @@ sh_bios_console_write(s, count); } -/* - * Receive character from the serial port - */ -static int sh_console_wait_key(struct console *co) -{ - /* Not implemented yet */ - return 0; -} - static kdev_t sh_console_device(struct console *c) { /* TODO: this is totally bogus */ @@ -183,7 +174,6 @@ name: "bios", write: sh_console_write, device: sh_console_device, - wait_key: sh_console_wait_key, setup: sh_console_setup, flags: CON_PRINTBUFFER, index: -1, diff -u --recursive --new-file v2.5.1/linux/drivers/block/block_ioctl.c linux/drivers/block/block_ioctl.c --- v2.5.1/linux/drivers/block/block_ioctl.c Tue Dec 18 14:56:35 2001 +++ linux/drivers/block/block_ioctl.c Thu Dec 20 08:50:13 2001 @@ -35,7 +35,7 @@ DECLARE_COMPLETION(wait); int err = 0; - rq->flags |= REQ_BARRIER; + rq->flags |= REQ_NOMERGE; rq->waiting = &wait; elv_add_request(q, rq, 1); generic_unplug_device(q); @@ -81,3 +81,5 @@ #endif return err; } + +EXPORT_SYMBOL(block_ioctl); diff -u --recursive --new-file v2.5.1/linux/drivers/block/elevator.c linux/drivers/block/elevator.c --- v2.5.1/linux/drivers/block/elevator.c Tue Dec 18 14:56:35 2001 +++ linux/drivers/block/elevator.c Thu Dec 20 08:50:13 2001 @@ -53,7 +53,7 @@ * if .next is a valid request */ next = rq->queuelist.next; - if (next == head) + if (unlikely(next == head)) return 0; next_rq = list_entry(next, struct request, queuelist); @@ -121,20 +121,52 @@ return 0; } -int elevator_linus_merge(request_queue_t *q, struct request **req, - struct list_head *head, struct bio *bio) +inline int elv_try_merge(struct request *__rq, struct bio *bio) { unsigned int count = bio_sectors(bio); - struct list_head *entry = &q->queue_head; - int ret = ELEVATOR_NO_MERGE; + + if (!elv_rq_merge_ok(__rq, bio)) + return ELEVATOR_NO_MERGE; + + /* + * we can merge and sequence is ok, check if it's possible + */ + if (__rq->sector + __rq->nr_sectors == bio->bi_sector) { + return ELEVATOR_BACK_MERGE; + } else if (__rq->sector - count == bio->bi_sector) { + __rq->elevator_sequence -= count; + return ELEVATOR_FRONT_MERGE; + } + + return ELEVATOR_NO_MERGE; +} + +int elevator_linus_merge(request_queue_t *q, struct request **req, + struct bio *bio) +{ + struct list_head *entry; struct request *__rq; + int ret; + + /* + * give a one-shot try to merging with the last touched + * request + */ + if (q->last_merge) { + __rq = list_entry_rq(q->last_merge); + BUG_ON(__rq->flags & REQ_STARTED); + + if ((ret = elv_try_merge(__rq, bio))) { + *req = __rq; + return ret; + } + } entry = &q->queue_head; - while ((entry = entry->prev) != head) { + ret = ELEVATOR_NO_MERGE; + while ((entry = entry->prev) != &q->queue_head) { __rq = list_entry_rq(entry); - prefetch(list_entry_rq(entry->prev)); - /* * simply "aging" of requests in queue */ @@ -144,26 +176,15 @@ break; if (!(__rq->flags & REQ_CMD)) continue; + if (__rq->elevator_sequence < 0) + break; if (!*req && bio_rq_in_between(bio, __rq, &q->queue_head)) *req = __rq; - if (!elv_rq_merge_ok(__rq, bio)) - continue; - - if (__rq->elevator_sequence < count) - break; - /* - * we can merge and sequence is ok, check if it's possible - */ - if (__rq->sector + __rq->nr_sectors == bio->bi_sector) { - ret = ELEVATOR_BACK_MERGE; - *req = __rq; - break; - } else if (__rq->sector - count == bio->bi_sector) { - ret = ELEVATOR_FRONT_MERGE; - __rq->elevator_sequence -= count; + if ((ret = elv_try_merge(__rq, bio))) { *req = __rq; + q->last_merge = &__rq->queuelist; break; } } @@ -183,7 +204,6 @@ entry = &req->queuelist; while ((entry = entry->next) != &q->queue_head) { struct request *tmp; - prefetch(list_entry_rq(entry->next)); tmp = list_entry_rq(entry); tmp->elevator_sequence -= count; } @@ -199,12 +219,20 @@ struct list_head *insert_here) { list_add(&rq->queuelist, insert_here); + + /* + * new merges must not precede this barrier + */ + if (rq->flags & REQ_BARRIER) + q->last_merge = NULL; + else if (!q->last_merge) + q->last_merge = &rq->queuelist; } struct request *elv_next_request_fn(request_queue_t *q) { if (!blk_queue_empty(q)) - return list_entry(q->queue_head.next, struct request, queuelist); + return list_entry_rq(q->queue_head.next); return NULL; } @@ -222,17 +250,24 @@ * See if we can find a request that this buffer can be coalesced with. */ int elevator_noop_merge(request_queue_t *q, struct request **req, - struct list_head *head, struct bio *bio) + struct bio *bio) { - unsigned int count = bio_sectors(bio); struct list_head *entry = &q->queue_head; struct request *__rq; + int ret; - entry = &q->queue_head; - while ((entry = entry->prev) != head) { - __rq = list_entry_rq(entry); + if (q->last_merge) { + __rq = list_entry_rq(q->last_merge); + BUG_ON(__rq->flags & REQ_STARTED); + + if ((ret = elv_try_merge(__rq, bio))) { + *req = __rq; + return ret; + } + } - prefetch(list_entry_rq(entry->prev)); + while ((entry = entry->prev) != &q->queue_head) { + __rq = list_entry_rq(entry); if (__rq->flags & (REQ_BARRIER | REQ_STARTED)) break; @@ -240,18 +275,10 @@ if (!(__rq->flags & REQ_CMD)) continue; - if (!elv_rq_merge_ok(__rq, bio)) - continue; - - /* - * we can merge and sequence is ok, check if it's possible - */ - if (__rq->sector + __rq->nr_sectors == bio->bi_sector) { - *req = __rq; - return ELEVATOR_BACK_MERGE; - } else if (__rq->sector - count == bio->bi_sector) { + if ((ret = elv_try_merge(__rq, bio))) { *req = __rq; - return ELEVATOR_FRONT_MERGE; + q->last_merge = &__rq->queuelist; + return ret; } } @@ -267,6 +294,7 @@ *e = type; INIT_LIST_HEAD(&q->queue_head); + q->last_merge = NULL; if (e->elevator_init_fn) return e->elevator_init_fn(q, e); diff -u --recursive --new-file v2.5.1/linux/drivers/block/ll_rw_blk.c linux/drivers/block/ll_rw_blk.c --- v2.5.1/linux/drivers/block/ll_rw_blk.c Tue Dec 18 14:56:35 2001 +++ linux/drivers/block/ll_rw_blk.c Thu Dec 20 08:50:13 2001 @@ -117,6 +117,11 @@ return &blk_dev[MAJOR(dev)].request_queue; } +void blk_queue_prep_rq(request_queue_t *q, prep_rq_fn *pfn) +{ + q->prep_rq_fn = pfn; +} + /** * blk_queue_make_request - define an alternate make_request function for a device * @q: the request queue for the device to be affected @@ -179,7 +184,6 @@ if (dma_addr == BLK_BOUNCE_ISA) { init_emergency_isa_pool(); q->bounce_gfp = GFP_NOIO | GFP_DMA; - printk("isa pfn %lu, max low %lu, max %lu\n", bounce_pfn, blk_max_low_pfn, blk_max_pfn); } else q->bounce_gfp = GFP_NOHIGHIO; @@ -319,7 +323,7 @@ /* * standard prep_rq_fn that builds 10 byte cmds */ -static int ll_10byte_cmd_build(request_queue_t *q, struct request *rq) +int ll_10byte_cmd_build(request_queue_t *q, struct request *rq) { int hard_sect = get_hardsect_size(rq->rq_dev); sector_t block = rq->hard_sector / (hard_sect >> 9); @@ -477,7 +481,7 @@ sg[nsegs - 1].length += nbytes; } else { new_segment: - sg[nsegs].address = NULL; + memset(&sg[nsegs],0,sizeof(struct scatterlist)); sg[nsegs].page = bvec->bv_page; sg[nsegs].length = nbytes; sg[nsegs].offset = bvec->bv_offset; @@ -540,11 +544,11 @@ { if (req->nr_sectors + bio_sectors(bio) > q->max_sectors) { req->flags |= REQ_NOMERGE; + q->last_merge = NULL; return 0; } - if (BIOVEC_VIRT_MERGEABLE(__BVEC_END(req->biotail), - __BVEC_START(bio))) + if (BIOVEC_VIRT_MERGEABLE(__BVEC_END(req->biotail), __BVEC_START(bio))) return ll_new_mergeable(q, req, bio); return ll_new_hw_segment(q, req, bio); @@ -555,11 +559,11 @@ { if (req->nr_sectors + bio_sectors(bio) > q->max_sectors) { req->flags |= REQ_NOMERGE; + q->last_merge = NULL; return 0; } - if (BIOVEC_VIRT_MERGEABLE(__BVEC_END(bio), - __BVEC_START(req->bio))) + if (BIOVEC_VIRT_MERGEABLE(__BVEC_END(bio), __BVEC_START(req->bio))) return ll_new_mergeable(q, req, bio); return ll_new_hw_segment(q, req, bio); @@ -568,7 +572,7 @@ static int ll_merge_requests_fn(request_queue_t *q, struct request *req, struct request *next) { - int total_phys_segments = req->nr_phys_segments + next->nr_phys_segments; + int total_phys_segments = req->nr_phys_segments +next->nr_phys_segments; int total_hw_segments = req->nr_hw_segments + next->nr_hw_segments; /* @@ -599,6 +603,9 @@ return 0; /* Merge is OK... */ + if (q->last_merge == &next->queuelist) + q->last_merge = NULL; + req->nr_phys_segments = total_phys_segments; req->nr_hw_segments = total_hw_segments; return 1; @@ -799,7 +806,7 @@ q->back_merge_fn = ll_back_merge_fn; q->front_merge_fn = ll_front_merge_fn; q->merge_requests_fn = ll_merge_requests_fn; - q->prep_rq_fn = ll_10byte_cmd_build; + q->prep_rq_fn = NULL; q->plug_tq.sync = 0; q->plug_tq.routine = &generic_unplug_device; q->plug_tq.data = q; @@ -1039,13 +1046,11 @@ attempt_merge(q, rq); } -static inline void attempt_front_merge(request_queue_t *q, - struct list_head *head, - struct request *rq) +static inline void attempt_front_merge(request_queue_t *q, struct request *rq) { struct list_head *prev = rq->queuelist.prev; - if (prev != head) + if (prev != &q->queue_head) attempt_merge(q, blkdev_entry_to_request(prev)); } @@ -1081,7 +1086,7 @@ { struct request *req, *freereq = NULL; int el_ret, latency = 0, rw, nr_sectors, cur_nr_sectors, barrier; - struct list_head *head, *insert_here; + struct list_head *insert_here = &q->queue_head; elevator_t *elevator = &q->elevator; sector_t sector; @@ -1102,24 +1107,18 @@ latency = elevator_request_latency(elevator, rw); barrier = test_bit(BIO_RW_BARRIER, &bio->bi_rw); -again: - req = NULL; - head = &q->queue_head; - spin_lock_irq(q->queue_lock); - insert_here = head->prev; if (blk_queue_empty(q) || barrier) { blk_plug_device(q); goto get_rq; - } else if ((req = __elv_next_request(q))) { - if (req->flags & REQ_STARTED) - head = head->next; - - req = NULL; } - el_ret = elevator->elevator_merge_fn(q, &req, head, bio); +again: + req = NULL; + insert_here = q->queue_head.prev; + + el_ret = elevator->elevator_merge_fn(q, &req, bio); switch (el_ret) { case ELEVATOR_BACK_MERGE: BUG_ON(req->flags & REQ_STARTED); @@ -1157,7 +1156,7 @@ req->sector = req->hard_sector = sector; req->nr_sectors = req->hard_nr_sectors += nr_sectors; drive_stat_acct(req, nr_sectors, 0); - attempt_front_merge(q, head, req); + attempt_front_merge(q, req); goto out; /* @@ -1188,7 +1187,6 @@ req = freereq; freereq = NULL; } else if ((req = get_request(q, rw)) == NULL) { - spin_unlock_irq(q->queue_lock); /* @@ -1200,6 +1198,7 @@ } freereq = get_request_wait(q, rw); + spin_lock_irq(q->queue_lock); goto again; } @@ -1755,3 +1754,6 @@ EXPORT_SYMBOL(blk_queue_assign_lock); EXPORT_SYMBOL(blk_phys_contig_segment); EXPORT_SYMBOL(blk_hw_contig_segment); + +EXPORT_SYMBOL(ll_10byte_cmd_build); +EXPORT_SYMBOL(blk_queue_prep_rq); diff -u --recursive --new-file v2.5.1/linux/drivers/block/loop.c linux/drivers/block/loop.c --- v2.5.1/linux/drivers/block/loop.c Tue Dec 18 14:56:35 2001 +++ linux/drivers/block/loop.c Mon Dec 17 14:10:42 2001 @@ -1007,6 +1007,7 @@ goto out_mem; blk_queue_make_request(BLK_DEFAULT_QUEUE(MAJOR_NR), loop_make_request); + blk_queue_bounce_limit(BLK_DEFAULT_QUEUE(MAJOR_NR), BLK_BOUNCE_HIGH); for (i = 0; i < max_loop; i++) { struct loop_device *lo = &loop_dev[i]; diff -u --recursive --new-file v2.5.1/linux/drivers/block/rd.c linux/drivers/block/rd.c --- v2.5.1/linux/drivers/block/rd.c Tue Dec 18 14:56:35 2001 +++ linux/drivers/block/rd.c Mon Dec 17 11:12:05 2001 @@ -43,26 +43,12 @@ */ #include <linux/config.h> -#include <linux/sched.h> -#include <linux/fs.h> -#include <linux/kernel.h> -#include <linux/hdreg.h> #include <linux/string.h> -#include <linux/mm.h> -#include <linux/mman.h> #include <linux/slab.h> -#include <linux/ioctl.h> -#include <linux/fd.h> #include <linux/module.h> #include <linux/init.h> #include <linux/devfs_fs_kernel.h> -#include <linux/smp_lock.h> - -#include <asm/system.h> #include <asm/uaccess.h> -#include <asm/byteorder.h> - -extern void wait_for_keypress(void); /* * 35 has been officially registered as the RAMDISK major number, but @@ -79,6 +65,8 @@ #ifdef CONFIG_BLK_DEV_INITRD static int initrd_users; static spinlock_t initrd_users_lock = SPIN_LOCK_UNLOCKED; +unsigned long initrd_start, initrd_end; +int initrd_below_start_ok; #endif /* Various static variables go here. Most are used only in the RAM disk code. @@ -111,70 +99,6 @@ */ int rd_blocksize = BLOCK_SIZE; /* blocksize of the RAM disks */ -#ifndef MODULE - -int rd_doload; /* 1 = load RAM disk, 0 = don't load */ -int rd_prompt = 1; /* 1 = prompt for RAM disk, 0 = don't prompt */ -int rd_image_start; /* starting block # of image */ -#ifdef CONFIG_BLK_DEV_INITRD -unsigned long initrd_start, initrd_end; -int mount_initrd = 1; /* zero if initrd should not be mounted */ -int initrd_below_start_ok; - -static int __init no_initrd(char *str) -{ - mount_initrd = 0; - return 1; -} - -__setup("noinitrd", no_initrd); - -#endif - -static int __init ramdisk_start_setup(char *str) -{ - rd_image_start = simple_strtol(str,NULL,0); - return 1; -} - -static int __init load_ramdisk(char *str) -{ - rd_doload = simple_strtol(str,NULL,0) & 3; - return 1; -} - -static int __init prompt_ramdisk(char *str) -{ - rd_prompt = simple_strtol(str,NULL,0) & 1; - return 1; -} - -static int __init ramdisk_size(char *str) -{ - rd_size = simple_strtol(str,NULL,0); - return 1; -} - -static int __init ramdisk_size2(char *str) -{ - return ramdisk_size(str); -} - -static int __init ramdisk_blocksize(char *str) -{ - rd_blocksize = simple_strtol(str,NULL,0); - return 1; -} - -__setup("ramdisk_start=", ramdisk_start_setup); -__setup("load_ramdisk=", load_ramdisk); -__setup("prompt_ramdisk=", prompt_ramdisk); -__setup("ramdisk=", ramdisk_size); -__setup("ramdisk_size=", ramdisk_size2); -__setup("ramdisk_blocksize=", ramdisk_blocksize); - -#endif - /* * Copyright (C) 2000 Linus Torvalds. * 2000 Transmeta Corp. @@ -492,7 +416,7 @@ } /* This is the registration and initialization section of the RAM disk driver */ -int __init rd_init (void) +static int __init rd_init (void) { int i; @@ -548,7 +472,28 @@ module_init(rd_init); module_exit(rd_cleanup); -/* loadable module support */ +/* options - nonmodular */ +#ifndef MODULE +static int __init ramdisk_size(char *str) +{ + rd_size = simple_strtol(str,NULL,0); + return 1; +} +static int __init ramdisk_size2(char *str) /* kludge */ +{ + return ramdisk_size(str); +} +static int __init ramdisk_blocksize(char *str) +{ + rd_blocksize = simple_strtol(str,NULL,0); + return 1; +} +__setup("ramdisk=", ramdisk_size); +__setup("ramdisk_size=", ramdisk_size2); +__setup("ramdisk_blocksize=", ramdisk_blocksize); +#endif + +/* options - modular */ MODULE_PARM (rd_size, "1i"); MODULE_PARM_DESC(rd_size, "Size of each RAM disk in kbytes."); MODULE_PARM (rd_blocksize, "i"); diff -u --recursive --new-file v2.5.1/linux/drivers/char/console.c linux/drivers/char/console.c --- v2.5.1/linux/drivers/char/console.c Mon Oct 15 14:00:43 2001 +++ linux/drivers/char/console.c Tue Dec 25 15:39:20 2001 @@ -2178,7 +2178,6 @@ name: "tty", write: vt_console_print, device: vt_console_device, - wait_key: keyboard_wait_for_keypress, unblank: unblank_screen, flags: CON_PRINTBUFFER, index: -1, diff -u --recursive --new-file v2.5.1/linux/drivers/char/dz.c linux/drivers/char/dz.c --- v2.5.1/linux/drivers/char/dz.c Sun Sep 9 10:43:02 2001 +++ linux/drivers/char/dz.c Tue Dec 25 15:39:20 2001 @@ -75,7 +75,6 @@ DECLARE_TASK_QUEUE(tq_serial); -extern wait_queue_head_t keypress_wait; static struct dz_serial *lines[4]; static unsigned char tmp_buffer[256]; @@ -228,8 +227,6 @@ if (info->is_console) { if (ch == 0) return; /* it's a break ... */ - - wake_up (&keypress_wait); /* It is a 'keyboard interrupt' ;-) */ } #endif @@ -1509,11 +1506,6 @@ } } -static int dz_console_wait_key(struct console *co) -{ - return 0; -} - static kdev_t dz_console_device(struct console *c) { return MKDEV(TTY_MAJOR, 64 + c->index); @@ -1614,7 +1606,6 @@ name: "ttyS", write: dz_console_print, device: dz_console_device, - wait_key: dz_console_wait_key, setup: dz_console_setup, flags: CON_CONSDEV | CON_PRINTBUFFER, index: CONSOLE_LINE, diff -u --recursive --new-file v2.5.1/linux/drivers/char/keyboard.c linux/drivers/char/keyboard.c --- v2.5.1/linux/drivers/char/keyboard.c Tue Sep 18 13:39:51 2001 +++ linux/drivers/char/keyboard.c Tue Dec 25 15:39:20 2001 @@ -67,15 +67,8 @@ extern void ctrl_alt_del(void); -DECLARE_WAIT_QUEUE_HEAD(keypress_wait); struct console; -int keyboard_wait_for_keypress(struct console *co) -{ - sleep_on(&keypress_wait); - return 0; -} - /* * global state includes the following, and various static variables * in this module: prev_scancode, shift_state, diacr, npadch, dead_key_next. @@ -334,7 +327,6 @@ void put_queue(int ch) { - wake_up(&keypress_wait); if (tty) { tty_insert_flip_char(tty, ch, 0); con_schedule_flip(tty); @@ -343,7 +335,6 @@ static void puts_queue(char *cp) { - wake_up(&keypress_wait); if (!tty) return; diff -u --recursive --new-file v2.5.1/linux/drivers/char/serial.c linux/drivers/char/serial.c --- v2.5.1/linux/drivers/char/serial.c Tue Dec 18 14:56:35 2001 +++ linux/drivers/char/serial.c Tue Dec 25 15:39:20 2001 @@ -5825,35 +5825,6 @@ serial_out(info, UART_IER, ier); } -/* - * Receive character from the serial port - */ -static int serial_console_wait_key(struct console *co) -{ - static struct async_struct *info; - int ier, c; - - info = &async_sercons; - - /* - * First save the IER then disable the interrupts so - * that the real driver for the port does not get the - * character. - */ - ier = serial_in(info, UART_IER); - serial_out(info, UART_IER, 0x00); - - while ((serial_in(info, UART_LSR) & UART_LSR_DR) == 0); - c = serial_in(info, UART_RX); - - /* - * Restore the interrupts - */ - serial_out(info, UART_IER, ier); - - return c; -} - static kdev_t serial_console_device(struct console *c) { return MKDEV(TTY_MAJOR, 64 + c->index); @@ -5994,7 +5965,6 @@ name: "ttyS", write: serial_console_write, device: serial_console_device, - wait_key: serial_console_wait_key, setup: serial_console_setup, flags: CON_PRINTBUFFER, index: -1, diff -u --recursive --new-file v2.5.1/linux/drivers/char/serial167.c linux/drivers/char/serial167.c --- v2.5.1/linux/drivers/char/serial167.c Sun Sep 16 21:23:07 2001 +++ linux/drivers/char/serial167.c Tue Dec 25 15:39:20 2001 @@ -2806,58 +2806,6 @@ restore_flags(flags); } -/* This is a hack; if there are multiple chars waiting in the chip we - * discard all but the last one, and return that. The cd2401 is not really - * designed to be driven in polled mode. - */ - -int serial167_console_wait_key(struct console *co) -{ - volatile unsigned char *base_addr = (u_char *)BASE_ADDR; - unsigned long flags; - volatile u_char sink; - u_char ier; - int port; - int keypress = 0; - - save_flags(flags); cli(); - - /* Ensure receiver is enabled! */ - - port = 0; - base_addr[CyCAR] = (u_char)port; - while (base_addr[CyCCR]) - ; - base_addr[CyCCR] = CyENB_RCVR; - ier = base_addr[CyIER]; - base_addr[CyIER] = CyRxData; - - while (!keypress) { - if (pcc2chip[PccSCCRICR] & 0x20) - { - /* We have an Rx int. Acknowledge it */ - sink = pcc2chip[PccRPIACKR]; - if ((base_addr[CyLICR] >> 2) == port) { - int cnt = base_addr[CyRFOC]; - while (cnt-- > 0) - { - keypress = base_addr[CyRDR]; - } - base_addr[CyREOIR] = 0; - } - else - base_addr[CyREOIR] = CyNOTRANS; - } - } - - base_addr[CyIER] = ier; - - restore_flags(flags); - - return keypress; -} - - static kdev_t serial167_console_device(struct console *c) { return MKDEV(TTY_MAJOR, 64 + c->index); @@ -2874,7 +2822,6 @@ name: "ttyS", write: serial167_console_write, device: serial167_console_device, - wait_key: serial167_console_wait_key, setup: serial167_console_setup, flags: CON_PRINTBUFFER, index: -1, diff -u --recursive --new-file v2.5.1/linux/drivers/char/serial_21285.c linux/drivers/char/serial_21285.c --- v2.5.1/linux/drivers/char/serial_21285.c Thu Sep 13 15:21:32 2001 +++ linux/drivers/char/serial_21285.c Tue Dec 25 15:39:20 2001 @@ -389,17 +389,6 @@ enable_irq(IRQ_CONTX); } -static int rs285_console_wait_key(struct console *co) -{ - int c; - - disable_irq(IRQ_CONRX); - while (*CSR_UARTFLG & 0x10); - c = *CSR_UARTDR; - enable_irq(IRQ_CONRX); - return c; -} - static kdev_t rs285_console_device(struct console *c) { return MKDEV(SERIAL_21285_MAJOR, SERIAL_21285_MINOR); @@ -493,7 +482,6 @@ name: SERIAL_21285_NAME, write: rs285_console_write, device: rs285_console_device, - wait_key: rs285_console_wait_key, setup: rs285_console_setup, flags: CON_PRINTBUFFER, index: -1, diff -u --recursive --new-file v2.5.1/linux/drivers/char/serial_amba.c linux/drivers/char/serial_amba.c --- v2.5.1/linux/drivers/char/serial_amba.c Sun Sep 16 21:23:14 2001 +++ linux/drivers/char/serial_amba.c Tue Dec 25 15:39:20 2001 @@ -1921,22 +1921,6 @@ UART_PUT_CR(port, old_cr); } -/* - * Receive character from the serial port - */ -static int ambauart_console_wait_key(struct console *co) -{ - struct amba_port *port = &amba_ports[co->index]; - unsigned int status; - int c; - - do { - status = UART_GET_FR(port); - } while (!UART_RX_DATA(status)); - c = UART_GET_CHAR(port); - return c; -} - static kdev_t ambauart_console_device(struct console *c) { return MKDEV(SERIAL_AMBA_MAJOR, SERIAL_AMBA_MINOR + c->index); @@ -2015,7 +1999,6 @@ read: ambauart_console_read, #endif device: ambauart_console_device, - wait_key: ambauart_console_wait_key, setup: ambauart_console_setup, flags: CON_PRINTBUFFER, index: -1, diff -u --recursive --new-file v2.5.1/linux/drivers/char/serial_tx3912.c linux/drivers/char/serial_tx3912.c --- v2.5.1/linux/drivers/char/serial_tx3912.c Fri Nov 9 14:01:21 2001 +++ linux/drivers/char/serial_tx3912.c Tue Dec 25 15:39:20 2001 @@ -993,21 +993,6 @@ IntEnable2 = int2; } -static int serial_console_wait_key(struct console *co) -{ - unsigned int int2, res; - - int2 = IntEnable2; - IntEnable2 = 0; - - while (!(UartA_Ctrl1 & UART_RX_HOLD_FULL)); - res = UartA_Data; - udelay(10); - - IntEnable2 = int2; - return res; -} - static void serial_console_write(struct console *co, const char *s, unsigned count) { @@ -1065,7 +1050,6 @@ name: "ttyS", write: serial_console_write, device: serial_console_device, - wait_key: serial_console_wait_key, setup: serial_console_setup, flags: CON_PRINTBUFFER, index: -1 diff -u --recursive --new-file v2.5.1/linux/drivers/char/sh-sci.c linux/drivers/char/sh-sci.c --- v2.5.1/linux/drivers/char/sh-sci.c Mon Oct 15 13:36:48 2001 +++ linux/drivers/char/sh-sci.c Tue Dec 25 15:39:20 2001 @@ -1181,15 +1181,6 @@ put_string(sercons_port, s, count); } -/* - * Receive character from the serial port - */ -static int serial_console_wait_key(struct console *co) -{ - /* Not implemented yet */ - return 0; -} - static kdev_t serial_console_device(struct console *c) { return MKDEV(SCI_MAJOR, SCI_MINOR_START + c->index); @@ -1273,7 +1264,6 @@ name: "ttySC", write: serial_console_write, device: serial_console_device, - wait_key: serial_console_wait_key, setup: serial_console_setup, flags: CON_PRINTBUFFER, index: -1, diff -u --recursive --new-file v2.5.1/linux/drivers/char/tty_io.c linux/drivers/char/tty_io.c --- v2.5.1/linux/drivers/char/tty_io.c Tue Dec 18 14:56:35 2001 +++ linux/drivers/char/tty_io.c Tue Dec 25 15:39:20 2001 @@ -597,12 +597,6 @@ read_unlock(&tasklist_lock); } -void wait_for_keypress(void) -{ - struct console *c = console_drivers; - if (c) c->wait_key(c); -} - void stop_tty(struct tty_struct *tty) { if (tty->stopped) diff -u --recursive --new-file v2.5.1/linux/drivers/char/vme_scc.c linux/drivers/char/vme_scc.c --- v2.5.1/linux/drivers/char/vme_scc.c Sun Sep 16 21:22:50 2001 +++ linux/drivers/char/vme_scc.c Tue Dec 25 15:39:20 2001 @@ -1065,41 +1065,6 @@ restore_flags(flags); } - -static int scc_console_wait_key(struct console *co) -{ - unsigned long flags; - volatile char *p = NULL; - int c; - -#ifdef CONFIG_MVME147_SCC - if (MACH_IS_MVME147) - p = (volatile char *)M147_SCC_A_ADDR; -#endif -#ifdef CONFIG_MVME162_SCC - if (MACH_IS_MVME16x) - p = (volatile char *)MVME_SCC_A_ADDR; -#endif -#ifdef CONFIG_BVME6000_SCC - if (MACH_IS_BVME6000) - p = (volatile char *)BVME_SCC_A_ADDR; -#endif - - save_flags(flags); - cli(); - - /* wait for rx buf filled */ - while ((*p & 0x01) == 0) - ; - - *p = 8; - scc_delay(); - c = *p; - restore_flags(flags); - return c; -} - - static kdev_t scc_console_device(struct console *c) { return MKDEV(TTY_MAJOR, SCC_MINOR_BASE + c->index); @@ -1116,7 +1081,6 @@ name: "ttyS", write: scc_console_write, device: scc_console_device, - wait_key: scc_console_wait_key, setup: scc_console_setup, flags: CON_PRINTBUFFER, index: -1, diff -u --recursive --new-file v2.5.1/linux/drivers/ide/amd74xx.c linux/drivers/ide/amd74xx.c --- v2.5.1/linux/drivers/ide/amd74xx.c Tue Dec 18 14:56:36 2001 +++ linux/drivers/ide/amd74xx.c Mon Dec 17 10:49:31 2001 @@ -459,6 +459,8 @@ hwif->tuneproc = &amd74xx_tune_drive; hwif->speedproc = &amd74xx_tune_chipset; + hwif->highmem = 1; + #ifndef CONFIG_BLK_DEV_IDEDMA hwif->drives[0].autotune = 1; hwif->drives[1].autotune = 1; diff -u --recursive --new-file v2.5.1/linux/drivers/ide/ide-cd.c linux/drivers/ide/ide-cd.c --- v2.5.1/linux/drivers/ide/ide-cd.c Tue Dec 18 14:56:36 2001 +++ linux/drivers/ide/ide-cd.c Thu Dec 20 08:50:13 2001 @@ -594,7 +594,7 @@ cdrom_end_request (1, drive); *startstop = ide_error (drive, "request sense failure", stat); return 1; - } else if (rq->flags & REQ_PC) { + } else if (rq->flags & (REQ_PC | REQ_BLOCK_PC)) { /* All other functions, except for READ. */ struct completion *wait = NULL; pc = (struct packet_command *) rq->special; @@ -2674,6 +2674,8 @@ set_device_ro(MKDEV(HWIF(drive)->major, minor), 1); set_blocksize(MKDEV(HWIF(drive)->major, minor), CD_FRAMESIZE); blk_queue_hardsect_size(&drive->queue, CD_FRAMESIZE); + + blk_queue_prep_rq(&drive->queue, ll_10byte_cmd_build); drive->special.all = 0; drive->ready_stat = 0; diff -u --recursive --new-file v2.5.1/linux/drivers/ide/ide-floppy.c linux/drivers/ide/ide-floppy.c --- v2.5.1/linux/drivers/ide/ide-floppy.c Tue Dec 18 14:56:36 2001 +++ linux/drivers/ide/ide-floppy.c Mon Dec 17 10:49:31 2001 @@ -336,23 +336,7 @@ #define IDEFLOPPY_IOCTL_FORMAT_START 0x4602 #define IDEFLOPPY_IOCTL_FORMAT_GET_PROGRESS 0x4603 -/* - * Special requests for our block device strategy routine. - */ -#define IDEFLOPPY_FIRST_RQ 90 - -/* - * IDEFLOPPY_PC_RQ is used to queue a packet command in the request queue. - */ -#define IDEFLOPPY_PC_RQ 90 - -#define IDEFLOPPY_LAST_RQ 90 - -/* - * A macro which can be used to check if a given request command - * originated in the driver or in the buffer cache layer. - */ -#define IDEFLOPPY_RQ_CMD(cmd) ((cmd >= IDEFLOPPY_FIRST_RQ) && (cmd <= IDEFLOPPY_LAST_RQ)) +#define IDEFLOPPY_RQ (REQ_SPECIAL) /* * Error codes which are returned in rq->errors to the higher part @@ -696,7 +680,7 @@ /* Why does this happen? */ if (!rq) return; - if (!IDEFLOPPY_RQ_CMD (rq->cmd)) { + if (rq->flags & IDEFLOPPY_RQ) { ide_end_request (uptodate, hwgroup); return; } @@ -776,7 +760,7 @@ { ide_init_drive_cmd (rq); rq->buffer = (char *) pc; - rq->cmd = IDEFLOPPY_PC_RQ; + rq->flags = IDEFLOPPY_RQ; (void) ide_do_drive_cmd (drive, rq, ide_preempt); } @@ -1192,6 +1176,7 @@ { int block = sector / floppy->bs_factor; int blocks = rq->nr_sectors / floppy->bs_factor; + int cmd = rq_data_dir(rq); #if IDEFLOPPY_DEBUG_LOG printk ("create_rw1%d_cmd: block == %d, blocks == %d\n", @@ -1200,18 +1185,18 @@ idefloppy_init_pc (pc); if (test_bit (IDEFLOPPY_USE_READ12, &floppy->flags)) { - pc->c[0] = rq->cmd == READ ? IDEFLOPPY_READ12_CMD : IDEFLOPPY_WRITE12_CMD; + pc->c[0] = cmd == READ ? IDEFLOPPY_READ12_CMD : IDEFLOPPY_WRITE12_CMD; put_unaligned (htonl (blocks), (unsigned int *) &pc->c[6]); } else { - pc->c[0] = rq->cmd == READ ? IDEFLOPPY_READ10_CMD : IDEFLOPPY_WRITE10_CMD; + pc->c[0] = cmd == READ ? IDEFLOPPY_READ10_CMD : IDEFLOPPY_WRITE10_CMD; put_unaligned (htons (blocks), (unsigned short *) &pc->c[7]); } put_unaligned (htonl (block), (unsigned int *) &pc->c[2]); pc->callback = &idefloppy_rw_callback; pc->rq = rq; pc->b_data = rq->buffer; - pc->b_count = rq->cmd == READ ? 0 : rq->bio->bi_size; - if (rq->cmd == WRITE) + pc->b_count = cmd == READ ? 0 : rq->bio->bi_size; + if (rq->flags & REQ_RW) set_bit (PC_WRITING, &pc->flags); pc->buffer = NULL; pc->request_transfer = pc->buffer_size = blocks * floppy->block_size; @@ -1227,8 +1212,8 @@ idefloppy_pc_t *pc; #if IDEFLOPPY_DEBUG_LOG - printk (KERN_INFO "rq_status: %d, rq_dev: %u, cmd: %d, errors: %d\n",rq->rq_status,(unsigned int) rq->rq_dev,rq->cmd,rq->errors); - printk (KERN_INFO "sector: %ld, nr_sectors: %ld, current_nr_sectors: %ld\n",rq->sector,rq->nr_sectors,rq->current_nr_sectors); + printk (KERN_INFO "rq_status: %d, rq_dev: %u, flags: %lx, errors: %d\n",rq->rq_status,(unsigned int) rq->rq_dev,rq->flags,rq->errors); + printk (KERN_INFO "sector: %ld, nr_sectors: %ld, current_nr_sectors: %d\n",rq->sector,rq->nr_sectors,rq->current_nr_sectors); #endif /* IDEFLOPPY_DEBUG_LOG */ if (rq->errors >= ERROR_MAX) { @@ -1240,24 +1225,20 @@ idefloppy_end_request (0, HWGROUP(drive)); return ide_stopped; } - switch (rq->cmd) { - case READ: - case WRITE: - if (rq->sector % floppy->bs_factor || rq->nr_sectors % floppy->bs_factor) { - printk ("%s: unsupported r/w request size\n", drive->name); - idefloppy_end_request (0, HWGROUP(drive)); - return ide_stopped; - } - pc = idefloppy_next_pc_storage (drive); - idefloppy_create_rw_cmd (floppy, pc, rq, block); - break; - case IDEFLOPPY_PC_RQ: - pc = (idefloppy_pc_t *) rq->buffer; - break; - default: - printk (KERN_ERR "ide-floppy: unsupported command %x in request queue\n", rq->cmd); - idefloppy_end_request (0,HWGROUP (drive)); + if (rq->flags & REQ_CMD) { + if (rq->sector % floppy->bs_factor || rq->nr_sectors % floppy->bs_factor) { + printk ("%s: unsupported r/w request size\n", drive->name); + idefloppy_end_request (0, HWGROUP(drive)); return ide_stopped; + } + pc = idefloppy_next_pc_storage (drive); + idefloppy_create_rw_cmd (floppy, pc, rq, block); + } else if (rq->flags & IDEFLOPPY_RQ) { + pc = (idefloppy_pc_t *) rq->buffer; + } else { + blk_dump_rq_flags(rq, "ide-floppy: unsupported command in queue"); + idefloppy_end_request (0,HWGROUP (drive)); + return ide_stopped; } pc->rq = rq; return idefloppy_issue_pc (drive, pc); @@ -1273,7 +1254,7 @@ ide_init_drive_cmd (&rq); rq.buffer = (char *) pc; - rq.cmd = IDEFLOPPY_PC_RQ; + rq.flags = IDEFLOPPY_RQ; return ide_do_drive_cmd (drive, &rq, ide_wait); } diff -u --recursive --new-file v2.5.1/linux/drivers/macintosh/mac_keyb.c linux/drivers/macintosh/mac_keyb.c --- v2.5.1/linux/drivers/macintosh/mac_keyb.c Tue Sep 18 14:23:14 2001 +++ linux/drivers/macintosh/mac_keyb.c Tue Dec 25 15:39:20 2001 @@ -256,7 +256,6 @@ #endif extern struct kbd_struct kbd_table[]; -extern wait_queue_head_t keypress_wait; extern void handle_scancode(unsigned char, int); @@ -423,7 +422,6 @@ struct tty_struct *tty; tty = console_driver.table? console_driver.table[fg_console]: NULL; - wake_up(&keypress_wait); if (tty) { tty_insert_flip_char(tty, ch, 0); con_schedule_flip(tty); diff -u --recursive --new-file v2.5.1/linux/drivers/macintosh/macserial.c linux/drivers/macintosh/macserial.c --- v2.5.1/linux/drivers/macintosh/macserial.c Mon Oct 15 13:43:24 2001 +++ linux/drivers/macintosh/macserial.c Tue Dec 25 15:39:20 2001 @@ -2792,30 +2792,6 @@ /* Don't disable the transmitter. */ } -/* - * Receive character from the serial port - */ -static int serial_console_wait_key(struct console *co) -{ - struct mac_serial *info = zs_soft + co->index; - int val; - - /* Turn of interrupts and enable the transmitter. */ - write_zsreg(info->zs_channel, R1, info->curregs[1] & ~INT_ALL_Rx); - write_zsreg(info->zs_channel, R3, info->curregs[3] | RxENABLE); - - /* Wait for something in the receive buffer. */ - while((read_zsreg(info->zs_channel, 0) & Rx_CH_AV) == 0) - eieio(); - val = read_zsdata(info->zs_channel); - - /* Restore the values in the registers. */ - write_zsreg(info->zs_channel, R1, info->curregs[1]); - write_zsreg(info->zs_channel, R3, info->curregs[3]); - - return val; -} - static kdev_t serial_console_device(struct console *c) { return MKDEV(TTY_MAJOR, 64 + c->index); @@ -3006,7 +2982,6 @@ name: "ttyS", write: serial_console_write, device: serial_console_device, - wait_key: serial_console_wait_key, setup: serial_console_setup, flags: CON_PRINTBUFFER, index: -1, diff -u --recursive --new-file v2.5.1/linux/drivers/md/md.c linux/drivers/md/md.c --- v2.5.1/linux/drivers/md/md.c Tue Dec 18 14:56:36 2001 +++ linux/drivers/md/md.c Wed Dec 19 15:18:11 2001 @@ -66,7 +66,7 @@ /* * Current RAID-1,4,5 parallel reconstruction 'guaranteed speed limit' - * is 100 KB/sec, so the extra system load does not show up that much. + * is 1000 KB/sec, so the extra system load does not show up that much. * Increase it if you want to have more _guaranteed_ speed. Note that * the RAID driver will use the maximum available bandwith if the IO * subsystem is idle. There is also an 'absolute maximum' reconstruction @@ -76,8 +76,8 @@ * you can change it via /proc/sys/dev/raid/speed_limit_min and _max. */ -static int sysctl_speed_limit_min = 100; -static int sysctl_speed_limit_max = 100000; +static int sysctl_speed_limit_min = 1000; +static int sysctl_speed_limit_max = 200000; static struct ctl_table_header *raid_table_header; @@ -3336,7 +3336,7 @@ int md_do_sync(mddev_t *mddev, mdp_disk_t *spare) { mddev_t *mddev2; - unsigned int max_sectors, currspeed, + unsigned int max_sectors, currspeed = 0, j, window, err, serialize; unsigned long mark[SYNC_MARKS]; unsigned long mark_cnt[SYNC_MARKS]; @@ -3376,8 +3376,7 @@ max_sectors = mddev->sb->size << 1; printk(KERN_INFO "md: syncing RAID array md%d\n", mdidx(mddev)); - printk(KERN_INFO "md: minimum _guaranteed_ reconstruction speed: %d KB/sec/disc.\n", - sysctl_speed_limit_min); + printk(KERN_INFO "md: minimum _guaranteed_ reconstruction speed: %d KB/sec/disc.\n", sysctl_speed_limit_min); printk(KERN_INFO "md: using maximum available idle IO bandwith " "(but not more than %d KB/sec) for reconstruction.\n", sysctl_speed_limit_max); @@ -3409,7 +3408,7 @@ for (j = 0; j < max_sectors;) { int sectors; - sectors = mddev->pers->sync_request(mddev, j); + sectors = mddev->pers->sync_request(mddev, j, currspeed < sysctl_speed_limit_min); if (sectors < 0) { err = sectors; goto out; diff -u --recursive --new-file v2.5.1/linux/drivers/md/raid1.c linux/drivers/md/raid1.c --- v2.5.1/linux/drivers/md/raid1.c Tue Dec 18 14:56:36 2001 +++ linux/drivers/md/raid1.c Wed Dec 19 15:18:11 2001 @@ -28,7 +28,6 @@ #define MD_DRIVER #define MD_PERSONALITY -#define MAX_WORK_PER_DISK 128 /* * Number of guaranteed r1bios in case of extreme VM load: */ @@ -38,24 +37,6 @@ static spinlock_t retry_list_lock = SPIN_LOCK_UNLOCKED; static LIST_HEAD(retry_list_head); -static inline void check_all_w_bios_empty(r1bio_t *r1_bio) -{ - int i; - - return; - for (i = 0; i < MD_SB_DISKS; i++) - if (r1_bio->write_bios[i]) - BUG(); -} - -static inline void check_all_bios_empty(r1bio_t *r1_bio) -{ - return; - if (r1_bio->read_bio) - BUG(); - check_all_w_bios_empty(r1_bio); -} - static void * r1bio_pool_alloc(int gfp_flags, void *data) { r1bio_t *r1_bio; @@ -69,11 +50,11 @@ static void r1bio_pool_free(void *r1_bio, void *data) { - check_all_bios_empty(r1_bio); kfree(r1_bio); } #define RESYNC_BLOCK_SIZE (64*1024) +#define RESYNC_SECTORS (RESYNC_BLOCK_SIZE >> 9) #define RESYNC_PAGES ((RESYNC_BLOCK_SIZE + PAGE_SIZE-1) / PAGE_SIZE) #define RESYNC_WINDOW (2048*1024) @@ -86,7 +67,6 @@ int i, j; r1_bio = mempool_alloc(conf->r1bio_pool, gfp_flags); - check_all_bios_empty(r1_bio); bio = bio_alloc(gfp_flags, RESYNC_PAGES); if (!bio) @@ -131,7 +111,6 @@ r1bio_t *r1bio = __r1_bio; struct bio *bio = r1bio->master_bio; - check_all_bios_empty(r1bio); if (atomic_read(&bio->bi_cnt) != 1) BUG(); for (i = 0; i < RESYNC_PAGES; i++) { @@ -163,13 +142,25 @@ } *bio = NULL; } - check_all_bios_empty(r1_bio); } static inline void free_r1bio(r1bio_t *r1_bio) { + unsigned long flags; + conf_t *conf = mddev_to_conf(r1_bio->mddev); + /* + * Wake up any possible resync thread that waits for the device + * to go idle. + */ + spin_lock_irqsave(&conf->resync_lock, flags); + if (!--conf->nr_pending) { + wake_up(&conf->wait_idle); + wake_up(&conf->wait_resume); + } + spin_unlock_irqrestore(&conf->resync_lock, flags); + put_all_bios(conf, r1_bio); mempool_free(r1_bio, conf->r1bio_pool); } @@ -178,7 +169,14 @@ { conf_t *conf = mddev_to_conf(r1_bio->mddev); struct bio *bio = r1_bio->master_bio; + unsigned long flags; + spin_lock_irqsave(&conf->resync_lock, flags); + if (!--conf->nr_pending) { + wake_up(&conf->wait_idle); + wake_up(&conf->wait_resume); + } + spin_unlock_irqrestore(&conf->resync_lock, flags); /* * undo any possible partial request fixup magic: */ @@ -222,37 +220,6 @@ md_wakeup_thread(conf->thread); } - -static void inline raid_request_done(unsigned long sector, conf_t *conf, int phase) -{ - unsigned long flags; - spin_lock_irqsave(&conf->segment_lock, flags); - if (sector < conf->start_active) - conf->cnt_done--; - else if (sector >= conf->start_future && conf->phase == phase) - conf->cnt_future--; - else if (!--conf->cnt_pending) - wake_up(&conf->wait_ready); - - spin_unlock_irqrestore(&conf->segment_lock, flags); -} - -static void inline sync_request_done(sector_t sector, conf_t *conf) -{ - unsigned long flags; - - spin_lock_irqsave(&conf->segment_lock, flags); - if (sector >= conf->start_ready) - --conf->cnt_ready; - else if (sector >= conf->start_active) { - if (!--conf->cnt_active) { - conf->start_active = conf->start_ready; - wake_up(&conf->wait_done); - } - } - spin_unlock_irqrestore(&conf->segment_lock, flags); -} - /* * raid_end_bio_io() is called when we have finished servicing a mirrored * operation and are ready to return a success/failure code to the buffer @@ -262,19 +229,29 @@ { struct bio *bio = r1_bio->master_bio; - raid_request_done(bio->bi_sector, mddev_to_conf(r1_bio->mddev), - test_bit(R1BIO_SyncPhase, &r1_bio->state)); - bio_endio(bio, uptodate, nr_sectors); free_r1bio(r1_bio); return 0; } +/* + * Update disk head position estimator based on IRQ completion info. + */ +static void inline update_head_pos(int disk, r1bio_t *r1_bio) +{ + conf_t *conf = mddev_to_conf(r1_bio->mddev); + + conf->mirrors[disk].head_position = + r1_bio->sector + (r1_bio->master_bio->bi_size >> 9); + atomic_dec(&conf->mirrors[disk].nr_pending); +} + static int end_request(struct bio *bio, int nr_sectors) { int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags); r1bio_t * r1_bio = (r1bio_t *)(bio->bi_private); + int i; /* * this branch is our 'one mirror IO has finished' event handler: @@ -287,20 +264,16 @@ * we will return a good error code for to the higher * levels even if IO on some other mirrored buffer fails. * - * The 'master' represents the complex operation to + * The 'master' represents the composite IO operation to * user-side. So if something waits for IO, then it will * wait for the 'master' bio. */ set_bit(R1BIO_Uptodate, &r1_bio->state); - /* - * We split up the read and write side, imho they are - * conceptually different. - */ - if ((r1_bio->cmd == READ) || (r1_bio->cmd == READA)) { if (!r1_bio->read_bio) BUG(); + update_head_pos(r1_bio->read_disk, r1_bio); /* * we have only one bio on the read side */ @@ -322,6 +295,14 @@ /* * WRITE: * + * First, find the disk this bio belongs to. + */ + for (i = 0; i < MD_SB_DISKS; i++) + if (r1_bio->write_bios[i] == bio) { + update_head_pos(i, r1_bio); + break; + } + /* * Let's see if all mirrored write operations have finished * already. */ @@ -332,44 +313,49 @@ /* * This routine returns the disk from which the requested read should - * be done. It bookkeeps the last read position for every disk - * in array and when new read requests come, the disk which last - * position is nearest to the request, is chosen. + * be done. There is a per-array 'next expected sequential IO' sector + * number - if this matches on the next IO then we use the last disk. + * There is also a per-disk 'last know head position' sector that is + * maintained from IRQ contexts, both the normal and the resync IO + * completion handlers update this position correctly. If there is no + * perfect sequential match then we pick the disk whose head is closest. * - * TODO: now if there are 2 mirrors in the same 2 devices, performance - * degrades dramatically because position is mirror, not device based. - * This should be changed to be device based. Also atomic sequential - * reads should be somehow balanced. + * If there are 2 mirrors in the same 2 devices, performance degrades + * because position is mirror, not device based. */ - static int read_balance(conf_t *conf, struct bio *bio, r1bio_t *r1_bio) { - const int sectors = bio->bi_size >> 9; const unsigned long this_sector = r1_bio->sector; - unsigned long new_distance, current_distance; int new_disk = conf->last_used, disk = new_disk; + const int sectors = bio->bi_size >> 9; + sector_t new_distance, current_distance; /* - * Check if it is sane at all to balance - */ - - if (conf->resync_mirrors) + * Check if it if we can balance. We can balance on the whole + * device if no resync is going on, or below the resync window. + * We take the first readable disk when above the resync window. + */ + if (conf->resync_mirrors && (this_sector + sectors >= conf->next_resync)) { + /* make sure that disk is operational */ + new_disk = 0; + while (!conf->mirrors[new_disk].operational || conf->mirrors[new_disk].write_only) { + new_disk++; + if (new_disk == conf->raid_disks) { + new_disk = 0; + break; + } + } goto rb_out; + } - /* make sure that disk is operational */ - while( !conf->mirrors[new_disk].operational) { - if (new_disk <= 0) new_disk = conf->raid_disks; + /* make sure the disk is operational */ + while (!conf->mirrors[new_disk].operational) { + if (new_disk <= 0) + new_disk = conf->raid_disks; new_disk--; if (new_disk == disk) { - /* - * This means no working disk was found - * Nothing much to do, lets not change anything - * and hope for the best... - */ - new_disk = conf->last_used; - goto rb_out; } } @@ -377,36 +363,16 @@ /* now disk == new_disk == starting point for search */ /* - * Don't touch anything for sequential reads. + * Don't change to another disk for sequential reads: */ - if (this_sector == conf->mirrors[new_disk].head_position) + if (conf->next_seq_sect == this_sector) goto rb_out; - - /* - * If reads have been done only on a single disk - * for a time, lets give another disk a change. - * This is for kicking those idling disks so that - * they would find work near some hotspot. - */ - if (conf->sect_count >= conf->mirrors[new_disk].sect_limit) { - conf->sect_count = 0; - - do { - if (new_disk <= 0) - new_disk = conf->raid_disks; - new_disk--; - if (new_disk == disk) - break; - } while ((conf->mirrors[new_disk].write_only) || - (!conf->mirrors[new_disk].operational)); - + if (this_sector == conf->mirrors[new_disk].head_position) goto rb_out; - } - current_distance = abs(this_sector - - conf->mirrors[disk].head_position); + current_distance = abs(this_sector - conf->mirrors[disk].head_position); - /* Find the disk which is closest */ + /* Find the disk whose head is closest */ do { if (disk <= 0) @@ -417,44 +383,56 @@ (!conf->mirrors[disk].operational)) continue; - new_distance = abs(this_sector - - conf->mirrors[disk].head_position); - + if (!atomic_read(&conf->mirrors[disk].nr_pending)) { + new_disk = disk; + break; + } + new_distance = abs(this_sector - conf->mirrors[disk].head_position); if (new_distance < current_distance) { - conf->sect_count = 0; current_distance = new_distance; new_disk = disk; } } while (disk != conf->last_used); rb_out: - conf->mirrors[new_disk].head_position = this_sector + sectors; + r1_bio->read_disk = new_disk; + conf->next_seq_sect = this_sector + sectors; conf->last_used = new_disk; - conf->sect_count += sectors; return new_disk; } /* - * Wait if the reconstruction state machine puts up a bar for - * new requests in this sector range: + * Throttle resync depth, so that we can both get proper overlapping of + * requests, but are still able to handle normal requests quickly. */ -static inline void new_request(conf_t *conf, r1bio_t *r1_bio) +#define RESYNC_DEPTH 32 + +static void device_barrier(conf_t *conf, sector_t sect) { - spin_lock_irq(&conf->segment_lock); - wait_event_lock_irq(conf->wait_done, - r1_bio->sector < conf->start_active || - r1_bio->sector >= conf->start_future, - conf->segment_lock); - if (r1_bio->sector < conf->start_active) - conf->cnt_done++; - else { - conf->cnt_future++; - if (conf->phase) - set_bit(R1BIO_SyncPhase, &r1_bio->state); + spin_lock_irq(&conf->resync_lock); + wait_event_lock_irq(conf->wait_idle, !waitqueue_active(&conf->wait_resume), conf->resync_lock); + + if (!conf->barrier++) { + wait_event_lock_irq(conf->wait_idle, !conf->nr_pending, conf->resync_lock); + if (conf->nr_pending) + BUG(); } - spin_unlock_irq(&conf->segment_lock); + wait_event_lock_irq(conf->wait_resume, conf->barrier < RESYNC_DEPTH, conf->resync_lock); + conf->next_resync = sect; + spin_unlock_irq(&conf->resync_lock); +} + +static void resume_device(conf_t *conf) +{ + spin_lock_irq(&conf->resync_lock); + if (!conf->barrier) + BUG(); + --conf->barrier; + wake_up(&conf->wait_resume); + wake_up(&conf->wait_idle); + spin_unlock_irq(&conf->resync_lock); } static int make_request(mddev_t *mddev, int rw, struct bio * bio) @@ -466,6 +444,16 @@ int i, sum_bios = 0, disks = MD_SB_DISKS; /* + * Register the new request and wait if the reconstruction + * thread has put up a bar for new requests. + * Continue immediately if no resync is active currently. + */ + spin_lock_irq(&conf->resync_lock); + wait_event_lock_irq(conf->wait_resume, !conf->barrier, conf->resync_lock); + conf->nr_pending++; + spin_unlock_irq(&conf->resync_lock); + + /* * make_request() can abort the operation when READA is being * used and no empty request is available. * @@ -475,7 +463,6 @@ rw = READ; r1_bio = mempool_alloc(conf->r1bio_pool, GFP_NOIO); - check_all_bios_empty(r1_bio); r1_bio->master_bio = bio; @@ -483,8 +470,6 @@ r1_bio->sector = bio->bi_sector; r1_bio->cmd = rw; - new_request(conf, r1_bio); - if (rw == READ) { /* * read balancing logic: @@ -503,15 +488,13 @@ read_bio->bi_private = r1_bio; generic_make_request(read_bio); + atomic_inc(&conf->mirrors[r1_bio->read_disk].nr_pending); return 0; } /* * WRITE: */ - - check_all_w_bios_empty(r1_bio); - for (i = 0; i < disks; i++) { struct bio *mbio; if (!conf->mirrors[i].operational) @@ -542,14 +525,13 @@ /* * We have to be a bit careful about the semaphore above, thats - * why we start the requests separately. Since kmalloc() could - * fail, sleep and make_request() can sleep too, this is the - * safer solution. Imagine, end_request decreasing the semaphore - * before we could have set it up ... We could play tricks with - * the semaphore (presetting it and correcting at the end if - * sum_bios is not 'n' but we have to do end_request by hand if - * all requests finish until we had a chance to set up the - * semaphore correctly ... lots of races). + * why we start the requests separately. Since generic_make_request() + * can sleep, this is the safer solution. Imagine, end_request + * decreasing the semaphore before we could have set it up ... + * We could play tricks with the semaphore (presetting it and + * correcting at the end if sum_bios is not 'n' but we have to + * do end_request by hand if all requests finish until we had a + * chance to set up the semaphore correctly ... lots of races). */ for (i = 0; i < disks; i++) { struct bio *mbio; @@ -558,6 +540,7 @@ continue; generic_make_request(mbio); + atomic_inc(&conf->mirrors[i].nr_pending); } return 0; } @@ -610,8 +593,7 @@ md_wakeup_thread(conf->thread); if (!mirror->write_only) conf->working_disks--; - printk(DISK_FAILED, partition_name(mirror->dev), - conf->working_disks); + printk(DISK_FAILED, partition_name(mirror->dev), conf->working_disks); } static int error(mddev_t *mddev, kdev_t dev) @@ -644,12 +626,6 @@ return 0; } -#undef LAST_DISK -#undef NO_SPARE_DISK -#undef DISK_FAILED -#undef START_SYNCING - - static void print_conf(conf_t *conf) { int i; @@ -674,29 +650,13 @@ static void close_sync(conf_t *conf) { - mddev_t *mddev = conf->mddev; - /* - * If reconstruction was interrupted, we need to close the "active" - * and "pending" holes. - * we know that there are no active rebuild requests, - * os cnt_active == cnt_ready == 0 - */ - spin_lock_irq(&conf->segment_lock); - conf->start_active = conf->start_pending; - conf->start_ready = conf->start_pending; - wait_event_lock_irq(conf->wait_ready, !conf->cnt_pending, conf->segment_lock); - conf->start_active = conf->start_ready = conf->start_pending = conf->start_future; - conf->start_future = mddev->sb->size+1; - conf->cnt_pending = conf->cnt_future; - conf->cnt_future = 0; - conf->phase = conf->phase ^1; - wait_event_lock_irq(conf->wait_ready, !conf->cnt_pending, conf->segment_lock); - conf->start_active = conf->start_ready = conf->start_pending = conf->start_future = 0; - conf->phase = 0; - conf->cnt_future = conf->cnt_done;; - conf->cnt_done = 0; - spin_unlock_irq(&conf->segment_lock); - wake_up(&conf->wait_done); + spin_lock_irq(&conf->resync_lock); + wait_event_lock_irq(conf->wait_resume, !conf->barrier, conf->resync_lock); + spin_unlock_irq(&conf->resync_lock); + + if (conf->barrier) BUG(); + if (waitqueue_active(&conf->wait_idle)) BUG(); + if (waitqueue_active(&conf->wait_resume)) BUG(); } static int diskop(mddev_t *mddev, mdp_disk_t **d, int state) @@ -975,9 +935,9 @@ int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags); r1bio_t * r1_bio = (r1bio_t *)(bio->bi_private); - check_all_w_bios_empty(r1_bio); if (r1_bio->read_bio != bio) BUG(); + update_head_pos(r1_bio->read_disk, r1_bio); /* * we have read a block, now it needs to be re-written, * or re-read if the read failed. @@ -997,13 +957,21 @@ int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags); r1bio_t * r1_bio = (r1bio_t *)(bio->bi_private); mddev_t *mddev = r1_bio->mddev; + int i; if (!uptodate) md_error(mddev, bio->bi_dev); + for (i = 0; i < MD_SB_DISKS; i++) + if (r1_bio->write_bios[i] == bio) { + update_head_pos(i, r1_bio); + break; + } + if (atomic_dec_and_test(&r1_bio->remaining)) { - sync_request_done(r1_bio->sector, mddev_to_conf(mddev)); + conf_t *conf = mddev_to_conf(mddev); md_done_sync(mddev, r1_bio->master_bio->bi_size >> 9, uptodate); + resume_device(conf); put_buf(r1_bio); } return 0; @@ -1029,11 +997,11 @@ */ printk(IO_ERROR, partition_name(bio->bi_dev), r1_bio->sector); md_done_sync(mddev, r1_bio->master_bio->bi_size >> 9, 0); + resume_device(conf); + put_buf(r1_bio); return; } - check_all_w_bios_empty(r1_bio); - for (i = 0; i < disks ; i++) { if (!conf->mirrors[i].operational) continue; @@ -1071,8 +1039,8 @@ * must be done */ printk(IO_ERROR, partition_name(bio->bi_dev), r1_bio->sector); - sync_request_done(r1_bio->sector, conf); md_done_sync(mddev, r1_bio->master_bio->bi_size >> 9, 0); + resume_device(conf); put_buf(r1_bio); return; } @@ -1083,6 +1051,7 @@ md_sync_acct(mbio->bi_dev, mbio->bi_size >> 9); generic_make_request(mbio); + atomic_inc(&conf->mirrors[i].nr_pending); } } @@ -1101,6 +1070,7 @@ struct bio *bio; unsigned long flags; mddev_t *mddev; + conf_t *conf; kdev_t dev; @@ -1111,9 +1081,9 @@ r1_bio = list_entry(head->prev, r1bio_t, retry_list); list_del(head->prev); spin_unlock_irqrestore(&retry_list_lock, flags); - check_all_w_bios_empty(r1_bio); mddev = r1_bio->mddev; + conf = mddev_to_conf(mddev); if (mddev->sb_dirty) { printk(KERN_INFO "raid1: dirty sb detected, updating.\n"); mddev->sb_dirty = 0; @@ -1139,13 +1109,12 @@ bio->bi_rw = r1_bio->cmd; generic_make_request(bio); + atomic_inc(&conf->mirrors[r1_bio->read_disk].nr_pending); break; } } spin_unlock_irqrestore(&retry_list_lock, flags); } -#undef IO_ERROR -#undef REDIRECT_SECTOR /* * Private kernel thread to reconstruct mirrors after an unclean @@ -1177,101 +1146,27 @@ { int buffs; - conf->start_active = 0; - conf->start_ready = 0; - conf->start_pending = 0; - conf->start_future = 0; - conf->phase = 0; - buffs = RESYNC_WINDOW / RESYNC_BLOCK_SIZE; if (conf->r1buf_pool) BUG(); conf->r1buf_pool = mempool_create(buffs, r1buf_pool_alloc, r1buf_pool_free, conf); if (!conf->r1buf_pool) return -ENOMEM; - conf->window = 2048; - conf->cnt_future += conf->cnt_done+conf->cnt_pending; - conf->cnt_done = conf->cnt_pending = 0; - if (conf->cnt_ready || conf->cnt_active) - MD_BUG(); + conf->next_resync = 0; return 0; } -static void wait_sync_pending(conf_t *conf, sector_t sector_nr) -{ - spin_lock_irq(&conf->segment_lock); - while (sector_nr >= conf->start_pending) { -// printk("wait .. sect=%lu start_active=%d ready=%d pending=%d future=%d, cnt_done=%d active=%d ready=%d pending=%d future=%d\n", sector_nr, conf->start_active, conf->start_ready, conf->start_pending, conf->start_future, conf->cnt_done, conf->cnt_active, conf->cnt_ready, conf->cnt_pending, conf->cnt_future); - wait_event_lock_irq(conf->wait_done, !conf->cnt_active, - conf->segment_lock); - wait_event_lock_irq(conf->wait_ready, !conf->cnt_pending, - conf->segment_lock); - conf->start_active = conf->start_ready; - conf->start_ready = conf->start_pending; - conf->start_pending = conf->start_future; - conf->start_future = conf->start_future+conf->window; - - // Note: falling off the end is not a problem - conf->phase = conf->phase ^1; - conf->cnt_active = conf->cnt_ready; - conf->cnt_ready = 0; - conf->cnt_pending = conf->cnt_future; - conf->cnt_future = 0; - wake_up(&conf->wait_done); - } - conf->cnt_ready++; - spin_unlock_irq(&conf->segment_lock); -} - /* * perform a "sync" on one "block" * * We need to make sure that no normal I/O request - particularly write * requests - conflict with active sync requests. - * This is achieved by conceptually dividing the block space into a - * number of sections: - * DONE: 0 .. a-1 These blocks are in-sync - * ACTIVE: a.. b-1 These blocks may have active sync requests, but - * no normal IO requests - * READY: b .. c-1 These blocks have no normal IO requests - sync - * request may be happening - * PENDING: c .. d-1 These blocks may have IO requests, but no new - * ones will be added - * FUTURE: d .. end These blocks are not to be considered yet. IO may - * be happening, but not sync - * - * We keep a - * phase which flips (0 or 1) each time d moves and - * a count of: - * z = active io requests in FUTURE since d moved - marked with - * current phase - * y = active io requests in FUTURE before d moved, or PENDING - - * marked with previous phase - * x = active sync requests in READY - * w = active sync requests in ACTIVE - * v = active io requests in DONE * - * Normally, a=b=c=d=0 and z= active io requests - * or a=b=c=d=END and v= active io requests - * Allowed changes to a,b,c,d: - * A: c==d && y==0 -> d+=window, y=z, z=0, phase=!phase - * B: y==0 -> c=d - * C: b=c, w+=x, x=0 - * D: w==0 -> a=b - * E: a==b==c==d==end -> a=b=c=d=0, z=v, v=0 - * - * At start of sync we apply A. - * When y reaches 0, we apply B then A then being sync requests - * When sync point reaches c-1, we wait for y==0, and W==0, and - * then apply apply B then A then D then C. - * Finally, we apply E - * - * The sync request simply issues a "read" against a working drive - * This is marked so that on completion the raid1d thread is woken to - * issue suitable write requests + * This is achieved by tracking pending requests and a 'barrier' concept + * that can be installed to exclude normal IO requests. */ -static int sync_request(mddev_t *mddev, sector_t sector_nr) +static int sync_request(mddev_t *mddev, sector_t sector_nr, int go_faster) { conf_t *conf = mddev_to_conf(mddev); mirror_info_t *mirror; @@ -1283,8 +1178,13 @@ if (!sector_nr) if (init_resync(conf)) return -ENOMEM; - - wait_sync_pending(conf, sector_nr); + /* + * If there is non-resync activity waiting for us then + * put in a delay to throttle resync. + */ + if (!go_faster && waitqueue_active(&conf->wait_resume)) + schedule_timeout(HZ); + device_barrier(conf, sector_nr + RESYNC_SECTORS); /* * If reconstructing, and >1 working disc, @@ -1302,10 +1202,13 @@ } conf->last_used = disk; - mirror = conf->mirrors+conf->last_used; + mirror = conf->mirrors + conf->last_used; r1_bio = mempool_alloc(conf->r1buf_pool, GFP_NOIO); - check_all_bios_empty(r1_bio); + + spin_lock_irq(&conf->resync_lock); + conf->nr_pending++; + spin_unlock_irq(&conf->resync_lock); r1_bio->mddev = mddev; r1_bio->sector = sector_nr; @@ -1344,6 +1247,7 @@ md_sync_acct(read_bio->bi_dev, nr_sectors); generic_make_request(read_bio); + atomic_inc(&conf->mirrors[conf->last_used].nr_pending); return nr_sectors; } @@ -1447,7 +1351,6 @@ disk->number = descriptor->number; disk->raid_disk = disk_idx; disk->dev = rdev->dev; - disk->sect_limit = MAX_WORK_PER_DISK; disk->operational = 0; disk->write_only = 0; disk->spare = 0; @@ -1479,7 +1382,6 @@ disk->number = descriptor->number; disk->raid_disk = disk_idx; disk->dev = rdev->dev; - disk->sect_limit = MAX_WORK_PER_DISK; disk->operational = 1; disk->write_only = 0; disk->spare = 0; @@ -1494,7 +1396,6 @@ disk->number = descriptor->number; disk->raid_disk = disk_idx; disk->dev = rdev->dev; - disk->sect_limit = MAX_WORK_PER_DISK; disk->operational = 0; disk->write_only = 0; disk->spare = 1; @@ -1507,9 +1408,9 @@ conf->mddev = mddev; conf->device_lock = SPIN_LOCK_UNLOCKED; - conf->segment_lock = SPIN_LOCK_UNLOCKED; - init_waitqueue_head(&conf->wait_done); - init_waitqueue_head(&conf->wait_ready); + conf->resync_lock = SPIN_LOCK_UNLOCKED; + init_waitqueue_head(&conf->wait_idle); + init_waitqueue_head(&conf->wait_resume); if (!conf->working_disks) { printk(NONE_OPERATIONAL, mdidx(mddev)); @@ -1610,17 +1511,6 @@ MOD_DEC_USE_COUNT; return -EIO; } - -#undef INVALID_LEVEL -#undef NO_SB -#undef ERRORS -#undef NOT_IN_SYNC -#undef INCONSISTENT -#undef ALREADY_RUNNING -#undef OPERATIONAL -#undef SPARE -#undef NONE_OPERATIONAL -#undef ARRAY_IS_ACTIVE static int stop_resync(mddev_t *mddev) { diff -u --recursive --new-file v2.5.1/linux/drivers/mtd/nftlmount.c linux/drivers/mtd/nftlmount.c --- v2.5.1/linux/drivers/mtd/nftlmount.c Thu Oct 4 15:13:18 2001 +++ linux/drivers/mtd/nftlmount.c Thu Dec 20 08:53:28 2001 @@ -4,7 +4,7 @@ * Author: Fabrice Bellard (fabrice.bellard@netgem.com) * Copyright (C) 2000 Netgem S.A. * - * $Id: nftlmount.c,v 1.23 2001/09/19 21:42:32 dwmw2 Exp $ + * $Id: nftlmount.c,v 1.25 2001/11/30 16:46:27 dwmw2 Exp $ * * This program is free software; you can redistribute it and/or modify * it under the terms of the GNU General Public License as published by @@ -39,7 +39,7 @@ #define SECTORSIZE 512 -char nftlmountrev[]="$Revision: 1.23 $"; +char nftlmountrev[]="$Revision: 1.25 $"; /* find_boot_record: Find the NFTL Media Header and its Spare copy which contains the * various device information of the NFTL partition and Bad Unit Table. Update @@ -94,11 +94,11 @@ continue; } -#if 1 /* Some people seem to have devices without ECC or erase marks +#if 0 /* Some people seem to have devices without ECC or erase marks on the Media Header blocks. There are enough other sanity checks in here that we can probably do without it. */ - if (le16_to_cpu ((h1.EraseMark | h1.EraseMark1) != ERASE_MARK)) { + if (le16_to_cpu(h1.EraseMark | h1.EraseMark1) != ERASE_MARK) { printk(KERN_NOTICE "ANAND header found at 0x%x in mtd%d, but erase mark not present (0x%04x,0x%04x instead)\n", block * nftl->EraseSize, nftl->mtd->index, le16_to_cpu(h1.EraseMark), le16_to_cpu(h1.EraseMark1)); diff -u --recursive --new-file v2.5.1/linux/drivers/s390/char/tuball.c linux/drivers/s390/char/tuball.c --- v2.5.1/linux/drivers/s390/char/tuball.c Thu Oct 11 09:43:29 2001 +++ linux/drivers/s390/char/tuball.c Tue Dec 25 15:39:20 2001 @@ -84,7 +84,6 @@ tub3270_con_write, /* write */ NULL, /* read */ tub3270_con_device, /* device */ - NULL, /* wait_key */ tub3270_con_unblank, /* unblank */ NULL, /* setup */ CON_PRINTBUFFER, /* flags */ diff -u --recursive --new-file v2.5.1/linux/drivers/sbus/char/sab82532.c linux/drivers/sbus/char/sab82532.c --- v2.5.1/linux/drivers/sbus/char/sab82532.c Wed Oct 17 14:16:39 2001 +++ linux/drivers/sbus/char/sab82532.c Tue Dec 25 15:39:20 2001 @@ -361,10 +361,6 @@ writeb(SAB82532_CMDR_RMC, &info->regs->w.cmdr); } -#ifdef CONFIG_SERIAL_CONSOLE - if (info->is_console) - wake_up(&keypress_wait); -#endif if (!tty) return; @@ -2541,13 +2537,6 @@ sab82532_tec_wait(info); } -static int -sab82532_console_wait_key(struct console *con) -{ - sleep_on(&keypress_wait); - return 0; -} - static kdev_t sab82532_console_device(struct console *con) { @@ -2622,7 +2611,6 @@ name: "ttyS", write: sab82532_console_write, device: sab82532_console_device, - wait_key: sab82532_console_wait_key, setup: sab82532_console_setup, flags: CON_PRINTBUFFER, index: -1, diff -u --recursive --new-file v2.5.1/linux/drivers/sbus/char/su.c linux/drivers/sbus/char/su.c --- v2.5.1/linux/drivers/sbus/char/su.c Tue Nov 13 09:16:05 2001 +++ linux/drivers/sbus/char/su.c Tue Dec 25 15:39:20 2001 @@ -2854,40 +2854,6 @@ su_outb(info, UART_IER, ier); } -/* - * Receive character from the serial port - */ -static int -serial_console_wait_key(struct console *co) -{ - struct su_struct *info; - int ier; - int lsr; - int c; - - info = su_table + co->index; - - /* - * First save the IER then disable the interrupts so - * that the real driver for the port does not get the - * character. - */ - ier = su_inb(info, UART_IER); - su_outb(info, UART_IER, 0x00); - - do { - lsr = su_inb(info, UART_LSR); - } while (!(lsr & UART_LSR_DR)); - c = su_inb(info, UART_RX); - - /* - * Restore the interrupts - */ - su_outb(info, UART_IER, ier); - - return c; -} - static kdev_t serial_console_device(struct console *c) { @@ -3013,7 +2979,6 @@ name: "ttyS", write: serial_console_write, device: serial_console_device, - wait_key: serial_console_wait_key, setup: serial_console_setup, flags: CON_PRINTBUFFER, index: -1, diff -u --recursive --new-file v2.5.1/linux/drivers/sbus/char/sunkbd.c linux/drivers/sbus/char/sunkbd.c --- v2.5.1/linux/drivers/sbus/char/sunkbd.c Sun Sep 16 21:22:50 2001 +++ linux/drivers/sbus/char/sunkbd.c Tue Dec 25 15:39:20 2001 @@ -82,16 +82,6 @@ struct l1a_kbd_state l1a_state; -#ifndef CONFIG_PCI -DECLARE_WAIT_QUEUE_HEAD(keypress_wait); -#endif - -int keyboard_wait_for_keypress(struct console *co) -{ - sleep_on(&keypress_wait); - return 0; -} - static spinlock_t sunkbd_lock = SPIN_LOCK_UNLOCKED; /* @@ -629,7 +619,6 @@ static void put_queue(int ch) { - wake_up(&keypress_wait); if (tty) { tty_insert_flip_char(tty, ch, 0); con_schedule_flip(tty); @@ -638,7 +627,6 @@ static void puts_queue(char *cp) { - wake_up(&keypress_wait); if (!tty) return; diff -u --recursive --new-file v2.5.1/linux/drivers/sbus/char/zs.c linux/drivers/sbus/char/zs.c --- v2.5.1/linux/drivers/sbus/char/zs.c Tue Oct 30 15:08:11 2001 +++ linux/drivers/sbus/char/zs.c Tue Dec 25 15:39:20 2001 @@ -532,8 +532,6 @@ /* Continue execution... */ return; } - /* It is a 'keyboard interrupt' ;-) */ - wake_up(&keypress_wait); } #ifndef __sparc_v9__ /* Look for kgdb 'stop' character, consult the gdb @@ -2761,13 +2759,6 @@ #endif } -static int -zs_console_wait_key(struct console *con) -{ - sleep_on(&keypress_wait); - return 0; -} - static kdev_t zs_console_device(struct console *con) { return MKDEV(TTY_MAJOR, 64 + con->index); @@ -2845,7 +2836,6 @@ name: "ttyS", write: zs_console_write, device: zs_console_device, - wait_key: zs_console_wait_key, setup: zs_console_setup, flags: CON_PRINTBUFFER, index: -1, diff -u --recursive --new-file v2.5.1/linux/drivers/scsi/qlogicfas.c linux/drivers/scsi/qlogicfas.c --- v2.5.1/linux/drivers/scsi/qlogicfas.c Thu Oct 25 13:53:51 2001 +++ linux/drivers/scsi/qlogicfas.c Mon Dec 17 10:49:31 2001 @@ -467,10 +467,11 @@ static void do_ql_ihandl(int irq, void *dev_id, struct pt_regs * regs) { unsigned long flags; + struct Scsi_Host *host = dev_id; - spin_lock_irqsave(&io_request_lock, flags); + spin_lock_irqsave(&host->host_lock, flags); ql_ihandl(irq, dev_id, regs); - spin_unlock_irqrestore(&io_request_lock, flags); + spin_unlock_irqrestore(&host->host_lock, flags); } #endif diff -u --recursive --new-file v2.5.1/linux/drivers/scsi/scsi.h linux/drivers/scsi/scsi.h --- v2.5.1/linux/drivers/scsi/scsi.h Tue Dec 18 14:56:38 2001 +++ linux/drivers/scsi/scsi.h Tue Dec 25 15:59:19 2001 @@ -548,10 +548,8 @@ */ struct scsi_device { -/* private: */ /* - * This information is private to the scsi mid-layer. Wrapping it in a - * struct private is a way of marking it in a sort of C++ type of way. + * This information is private to the scsi mid-layer. */ struct scsi_device *next; /* Used for linked list */ struct scsi_device *prev; /* Used for linked list */ @@ -563,7 +561,6 @@ volatile unsigned short device_busy; /* commands actually active on low-level */ Scsi_Cmnd *device_queue; /* queue of SCSI Command structures */ -/* public: */ unsigned int id, lun, channel; unsigned int manufacturer; /* Manufacturer of device, for using @@ -681,11 +678,7 @@ */ struct scsi_cmnd { int sc_magic; -/* private: */ - /* - * This information is private to the scsi mid-layer. Wrapping it in a - * struct private is a way of marking it in a sort of C++ type of way. - */ + struct Scsi_Host *host; unsigned short state; unsigned short owner; @@ -726,8 +719,6 @@ unsigned volatile char internal_timeout; struct scsi_cmnd *bh_next; /* To enumerate the commands waiting to be processed. */ - -/* public: */ unsigned int target; unsigned int lun; diff -u --recursive --new-file v2.5.1/linux/drivers/scsi/scsi_merge.c linux/drivers/scsi/scsi_merge.c --- v2.5.1/linux/drivers/scsi/scsi_merge.c Tue Dec 18 14:56:38 2001 +++ linux/drivers/scsi/scsi_merge.c Thu Dec 20 08:50:13 2001 @@ -59,12 +59,10 @@ */ int scsi_init_io(Scsi_Cmnd *SCpnt) { - struct request *req; + struct request *req = &SCpnt->request; struct scatterlist *sgpnt; int count, gfp_mask; - req = &SCpnt->request; - /* * First we need to know how many scatter gather segments are needed. */ @@ -85,14 +83,13 @@ BUG_ON(!sgpnt); SCpnt->request_buffer = (char *) sgpnt; - SCpnt->request_bufflen = 0; + SCpnt->request_bufflen = req->nr_sectors << 9; req->buffer = NULL; /* * Next, walk the list, and fill in the addresses and sizes of * each segment. */ - SCpnt->request_bufflen = req->nr_sectors << 9; count = blk_rq_map_sg(req->q, req, SCpnt->request_buffer); /* @@ -142,8 +139,7 @@ bounce_limit = BLK_BOUNCE_ANY; else bounce_limit = SHpnt->pci_dev->dma_mask; - } - if (SHpnt->unchecked_isa_dma) + } else if (SHpnt->unchecked_isa_dma) bounce_limit = BLK_BOUNCE_ISA; blk_queue_bounce_limit(q, bounce_limit); diff -u --recursive --new-file v2.5.1/linux/drivers/scsi/sr.c linux/drivers/scsi/sr.c --- v2.5.1/linux/drivers/scsi/sr.c Tue Dec 18 14:56:38 2001 +++ linux/drivers/scsi/sr.c Thu Dec 20 08:50:13 2001 @@ -287,9 +287,6 @@ return 0; } - if (rq_data_dir(&SCpnt->request) == WRITE && !scsi_CDs[dev].device->writeable) - return 0; - /* * we do lazy blocksize switching (when reading XA sectors, * see CDROMREADMODE2 ioctl) diff -u --recursive --new-file v2.5.1/linux/drivers/sgi/char/sgiserial.c linux/drivers/sgi/char/sgiserial.c --- v2.5.1/linux/drivers/sgi/char/sgiserial.c Mon Aug 27 08:56:31 2001 +++ linux/drivers/sgi/char/sgiserial.c Tue Dec 25 15:39:20 2001 @@ -49,8 +49,6 @@ #define NUM_SERIAL 1 /* One chip on board. */ #define NUM_CHANNELS (NUM_SERIAL * 2) -extern wait_queue_head_t keypress_wait; - struct sgi_zslayout *zs_chips[NUM_SERIAL] = { 0, }; struct sgi_zschannel *zs_channels[NUM_CHANNELS] = { 0, 0, }; struct sgi_zschannel *zs_conschan; @@ -428,8 +426,6 @@ show_buffers(); return; } - /* It is a 'keyboard interrupt' ;-) */ - wake_up(&keypress_wait); } /* Look for kgdb 'stop' character, consult the gdb documentation * for remote target debugging and arch/sparc/kernel/sparc-stub.c @@ -2117,12 +2113,6 @@ rs_fair_output(); } -static int zs_console_wait_key(struct console *con) -{ - sleep_on(&keypress_wait); - return 0; -} - static kdev_t zs_console_device(struct console *con) { return MKDEV(TTY_MAJOR, 64 + con->index); @@ -2274,7 +2264,6 @@ name: "ttyS", write: zs_console_write, device: zs_console_device, - wait_key: zs_console_wait_key, setup: zs_console_setup, flags: CON_PRINTBUFFER, index: -1, diff -u --recursive --new-file v2.5.1/linux/drivers/tc/zs.c linux/drivers/tc/zs.c --- v2.5.1/linux/drivers/tc/zs.c Mon Aug 27 08:56:31 2001 +++ linux/drivers/tc/zs.c Tue Dec 25 15:39:20 2001 @@ -2125,18 +2125,6 @@ } } -/* - * Receive character from the serial port - */ -static int serial_console_wait_key(struct console *co) -{ - struct dec_serial *info; - - info = zs_soft + co->index; - - return zs_poll_rx_char(info); -} - static kdev_t serial_console_device(struct console *c) { return MKDEV(TTY_MAJOR, 64 + c->index); @@ -2271,7 +2259,6 @@ name: "ttyS", write: serial_console_write, device: serial_console_device, - wait_key: serial_console_wait_key, setup: serial_console_setup, flags: CON_PRINTBUFFER, index: -1, diff -u --recursive --new-file v2.5.1/linux/fs/Config.in linux/fs/Config.in --- v2.5.1/linux/fs/Config.in Mon Nov 12 09:34:16 2001 +++ linux/fs/Config.in Tue Dec 25 15:39:20 2001 @@ -45,7 +45,7 @@ fi tristate 'Compressed ROM file system support' CONFIG_CRAMFS bool 'Virtual memory file system support (former shm fs)' CONFIG_TMPFS -tristate 'Simple RAM-based file system support' CONFIG_RAMFS +define_bool CONFIG_RAMFS y tristate 'ISO 9660 CDROM file system support' CONFIG_ISO9660_FS dep_mbool ' Microsoft Joliet CDROM extensions' CONFIG_JOLIET $CONFIG_ISO9660_FS diff -u --recursive --new-file v2.5.1/linux/fs/adfs/super.c linux/fs/adfs/super.c --- v2.5.1/linux/fs/adfs/super.c Tue Dec 18 14:56:39 2001 +++ linux/fs/adfs/super.c Tue Dec 25 15:39:20 2001 @@ -319,8 +319,7 @@ if (parse_options(sb, data)) goto error; - sb->s_blocksize = BLOCK_SIZE; - set_blocksize(dev, BLOCK_SIZE); + sb_set_blocksize(sb, BLOCK_SIZE); if (!(bh = sb_bread(sb, ADFS_DISCRECORD / BLOCK_SIZE))) { adfs_error(sb, "unable to read superblock"); goto error; @@ -347,14 +346,8 @@ goto error_free_bh; } - sb->s_blocksize_bits = dr->log2secsize; - sb->s_blocksize = 1 << sb->s_blocksize_bits; - if (sb->s_blocksize != BLOCK_SIZE && - (sb->s_blocksize == 512 || sb->s_blocksize == 1024 || - sb->s_blocksize == 2048 || sb->s_blocksize == 4096)) { - - brelse(bh); - set_blocksize(dev, sb->s_blocksize); + brelse(bh); + if (sb_set_blocksize(sb, 1 << dr->log2secsize)) { bh = sb_bread(sb, ADFS_DISCRECORD / sb->s_blocksize); if (!bh) { adfs_error(sb, "couldn't read superblock on " @@ -367,12 +360,11 @@ goto error_free_bh; } dr = (struct adfs_discrecord *)(b_data + ADFS_DR_OFFSET); - } - if (sb->s_blocksize != bh->b_size) { + } else { if (!silent) printk(KERN_ERR "VFS: Unsupported blocksize on dev " "%s.\n", bdevname(dev)); - goto error_free_bh; + goto error; } /* diff -u --recursive --new-file v2.5.1/linux/fs/affs/super.c linux/fs/affs/super.c --- v2.5.1/linux/fs/affs/super.c Tue Dec 18 14:56:39 2001 +++ linux/fs/affs/super.c Tue Dec 25 15:39:20 2001 @@ -327,10 +327,6 @@ got_root: root_block = AFFS_SB->s_root_block; - sb->s_blocksize_bits = blocksize == 512 ? 9 : - blocksize == 1024 ? 10 : - blocksize == 2048 ? 11 : 12; - /* Find out which kind of FS we have */ boot_bh = sb_bread(sb, 0); if (!boot_bh) { diff -u --recursive --new-file v2.5.1/linux/fs/bfs/inode.c linux/fs/bfs/inode.c --- v2.5.1/linux/fs/bfs/inode.c Tue Dec 18 14:56:39 2001 +++ linux/fs/bfs/inode.c Tue Dec 25 15:39:20 2001 @@ -248,9 +248,7 @@ int i, imap_len; dev = s->s_dev; - set_blocksize(dev, BFS_BSIZE); - s->s_blocksize = BFS_BSIZE; - s->s_blocksize_bits = BFS_BSIZE_BITS; + sb_set_blocksize(s, BFS_BSIZE); bh = sb_bread(s, 0); if(!bh) diff -u --recursive --new-file v2.5.1/linux/fs/bio.c linux/fs/bio.c --- v2.5.1/linux/fs/bio.c Tue Dec 18 14:56:39 2001 +++ linux/fs/bio.c Mon Dec 17 10:49:31 2001 @@ -142,6 +142,7 @@ bio->bi_io_vec = bvl; return bio; } + mempool_free(bio, bio_pool); return NULL; } @@ -310,28 +311,6 @@ mempool_free(b, bio_pool); return NULL; } - -#ifdef BIO_PAGEIO -static int bio_end_io_page(struct bio *bio) -{ - struct page *page = bio_page(bio); - - if (!test_bit(BIO_UPTODATE, &bio->bi_flags)) - SetPageError(page); - if (!PageError(page)) - SetPageUptodate(page); - - /* - * Run the hooks that have to be done when a page I/O has completed. - */ - if (PageTestandClearDecrAfter(page)) - atomic_dec(&nr_async_pages); - - UnlockPage(page); - bio_put(bio); - return 1; -} -#endif static int bio_end_io_kio(struct bio *bio, int nr_sectors) { diff -u --recursive --new-file v2.5.1/linux/fs/block_dev.c linux/fs/block_dev.c --- v2.5.1/linux/fs/block_dev.c Tue Dec 18 14:56:39 2001 +++ linux/fs/block_dev.c Tue Dec 25 15:39:20 2001 @@ -102,6 +102,26 @@ return 0; } +int sb_set_blocksize(struct super_block *sb, int size) +{ + int bits; + if (set_blocksize(sb->s_dev, size) < 0) + return 0; + sb->s_blocksize = size; + for (bits = 9, size >>= 9; size >>= 1; bits++) + ; + sb->s_blocksize_bits = bits; + return sb->s_blocksize; +} + +int sb_min_blocksize(struct super_block *sb, int size) +{ + int minsize = get_hardsect_size(sb->s_dev); + if (size < minsize) + size = minsize; + return sb_set_blocksize(sb, size); +} + static int blkdev_get_block(struct inode * inode, sector_t iblock, struct buffer_head * bh, int create) { if (iblock >= max_block(inode->i_rdev)) diff -u --recursive --new-file v2.5.1/linux/fs/cramfs/inode.c linux/fs/cramfs/inode.c --- v2.5.1/linux/fs/cramfs/inode.c Tue Dec 18 14:56:39 2001 +++ linux/fs/cramfs/inode.c Tue Dec 25 15:39:20 2001 @@ -195,9 +195,7 @@ unsigned long root_offset; struct super_block * retval = NULL; - set_blocksize(sb->s_dev, PAGE_CACHE_SIZE); - sb->s_blocksize = PAGE_CACHE_SIZE; - sb->s_blocksize_bits = PAGE_CACHE_SHIFT; + sb_set_blocksize(sb, PAGE_CACHE_SIZE); /* Invalidate the read buffers on mount: think disk change.. */ for (i = 0; i < READ_BUFFERS; i++) diff -u --recursive --new-file v2.5.1/linux/fs/devfs/util.c linux/fs/devfs/util.c --- v2.5.1/linux/fs/devfs/util.c Thu Oct 11 09:43:30 2001 +++ linux/fs/devfs/util.c Tue Dec 25 15:39:20 2001 @@ -61,67 +61,6 @@ /* Private functions follow */ /** - * _devfs_convert_name - Convert from an old style location-based name to new style. - * @new: The new name will be written here. - * @old: The old name. - * @disc: If true, disc partitioning information should be processed. - */ - -static void __init _devfs_convert_name (char *new, const char *old, int disc) -{ - int host, bus, target, lun; - char *ptr; - char part[8]; - - /* Decode "c#b#t#u#" */ - if (old[0] != 'c') return; - host = simple_strtol (old + 1, &ptr, 10); - if (ptr[0] != 'b') return; - bus = simple_strtol (ptr + 1, &ptr, 10); - if (ptr[0] != 't') return; - target = simple_strtol (ptr + 1, &ptr, 10); - if (ptr[0] != 'u') return; - lun = simple_strtol (ptr + 1, &ptr, 10); - if (disc) - { - /* Decode "p#" */ - if (ptr[0] == 'p') sprintf (part, "part%s", ptr + 1); - else strcpy (part, "disc"); - } - else part[0] = '\0'; - sprintf (new, "/host%d/bus%d/target%d/lun%d/%s", - host, bus, target, lun, part); -} /* End Function _devfs_convert_name */ - - -/* Public functions follow */ - -/** - * devfs_make_root - Create the root FS device entry if required. - * @name: The name of the root FS device, as passed by "root=". - */ - -void __init devfs_make_root (const char *name) -{ - char dest[64]; - - if ( (strncmp (name, "sd/", 3) == 0) || (strncmp (name, "sr/", 3) == 0) ) - { - strcpy (dest, "../scsi"); - _devfs_convert_name (dest + 7, name + 3, (name[1] == 'd') ? 1 : 0); - } - else if ( (strncmp (name, "ide/hd/", 7) == 0) || - (strncmp (name, "ide/cd/", 7) == 0) ) - { - strcpy (dest, ".."); - _devfs_convert_name (dest + 2, name + 7, (name[4] == 'h') ? 1 : 0); - } - else return; - devfs_mk_symlink (NULL, name, DEVFS_FL_DEFAULT, dest, NULL, NULL); -} /* End Function devfs_make_root */ - - -/** * devfs_register_tape - Register a tape device in the "/dev/tapes" hierarchy. * @de: Any tape device entry in the device directory. */ diff -u --recursive --new-file v2.5.1/linux/fs/efs/super.c linux/fs/efs/super.c --- v2.5.1/linux/fs/efs/super.c Tue Dec 18 14:56:39 2001 +++ linux/fs/efs/super.c Tue Dec 25 15:39:20 2001 @@ -132,16 +132,13 @@ } struct super_block *efs_read_super(struct super_block *s, void *d, int silent) { - kdev_t dev = s->s_dev; struct efs_sb_info *sb; struct buffer_head *bh; sb = SUPER_INFO(s); s->s_magic = EFS_SUPER_MAGIC; - s->s_blocksize = EFS_BLOCKSIZE; - s->s_blocksize_bits = EFS_BLOCKSIZE_BITS; - set_blocksize(dev, EFS_BLOCKSIZE); + sb_set_blocksize(s, EFS_BLOCKSIZE); /* read the vh (volume header) block */ bh = sb_bread(s, 0); diff -u --recursive --new-file v2.5.1/linux/fs/ext2/super.c linux/fs/ext2/super.c --- v2.5.1/linux/fs/ext2/super.c Tue Dec 18 14:56:39 2001 +++ linux/fs/ext2/super.c Tue Dec 25 15:39:20 2001 @@ -418,9 +418,6 @@ * This is important for devices that have a hardware * sectorsize that is larger than the default. */ - blocksize = get_hardsect_size(dev); - if(blocksize < BLOCK_SIZE ) - blocksize = BLOCK_SIZE; sb->u.ext2_sb.s_mount_opt = 0; if (!parse_options ((char *) data, &sb_block, &resuid, &resgid, @@ -428,11 +425,11 @@ return NULL; } - if (set_blocksize(dev, blocksize) < 0) { - printk ("EXT2-fs: unable to set blocksize %d\n", blocksize); + blocksize = sb_min_blocksize(sb, BLOCK_SIZE); + if (!blocksize) { + printk ("EXT2-fs: unable to set blocksize\n"); return NULL; } - sb->s_blocksize = blocksize; /* * If the superblock doesn't start on a sector boundary, @@ -485,18 +482,12 @@ bdevname(dev), i); goto failed_mount; } - sb->s_blocksize_bits = - le32_to_cpu(EXT2_SB(sb)->s_es->s_log_block_size) + 10; - sb->s_blocksize = 1 << sb->s_blocksize_bits; - - sb->s_maxbytes = ext2_max_size(sb->s_blocksize_bits); - + blocksize = BLOCK_SIZE << le32_to_cpu(EXT2_SB(sb)->s_es->s_log_block_size); /* If the blocksize doesn't match, re-read the thing.. */ if (sb->s_blocksize != blocksize) { - blocksize = sb->s_blocksize; brelse(bh); - if (set_blocksize(dev, blocksize) < 0) { + if (!sb_set_blocksize(sb, blocksize)) { printk(KERN_ERR "EXT2-fs: blocksize too small for device.\n"); return NULL; } @@ -516,6 +507,8 @@ goto failed_mount; } } + + sb->s_maxbytes = ext2_max_size(sb->s_blocksize_bits); if (le32_to_cpu(es->s_rev_level) == EXT2_GOOD_OLD_REV) { sb->u.ext2_sb.s_inode_size = EXT2_GOOD_OLD_INODE_SIZE; diff -u --recursive --new-file v2.5.1/linux/fs/ext3/super.c linux/fs/ext3/super.c --- v2.5.1/linux/fs/ext3/super.c Tue Dec 18 14:56:39 2001 +++ linux/fs/ext3/super.c Tue Dec 25 15:39:20 2001 @@ -912,10 +912,6 @@ * This is important for devices that have a hardware * sectorsize that is larger than the default. */ - blocksize = EXT3_MIN_BLOCK_SIZE; - hblock = get_hardsect_size(dev); - if (blocksize < hblock) - blocksize = hblock; sbi->s_mount_opt = 0; sbi->s_resuid = EXT3_DEF_RESUID; @@ -925,8 +921,7 @@ goto out_fail; } - sb->s_blocksize = blocksize; - set_blocksize (dev, blocksize); + blocksize = sb_min_blocksize(sb, EXT3_MIN_BLOCK_SIZE); /* * The ext3 superblock will not be buffer aligned for other than 1kB @@ -980,11 +975,10 @@ bdevname(dev), i); goto failed_mount; } - sb->s_blocksize_bits = le32_to_cpu(es->s_log_block_size) + 10; - sb->s_blocksize = 1 << sb->s_blocksize_bits; + blocksize = BLOCK_SIZE << le32_to_cpu(es->s_log_block_size); - if (sb->s_blocksize < EXT3_MIN_BLOCK_SIZE || - sb->s_blocksize > EXT3_MAX_BLOCK_SIZE) { + if (blocksize < EXT3_MIN_BLOCK_SIZE || + blocksize > EXT3_MAX_BLOCK_SIZE) { printk(KERN_ERR "EXT3-fs: Unsupported filesystem blocksize %d on %s.\n", blocksize, bdevname(dev)); @@ -993,21 +987,20 @@ sb->s_maxbytes = ext3_max_size(sb->s_blocksize_bits); + hblock = get_hardsect_size(dev); if (sb->s_blocksize != blocksize) { - blocksize = sb->s_blocksize; - /* * Make sure the blocksize for the filesystem is larger * than the hardware sectorsize for the machine. */ - if (sb->s_blocksize < hblock) { + if (blocksize < hblock) { printk(KERN_ERR "EXT3-fs: blocksize %d too small for " "device blocksize %d.\n", blocksize, hblock); goto failed_mount; } brelse (bh); - set_blocksize (dev, sb->s_blocksize); + sb_set_blocksize(sb, blocksize); logic_sb_block = (sb_block * EXT3_MIN_BLOCK_SIZE) / blocksize; offset = (sb_block * EXT3_MIN_BLOCK_SIZE) % blocksize; bh = sb_bread(sb, logic_sb_block); diff -u --recursive --new-file v2.5.1/linux/fs/fat/inode.c linux/fs/fat/inode.c --- v2.5.1/linux/fs/fat/inode.c Tue Dec 18 14:56:39 2001 +++ linux/fs/fat/inode.c Tue Dec 25 15:39:20 2001 @@ -569,10 +569,6 @@ sb->s_maxbytes = MAX_NON_LFS; sb->s_op = &fat_sops; - hard_blksize = get_hardsect_size(sb->s_dev); - if (!hard_blksize) - hard_blksize = 512; - opts.isvfat = sbi->options.isvfat; if (!parse_options((char *) data, &fat, &debug, &opts, cvf_format, cvf_options)) @@ -582,8 +578,7 @@ fat_cache_init(); - sb->s_blocksize = hard_blksize; - set_blocksize(sb->s_dev, hard_blksize); + sb_min_blocksize(sb, 512); bh = sb_bread(sb, 0); if (bh == NULL) { printk("FAT: unable to read boot sector\n"); @@ -625,13 +620,15 @@ goto out_invalid; } - if (logical_sector_size < hard_blksize) { + if (logical_sector_size < sb->s_blocksize) { printk("FAT: logical sector size too small for device" " (logical sector size = %d)\n", logical_sector_size); brelse(bh); goto out_invalid; } + hard_blksize = sb->s_blocksize; + sbi->cluster_bits = ffs(logical_sector_size * sbi->cluster_size) - 1; sbi->fats = b->fats; sbi->fat_start = CF_LE_W(b->reserved); @@ -716,9 +713,7 @@ if (error) goto out_invalid; - sb->s_blocksize = logical_sector_size; - sb->s_blocksize_bits = ffs(logical_sector_size) - 1; - set_blocksize(sb->s_dev, sb->s_blocksize); + sb_set_blocksize(sb, logical_sector_size); sbi->cvf_format = &default_cvf; if (!strcmp(cvf_format, "none")) i = -1; @@ -802,7 +797,7 @@ out_invalid: if (!silent) { printk("VFS: Can't find a valid FAT filesystem on dev %s.\n", - kdevname(sb->s_dev)); + bdevname(sb->s_dev)); } out_fail: if (opts.iocharset) { diff -u --recursive --new-file v2.5.1/linux/fs/hfs/super.c linux/fs/hfs/super.c --- v2.5.1/linux/fs/hfs/super.c Tue Dec 18 14:56:39 2001 +++ linux/fs/hfs/super.c Tue Dec 25 15:39:20 2001 @@ -401,9 +401,7 @@ } /* set the device driver to 512-byte blocks */ - set_blocksize(dev, HFS_SECTOR_SIZE); - s->s_blocksize_bits = HFS_SECTOR_SIZE_BITS; - s->s_blocksize = HFS_SECTOR_SIZE; + sb_set_blocksize(s, HFS_SECTOR_SIZE); #ifdef CONFIG_MAC_PARTITION /* check to see if we're in a partition */ diff -u --recursive --new-file v2.5.1/linux/fs/hpfs/super.c linux/fs/hpfs/super.c --- v2.5.1/linux/fs/hpfs/super.c Thu Oct 25 00:02:26 2001 +++ linux/fs/hpfs/super.c Tue Dec 25 15:39:20 2001 @@ -409,7 +409,7 @@ /*s->s_hpfs_mounting = 1;*/ dev = s->s_dev; - set_blocksize(dev, 512); + sb_set_blocksize(s, 512); s->s_hpfs_fs_size = -1; if (!(bootblock = hpfs_map_sector(s, 0, &bh0, 0))) goto bail1; if (!(superblock = hpfs_map_sector(s, 16, &bh1, 1))) goto bail2; @@ -436,8 +436,6 @@ /* Fill superblock stuff */ s->s_magic = HPFS_SUPER_MAGIC; - s->s_blocksize = 512; - s->s_blocksize_bits = 9; s->s_op = &hpfs_sops; s->s_hpfs_root = superblock->root; diff -u --recursive --new-file v2.5.1/linux/fs/isofs/inode.c linux/fs/isofs/inode.c --- v2.5.1/linux/fs/isofs/inode.c Tue Dec 18 14:56:39 2001 +++ linux/fs/isofs/inode.c Tue Dec 25 15:39:20 2001 @@ -480,7 +480,6 @@ int iso_blknum, block; int orig_zonesize; int table; - unsigned int blocksize, blocksize_bits; unsigned int vol_desc_start; unsigned long first_data_zone; struct inode * inode; @@ -508,26 +507,10 @@ * larger than the blocksize the user specified, then use * that value. */ - blocksize = get_hardsect_size(dev); - if(blocksize > opt.blocksize) { - /* - * Force the blocksize we are going to use to be the - * hardware blocksize. - */ - opt.blocksize = blocksize; - } - - blocksize_bits = 0; - { - int i = opt.blocksize; - while (i != 1){ - blocksize_bits++; - i >>=1; - } - } - - set_blocksize(dev, opt.blocksize); - s->s_blocksize = opt.blocksize; + /* + * What if bugger tells us to go beyond page size? + */ + opt.blocksize = sb_min_blocksize(s, opt.blocksize); s->u.isofs_sb.s_high_sierra = high_sierra = 0; /* default is iso9660 */ @@ -540,7 +523,7 @@ struct hs_volume_descriptor * hdp; struct iso_volume_descriptor * vdp; - block = iso_blknum << (ISOFS_BLOCK_BITS-blocksize_bits); + block = iso_blknum << (ISOFS_BLOCK_BITS - s->s_blocksize_bits); if (!(bh = sb_bread(s, block))) goto out_no_read; @@ -651,7 +634,7 @@ * blocks that were 512 bytes (which should only very rarely * happen.) */ - if(blocksize != 0 && orig_zonesize < blocksize) + if(orig_zonesize < opt.blocksize) goto out_bad_size; /* RDE: convert log zone size to bit shift */ @@ -734,15 +717,7 @@ * entries. By forcing the blocksize in this way, we ensure * that we will never be required to do this. */ - if ( orig_zonesize != opt.blocksize ) { - set_blocksize(dev, orig_zonesize); -#ifndef BEQUIET - printk(KERN_DEBUG - "ISOFS: Forcing new log zone size:%d\n", orig_zonesize); -#endif - } - s->s_blocksize = orig_zonesize; - s->s_blocksize_bits = s -> u.isofs_sb.s_log_zone_size; + sb_set_blocksize(s, orig_zonesize); s->u.isofs_sb.s_nls_iocharset = NULL; @@ -853,7 +828,7 @@ goto out_freebh; out_bad_size: printk(KERN_WARNING "Logical zone size(%d) < hardware blocksize(%u)\n", - orig_zonesize, blocksize); + orig_zonesize, opt.blocksize); goto out_freebh; #ifndef IGNORE_WRONG_MULTI_VOLUME_SPECS out_no_support: diff -u --recursive --new-file v2.5.1/linux/fs/minix/inode.c linux/fs/minix/inode.c --- v2.5.1/linux/fs/minix/inode.c Tue Dec 18 14:56:39 2001 +++ linux/fs/minix/inode.c Tue Dec 25 15:39:20 2001 @@ -126,9 +126,7 @@ struct buffer_head **map; struct minix_super_block *ms; int i, block; - kdev_t dev = s->s_dev; struct inode *root_inode; - unsigned int hblock; struct minix_sb_info *sbi = &s->u.minix_sb; /* N.B. These should be compile-time tests. @@ -138,13 +136,9 @@ if (64 != sizeof(struct minix2_inode)) panic("bad V2 i-node size"); - hblock = get_hardsect_size(dev); - if (hblock > BLOCK_SIZE) + if (!sb_set_blocksize(s, BLOCK_SIZE)) goto out_bad_hblock; - set_blocksize(dev, BLOCK_SIZE); - s->s_blocksize = BLOCK_SIZE; - s->s_blocksize_bits = BLOCK_SIZE_BITS; if (!(bh = sb_bread(s, 1))) goto out_bad_sb; @@ -264,7 +258,7 @@ out_no_fs: if (!silent) printk("VFS: Can't find a Minix or Minix V2 filesystem on device " - "%s.\n", kdevname(dev)); + "%s.\n", bdevname(s->s_dev)); out_release: brelse(bh); goto out; diff -u --recursive --new-file v2.5.1/linux/fs/namespace.c linux/fs/namespace.c --- v2.5.1/linux/fs/namespace.c Tue Dec 18 14:56:39 2001 +++ linux/fs/namespace.c Tue Dec 25 15:39:20 2001 @@ -16,10 +16,11 @@ #include <linux/acct.h> #include <linux/module.h> #include <linux/seq_file.h> +#include <linux/namespace.h> #include <asm/uaccess.h> -struct vfsmount *do_kern_mount(char *type, int flags, char *name, void *data); +struct vfsmount *do_kern_mount(const char *type, int flags, char *name, void *data); int do_remount_sb(struct super_block *sb, int flags, void * data); void kill_super(struct super_block *sb); @@ -27,10 +28,6 @@ static int hash_mask, hash_bits; static kmem_cache_t *mnt_cache; -static LIST_HEAD(vfsmntlist); -static DECLARE_MUTEX(mount_sem); -static struct vfsmount *root_vfsmnt; - static inline unsigned long hash(struct vfsmount *mnt, struct dentry *dentry) { unsigned long tmp = ((unsigned long) mnt / L1_CACHE_BYTES); @@ -96,7 +93,7 @@ while (mnt->mnt_parent != mnt) mnt = mnt->mnt_parent; spin_unlock(&dcache_lock); - return mnt == root_vfsmnt; + return mnt == current->namespace->root; } static void detach_mnt(struct vfsmount *mnt, struct nameidata *old_nd) @@ -164,26 +161,29 @@ /* iterator */ static void *m_start(struct seq_file *m, loff_t *pos) { + struct namespace *n = m->private; struct list_head *p; - loff_t n = *pos; + loff_t l = *pos; - down(&mount_sem); - list_for_each(p, &vfsmntlist) - if (!n--) + down_read(&n->sem); + list_for_each(p, &n->list) + if (!l--) return list_entry(p, struct vfsmount, mnt_list); return NULL; } static void *m_next(struct seq_file *m, void *v, loff_t *pos) { + struct namespace *n = m->private; struct list_head *p = ((struct vfsmount *)v)->mnt_list.next; (*pos)++; - return p==&vfsmntlist ? NULL : list_entry(p, struct vfsmount, mnt_list); + return p==&n->list ? NULL : list_entry(p, struct vfsmount, mnt_list); } static void m_stop(struct seq_file *m, void *v) { - up(&mount_sem); + struct namespace *n = m->private; + up_read(&n->sem); } static inline void mangle(struct seq_file *m, const char *s) @@ -329,7 +329,7 @@ return retval; } - down(&mount_sem); + down_write(¤t->namespace->sem); spin_lock(&dcache_lock); if (atomic_read(&sb->s_active) == 1) { @@ -348,7 +348,7 @@ retval = 0; } spin_unlock(&dcache_lock); - up(&mount_sem); + up_write(¤t->namespace->sem); return retval; } @@ -478,7 +478,7 @@ struct list_head head; attach_mnt(mnt, nd); list_add_tail(&head, &mnt->mnt_list); - list_splice(&head, vfsmntlist.prev); + list_splice(&head, current->namespace->list.prev); mntget(mnt); err = 0; } @@ -505,7 +505,7 @@ if (err) return err; - down(&mount_sem); + down_write(¤t->namespace->sem); err = -EINVAL; if (check_mnt(nd->mnt) && (!recurse || check_mnt(old_nd.mnt))) { err = -ENOMEM; @@ -525,7 +525,7 @@ mntput(mnt); } - up(&mount_sem); + up_write(¤t->namespace->sem); path_release(&old_nd); return err; } @@ -572,7 +572,7 @@ if (err) return err; - down(&mount_sem); + down_write(¤t->namespace->sem); while(d_mountpoint(nd->dentry) && follow_down(&nd->mnt, &nd->dentry)) ; err = -EINVAL; @@ -612,7 +612,7 @@ out1: up(&nd->dentry->d_inode->i_zombie); out: - up(&mount_sem); + up_write(¤t->namespace->sem); if (!err) path_release(&parent_nd); path_release(&old_nd); @@ -622,13 +622,22 @@ static int do_add_mount(struct nameidata *nd, char *type, int flags, int mnt_flags, char *name, void *data) { - struct vfsmount *mnt = do_kern_mount(type, flags, name, data); - int err = PTR_ERR(mnt); + struct vfsmount *mnt; + int err; + + if (!type || !memchr(type, 0, PAGE_SIZE)) + return -EINVAL; + /* we need capabilities... */ + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + mnt = do_kern_mount(type, flags, name, data); + err = PTR_ERR(mnt); if (IS_ERR(mnt)) goto out; - down(&mount_sem); + down_write(¤t->namespace->sem); /* Something was mounted here while we slept */ while(d_mountpoint(nd->dentry) && follow_down(&nd->mnt, &nd->dentry)) ; @@ -644,7 +653,7 @@ mnt->mnt_flags = mnt_flags; err = graft_tree(mnt, nd); unlock: - up(&mount_sem); + up_write(¤t->namespace->sem); mntput(mnt); out: return err; @@ -744,6 +753,86 @@ return retval; } +int copy_namespace(int flags, struct task_struct *tsk) +{ + struct namespace *namespace = tsk->namespace; + struct namespace *new_ns; + struct vfsmount *rootmnt = NULL, *pwdmnt = NULL, *altrootmnt = NULL; + struct fs_struct *fs = tsk->fs; + + if (!namespace) + return 0; + + get_namespace(namespace); + + if (! (flags & CLONE_NEWNS)) + return 0; + + if (!capable(CAP_SYS_ADMIN)) { + put_namespace(namespace); + return -EPERM; + } + + new_ns = kmalloc(sizeof(struct namespace *), GFP_KERNEL); + if (!new_ns) + goto out; + + atomic_set(&new_ns->count, 1); + init_rwsem(&new_ns->sem); + new_ns->root = NULL; + INIT_LIST_HEAD(&new_ns->list); + + down_write(&tsk->namespace->sem); + /* First pass: copy the tree topology */ + new_ns->root = copy_tree(namespace->root, namespace->root->mnt_root); + spin_lock(&dcache_lock); + list_add_tail(&new_ns->list, &new_ns->root->mnt_list); + spin_unlock(&dcache_lock); + + /* Second pass: switch the tsk->fs->* elements */ + if (fs) { + struct vfsmount *p, *q; + write_lock(&fs->lock); + + p = namespace->root; + q = new_ns->root; + while (p) { + if (p == fs->rootmnt) { + rootmnt = p; + fs->rootmnt = mntget(q); + } + if (p == fs->pwdmnt) { + pwdmnt = p; + fs->pwdmnt = mntget(q); + } + if (p == fs->altrootmnt) { + altrootmnt = p; + fs->altrootmnt = mntget(q); + } + p = next_mnt(p, namespace->root); + q = next_mnt(q, new_ns->root); + } + write_unlock(&fs->lock); + } + up_write(&tsk->namespace->sem); + + tsk->namespace = new_ns; + + if (rootmnt) + mntput(rootmnt); + if (pwdmnt) + mntput(pwdmnt); + if (altrootmnt) + mntput(altrootmnt); + + put_namespace(namespace); + return 0; + +out: + put_namespace(namespace); + return -ENOMEM; +} + asmlinkage long sys_mount(char * dev_name, char * dir_name, char * type, unsigned long flags, void * data) { @@ -862,7 +951,7 @@ user_nd.mnt = mntget(current->fs->rootmnt); user_nd.dentry = dget(current->fs->root); read_unlock(¤t->fs->lock); - down(&mount_sem); + down_write(¤t->namespace->sem); down(&old_nd.dentry->d_inode->i_zombie); error = -EINVAL; if (!check_mnt(user_nd.mnt)) @@ -907,7 +996,7 @@ path_release(&parent_nd); out2: up(&old_nd.dentry->d_inode->i_zombie); - up(&mount_sem); + up_write(¤t->namespace->sem); path_release(&user_nd); path_release(&old_nd); out1: @@ -920,95 +1009,34 @@ goto out2; } -/* - * Absolutely minimal fake fs - only empty root directory and nothing else. - * In 2.5 we'll use ramfs or tmpfs, but for now it's all we need - just - * something to go with root vfsmount. - */ -static struct inode_operations rootfs_dir_inode_operations; -static struct file_operations rootfs_dir_operations; -static int rootfs_mkdir(struct inode *dir, struct dentry *dentry, int mode) -{ - struct inode * inode = new_inode(dir->i_sb); - int error = -ENOSPC; - if (inode) { - inode->i_mode = S_IFDIR|mode; - inode->i_uid = current->fsuid; - inode->i_gid = current->fsgid; - inode->i_op = &rootfs_dir_inode_operations; - inode->i_fop = &rootfs_dir_operations; - d_instantiate(dentry, inode); - dget(dentry); - error = 0; - } - return error; -} -static int rootfs_mknod(struct inode *dir, struct dentry *dentry, int mode, int dev) -{ - struct inode * inode = new_inode(dir->i_sb); - int error = -ENOSPC; - if (inode) { - inode->i_uid = current->fsuid; - inode->i_gid = current->fsgid; - init_special_inode(inode, mode, dev); - d_instantiate(dentry, inode); - dget(dentry); - error = 0; - } - return error; -} -static int rootfs_unlink(struct inode * dir, struct dentry *dentry) -{ - dentry->d_inode->i_nlink--; - dput(dentry); - return 0; -} -static struct dentry *rootfs_lookup(struct inode *dir, struct dentry *dentry) -{ - d_add(dentry, NULL); - return NULL; -} -static struct file_operations rootfs_dir_operations = { - read: generic_read_dir, - readdir: dcache_readdir, -}; -static struct inode_operations rootfs_dir_inode_operations = { - lookup: rootfs_lookup, - mkdir: rootfs_mkdir, - mknod: rootfs_mknod, - unlink: rootfs_unlink, -}; -static struct super_block *rootfs_read_super(struct super_block * sb, void * data, int silent) +static void __init init_mount_tree(void) { - struct inode * inode; - struct dentry * root; - static struct super_operations s_ops = {}; - sb->s_op = &s_ops; - inode = new_inode(sb); - if (!inode) - return NULL; - inode->i_mode = S_IFDIR|0555; - inode->i_uid = inode->i_gid = 0; - inode->i_op = &rootfs_dir_inode_operations; - inode->i_fop = &rootfs_dir_operations; - root = d_alloc_root(inode); - if (!root) { - iput(inode); - return NULL; + struct vfsmount *mnt; + struct namespace *namespace; + struct task_struct *p; + + mnt = do_kern_mount("rootfs", 0, "rootfs", NULL); + if (IS_ERR(mnt)) + panic("Can't create rootfs"); + namespace = kmalloc(sizeof(*namespace), GFP_KERNEL); + if (!namespace) + panic("Can't allocate initial namespace"); + atomic_set(&namespace->count, 1); + INIT_LIST_HEAD(&namespace->list); + init_rwsem(&namespace->sem); + list_add(&mnt->mnt_list, &namespace->list); + namespace->root = mnt; + + init_task.namespace = namespace; + read_lock(&tasklist_lock); + for_each_task(p) { + get_namespace(namespace); + p->namespace = namespace; } - sb->s_root = root; - return sb; -} -static DECLARE_FSTYPE(root_fs_type, "rootfs", rootfs_read_super, FS_NOMOUNT); + read_unlock(&tasklist_lock); -static void __init init_mount_tree(void) -{ - register_filesystem(&root_fs_type); - root_vfsmnt = do_kern_mount("rootfs", 0, "rootfs", NULL); - if (IS_ERR(root_vfsmnt)) - panic("can't allocate root vfsmount"); - set_fs_pwd(current->fs, root_vfsmnt, root_vfsmnt->mnt_root); - set_fs_root(current->fs, root_vfsmnt, root_vfsmnt->mnt_root); + set_fs_pwd(current->fs, namespace->root, namespace->root->mnt_root); + set_fs_root(current->fs, namespace->root, namespace->root->mnt_root); } void __init mnt_init(unsigned long mempages) @@ -1066,5 +1094,6 @@ d++; i--; } while (i); + init_rootfs(); init_mount_tree(); } diff -u --recursive --new-file v2.5.1/linux/fs/nfs/file.c linux/fs/nfs/file.c --- v2.5.1/linux/fs/nfs/file.c Sun Sep 23 09:48:01 2001 +++ linux/fs/nfs/file.c Thu Dec 20 08:57:31 2001 @@ -161,15 +161,10 @@ static int nfs_commit_write(struct file *file, struct page *page, unsigned offset, unsigned to) { long status; - loff_t pos = ((loff_t)page->index<<PAGE_CACHE_SHIFT) + to; - struct inode *inode = page->mapping->host; lock_kernel(); status = nfs_updatepage(file, page, offset, to-offset); unlock_kernel(); - /* most likely it's already done. CHECKME */ - if (pos > inode->i_size) - inode->i_size = pos; return status; } diff -u --recursive --new-file v2.5.1/linux/fs/nfs/inode.c linux/fs/nfs/inode.c --- v2.5.1/linux/fs/nfs/inode.c Tue Dec 18 14:56:40 2001 +++ linux/fs/nfs/inode.c Thu Dec 20 08:57:31 2001 @@ -107,17 +107,10 @@ inode->i_rdev = 0; /* We can't support UPDATE_ATIME(), since the server will reset it */ inode->i_flags |= S_NOATIME; - NFS_FILEID(inode) = 0; - NFS_FSID(inode) = 0; - NFS_FLAGS(inode) = 0; INIT_LIST_HEAD(&inode->u.nfs_i.read); INIT_LIST_HEAD(&inode->u.nfs_i.dirty); INIT_LIST_HEAD(&inode->u.nfs_i.commit); INIT_LIST_HEAD(&inode->u.nfs_i.writeback); - inode->u.nfs_i.nread = 0; - inode->u.nfs_i.ndirty = 0; - inode->u.nfs_i.ncommit = 0; - inode->u.nfs_i.npages = 0; NFS_CACHEINV(inode); NFS_ATTRTIMEO(inode) = NFS_MINATTRTIMEO(inode); NFS_ATTRTIMEO_UPDATE(inode) = jiffies; @@ -655,19 +648,6 @@ inode->i_op = &nfs_symlink_inode_operations; else init_special_inode(inode, inode->i_mode, fattr->rdev); - /* - * Preset the size and mtime, as there's no need - * to invalidate the caches. - */ - inode->i_size = nfs_size_to_loff_t(fattr->size); - inode->i_mtime = nfs_time_to_secs(fattr->mtime); - inode->i_atime = nfs_time_to_secs(fattr->atime); - inode->i_ctime = nfs_time_to_secs(fattr->ctime); - NFS_CACHE_CTIME(inode) = fattr->ctime; - NFS_CACHE_MTIME(inode) = fattr->mtime; - NFS_CACHE_ISIZE(inode) = fattr->size; - NFS_ATTRTIMEO(inode) = NFS_MINATTRTIMEO(inode); - NFS_ATTRTIMEO_UPDATE(inode) = jiffies; memcpy(&inode->u.nfs_i.fh, fh, sizeof(inode->u.nfs_i.fh)); } nfs_refresh_inode(inode, fattr); @@ -697,6 +677,9 @@ return 0; if (memcmp(&inode->u.nfs_i.fh, fh, sizeof(inode->u.nfs_i.fh)) != 0) return 0; + /* Force an attribute cache update if inode->i_count == 0 */ + if (!atomic_read(&inode->i_count)) + NFS_CACHEINV(inode); return 1; } @@ -797,7 +780,9 @@ if (!S_ISREG(inode->i_mode)) attr->ia_valid &= ~ATTR_SIZE; + filemap_fdatasync(inode->i_mapping); error = nfs_wb_all(inode); + filemap_fdatawait(inode->i_mapping); if (error) goto out; @@ -825,6 +810,8 @@ fattr.pre_ctime = NFS_CACHE_CTIME(inode); fattr.valid |= NFS_ATTR_WCC; } + /* Force an attribute cache update */ + NFS_CACHEINV(inode); error = nfs_refresh_inode(inode, &fattr); out: return error; @@ -966,6 +953,34 @@ } /* + * nfs_fattr_obsolete - Test if attribute data is newer than cached data + * @inode: inode + * @fattr: attributes to test + * + * Avoid stuffing the attribute cache with obsolete information. + * We always accept updates if the attribute cache timed out, or if + * fattr->ctime is newer than our cached value. + * If fattr->ctime matches the cached value, we still accept the update + * if it increases the file size. + */ +static inline +int nfs_fattr_obsolete(struct inode *inode, struct nfs_fattr *fattr) +{ + s64 cdif; + + if (time_after(jiffies, NFS_READTIME(inode)+NFS_ATTRTIMEO(inode))) + goto out_valid; + if ((cdif = (s64)fattr->ctime - (s64)NFS_CACHE_CTIME(inode)) > 0) + goto out_valid; + /* Ugh... */ + if (cdif == 0 && fattr->size > NFS_CACHE_ISIZE(inode)) + goto out_valid; + return -1; + out_valid: + return 0; +} + +/* * Many nfs protocol calls return the new file attributes after * an operation. Here we update the inode to reflect the state * of the server's inode. @@ -982,6 +997,7 @@ { __u64 new_size, new_mtime; loff_t new_isize; + time_t new_atime; int invalid = 0; dfprintk(VFS, "NFS: refresh_inode(%x/%ld ct=%d info=0x%x)\n", @@ -1007,6 +1023,11 @@ new_size = fattr->size; new_isize = nfs_size_to_loff_t(fattr->size); + new_atime = nfs_time_to_secs(fattr->atime); + /* Avoid races */ + if (nfs_fattr_obsolete(inode, fattr)) + goto out_nochange; + /* * Update the read time so we don't revalidate too often. */ @@ -1056,7 +1077,7 @@ NFS_CACHE_CTIME(inode) = fattr->ctime; inode->i_ctime = nfs_time_to_secs(fattr->ctime); - inode->i_atime = nfs_time_to_secs(fattr->atime); + inode->i_atime = new_atime; NFS_CACHE_MTIME(inode) = new_mtime; inode->i_mtime = nfs_time_to_secs(new_mtime); @@ -1093,7 +1114,10 @@ if (invalid) nfs_zap_caches(inode); return 0; - + out_nochange: + if (new_atime - inode->i_atime > 0) + inode->i_atime = new_atime; + return 0; out_changed: /* * Big trouble! The inode has become a different object. diff -u --recursive --new-file v2.5.1/linux/fs/nfs/nfs2xdr.c linux/fs/nfs/nfs2xdr.c --- v2.5.1/linux/fs/nfs/nfs2xdr.c Fri Nov 2 17:40:09 2001 +++ linux/fs/nfs/nfs2xdr.c Thu Dec 20 08:57:31 2001 @@ -270,14 +270,12 @@ count = ntohl(*p++); hdrlen = (u8 *) p - (u8 *) iov->iov_base; - recvd = req->rq_rlen - hdrlen; - if (p != iov[req->rq_rnr-1].iov_base) { - /* Unexpected reply header size. Punt. - * XXX: Move iovec contents to align data on page - * boundary and adjust RPC header size guess */ - printk(KERN_WARNING "NFS: Odd RPC header size in read reply: %d\n", hdrlen); - return -errno_NFSERR_IO; + if (iov->iov_len > hdrlen) { + dprintk("NFS: READ header is short. iovec will be shifted.\n"); + xdr_shift_iovec(iov, req->rq_rnr, iov->iov_len - hdrlen); } + + recvd = req->rq_rlen - hdrlen; if (count > recvd) { printk(KERN_WARNING "NFS: server cheating in read reply: " "count %d > recvd %d\n", count, recvd); @@ -448,27 +446,23 @@ nfs_xdr_readdirres(struct rpc_rqst *req, u32 *p, struct nfs_readdirres *res) { struct iovec *iov = req->rq_rvec; + int hdrlen; int status, nr; u32 *end, *entry, len; if ((status = ntohl(*p++))) return -nfs_stat_to_errno(status); - if ((void *) p != ((u8 *) iov->iov_base+iov->iov_len)) { - /* Unexpected reply header size. Punt. */ - printk(KERN_WARNING "NFS: Odd RPC header size in readdirres reply\n"); - return -errno_NFSERR_IO; + + hdrlen = (u8 *) p - (u8 *) iov->iov_base; + if (iov->iov_len > hdrlen) { + dprintk("NFS: READDIR header is short. iovec will be shifted.\n"); + xdr_shift_iovec(iov, req->rq_rnr, iov->iov_len - hdrlen); } + /* Get start and end address of XDR data */ p = (u32 *) iov[1].iov_base; end = (u32 *) ((u8 *) p + iov[1].iov_len); - - /* Get start and end of dirent buffer */ - if (res->buffer != p) { - printk(KERN_ERR "NFS: Bad result buffer in readdir\n"); - return -errno_NFSERR_IO; - } - for (nr = 0; *p++; nr++) { entry = p - 1; if (p + 2 > end) @@ -598,13 +592,21 @@ static int nfs_xdr_readlinkres(struct rpc_rqst *req, u32 *p, struct nfs_readlinkres *res) { + struct iovec *iov = req->rq_rvec; u32 *strlen; char *string; + int hdrlen; int status; unsigned int len; if ((status = ntohl(*p++))) return -nfs_stat_to_errno(status); + hdrlen = (u8 *) p - (u8 *) iov->iov_base; + if (iov->iov_len > hdrlen) { + dprintk("NFS: READLINK header is short. iovec will be shifted.\n"); + xdr_shift_iovec(iov, req->rq_rnr, iov->iov_len - hdrlen); + } + strlen = (u32*)res->buffer; /* Convert length of symlink */ len = ntohl(*strlen); diff -u --recursive --new-file v2.5.1/linux/fs/nfs/read.c linux/fs/nfs/read.c --- v2.5.1/linux/fs/nfs/read.c Fri Nov 9 14:28:15 2001 +++ linux/fs/nfs/read.c Thu Dec 20 08:57:31 2001 @@ -397,7 +397,7 @@ { struct nfs_read_data *data = (struct nfs_read_data *) task->tk_calldata; struct inode *inode = data->inode; - int count = data->res.count; + unsigned int count = data->res.count; dprintk("NFS: %4d nfs_readpage_result, (status %d)\n", task->tk_pid, task->tk_status); @@ -408,9 +408,15 @@ struct page *page = req->wb_page; nfs_list_remove_request(req); - if (task->tk_status >= 0 && count >= 0) { + if (task->tk_status >= 0) { + if (count < PAGE_CACHE_SIZE) { + char *p = kmap(page); + memset(p + count, 0, PAGE_CACHE_SIZE - count); + kunmap(page); + count = 0; + } else + count -= PAGE_CACHE_SIZE; SetPageUptodate(page); - count -= PAGE_CACHE_SIZE; } else SetPageError(page); flush_dcache_page(page); diff -u --recursive --new-file v2.5.1/linux/fs/nfs/write.c linux/fs/nfs/write.c --- v2.5.1/linux/fs/nfs/write.c Tue Nov 20 14:18:50 2001 +++ linux/fs/nfs/write.c Thu Dec 20 08:57:31 2001 @@ -213,6 +213,7 @@ unsigned int offset, unsigned int count) { struct nfs_page *req; + loff_t end; int status; req = nfs_update_request(file, inode, page, offset, count); @@ -223,6 +224,10 @@ req->wb_cred = get_rpccred(NFS_I(inode)->mm_cred); nfs_unlock_request(req); nfs_strategy(inode); + end = ((loff_t)page->index<<PAGE_CACHE_SHIFT) + (loff_t)(offset + count); + if (inode->i_size < end) + inode->i_size = end; + out: return status; } @@ -781,6 +786,7 @@ struct dentry *dentry = file->f_dentry; struct inode *inode = dentry->d_inode; struct nfs_page *req; + loff_t end; int status = 0; dprintk("NFS: nfs_updatepage(%s/%s %d@%Ld)\n", @@ -812,6 +818,10 @@ goto done; status = 0; + end = ((loff_t)page->index<<PAGE_CACHE_SHIFT) + (loff_t)(offset + count); + if (inode->i_size < end) + inode->i_size = end; + /* If we wrote past the end of the page. * Call the strategy routine so it can send out a bunch * of requests. diff -u --recursive --new-file v2.5.1/linux/fs/ntfs/fs.c linux/fs/ntfs/fs.c --- v2.5.1/linux/fs/ntfs/fs.c Tue Dec 18 14:56:40 2001 +++ linux/fs/ntfs/fs.c Tue Dec 25 15:39:20 2001 @@ -1016,14 +1016,11 @@ init_ntfs_super_block(vol); if (!parse_options(vol, (char*)options)) goto ntfs_read_super_vol; - blocksize = get_hardsect_size(sb->s_dev); - if (blocksize < 512) - blocksize = 512; - if (set_blocksize(sb->s_dev, blocksize) < 0) { - ntfs_error("Unable to set blocksize %d.\n", blocksize); + blocksize = sb_min_blocksize(sb, 512); + if (!blocksize) { + ntfs_error("Unable to set blocksize.\n"); goto ntfs_read_super_vol; } - sb->s_blocksize = blocksize; /* Read the super block (boot block). */ if (!(bh = sb_bread(sb, 0))) { ntfs_error("Reading super block failed\n"); @@ -1052,10 +1049,7 @@ } ntfs_debug(DEBUG_OTHER, "Done to init volume\n"); /* Inform the kernel that a device block is a NTFS cluster. */ - sb->s_blocksize = vol->cluster_size; - sb->s_blocksize_bits = vol->cluster_size_bits; - if (blocksize != vol->cluster_size && - set_blocksize(sb->s_dev, sb->s_blocksize) < 0) { + if (!sb_set_blocksize(sb, vol->cluster_size)) { ntfs_error("Cluster size too small for device.\n"); goto ntfs_read_super_unl; } diff -u --recursive --new-file v2.5.1/linux/fs/proc/base.c linux/fs/proc/base.c --- v2.5.1/linux/fs/proc/base.c Wed Oct 10 23:42:47 2001 +++ linux/fs/proc/base.c Tue Dec 25 15:39:20 2001 @@ -23,6 +23,8 @@ #include <linux/init.h> #include <linux/file.h> #include <linux/string.h> +#include <linux/seq_file.h> +#include <linux/namespace.h> /* * For hysterical raisins we keep the same inumbers as in the old procfs. @@ -246,6 +248,46 @@ read: pid_maps_read, }; +extern struct seq_operations mounts_op; +static int mounts_open(struct inode *inode, struct file *file) +{ + struct task_struct *task = inode->u.proc_i.task; + int ret = seq_open(file, &mounts_op); + + if (!ret) { + struct seq_file *m = file->private_data; + struct namespace *namespace; + task_lock(task); + namespace = task->namespace; + if (namespace) + get_namespace(namespace); + task_unlock(task); + + if (namespace) + m->private = namespace; + else { + seq_release(inode, file); + ret = -EINVAL; + } + } + return ret; +} + +static int mounts_release(struct inode *inode, struct file *file) +{ + struct seq_file *m = file->private_data; + struct namespace *namespace = m->private; + put_namespace(namespace); + return seq_release(inode, file); +} + +static struct file_operations proc_mounts_operations = { + open: mounts_open, + read: seq_read, + llseek: seq_lseek, + release: mounts_release, +}; + #define PROC_BLOCK_SIZE (3*1024) /* 4K page size but our output routines use some slack for overruns */ static ssize_t proc_info_read(struct file * file, char * buf, @@ -497,6 +539,7 @@ PROC_PID_STATM, PROC_PID_MAPS, PROC_PID_CPU, + PROC_PID_MOUNTS, PROC_PID_FD_DIR = 0x8000, /* 0x8000-0xffff */ }; @@ -516,6 +559,7 @@ E(PROC_PID_CWD, "cwd", S_IFLNK|S_IRWXUGO), E(PROC_PID_ROOT, "root", S_IFLNK|S_IRWXUGO), E(PROC_PID_EXE, "exe", S_IFLNK|S_IRWXUGO), + E(PROC_PID_MOUNTS, "mounts", S_IFREG|S_IRUGO), {0,0,NULL,0} }; #undef E @@ -874,6 +918,9 @@ case PROC_PID_MEM: inode->i_op = &proc_mem_inode_operations; inode->i_fop = &proc_mem_operations; + break; + case PROC_PID_MOUNTS: + inode->i_fop = &proc_mounts_operations; break; default: printk("procfs: impossible type (%d)",p->type); diff -u --recursive --new-file v2.5.1/linux/fs/proc/proc_misc.c linux/fs/proc/proc_misc.c --- v2.5.1/linux/fs/proc/proc_misc.c Tue Dec 18 14:56:40 2001 +++ linux/fs/proc/proc_misc.c Tue Dec 25 15:39:20 2001 @@ -500,18 +500,6 @@ write: write_profile, }; -extern struct seq_operations mounts_op; -static int mounts_open(struct inode *inode, struct file *file) -{ - return seq_open(file, &mounts_op); -} -static struct file_operations proc_mounts_operations = { - open: mounts_open, - read: seq_read, - llseek: seq_lseek, - release: seq_release, -}; - struct proc_dir_entry *proc_root_kcore; static void create_seq_entry(char *name, mode_t mode, struct file_operations *f) @@ -555,11 +543,12 @@ for (p = simple_ones; p->name; p++) create_proc_read_entry(p->name, 0, NULL, p->read_proc, NULL); + proc_symlink("mounts", NULL, "self/mounts"); + /* And now for trickier ones */ entry = create_proc_entry("kmsg", S_IRUSR, &proc_root); if (entry) entry->proc_fops = &proc_kmsg_operations; - create_seq_entry("mounts", 0, &proc_mounts_operations); create_seq_entry("cpuinfo", 0, &proc_cpuinfo_operations); create_seq_entry("interrupts", 0, &proc_interrupts_operations); #ifdef CONFIG_MODULES diff -u --recursive --new-file v2.5.1/linux/fs/qnx4/inode.c linux/fs/qnx4/inode.c --- v2.5.1/linux/fs/qnx4/inode.c Tue Dec 18 14:56:40 2001 +++ linux/fs/qnx4/inode.c Tue Dec 25 15:39:20 2001 @@ -339,13 +339,10 @@ void *data, int silent) { struct buffer_head *bh; - kdev_t dev = s->s_dev; struct inode *root; const char *errmsg; - set_blocksize(dev, QNX4_BLOCK_SIZE); - s->s_blocksize = QNX4_BLOCK_SIZE; - s->s_blocksize_bits = QNX4_BLOCK_SIZE_BITS; + sb_set_blocksize(s, QNX4_BLOCK_SIZE); /* Check the boot signature. Since the qnx4 code is dangerous, we should leave as quickly as possible diff -u --recursive --new-file v2.5.1/linux/fs/ramfs/inode.c linux/fs/ramfs/inode.c --- v2.5.1/linux/fs/ramfs/inode.c Wed Oct 24 15:19:18 2001 +++ linux/fs/ramfs/inode.c Tue Dec 25 15:39:20 2001 @@ -325,6 +325,7 @@ } static DECLARE_FSTYPE(ramfs_fs_type, "ramfs", ramfs_read_super, FS_LITTER); +static DECLARE_FSTYPE(rootfs_fs_type, "rootfs", ramfs_read_super, FS_NOMOUNT|FS_LITTER); static int __init init_ramfs_fs(void) { @@ -338,5 +339,11 @@ module_init(init_ramfs_fs) module_exit(exit_ramfs_fs) + +int __init init_rootfs(void) +{ + return register_filesystem(&rootfs_fs_type); +} + MODULE_LICENSE("GPL"); diff -u --recursive --new-file v2.5.1/linux/fs/romfs/inode.c linux/fs/romfs/inode.c --- v2.5.1/linux/fs/romfs/inode.c Tue Dec 18 14:56:40 2001 +++ linux/fs/romfs/inode.c Tue Dec 25 15:39:20 2001 @@ -102,9 +102,7 @@ /* I would parse the options here, but there are none.. :) */ - set_blocksize(dev, ROMBSIZE); - s->s_blocksize = ROMBSIZE; - s->s_blocksize_bits = ROMBSBITS; + sb_set_blocksize(s, ROMBSIZE); s->u.generic_sbp = (void *) 0; s->s_maxbytes = 0xFFFFFFFF; diff -u --recursive --new-file v2.5.1/linux/fs/super.c linux/fs/super.c --- v2.5.1/linux/fs/super.c Tue Dec 18 14:56:40 2001 +++ linux/fs/super.c Mon Dec 17 19:24:59 2001 @@ -34,8 +34,6 @@ #define __NO_VERSION__ #include <linux/module.h> -int do_remount_sb(struct super_block *sb, int flags, void * data); - LIST_HEAD(super_blocks); spinlock_t sb_lock = SPIN_LOCK_UNLOCKED; @@ -377,7 +375,7 @@ get_filesystem(type); } -void put_unnamed_dev(kdev_t dev); /* should become static */ +static void put_anon_dev(kdev_t dev); /** * remove_super - makes superblock unreachable @@ -407,7 +405,7 @@ if (bdev) blkdev_put(bdev, BDEV_FS); else - put_unnamed_dev(dev); + put_anon_dev(dev); } struct vfsmount *alloc_vfsmnt(void); @@ -530,36 +528,122 @@ return err; } +/** + * do_remount_sb - asks filesystem to change mount options. + * @sb: superblock in question + * @flags: numeric part of options + * @data: the rest of options + * + * Alters the mount options of a mounted file system. + */ +int do_remount_sb(struct super_block *sb, int flags, void *data) +{ + int retval; + + if (!(flags & MS_RDONLY) && sb->s_dev && is_read_only(sb->s_dev)) + return -EACCES; + /*flags |= MS_RDONLY;*/ + if (flags & MS_RDONLY) + acct_auto_close(sb->s_dev); + shrink_dcache_sb(sb); + fsync_super(sb); + /* If we are remounting RDONLY, make sure there are no rw files open */ + if ((flags & MS_RDONLY) && !(sb->s_flags & MS_RDONLY)) + if (!fs_may_remount_ro(sb)) + return -EBUSY; + if (sb->s_op && sb->s_op->remount_fs) { + lock_super(sb); + retval = sb->s_op->remount_fs(sb, &flags, data); + unlock_super(sb); + if (retval) + return retval; + } + sb->s_flags = (sb->s_flags & ~MS_RMT_MASK) | (flags & MS_RMT_MASK); + return 0; +} + /* * Unnamed block devices are dummy devices used by virtual * filesystems which don't use real block-devices. -- jrs */ -static unsigned long unnamed_dev_in_use[256/(8*sizeof(unsigned long))]; +enum {Max_anon = 256}; +static unsigned long unnamed_dev_in_use[Max_anon/(8*sizeof(unsigned long))]; +static spinlock_t unnamed_dev_lock = SPIN_LOCK_UNLOCKED;/* protects the above */ -kdev_t get_unnamed_dev(void) +/** + * put_anon_dev - release anonymous device number. + * @dev: device in question + */ +static void put_anon_dev(kdev_t dev) { - int i; - - for (i = 1; i < 256; i++) { - if (!test_and_set_bit(i,unnamed_dev_in_use)) - return MKDEV(UNNAMED_MAJOR, i); - } - return 0; + spin_lock(&unnamed_dev_lock); + clear_bit(MINOR(dev), unnamed_dev_in_use); + spin_unlock(&unnamed_dev_lock); } -void put_unnamed_dev(kdev_t dev) +/** + * get_anon_super - allocate a superblock for non-device fs + * @type: filesystem type + * @compare: check if existing superblock is what we want + * @data: argument for @compare. + * + * get_anon_super is a helper for non-blockdevice filesystems. + * It either finds and returns one of the superblocks of given type + * (if it can find one that would satisfy caller) or creates a new + * one. In the either case we return an active reference to superblock + * with ->s_umount locked. If superblock is new it gets a new + * anonymous device allocated for it and is inserted into lists - + * other initialization is left to caller. + * + * Rather than duplicating all that logics every time when + * we want something that doesn't fit "nodev" and "single" we pull + * the relevant code into common helper and let get_sb_...() call + * it. + * + * NB: get_sb_...() is going to become an fs type method, with + * current ->read_super() becoming a callback used by common instances. + */ +struct super_block *get_anon_super(struct file_system_type *type, + int (*compare)(struct super_block *,void *), void *data) { - if (!dev || MAJOR(dev) != UNNAMED_MAJOR) - return; - if (test_and_clear_bit(MINOR(dev), unnamed_dev_in_use)) - return; - printk("VFS: put_unnamed_dev: freeing unused device %s\n", - kdevname(dev)); + struct super_block *s = alloc_super(); + kdev_t dev; + struct list_head *p; + + if (!s) + return ERR_PTR(-ENOMEM); + + spin_lock(&unnamed_dev_lock); + dev = find_first_zero_bit(unnamed_dev_in_use, Max_anon); + if (dev == Max_anon) { + spin_unlock(&unnamed_dev_lock); + destroy_super(s); + return ERR_PTR(-EMFILE); + } + set_bit(dev, unnamed_dev_in_use); + spin_unlock(&unnamed_dev_lock); + +retry: + spin_lock(&sb_lock); + if (compare) list_for_each(p, &type->fs_supers) { + struct super_block *old; + old = list_entry(p, struct super_block, s_instances); + if (!compare(old, data)) + continue; + if (!grab_super(old)) + goto retry; + destroy_super(s); + return old; + } + + s->s_dev = dev; + insert_super(s, type); + return s; } static struct super_block *get_sb_bdev(struct file_system_type *fs_type, - char *dev_name, int flags, void * data) + int flags, char *dev_name, void * data) { struct inode *inode; struct block_device *bdev; @@ -598,17 +682,13 @@ goto out; check_disk_change(dev); error = -EACCES; - if (!(flags & MS_RDONLY) && is_read_only(dev)) { - blkdev_put(bdev, BDEV_FS); - goto out; - } + if (!(flags & MS_RDONLY) && is_read_only(dev)) + goto out1; error = -ENOMEM; s = alloc_super(); - if (!s) { - blkdev_put(bdev, BDEV_FS); - goto out; - } + if (!s) + goto out1; error = -EBUSY; restart: @@ -622,8 +702,7 @@ ((flags ^ old->s_flags) & MS_RDONLY)) { spin_unlock(&sb_lock); destroy_super(s); - blkdev_put(bdev, BDEV_FS); - goto out; + goto out1; } if (!grab_super(old)) goto restart; @@ -636,97 +715,103 @@ s->s_bdev = bdev; s->s_flags = flags; insert_super(s, fs_type); - - error = -EINVAL; - lock_super(s); if (!fs_type->read_super(s, data, flags & MS_VERBOSE ? 1 : 0)) - goto out_fail; + goto Einval; s->s_flags |= MS_ACTIVE; - unlock_super(s); path_release(&nd); return s; -out_fail: - unlock_super(s); +Einval: deactivate_super(s); remove_super(s); + error = -EINVAL; + goto out; +out1: + blkdev_put(bdev, BDEV_FS); out: path_release(&nd); return ERR_PTR(error); } static struct super_block *get_sb_nodev(struct file_system_type *fs_type, - int flags, void * data) + int flags, char *dev_name, void *data) { - struct super_block *s = alloc_super(); + struct super_block *s = get_anon_super(fs_type, NULL, NULL); + + if (IS_ERR(s)) + return s; - if (!s) - return ERR_PTR(-ENOMEM); - s->s_dev = get_unnamed_dev(); - if (!s->s_dev) { - destroy_super(s); - return ERR_PTR(-EMFILE); - } s->s_flags = flags; - spin_lock(&sb_lock); - insert_super(s, fs_type); - lock_super(s); - if (!fs_type->read_super(s, data, flags & MS_VERBOSE ? 1 : 0)) - goto out_fail; + if (!fs_type->read_super(s, data, flags & MS_VERBOSE ? 1 : 0)) { + deactivate_super(s); + remove_super(s); + return ERR_PTR(-EINVAL); + } s->s_flags |= MS_ACTIVE; - unlock_super(s); return s; +} -out_fail: - unlock_super(s); - deactivate_super(s); - remove_super(s); - return ERR_PTR(-EINVAL); +static int compare_single(struct super_block *s, void *p) +{ + return 1; } static struct super_block *get_sb_single(struct file_system_type *fs_type, - int flags, void *data) + int flags, char *dev_name, void *data) { - struct super_block * s = alloc_super(); - if (!s) - return ERR_PTR(-ENOMEM); - /* - * Get the superblock of kernel-wide instance, but - * keep the reference to fs_type. - */ -retry: - spin_lock(&sb_lock); - if (!list_empty(&fs_type->fs_supers)) { - struct super_block *old; - old = list_entry(fs_type->fs_supers.next, struct super_block, - s_instances); - if (!grab_super(old)) - goto retry; - destroy_super(s); - do_remount_sb(old, flags, data); - return old; - } else { - s->s_dev = get_unnamed_dev(); - if (!s->s_dev) { - spin_unlock(&sb_lock); - destroy_super(s); - return ERR_PTR(-EMFILE); - } + struct super_block *s = get_anon_super(fs_type, compare_single, NULL); + + if (IS_ERR(s)) + return s; + if (!s->s_root) { s->s_flags = flags; - insert_super(s, fs_type); - lock_super(s); - if (!fs_type->read_super(s, data, flags & MS_VERBOSE ? 1 : 0)) - goto out_fail; + if (!fs_type->read_super(s, data, flags & MS_VERBOSE ? 1 : 0)) { + deactivate_super(s); + remove_super(s); + return ERR_PTR(-EINVAL); + } s->s_flags |= MS_ACTIVE; - unlock_super(s); - return s; - - out_fail: - unlock_super(s); - deactivate_super(s); - remove_super(s); - return ERR_PTR(-EINVAL); } + do_remount_sb(s, flags, data); + return s; +} + +struct vfsmount * +do_kern_mount(const char *fstype, int flags, char *name, void *data) +{ + struct file_system_type *type = get_fs_type(fstype); + struct super_block *sb = ERR_PTR(-ENOMEM); + struct vfsmount *mnt; + + if (!type) + return ERR_PTR(-ENODEV); + + mnt = alloc_vfsmnt(); + if (!mnt) + goto out; + set_devname(mnt, name); + if (type->fs_flags & FS_REQUIRES_DEV) + sb = get_sb_bdev(type, flags, name, data); + else if (type->fs_flags & FS_SINGLE) + sb = get_sb_single(type, flags, name, data); + else + sb = get_sb_nodev(type, flags, name, data); + if (IS_ERR(sb)) + goto out_mnt; + if (type->fs_flags & FS_NOMOUNT) + sb->s_flags |= MS_NOUSER; + mnt->mnt_sb = sb; + mnt->mnt_root = dget(sb->s_root); + mnt->mnt_mountpoint = sb->s_root; + mnt->mnt_parent = mnt; + up_write(&sb->s_umount); + put_filesystem(type); + return mnt; +out_mnt: + free_vfsmnt(mnt); +out: + put_filesystem(type); + return (struct vfsmount *)sb; } void kill_super(struct super_block *sb) @@ -739,7 +824,6 @@ return; down_write(&sb->s_umount); - lock_kernel(); sb->s_root = NULL; /* Need to clean after the sucker */ if (fs->fs_flags & FS_LITTER) @@ -748,6 +832,7 @@ dput(root); fsync_super(sb); lock_super(sb); + lock_kernel(); sb->s_flags &= ~MS_ACTIVE; invalidate_inodes(sb); /* bad name - it should be evict_inodes() */ if (sop) { @@ -768,96 +853,7 @@ remove_super(sb); } -/* - * Alters the mount flags of a mounted file system. Only the mount point - * is used as a reference - file system type and the device are ignored. - */ - -int do_remount_sb(struct super_block *sb, int flags, void *data) -{ - int retval; - - if (!(flags & MS_RDONLY) && sb->s_dev && is_read_only(sb->s_dev)) - return -EACCES; - /*flags |= MS_RDONLY;*/ - if (flags & MS_RDONLY) - acct_auto_close(sb->s_dev); - shrink_dcache_sb(sb); - fsync_super(sb); - /* If we are remounting RDONLY, make sure there are no rw files open */ - if ((flags & MS_RDONLY) && !(sb->s_flags & MS_RDONLY)) - if (!fs_may_remount_ro(sb)) - return -EBUSY; - if (sb->s_op && sb->s_op->remount_fs) { - lock_super(sb); - retval = sb->s_op->remount_fs(sb, &flags, data); - unlock_super(sb); - if (retval) - return retval; - } - sb->s_flags = (sb->s_flags & ~MS_RMT_MASK) | (flags & MS_RMT_MASK); - - /* - * We can't invalidate inodes as we can loose data when remounting - * (someone might manage to alter data while we are waiting in lock_super() - * or in foo_remount_fs())) - */ - - return 0; -} - -struct vfsmount *do_kern_mount(char *type, int flags, char *name, void *data) -{ - struct file_system_type * fstype; - struct vfsmount *mnt = NULL; - struct super_block *sb; - - if (!type || !memchr(type, 0, PAGE_SIZE)) - return ERR_PTR(-EINVAL); - - /* we need capabilities... */ - if (!capable(CAP_SYS_ADMIN)) - return ERR_PTR(-EPERM); - - /* ... filesystem driver... */ - fstype = get_fs_type(type); - if (!fstype) - return ERR_PTR(-ENODEV); - - /* ... allocated vfsmount... */ - mnt = alloc_vfsmnt(); - if (!mnt) { - mnt = ERR_PTR(-ENOMEM); - goto fs_out; - } - set_devname(mnt, name); - /* get locked superblock */ - if (fstype->fs_flags & FS_REQUIRES_DEV) - sb = get_sb_bdev(fstype, name, flags, data); - else if (fstype->fs_flags & FS_SINGLE) - sb = get_sb_single(fstype, flags, data); - else - sb = get_sb_nodev(fstype, flags, data); - - if (IS_ERR(sb)) { - free_vfsmnt(mnt); - mnt = (struct vfsmount *)sb; - goto fs_out; - } - if (fstype->fs_flags & FS_NOMOUNT) - sb->s_flags |= MS_NOUSER; - - mnt->mnt_sb = sb; - mnt->mnt_root = dget(sb->s_root); - mnt->mnt_mountpoint = mnt->mnt_root; - mnt->mnt_parent = mnt; - up_write(&sb->s_umount); -fs_out: - put_filesystem(fstype); - return mnt; -} - struct vfsmount *kern_mount(struct file_system_type *type) { - return do_kern_mount((char *)type->name, 0, (char *)type->name, NULL); + return do_kern_mount(type->name, 0, (char *)type->name, NULL); } diff -u --recursive --new-file v2.5.1/linux/fs/sysv/super.c linux/fs/sysv/super.c --- v2.5.1/linux/fs/sysv/super.c Tue Dec 18 14:56:40 2001 +++ linux/fs/sysv/super.c Tue Dec 25 15:39:20 2001 @@ -315,8 +315,6 @@ sb->sv_ninodes = (sb->sv_firstdatazone - sb->sv_firstinodezone) << sb->sv_inodes_per_block_bits; - sb->s_blocksize = bsize; - sb->s_blocksize_bits = n_bits; if (!silent) printk("VFS: Found a %s FS (block size = %ld) on device %s\n", found, sb->s_blocksize, bdevname(sb->s_dev)); @@ -361,8 +359,7 @@ panic("Coherent FS: bad super-block size"); if (64 != sizeof (struct sysv_inode)) panic("sysv fs: bad i-node size"); - set_blocksize(dev,BLOCK_SIZE); - sb->s_blocksize = BLOCK_SIZE; + sb_set_blocksize(sb, BLOCK_SIZE); sb->sv_block_base = 0; for (i = 0; i < sizeof(flavours)/sizeof(flavours[0]) && !size; i++) { @@ -380,8 +377,7 @@ case 1: blocknr = bh->b_blocknr << 1; brelse(bh); - set_blocksize(dev, 512); - sb->s_blocksize = 512; + sb_set_blocksize(sb, 512); bh1 = sb_bread(sb, blocknr); bh = sb_bread(sb, blocknr + 1); break; @@ -391,8 +387,7 @@ case 3: blocknr = bh->b_blocknr >> 1; brelse(bh); - set_blocksize(dev, 2048); - sb->s_blocksize = 2048; + sb_set_blocksize(sb, 2048); bh1 = bh = sb_bread(sb, blocknr); break; default: @@ -408,7 +403,7 @@ brelse(bh1); brelse(bh); - set_blocksize(sb->s_dev,BLOCK_SIZE); + sb_set_blocksize(sb, BLOCK_SIZE); printk("oldfs: cannot read superblock\n"); failed: return NULL; @@ -443,8 +438,7 @@ sb->sv_type = FSTYPE_V7; sb->sv_bytesex = BYTESEX_PDP; - set_blocksize(dev, 512); - sb->s_blocksize = 512; + sb_set_blocksize(sb, 512); if ((bh = sb_bread(sb, 1)) == NULL) { if (!silent) diff -u --recursive --new-file v2.5.1/linux/fs/udf/super.c linux/fs/udf/super.c --- v2.5.1/linux/fs/udf/super.c Tue Dec 18 14:56:40 2001 +++ linux/fs/udf/super.c Tue Dec 25 15:39:20 2001 @@ -357,30 +357,11 @@ static int udf_set_blocksize(struct super_block *sb, int bsize) { - /* Use specified block size if specified */ - if (bsize) - sb->s_blocksize = bsize; - if (get_hardsect_size(sb->s_dev) > sb->s_blocksize) - sb->s_blocksize = get_hardsect_size(sb->s_dev); - - /* Block size must be an even multiple of 512 */ - switch (sb->s_blocksize) - { - case 512: sb->s_blocksize_bits = 9; break; - case 1024: sb->s_blocksize_bits = 10; break; - case 2048: sb->s_blocksize_bits = 11; break; - case 4096: sb->s_blocksize_bits = 12; break; - case 8192: sb->s_blocksize_bits = 13; break; - default: - { - udf_debug("Bad block size (%ld)\n", sb->s_blocksize); - printk(KERN_ERR "udf: bad block size (%ld)\n", sb->s_blocksize); - return 0; - } + if (!sb_min_blocksize(sb, bsize)) { + udf_debug("Bad block size (%d)\n", bsize); + printk(KERN_ERR "udf: bad block size (%d)\n", bsize); + return 0; } - - /* Set the block size */ - set_blocksize(sb->s_dev, sb->s_blocksize); return sb->s_blocksize; } diff -u --recursive --new-file v2.5.1/linux/fs/ufs/super.c linux/fs/ufs/super.c --- v2.5.1/linux/fs/ufs/super.c Tue Dec 18 14:56:40 2001 +++ linux/fs/ufs/super.c Tue Dec 25 15:39:20 2001 @@ -596,8 +596,7 @@ } again: - set_blocksize (sb->s_dev, block_size); - sb->s_blocksize = block_size; + sb_set_blocksize(sb, block_size); /* * read ufs super block from device @@ -716,8 +715,6 @@ /* * Read ufs_super_block into internal data structures */ - sb->s_blocksize = fs32_to_cpu(sb, usb1->fs_fsize); - sb->s_blocksize_bits = fs32_to_cpu(sb, usb1->fs_fshift); sb->s_op = &ufs_super_ops; sb->dq_op = NULL; /***/ sb->s_magic = fs32_to_cpu(sb, usb3->fs_magic); diff -u --recursive --new-file v2.5.1/linux/include/asm-sparc/keyboard.h linux/include/asm-sparc/keyboard.h --- v2.5.1/linux/include/asm-sparc/keyboard.h Tue Aug 28 07:09:44 2001 +++ linux/include/asm-sparc/keyboard.h Tue Dec 25 15:39:20 2001 @@ -40,7 +40,6 @@ #define kbd_init pcikbd_init #define compute_shiftstate pci_compute_shiftstate -#define keyboard_wait_for_keypress pci_wait_for_keypress #define getkeycode pci_getkeycode #define setkeycode pci_setkeycode #define getledstate pci_getledstate diff -u --recursive --new-file v2.5.1/linux/include/asm-sparc64/keyboard.h linux/include/asm-sparc64/keyboard.h --- v2.5.1/linux/include/asm-sparc64/keyboard.h Tue Aug 28 07:09:44 2001 +++ linux/include/asm-sparc64/keyboard.h Tue Dec 25 15:39:20 2001 @@ -38,7 +38,6 @@ #define kbd_init pcikbd_init #define compute_shiftstate pci_compute_shiftstate -#define keyboard_wait_for_keypress pci_wait_for_keypress #define getkeycode pci_getkeycode #define setkeycode pci_setkeycode #define getledstate pci_getledstate diff -u --recursive --new-file v2.5.1/linux/include/linux/amigaffs.h linux/include/linux/amigaffs.h --- v2.5.1/linux/include/linux/amigaffs.h Tue Dec 18 14:56:40 2001 +++ linux/include/linux/amigaffs.h Tue Dec 25 15:39:20 2001 @@ -23,8 +23,7 @@ static inline void affs_set_blocksize(struct super_block *sb, int size) { - set_blocksize(sb->s_dev, size); - sb->s_blocksize = size; + sb_set_blocksize(sb, size); } static inline struct buffer_head * affs_bread(struct super_block *sb, int block) diff -u --recursive --new-file v2.5.1/linux/include/linux/blk.h linux/include/linux/blk.h --- v2.5.1/linux/include/linux/blk.h Tue Dec 18 14:56:40 2001 +++ linux/include/linux/blk.h Tue Dec 25 15:57:29 2001 @@ -20,7 +20,6 @@ #define INITRD_MINOR 250 /* shouldn't collide with /dev/ram* too soon ... */ extern unsigned long initrd_start,initrd_end; -extern int mount_initrd; /* zero if initrd should not be mounted */ extern int initrd_below_start_ok; /* 1 if it is not an error if initrd_start < memory_start */ extern int rd_doload; /* 1 = load ramdisk, 0 = don't load */ extern int rd_prompt; /* 1 = prompt for ramdisk, 0 = don't prompt */ @@ -54,6 +53,9 @@ while ((rq = __elv_next_request(q))) { rq->flags |= REQ_STARTED; + + if (&rq->queuelist == q->last_merge) + q->last_merge = NULL; if ((rq->flags & REQ_DONTPREP) || !q->prep_rq_fn) break; diff -u --recursive --new-file v2.5.1/linux/include/linux/blkdev.h linux/include/linux/blkdev.h --- v2.5.1/linux/include/linux/blkdev.h Tue Dec 18 14:56:40 2001 +++ linux/include/linux/blkdev.h Tue Dec 25 15:56:47 2001 @@ -130,16 +130,17 @@ struct request_queue { /* - * the queue request freelist, one for reads and one for writes - */ - struct request_list rq[2]; - - /* * Together with queue_head for cacheline sharing */ struct list_head queue_head; + struct list_head *last_merge; elevator_t elevator; + /* + * the queue request freelist, one for reads and one for writes + */ + struct request_list rq[2]; + request_fn_proc *request_fn; merge_request_fn *back_merge_fn; merge_request_fn *front_merge_fn; @@ -213,27 +214,25 @@ extern unsigned long blk_max_low_pfn, blk_max_pfn; -#define BLK_BOUNCE_HIGH (blk_max_low_pfn << PAGE_SHIFT) -#define BLK_BOUNCE_ANY (blk_max_pfn << PAGE_SHIFT) -#define BLK_BOUNCE_ISA (ISA_DMA_THRESHOLD) - -#ifdef CONFIG_HIGHMEM +/* + * standard bounce addresses: + * + * BLK_BOUNCE_HIGH : bounce all highmem pages + * BLK_BOUNCE_ANY : don't bounce anything + * BLK_BOUNCE_ISA : bounce pages above ISA DMA boundary + */ +#define BLK_BOUNCE_HIGH ((blk_max_low_pfn + 1) << PAGE_SHIFT) +#define BLK_BOUNCE_ANY ((blk_max_pfn + 1) << PAGE_SHIFT) +#define BLK_BOUNCE_ISA (ISA_DMA_THRESHOLD) +extern int init_emergency_isa_pool(void); extern void create_bounce(unsigned long pfn, int gfp, struct bio **bio_orig); -extern void init_emergency_isa_pool(void); extern inline void blk_queue_bounce(request_queue_t *q, struct bio **bio) { create_bounce(q->bounce_pfn, q->bounce_gfp, bio); } -#else /* CONFIG_HIGHMEM */ - -#define blk_queue_bounce(q, bio) do { } while (0) -#define init_emergency_isa_pool() do { } while (0) - -#endif /* CONFIG_HIGHMEM */ - #define rq_for_each_bio(bio, rq) \ if ((rq->bio)) \ for (bio = (rq)->bio; bio; bio = bio->bi_next) @@ -275,9 +274,8 @@ extern void blk_recount_segments(request_queue_t *, struct bio *); extern inline int blk_phys_contig_segment(request_queue_t *q, struct bio *, struct bio *); extern inline int blk_hw_contig_segment(request_queue_t *q, struct bio *, struct bio *); -extern void blk_queue_assign_lock(request_queue_t *q, spinlock_t *); - extern int block_ioctl(kdev_t, unsigned int, unsigned long); +extern int ll_10byte_cmd_build(request_queue_t *, struct request *); /* * Access functions for manipulating queue properties @@ -292,6 +290,9 @@ extern void blk_queue_max_segment_size(request_queue_t *q, unsigned int); extern void blk_queue_hardsect_size(request_queue_t *q, unsigned short); extern void blk_queue_segment_boundary(request_queue_t *q, unsigned long); +extern void blk_queue_assign_lock(request_queue_t *q, spinlock_t *); +extern void blk_queue_prep_rq(request_queue_t *q, prep_rq_fn *pfn); + extern int blk_rq_map_sg(request_queue_t *, struct request *, struct scatterlist *); extern void blk_dump_rq_flags(struct request *, char *); extern void generic_unplug_device(void *); diff -u --recursive --new-file v2.5.1/linux/include/linux/blkdev.h.orig linux/include/linux/blkdev.h.orig --- v2.5.1/linux/include/linux/blkdev.h.orig Tue Dec 18 14:56:40 2001 +++ linux/include/linux/blkdev.h.orig Wed Dec 31 16:00:00 1969 @@ -1,371 +0,0 @@ -#ifndef _LINUX_BLKDEV_H -#define _LINUX_BLKDEV_H - -#include <linux/major.h> -#include <linux/sched.h> -#include <linux/genhd.h> -#include <linux/tqueue.h> -#include <linux/list.h> -#include <linux/mm.h> - -#include <asm/scatterlist.h> - -struct request_queue; -typedef struct request_queue request_queue_t; -struct elevator_s; -typedef struct elevator_s elevator_t; - -struct request_list { - unsigned int count; - struct list_head free; - wait_queue_head_t wait; -}; - -struct request { - struct list_head queuelist; /* looking for ->queue? you must _not_ - * access it directly, use - * blkdev_dequeue_request! */ - int elevator_sequence; - - unsigned char cmd[16]; - - unsigned long flags; /* see REQ_ bits below */ - - int rq_status; /* should split this into a few status bits */ - kdev_t rq_dev; - int errors; - sector_t sector; - unsigned long nr_sectors; - unsigned long hard_sector; /* the hard_* are block layer - * internals, no driver should - * touch them - */ - unsigned long hard_nr_sectors; - - /* Number of scatter-gather DMA addr+len pairs after - * physical address coalescing is performed. - */ - unsigned short nr_phys_segments; - - /* Number of scatter-gather addr+len pairs after - * physical and DMA remapping hardware coalescing is performed. - * This is the number of scatter-gather entries the driver - * will actually have to deal with after DMA mapping is done. - */ - unsigned short nr_hw_segments; - - unsigned int current_nr_sectors; - unsigned int hard_cur_sectors; - void *special; - char *buffer; - struct completion *waiting; - struct bio *bio, *biotail; - request_queue_t *q; - struct request_list *rl; -}; - -/* - * first three bits match BIO_RW* bits, important - */ -enum rq_flag_bits { - __REQ_RW, /* not set, read. set, write */ - __REQ_RW_AHEAD, /* READA */ - __REQ_BARRIER, /* may not be passed */ - __REQ_CMD, /* is a regular fs rw request */ - __REQ_NOMERGE, /* don't touch this for merging */ - __REQ_STARTED, /* drive already may have started this one */ - __REQ_DONTPREP, /* don't call prep for this one */ - /* - * for IDE - */ - __REQ_DRIVE_CMD, - __REQ_DRIVE_TASK, - - __REQ_PC, /* packet command (special) */ - __REQ_BLOCK_PC, /* queued down pc from block layer */ - __REQ_SENSE, /* sense retrival */ - - __REQ_SPECIAL, /* driver special command */ - - __REQ_NR_BITS, /* stops here */ -}; - -#define REQ_RW (1 << __REQ_RW) -#define REQ_RW_AHEAD (1 << __REQ_RW_AHEAD) -#define REQ_BARRIER (1 << __REQ_BARRIER) -#define REQ_CMD (1 << __REQ_CMD) -#define REQ_NOMERGE (1 << __REQ_NOMERGE) -#define REQ_STARTED (1 << __REQ_STARTED) -#define REQ_DONTPREP (1 << __REQ_DONTPREP) -#define REQ_DRIVE_CMD (1 << __REQ_DRIVE_CMD) -#define REQ_DRIVE_TASK (1 << __REQ_DRIVE_TASK) -#define REQ_PC (1 << __REQ_PC) -#define REQ_SENSE (1 << __REQ_SENSE) -#define REQ_BLOCK_PC (1 << __REQ_BLOCK_PC) -#define REQ_SPECIAL (1 << __REQ_SPECIAL) - -#include <linux/elevator.h> - -typedef int (merge_request_fn) (request_queue_t *, struct request *, - struct bio *); -typedef int (merge_requests_fn) (request_queue_t *, struct request *, - struct request *); -typedef void (request_fn_proc) (request_queue_t *q); -typedef request_queue_t * (queue_proc) (kdev_t dev); -typedef int (make_request_fn) (request_queue_t *q, struct bio *bio); -typedef int (prep_rq_fn) (request_queue_t *, struct request *); -typedef void (unplug_device_fn) (void *q); - -enum blk_queue_state { - Queue_down, - Queue_up, -}; - -/* - * Default nr free requests per queue, ll_rw_blk will scale it down - * according to available RAM at init time - */ -#define QUEUE_NR_REQUESTS 8192 - -struct request_queue -{ - /* - * the queue request freelist, one for reads and one for writes - */ - struct request_list rq[2]; - - /* - * Together with queue_head for cacheline sharing - */ - struct list_head queue_head; - elevator_t elevator; - - request_fn_proc *request_fn; - merge_request_fn *back_merge_fn; - merge_request_fn *front_merge_fn; - merge_requests_fn *merge_requests_fn; - make_request_fn *make_request_fn; - prep_rq_fn *prep_rq_fn; - - /* - * The queue owner gets to use this for whatever they like. - * ll_rw_blk doesn't touch it. - */ - void *queuedata; - - /* - * queue needs bounce pages for pages above this limit - */ - unsigned long bounce_pfn; - int bounce_gfp; - - /* - * This is used to remove the plug when tq_disk runs. - */ - struct tq_struct plug_tq; - - /* - * various queue flags, see QUEUE_* below - */ - unsigned long queue_flags; - - /* - * protects queue structures from reentrancy - */ - spinlock_t *queue_lock; - - /* - * queue settings - */ - unsigned short max_sectors; - unsigned short max_phys_segments; - unsigned short max_hw_segments; - unsigned short hardsect_size; - unsigned int max_segment_size; - - unsigned long seg_boundary_mask; - - wait_queue_head_t queue_wait; -}; - -#define RQ_INACTIVE (-1) -#define RQ_ACTIVE 1 -#define RQ_SCSI_BUSY 0xffff -#define RQ_SCSI_DONE 0xfffe -#define RQ_SCSI_DISCONNECTING 0xffe0 - -#define QUEUE_FLAG_PLUGGED 0 /* queue is plugged */ -#define QUEUE_FLAG_NOSPLIT 1 /* can process bio over several goes */ -#define QUEUE_FLAG_CLUSTER 2 /* cluster several segments into 1 */ - -#define blk_queue_plugged(q) test_bit(QUEUE_FLAG_PLUGGED, &(q)->queue_flags) -#define blk_mark_plugged(q) set_bit(QUEUE_FLAG_PLUGGED, &(q)->queue_flags) -#define blk_queue_empty(q) elv_queue_empty(q) -#define list_entry_rq(ptr) list_entry((ptr), struct request, queuelist) - -#define rq_data_dir(rq) ((rq)->flags & 1) - -/* - * noop, requests are automagically marked as active/inactive by I/O - * scheduler -- see elv_next_request - */ -#define blk_queue_headactive(q, head_active) - -extern unsigned long blk_max_low_pfn, blk_max_pfn; - -#define BLK_BOUNCE_HIGH (blk_max_low_pfn << PAGE_SHIFT) -#define BLK_BOUNCE_ANY (blk_max_pfn << PAGE_SHIFT) -#define BLK_BOUNCE_ISA (ISA_DMA_THRESHOLD) - -#ifdef CONFIG_HIGHMEM - -extern void create_bounce(unsigned long pfn, int gfp, struct bio **bio_orig); -extern void init_emergency_isa_pool(void); - -extern inline void blk_queue_bounce(request_queue_t *q, struct bio **bio) -{ - create_bounce(q->bounce_pfn, q->bounce_gfp, bio); -} - -#else /* CONFIG_HIGHMEM */ - -#define blk_queue_bounce(q, bio) do { } while (0) -#define init_emergency_isa_pool() do { } while (0) - -#endif /* CONFIG_HIGHMEM */ - -#define rq_for_each_bio(bio, rq) \ - if ((rq->bio)) \ - for (bio = (rq)->bio; bio; bio = bio->bi_next) - -struct blk_dev_struct { - /* - * queue_proc has to be atomic - */ - request_queue_t request_queue; - queue_proc *queue; - void *data; -}; - -struct sec_size { - unsigned block_size; - unsigned block_size_bits; -}; - -/* - * Used to indicate the default queue for drivers that don't bother - * to implement multiple queues. We have this access macro here - * so as to eliminate the need for each and every block device - * driver to know about the internal structure of blk_dev[]. - */ -#define BLK_DEFAULT_QUEUE(_MAJOR) &blk_dev[_MAJOR].request_queue - -extern struct sec_size * blk_sec[MAX_BLKDEV]; -extern struct blk_dev_struct blk_dev[MAX_BLKDEV]; -extern void grok_partitions(kdev_t dev, long size); -extern int wipe_partitions(kdev_t dev); -extern void register_disk(struct gendisk *dev, kdev_t first, unsigned minors, struct block_device_operations *ops, long size); -extern void generic_make_request(struct bio *bio); -extern inline request_queue_t *blk_get_queue(kdev_t dev); -extern void blkdev_release_request(struct request *); -extern void blk_attempt_remerge(request_queue_t *, struct request *); -extern struct request *blk_get_request(request_queue_t *, int, int); -extern void blk_put_request(struct request *); -extern void blk_plug_device(request_queue_t *); -extern void blk_recount_segments(request_queue_t *, struct bio *); -extern inline int blk_phys_contig_segment(request_queue_t *q, struct bio *, struct bio *); -extern inline int blk_hw_contig_segment(request_queue_t *q, struct bio *, struct bio *); -extern void blk_queue_assign_lock(request_queue_t *q, spinlock_t *); - -extern int block_ioctl(kdev_t, unsigned int, unsigned long); - -/* - * Access functions for manipulating queue properties - */ -extern int blk_init_queue(request_queue_t *, request_fn_proc *, spinlock_t *); -extern void blk_cleanup_queue(request_queue_t *); -extern void blk_queue_make_request(request_queue_t *, make_request_fn *); -extern void blk_queue_bounce_limit(request_queue_t *, u64); -extern void blk_queue_max_sectors(request_queue_t *q, unsigned short); -extern void blk_queue_max_phys_segments(request_queue_t *q, unsigned short); -extern void blk_queue_max_hw_segments(request_queue_t *q, unsigned short); -extern void blk_queue_max_segment_size(request_queue_t *q, unsigned int); -extern void blk_queue_hardsect_size(request_queue_t *q, unsigned short); -extern void blk_queue_segment_boundary(request_queue_t *q, unsigned long); -extern int blk_rq_map_sg(request_queue_t *, struct request *, struct scatterlist *); -extern void blk_dump_rq_flags(struct request *, char *); -extern void generic_unplug_device(void *); - -extern int * blk_size[MAX_BLKDEV]; - -extern int * blksize_size[MAX_BLKDEV]; - -extern int * max_readahead[MAX_BLKDEV]; - -#define MAX_PHYS_SEGMENTS 128 -#define MAX_HW_SEGMENTS 128 -#define MAX_SECTORS 255 - -#define MAX_SEGMENT_SIZE 65536 - -/* read-ahead in pages.. */ -#define MAX_READAHEAD 31 -#define MIN_READAHEAD 3 - -#define blkdev_entry_to_request(entry) list_entry((entry), struct request, queuelist) -#define blkdev_entry_next_request(entry) blkdev_entry_to_request((entry)->next) -#define blkdev_entry_prev_request(entry) blkdev_entry_to_request((entry)->prev) -#define blkdev_next_request(req) blkdev_entry_to_request((req)->queuelist.next) -#define blkdev_prev_request(req) blkdev_entry_to_request((req)->queuelist.prev) - -extern void drive_stat_acct(struct request *, int, int); - -extern inline void blk_clear(int major) -{ - blk_size[major] = NULL; -#if 0 - blk_size_in_bytes[major] = NULL; -#endif - blksize_size[major] = NULL; - max_readahead[major] = NULL; - read_ahead[major] = 0; -} - -extern inline int get_hardsect_size(kdev_t dev) -{ - request_queue_t *q = blk_get_queue(dev); - int retval = 512; - - if (q && q->hardsect_size) - retval = q->hardsect_size; - - return retval; -} - -#define blk_finished_io(nsects) do { } while (0) -#define blk_started_io(nsects) do { } while (0) - -extern inline unsigned int blksize_bits(unsigned int size) -{ - unsigned int bits = 8; - do { - bits++; - size >>= 1; - } while (size > 256); - return bits; -} - -extern inline unsigned int block_size(kdev_t dev) -{ - int retval = BLOCK_SIZE; - int major = MAJOR(dev); - - if (blksize_size[major]) { - int minor = MINOR(dev); - if (blksize_size[major][minor]) - retval = blksize_size[major][minor]; - } - return retval; -} - -#endif diff -u --recursive --new-file v2.5.1/linux/include/linux/console.h linux/include/linux/console.h --- v2.5.1/linux/include/linux/console.h Thu Nov 22 11:46:19 2001 +++ linux/include/linux/console.h Tue Dec 25 15:56:40 2001 @@ -97,7 +97,6 @@ void (*write)(struct console *, const char *, unsigned); int (*read)(struct console *, const char *, unsigned); kdev_t (*device)(struct console *); - int (*wait_key)(struct console *); void (*unblank)(void); int (*setup)(struct console *, char *); short flags; diff -u --recursive --new-file v2.5.1/linux/include/linux/devfs_fs_kernel.h linux/include/linux/devfs_fs_kernel.h --- v2.5.1/linux/include/linux/devfs_fs_kernel.h Tue Dec 18 14:56:40 2001 +++ linux/include/linux/devfs_fs_kernel.h Tue Dec 25 15:56:40 2001 @@ -112,7 +112,6 @@ int number); extern void mount_devfs_fs (void); -extern void devfs_make_root (const char *name); #else /* CONFIG_DEVFS_FS */ @@ -288,10 +287,6 @@ } static inline void mount_devfs_fs (void) -{ - return; -} -static inline void devfs_make_root (const char *name) { return; } diff -u --recursive --new-file v2.5.1/linux/include/linux/elevator.h linux/include/linux/elevator.h --- v2.5.1/linux/include/linux/elevator.h Tue Dec 18 14:56:40 2001 +++ linux/include/linux/elevator.h Thu Dec 20 08:50:13 2001 @@ -1,12 +1,8 @@ #ifndef _LINUX_ELEVATOR_H #define _LINUX_ELEVATOR_H -typedef void (elevator_fn) (struct request *, elevator_t *, - struct list_head *, - struct list_head *, int); - typedef int (elevator_merge_fn) (request_queue_t *, struct request **, - struct list_head *, struct bio *); + struct bio *); typedef void (elevator_merge_cleanup_fn) (request_queue_t *, struct request *, int); @@ -21,8 +17,7 @@ struct elevator_s { - int read_latency; - int write_latency; + int latency[2]; elevator_merge_fn *elevator_merge_fn; elevator_merge_cleanup_fn *elevator_merge_cleanup_fn; @@ -35,11 +30,11 @@ elevator_exit_fn *elevator_exit_fn; }; -int elevator_noop_merge(request_queue_t *, struct request **, struct list_head *, struct bio *); +int elevator_noop_merge(request_queue_t *, struct request **, struct bio *); void elevator_noop_merge_cleanup(request_queue_t *, struct request *, int); void elevator_noop_merge_req(struct request *, struct request *); -int elevator_linus_merge(request_queue_t *, struct request **, struct list_head *, struct bio *); +int elevator_linus_merge(request_queue_t *, struct request **, struct bio *); void elevator_linus_merge_cleanup(request_queue_t *, struct request *, int); void elevator_linus_merge_req(struct request *, struct request *); int elv_linus_init(request_queue_t *, elevator_t *); @@ -69,32 +64,7 @@ #define ELEVATOR_FRONT_MERGE 1 #define ELEVATOR_BACK_MERGE 2 -/* - * This is used in the elevator algorithm. We don't prioritise reads - * over writes any more --- although reads are more time-critical than - * writes, by treating them equally we increase filesystem throughput. - * This turns out to give better overall performance. -- sct - */ -#define IN_ORDER(s1,s2) \ - ((((s1)->rq_dev == (s2)->rq_dev && \ - (s1)->sector < (s2)->sector)) || \ - (s1)->rq_dev < (s2)->rq_dev) - -#define BHRQ_IN_ORDER(bh, rq) \ - ((((bh)->b_rdev == (rq)->rq_dev && \ - (bh)->b_rsector < (rq)->sector)) || \ - (bh)->b_rdev < (rq)->rq_dev) - -static inline int elevator_request_latency(elevator_t * elevator, int rw) -{ - int latency; - - latency = elevator->read_latency; - if (rw != READ) - latency = elevator->write_latency; - - return latency; -} +#define elevator_request_latency(e, rw) ((e)->latency[(rw) & 1]) /* * will change once we move to a more complex data structure than a simple @@ -116,9 +86,7 @@ #define ELEVATOR_NOOP \ ((elevator_t) { \ - 0, /* read_latency */ \ - 0, /* write_latency */ \ - \ + { 0, 0}, \ elevator_noop_merge, /* elevator_merge_fn */ \ elevator_noop_merge_cleanup, /* elevator_merge_cleanup_fn */ \ elevator_noop_merge_req, /* elevator_merge_req_fn */ \ @@ -130,9 +98,7 @@ #define ELEVATOR_LINUS \ ((elevator_t) { \ - 8192, /* read passovers */ \ - 16384, /* write passovers */ \ - \ + { 8192, 16384 }, \ elevator_linus_merge, /* elevator_merge_fn */ \ elevator_linus_merge_cleanup, /* elevator_merge_cleanup_fn */ \ elevator_linus_merge_req, /* elevator_merge_req_fn */ \ diff -u --recursive --new-file v2.5.1/linux/include/linux/fs.h linux/include/linux/fs.h --- v2.5.1/linux/include/linux/fs.h Tue Dec 18 14:56:40 2001 +++ linux/include/linux/fs.h Tue Dec 25 15:56:40 2001 @@ -982,6 +982,7 @@ extern struct vfsmount *kern_mount(struct file_system_type *); extern int may_umount(struct vfsmount *); extern long do_mount(char *, char *, char *, unsigned long, void *); +extern void umount_tree(struct vfsmount *); #define kern_umount mntput @@ -1378,6 +1379,8 @@ __bforget(buf); } extern int set_blocksize(kdev_t, int); +extern int sb_set_blocksize(struct super_block *, int); +extern int sb_min_blocksize(struct super_block *, int); extern struct buffer_head * bread(kdev_t, int, int); static inline struct buffer_head * sb_bread(struct super_block *sb, int block) { diff -u --recursive --new-file v2.5.1/linux/include/linux/kbd_kern.h linux/include/linux/kbd_kern.h --- v2.5.1/linux/include/linux/kbd_kern.h Thu Nov 22 11:47:23 2001 +++ linux/include/linux/kbd_kern.h Tue Dec 25 15:57:37 2001 @@ -140,7 +140,6 @@ int getkeycode(unsigned int scancode); int setkeycode(unsigned int scancode, unsigned int keycode); void compute_shiftstate(void); -int keyboard_wait_for_keypress(struct console *); /* defkeymap.c */ diff -u --recursive --new-file v2.5.1/linux/include/linux/keyboard.h linux/include/linux/keyboard.h --- v2.5.1/linux/include/linux/keyboard.h Thu Nov 22 11:47:07 2001 +++ linux/include/linux/keyboard.h Tue Dec 25 15:57:22 2001 @@ -26,7 +26,6 @@ extern const int max_vals[]; extern unsigned short *key_maps[MAX_NR_KEYMAPS]; extern unsigned short plain_map[NR_KEYS]; -extern wait_queue_head_t keypress_wait; extern unsigned char keyboard_type; #endif diff -u --recursive --new-file v2.5.1/linux/include/linux/namespace.h linux/include/linux/namespace.h --- v2.5.1/linux/include/linux/namespace.h Wed Dec 31 16:00:00 1969 +++ linux/include/linux/namespace.h Tue Dec 25 15:39:20 2001 @@ -0,0 +1,42 @@ +#ifndef _NAMESPACE_H_ +#define _NAMESPACE_H_ +#ifdef __KERNEL__ + +struct namespace { + atomic_t count; + struct vfsmount * root; + struct list_head list; + struct rw_semaphore sem; +}; + +static inline void put_namespace(struct namespace *namespace) +{ + if (atomic_dec_and_test(&namespace->count)) { + down_write(&namespace->sem); + spin_lock(&dcache_lock); + umount_tree(namespace->root); + spin_unlock(&dcache_lock); + up_write(&namespace->sem); + kfree(namespace); + } +} + +static inline void exit_namespace(struct task_struct *p) +{ + struct namespace *namespace = p->namespace; + if (namespace) { + task_lock(p); + p->namespace = NULL; + task_unlock(p); + put_namespace(namespace); + } +} +extern int copy_namespace(int, struct task_struct *); + +static inline void get_namespace(struct namespace *namespace) +{ + atomic_inc(&namespace->count); +} + +#endif +#endif diff -u --recursive --new-file v2.5.1/linux/include/linux/raid/md_k.h linux/include/linux/raid/md_k.h --- v2.5.1/linux/include/linux/raid/md_k.h Tue Dec 18 14:56:40 2001 +++ linux/include/linux/raid/md_k.h Wed Dec 19 15:18:10 2001 @@ -240,7 +240,7 @@ int (*stop_resync)(mddev_t *mddev); int (*restart_resync)(mddev_t *mddev); - int (*sync_request)(mddev_t *mddev, sector_t sector_nr); + int (*sync_request)(mddev_t *mddev, sector_t sector_nr, int go_faster); }; diff -u --recursive --new-file v2.5.1/linux/include/linux/raid/raid1.h linux/include/linux/raid/raid1.h --- v2.5.1/linux/include/linux/raid/raid1.h Tue Dec 18 14:56:40 2001 +++ linux/include/linux/raid/raid1.h Wed Dec 19 15:18:10 2001 @@ -9,8 +9,8 @@ int number; int raid_disk; kdev_t dev; - int sect_limit; - int head_position; + sector_t head_position; + atomic_t nr_pending; /* * State bits: @@ -31,23 +31,21 @@ int raid_disks; int working_disks; int last_used; - sector_t next_sect; - int sect_count; + sector_t next_seq_sect; mdk_thread_t *thread, *resync_thread; int resync_mirrors; mirror_info_t *spare; spinlock_t device_lock; /* for use when syncing mirrors: */ - unsigned long start_active, start_ready, - start_pending, start_future; - int cnt_done, cnt_active, cnt_ready, - cnt_pending, cnt_future; - int phase; - int window; - wait_queue_head_t wait_done; - wait_queue_head_t wait_ready; - spinlock_t segment_lock; + + spinlock_t resync_lock; + int nr_pending; + int barrier; + sector_t next_resync; + + wait_queue_head_t wait_idle; + wait_queue_head_t wait_resume; mempool_t *r1bio_pool; mempool_t *r1buf_pool; @@ -62,7 +60,8 @@ #define mddev_to_conf(mddev) ((conf_t *) mddev->private) /* - * this is our 'private' 'collective' RAID1 buffer head. + * this is our 'private' RAID1 bio. + * * it contains information about what kind of IO operations were started * for this RAID1 operation, and about their status: */ @@ -83,6 +82,7 @@ * if the IO is in READ direction, then this bio is used: */ struct bio *read_bio; + int read_disk; /* * if the IO is in WRITE direction, then multiple bios are used: */ @@ -94,5 +94,5 @@ /* bits for r1bio.state */ #define R1BIO_Uptodate 1 -#define R1BIO_SyncPhase 2 + #endif diff -u --recursive --new-file v2.5.1/linux/include/linux/sched.h linux/include/linux/sched.h --- v2.5.1/linux/include/linux/sched.h Tue Dec 18 14:56:40 2001 +++ linux/include/linux/sched.h Tue Dec 25 15:56:40 2001 @@ -41,6 +41,7 @@ #define CLONE_VFORK 0x00004000 /* set if the parent wants the child to wake it up on mm_release */ #define CLONE_PARENT 0x00008000 /* set if we want to have the same parent as the cloner */ #define CLONE_THREAD 0x00010000 /* Same thread group? */ +#define CLONE_NEWNS 0x00020000 /* New namespace group? */ #define CLONE_SIGNAL (CLONE_SIGHAND | CLONE_THREAD) @@ -165,6 +166,7 @@ */ #define NR_OPEN_DEFAULT BITS_PER_LONG +struct namespace; /* * Open file table structure */ @@ -388,6 +390,8 @@ struct fs_struct *fs; /* open file information */ struct files_struct *files; +/* namespace */ + struct namespace *namespace; /* signal handlers */ spinlock_t sigmask_lock; /* Protects signal and blocked */ struct signal_struct *sig; diff -u --recursive --new-file v2.5.1/linux/include/linux/vt_kern.h linux/include/linux/vt_kern.h --- v2.5.1/linux/include/linux/vt_kern.h Thu Nov 22 11:47:07 2001 +++ linux/include/linux/vt_kern.h Tue Dec 25 15:57:22 2001 @@ -9,6 +9,7 @@ #include <linux/config.h> #include <linux/vt.h> #include <linux/kd.h> +#include <linux/tty.h> /* * Presently, a lot of graphics programs do not restore the contents of diff -u --recursive --new-file v2.5.1/linux/init/do_mounts.c linux/init/do_mounts.c --- v2.5.1/linux/init/do_mounts.c Tue Dec 18 14:56:40 2001 +++ linux/init/do_mounts.c Tue Dec 25 15:39:20 2001 @@ -3,13 +3,10 @@ #include <linux/slab.h> #include <linux/devfs_fs_kernel.h> #include <linux/unistd.h> -#include <linux/string.h> #include <linux/ctype.h> -#include <linux/init.h> -#include <linux/smp_lock.h> #include <linux/blk.h> -#include <linux/tty.h> #include <linux/fd.h> +#include <linux/tty.h> #include <linux/nfs_fs.h> #include <linux/nfs_fs_sb.h> @@ -18,12 +15,9 @@ #include <linux/ext2_fs.h> #include <linux/romfs_fs.h> -#include <asm/uaccess.h> - #define BUILD_CRAMDISK extern int get_filesystem_list(char * buf); -extern void wait_for_keypress(void); asmlinkage long sys_mount(char *dev_name, char *dir_name, char *type, unsigned long flags, void *data); @@ -38,12 +32,21 @@ #ifdef CONFIG_BLK_DEV_INITRD unsigned int real_root_dev; /* do_proc_dointvec cannot handle kdev_t */ -#endif -#ifdef CONFIG_BLK_DEV_RAM -extern int rd_doload; +static int __initdata mount_initrd = 1; + +static int __init no_initrd(char *str) +{ + mount_initrd = 0; + return 1; +} + +__setup("noinitrd", no_initrd); #else -static int rd_doload = 0; +static int __initdata mount_initrd = 0; #endif + +int __initdata rd_doload; /* 1 = load RAM disk, 0 = don't load */ + int root_mountflags = MS_RDONLY | MS_VERBOSE; static char root_device_name[64]; @@ -52,6 +55,13 @@ static int do_devfs = 0; +static int __init load_ramdisk(char *str) +{ + rd_doload = simple_strtol(str,NULL,0) & 3; + return 1; +} +__setup("load_ramdisk=", load_ramdisk); + static int __init readonly(char *str) { if (*str) @@ -353,8 +363,9 @@ static void __init change_floppy(char *fmt, ...) { - extern void wait_for_keypress(void); + struct termios termios; char buf[80]; + char c; int fd; va_list args; va_start(args, fmt); @@ -366,11 +377,38 @@ close(fd); } printk(KERN_NOTICE "VFS: Insert %s and press ENTER\n", buf); - wait_for_keypress(); + fd = open("/dev/console", O_RDWR, 0); + if (fd >= 0) { + sys_ioctl(fd, TCGETS, (long)&termios); + termios.c_lflag &= ~ICANON; + sys_ioctl(fd, TCSETSF, (long)&termios); + read(fd, &c, 1); + termios.c_lflag |= ICANON; + sys_ioctl(fd, TCSETSF, (long)&termios); + close(fd); + } } #ifdef CONFIG_BLK_DEV_RAM +int __initdata rd_prompt = 1; /* 1 = prompt for RAM disk, 0 = don't prompt */ + +static int __init prompt_ramdisk(char *str) +{ + rd_prompt = simple_strtol(str,NULL,0) & 1; + return 1; +} +__setup("prompt_ramdisk=", prompt_ramdisk); + +int __initdata rd_image_start; /* starting block # of image */ + +static int __init ramdisk_start_setup(char *str) +{ + rd_image_start = simple_strtol(str,NULL,0); + return 1; +} +__setup("ramdisk_start=", ramdisk_start_setup); + static int __init crd_load(int in_fd, int out_fd); /* @@ -588,7 +626,6 @@ static int __init rd_load_disk(int n) { #ifdef CONFIG_BLK_DEV_RAM - extern int rd_prompt; if (rd_prompt) change_floppy("root floppy disk to be loaded into RAM disk"); create_dev("/dev/ram", MKDEV(RAMDISK_MAJOR, n), NULL); @@ -596,6 +633,63 @@ return rd_load_image("/dev/root"); } +#ifdef CONFIG_DEVFS_FS + +static void __init convert_name(char *prefix, char *name, char *p, int part) +{ + int host, bus, target, lun; + char dest[64]; + char src[64]; + char *base = p - 1; + + /* Decode "c#b#t#u#" */ + if (*p++ != 'c') + return; + host = simple_strtol(p, &p, 10); + if (*p++ != 'b') + return; + bus = simple_strtol(p, &p, 10); + if (*p++ != 't') + return; + target = simple_strtol(p, &p, 10); + if (*p++ != 'u') + return; + lun = simple_strtol(p, &p, 10); + if (!part) + sprintf(dest, "%s/host%d/bus%d/target%d/lun%d", + prefix, host, bus, target, lun); + else if (*p++ == 'p') + sprintf(dest, "%s/host%d/bus%d/target%d/lun%d/part%s", + prefix, host, bus, target, lun, p); + else + sprintf(dest, "%s/host%d/bus%d/target%d/lun%d/disc", + prefix, host, bus, target, lun); + *base = '\0'; + sprintf(src, "/dev/%s", name); + sys_mkdir(src, 0755); + *base = '/'; + sprintf(src, "/dev/%s", name); + sys_symlink(dest, src); +} + +static void __init devfs_make_root(char *name) +{ + + if (!strncmp(name, "sd/", 3)) + convert_name("../scsi", name, name+3, 1); + else if (!strncmp(name, "sr/", 3)) + convert_name("../scsi", name, name+3, 0); + else if (!strncmp(name, "ide/hd/", 7)) + convert_name("..", name, name + 7, 1); + else if (!strncmp(name, "ide/cd/", 7)) + convert_name("..", name, name + 7, 0); +} +#else +static void __init devfs_make_root(char *name) +{ +} +#endif + static void __init mount_root(void) { #ifdef CONFIG_ROOT_NFS @@ -715,24 +809,22 @@ */ void prepare_namespace(void) { - int do_initrd = 0; int is_floppy = MAJOR(ROOT_DEV) == FLOPPY_MAJOR; #ifdef CONFIG_BLK_DEV_INITRD if (!initrd_start) mount_initrd = 0; - if (mount_initrd) - do_initrd = 1; real_root_dev = ROOT_DEV; #endif sys_mkdir("/dev", 0700); sys_mkdir("/root", 0700); + sys_mknod("/dev/console", S_IFCHR|0600, MKDEV(TTYAUX_MAJOR, 1)); #ifdef CONFIG_DEVFS_FS sys_mount("devfs", "/dev", "devfs", 0, NULL); do_devfs = 1; #endif create_dev("/dev/root", ROOT_DEV, NULL); - if (do_initrd) { + if (mount_initrd) { if (initrd_load() && ROOT_DEV != MKDEV(RAMDISK_MAJOR, 0)) { handle_initrd(); goto out; diff -u --recursive --new-file v2.5.1/linux/kernel/exit.c linux/kernel/exit.c --- v2.5.1/linux/kernel/exit.c Tue Dec 18 14:56:41 2001 +++ linux/kernel/exit.c Tue Dec 25 15:39:20 2001 @@ -12,6 +12,7 @@ #include <linux/completion.h> #include <linux/personality.h> #include <linux/tty.h> +#include <linux/namespace.h> #ifdef CONFIG_BSD_PROCESS_ACCT #include <linux/acct.h> #endif @@ -452,6 +453,7 @@ sem_exit(); __exit_files(tsk); __exit_fs(tsk); + exit_namespace(tsk); exit_sighand(tsk); exit_thread(); diff -u --recursive --new-file v2.5.1/linux/kernel/fork.c linux/kernel/fork.c --- v2.5.1/linux/kernel/fork.c Tue Dec 18 14:56:41 2001 +++ linux/kernel/fork.c Tue Dec 25 15:39:20 2001 @@ -19,6 +19,7 @@ #include <linux/module.h> #include <linux/vmalloc.h> #include <linux/completion.h> +#include <linux/namespace.h> #include <linux/personality.h> #include <asm/pgtable.h> @@ -566,6 +567,9 @@ struct task_struct *p; struct completion vfork; + if ((clone_flags & (CLONE_NEWNS|CLONE_FS)) == (CLONE_NEWNS|CLONE_FS)) + return -EINVAL; + retval = -EPERM; /* @@ -661,9 +665,11 @@ goto bad_fork_cleanup_fs; if (copy_mm(clone_flags, p)) goto bad_fork_cleanup_sighand; + if (copy_namespace(clone_flags, p)) + goto bad_fork_cleanup_mm; retval = copy_thread(0, clone_flags, stack_start, stack_size, p, regs); if (retval) - goto bad_fork_cleanup_mm; + goto bad_fork_cleanup_namespace; p->semundo = NULL; /* Our parent execution domain becomes current domain @@ -730,6 +736,8 @@ fork_out: return retval; +bad_fork_cleanup_namespace: + exit_namespace(p); bad_fork_cleanup_mm: exit_mm(p); bad_fork_cleanup_sighand: diff -u --recursive --new-file v2.5.1/linux/kernel/kmod.c linux/kernel/kmod.c --- v2.5.1/linux/kernel/kmod.c Tue Jul 17 18:23:50 2001 +++ linux/kernel/kmod.c Tue Dec 25 15:39:20 2001 @@ -24,6 +24,8 @@ #include <linux/unistd.h> #include <linux/kmod.h> #include <linux/smp_lock.h> +#include <linux/slab.h> +#include <linux/namespace.h> #include <linux/completion.h> #include <asm/uaccess.h> @@ -36,6 +38,7 @@ struct fs_struct *our_fs, *init_fs; struct dentry *root, *pwd; struct vfsmount *rootmnt, *pwdmnt; + struct namespace *our_ns, *init_ns; /* * Make modprobe's fs context be a copy of init's. @@ -55,6 +58,11 @@ */ init_fs = init_task.fs; + init_ns = init_task.namespace; + get_namespace(init_ns); + our_ns = current->namespace; + current->namespace = init_ns; + put_namespace(our_ns); read_lock(&init_fs->lock); rootmnt = mntget(init_fs->rootmnt); root = dget(init_fs->root); diff -u --recursive --new-file v2.5.1/linux/kernel/ksyms.c linux/kernel/ksyms.c --- v2.5.1/linux/kernel/ksyms.c Tue Dec 18 14:56:41 2001 +++ linux/kernel/ksyms.c Tue Dec 25 15:39:20 2001 @@ -116,11 +116,11 @@ EXPORT_SYMBOL(find_vma); EXPORT_SYMBOL(get_unmapped_area); EXPORT_SYMBOL(init_mm); +EXPORT_SYMBOL(create_bounce); #ifdef CONFIG_HIGHMEM EXPORT_SYMBOL(kmap_high); EXPORT_SYMBOL(kunmap_high); EXPORT_SYMBOL(highmem_start_page); -EXPORT_SYMBOL(create_bounce); EXPORT_SYMBOL(kmap_prot); EXPORT_SYMBOL(kmap_pte); #endif @@ -188,6 +188,8 @@ EXPORT_SYMBOL(write_inode_now); EXPORT_SYMBOL(notify_change); EXPORT_SYMBOL(set_blocksize); +EXPORT_SYMBOL(sb_set_blocksize); +EXPORT_SYMBOL(sb_min_blocksize); EXPORT_SYMBOL(getblk); EXPORT_SYMBOL(cdget); EXPORT_SYMBOL(cdput); diff -u --recursive --new-file v2.5.1/linux/kernel/signal.c linux/kernel/signal.c --- v2.5.1/linux/kernel/signal.c Tue Dec 18 14:56:41 2001 +++ linux/kernel/signal.c Thu Dec 20 08:50:19 2001 @@ -649,10 +649,8 @@ /* * kill_something_info() interprets pid in interesting ways just like kill(2). * - * POSIX (2001) specifies "If pid is -1, sig shall be sent to all processes - * (excluding an unspecified set of system processes) for which the process - * has permission to send that signal." - * So, probably the process should also signal itself. + * POSIX specifies that kill(-1,sig) is unspecified, but what we have + * is probably wrong. Should make it like BSD or SYSV. */ static int kill_something_info(int sig, struct siginfo *info, int pid) @@ -665,7 +663,7 @@ read_lock(&tasklist_lock); for_each_task(p) { - if (p->pid > 1) { + if (p->pid > 1 && p != current) { int err = send_sig_info(sig, info, p); ++count; if (err != -EPERM) diff -u --recursive --new-file v2.5.1/linux/mm/Makefile linux/mm/Makefile --- v2.5.1/linux/mm/Makefile Tue Dec 18 14:56:41 2001 +++ linux/mm/Makefile Mon Dec 17 10:49:31 2001 @@ -14,8 +14,6 @@ obj-y := memory.o mmap.o filemap.o mprotect.o mlock.o mremap.o \ vmalloc.o slab.o bootmem.o swap.o vmscan.o page_io.o \ page_alloc.o swap_state.o swapfile.o numa.o oom_kill.o \ - shmem.o mempool.o - -obj-$(CONFIG_HIGHMEM) += highmem.o + shmem.o highmem.o mempool.o include $(TOPDIR)/Rules.make diff -u --recursive --new-file v2.5.1/linux/mm/highmem.c linux/mm/highmem.c --- v2.5.1/linux/mm/highmem.c Tue Dec 18 14:56:41 2001 +++ linux/mm/highmem.c Thu Dec 20 08:50:19 2001 @@ -19,6 +19,19 @@ #include <linux/mm.h> #include <linux/pagemap.h> #include <linux/mempool.h> +#include <linux/blkdev.h> + +static mempool_t *page_pool, *isa_page_pool; + +static void *page_pool_alloc(int gfp_mask, void *data) +{ + return alloc_page(gfp_mask); +} + +static void page_pool_free(void *page, void *data) +{ + __free_page(page); +} /* * Virtual_count is not a pure "count". @@ -28,6 +41,7 @@ * since the last TLB flush - so we can't use it. * n means that there are (n-1) current users of it. */ +#ifdef CONFIG_HIGHMEM static int pkmap_count[LAST_PKMAP]; static unsigned int last_pkmap_nr; static spinlock_t kmap_lock = SPIN_LOCK_UNLOCKED; @@ -185,19 +199,6 @@ } #define POOL_SIZE 64 -#define ISA_POOL_SIZE 16 - -static mempool_t *page_pool, *isa_page_pool; - -static void *page_pool_alloc(int gfp_mask, void *data) -{ - return alloc_page(gfp_mask); -} - -static void page_pool_free(void *page, void *data) -{ - __free_page(page); -} static __init int init_emergency_pool(void) { @@ -211,11 +212,37 @@ page_pool = mempool_create(POOL_SIZE, page_pool_alloc, page_pool_free, NULL); if (!page_pool) BUG(); - printk("highmem bounce pool size: %d pages and bhs.\n", POOL_SIZE); + printk("highmem bounce pool size: %d pages\n", POOL_SIZE); return 0; } +__initcall(init_emergency_pool); + +/* + * highmem version, map in to vec + */ +static inline void bounce_copy_vec(struct bio_vec *to, unsigned char *vfrom) +{ + unsigned long flags; + unsigned char *vto; + + local_irq_save(flags); + vto = kmap_atomic(to->bv_page, KM_BOUNCE_READ); + memcpy(vto + to->bv_offset, vfrom, to->bv_len); + kunmap_atomic(vto, KM_BOUNCE_READ); + local_irq_restore(flags); +} + +#else /* CONFIG_HIGHMEM */ + +#define bounce_copy_vec(to, vfrom) \ + memcpy(page_address((to)->bv_page) + (to)->bv_offset, vfrom, (to)->bv_len) + +#endif + +#define ISA_POOL_SIZE 16 + /* * gets called "every" time someone init's a queue with BLK_BOUNCE_ISA * as the max address, so check if the pool has already been created. @@ -233,8 +260,6 @@ return 0; } -__initcall(init_emergency_pool); - /* * Simple bounce buffer support for highmem pages. Depending on the * queue gfp mask set, *to may or may not be a highmem page. kmap it @@ -242,8 +267,7 @@ */ static inline void copy_to_high_bio_irq(struct bio *to, struct bio *from) { - unsigned char *vto, *vfrom; - unsigned long flags; + unsigned char *vfrom; struct bio_vec *tovec, *fromvec; int i; @@ -258,11 +282,7 @@ vfrom = page_address(fromvec->bv_page) + fromvec->bv_offset; - local_irq_save(flags); - vto = kmap_atomic(tovec->bv_page, KM_BOUNCE_READ); - memcpy(vto + tovec->bv_offset, vfrom, tovec->bv_len); - kunmap_atomic(vto, KM_BOUNCE_READ); - local_irq_restore(flags); + bounce_copy_vec(tovec, vfrom); } } @@ -336,10 +356,25 @@ BUG_ON((*bio_orig)->bi_idx); + /* + * for non-isa bounce case, just check if the bounce pfn is equal + * to or bigger than the highest pfn in the system -- in that case, + * don't waste time iterating over bio segments + */ if (!(gfp & GFP_DMA)) { + if (pfn >= blk_max_pfn) + return; + +#ifndef CONFIG_HIGHMEM + /* + * should not hit for non-highmem case + */ + BUG(); +#endif bio_gfp = GFP_NOHIGHIO; pool = page_pool; } else { + BUG_ON(!isa_page_pool); bio_gfp = GFP_NOIO; pool = isa_page_pool; }