* [PATCH] md - 0 of 2 - Introduction
@ 2004-03-05 5:45 NeilBrown
2004-03-05 5:45 ` [PATCH] md - 1 of 2 - Use "shedule_timeout(2)" instead of yield() as it seems to wait for less time NeilBrown
` (2 more replies)
0 siblings, 3 replies; 7+ messages in thread
From: NeilBrown @ 2004-03-05 5:45 UTC (permalink / raw)
To: Andrew Morton; +Cc: linux-raid
Here are two patches for md that I believe are suitable
for 2.6.4.
1/ change yield() to schedule_timeout(2) as promised a while back.
2/ Support assembling partitionable raid at boot-time so that it is
possible to boot from a partition of a raid array without needing
an initramfs.
Thanks,
NeilBrown
^ permalink raw reply [flat|nested] 7+ messages in thread* [PATCH] md - 1 of 2 - Use "shedule_timeout(2)" instead of yield() as it seems to wait for less time. 2004-03-05 5:45 [PATCH] md - 0 of 2 - Introduction NeilBrown @ 2004-03-05 5:45 ` NeilBrown 2004-03-05 6:16 ` Andrew Morton 2004-03-05 5:45 ` [PATCH] md - 2 of 2 - Allow assembling of partitioned arrays at boot time NeilBrown 2004-03-05 8:42 ` [PATCH] md - 0 of 2 - Introduction Clemens Schwaighofer 2 siblings, 1 reply; 7+ messages in thread From: NeilBrown @ 2004-03-05 5:45 UTC (permalink / raw) To: Andrew Morton; +Cc: linux-raid ----------- Diffstat output ------------ ./drivers/md/raid5.c | 3 ++- ./drivers/md/raid6main.c | 3 ++- 2 files changed, 4 insertions(+), 2 deletions(-) diff ./drivers/md/raid5.c~current~ ./drivers/md/raid5.c --- ./drivers/md/raid5.c~current~ 2004-03-05 16:43:07.000000000 +1100 +++ ./drivers/md/raid5.c 2004-03-05 16:43:07.000000000 +1100 @@ -1409,7 +1409,8 @@ static int sync_request (mddev_t *mddev, /* make sure we don't swamp the stripe cache if someone else * is trying to get access */ - yield(); + set_current_state(TASK_UNINTERRUPTIBLE); + schedule_timeout(2); } spin_lock(&sh->lock); set_bit(STRIPE_SYNCING, &sh->state); diff ./drivers/md/raid6main.c~current~ ./drivers/md/raid6main.c --- ./drivers/md/raid6main.c~current~ 2004-03-05 16:43:07.000000000 +1100 +++ ./drivers/md/raid6main.c 2004-03-05 16:43:07.000000000 +1100 @@ -1571,7 +1571,8 @@ static int sync_request (mddev_t *mddev, /* make sure we don't swamp the stripe cache if someone else * is trying to get access */ - yield(); + set_current_state(TASK_UNINTERRUPTIBLE); + schedule_timeout(2); } spin_lock(&sh->lock); set_bit(STRIPE_SYNCING, &sh->state); ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH] md - 1 of 2 - Use "shedule_timeout(2)" instead of yield() as it seems to wait for less time. 2004-03-05 5:45 ` [PATCH] md - 1 of 2 - Use "shedule_timeout(2)" instead of yield() as it seems to wait for less time NeilBrown @ 2004-03-05 6:16 ` Andrew Morton 2004-03-10 2:26 ` Neil Brown 0 siblings, 1 reply; 7+ messages in thread From: Andrew Morton @ 2004-03-05 6:16 UTC (permalink / raw) To: NeilBrown; +Cc: linux-raid NeilBrown <neilb@cse.unsw.edu.au> wrote: > > + schedule_timeout(2); Are you sure we want to sleep for 20-30 milliseconds on a 100 Hz machine? Wouldn't it be better to do diff -puN drivers/md/raid5.c~md-use-schedule_timeout drivers/md/raid5.c --- 25/drivers/md/raid5.c~md-use-schedule_timeout 2004-03-04 22:15:14.000000000 -0800 +++ 25-akpm/drivers/md/raid5.c 2004-03-04 22:15:41.000000000 -0800 @@ -1409,7 +1409,8 @@ static int sync_request (mddev_t *mddev, /* make sure we don't swamp the stripe cache if someone else * is trying to get access */ - yield(); + set_current_state(TASK_UNINTERRUPTIBLE); + schedule_timeout(max(HZ/500, 1)); } spin_lock(&sh->lock); set_bit(STRIPE_SYNCING, &sh->state); diff -puN drivers/md/raid6main.c~md-use-schedule_timeout drivers/md/raid6main.c --- 25/drivers/md/raid6main.c~md-use-schedule_timeout 2004-03-04 22:15:14.000000000 -0800 +++ 25-akpm/drivers/md/raid6main.c 2004-03-04 22:15:50.000000000 -0800 @@ -1571,7 +1571,8 @@ static int sync_request (mddev_t *mddev, /* make sure we don't swamp the stripe cache if someone else * is trying to get access */ - yield(); + set_current_state(TASK_UNINTERRUPTIBLE); + schedule_timeout(max(HZ/500, 1)); } spin_lock(&sh->lock); set_bit(STRIPE_SYNCING, &sh->state); _ ? ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH] md - 1 of 2 - Use "shedule_timeout(2)" instead of yield() as it seems to wait for less time. 2004-03-05 6:16 ` Andrew Morton @ 2004-03-10 2:26 ` Neil Brown 0 siblings, 0 replies; 7+ messages in thread From: Neil Brown @ 2004-03-10 2:26 UTC (permalink / raw) To: Andrew Morton; +Cc: linux-raid On Thursday March 4, akpm@osdl.org wrote: > NeilBrown <neilb@cse.unsw.edu.au> wrote: > > > > + schedule_timeout(2); > > Are you sure we want to sleep for 20-30 milliseconds on a 100 Hz > machine? I'm not sure it really matters. This bit of code doesn't trigger very often. I'm not unhappy with your version, but I suspect that just schedule_timeout(1) would be ok. I think I used '2' because there is no lower bound on how long schedule_timeout(1) will sleep (until the next tick, which could be anytime - right?). But on reflection, that doesn't really matter. As long as it pauses for a little while at least once in a while, it will give any non-sync IO a chance, that once that starts happening, resync will automatically wind back and there won't be a problem anymore. So I think (and some testing seems to demonstrate) that it doesn't really matter. So leave it as it is I guess. Thank, NeilBrown > > Wouldn't it be better to do > > diff -puN drivers/md/raid5.c~md-use-schedule_timeout drivers/md/raid5.c > --- 25/drivers/md/raid5.c~md-use-schedule_timeout 2004-03-04 22:15:14.000000000 -0800 > +++ 25-akpm/drivers/md/raid5.c 2004-03-04 22:15:41.000000000 -0800 > @@ -1409,7 +1409,8 @@ static int sync_request (mddev_t *mddev, > /* make sure we don't swamp the stripe cache if someone else > * is trying to get access > */ > - yield(); > + set_current_state(TASK_UNINTERRUPTIBLE); > + schedule_timeout(max(HZ/500, 1)); > } > spin_lock(&sh->lock); > set_bit(STRIPE_SYNCING, &sh->state); > diff -puN drivers/md/raid6main.c~md-use-schedule_timeout drivers/md/raid6main.c > --- 25/drivers/md/raid6main.c~md-use-schedule_timeout 2004-03-04 22:15:14.000000000 -0800 > +++ 25-akpm/drivers/md/raid6main.c 2004-03-04 22:15:50.000000000 -0800 > @@ -1571,7 +1571,8 @@ static int sync_request (mddev_t *mddev, > /* make sure we don't swamp the stripe cache if someone else > * is trying to get access > */ > - yield(); > + set_current_state(TASK_UNINTERRUPTIBLE); > + schedule_timeout(max(HZ/500, 1)); > } > spin_lock(&sh->lock); > set_bit(STRIPE_SYNCING, &sh->state); > > _ > > ? ^ permalink raw reply [flat|nested] 7+ messages in thread
* [PATCH] md - 2 of 2 - Allow assembling of partitioned arrays at boot time. 2004-03-05 5:45 [PATCH] md - 0 of 2 - Introduction NeilBrown 2004-03-05 5:45 ` [PATCH] md - 1 of 2 - Use "shedule_timeout(2)" instead of yield() as it seems to wait for less time NeilBrown @ 2004-03-05 5:45 ` NeilBrown 2004-03-05 8:42 ` [PATCH] md - 0 of 2 - Introduction Clemens Schwaighofer 2 siblings, 0 replies; 7+ messages in thread From: NeilBrown @ 2004-03-05 5:45 UTC (permalink / raw) To: Andrew Morton; +Cc: linux-raid kernel parameters: raid=partitionable will make all auto-detected md arrays partitionable md=d.... will assemble an array as a partitionable array. ----------- Diffstat output ------------ ./drivers/md/md.c | 19 ++++++--- ./init/do_mounts_md.c | 99 +++++++++++++++++++++++++++++--------------------- 2 files changed, 70 insertions(+), 48 deletions(-) diff ./drivers/md/md.c~current~ ./drivers/md/md.c --- ./drivers/md/md.c~current~ 2004-03-05 16:43:23.000000000 +1100 +++ ./drivers/md/md.c 2004-03-05 16:43:24.000000000 +1100 @@ -57,7 +57,7 @@ #ifndef MODULE -static void autostart_arrays (void); +static void autostart_arrays (int part); #endif static mdk_personality_t *pers[MAX_PERSONALITY]; @@ -1792,7 +1792,7 @@ static void autorun_array(mddev_t *mddev * * If "unit" is allocated, then bump its reference count */ -static void autorun_devices(void) +static void autorun_devices(int part) { struct list_head candidates; struct list_head *tmp; @@ -1825,7 +1825,12 @@ static void autorun_devices(void) bdevname(rdev0->bdev, b), rdev0->preferred_minor); break; } - dev = MKDEV(MD_MAJOR, rdev0->preferred_minor); + if (part) + dev = MKDEV(mdp_major, + rdev0->preferred_minor << MdpMinorShift); + else + dev = MKDEV(MD_MAJOR, rdev0->preferred_minor); + md_probe(dev, NULL, NULL); mddev = mddev_find(dev); if (!mddev) { @@ -1922,7 +1927,7 @@ static int autostart_array(dev_t startde /* * possibly return codes */ - autorun_devices(); + autorun_devices(0); return 0; } @@ -2407,7 +2412,7 @@ static int md_ioctl(struct inode *inode, #ifndef MODULE case RAID_AUTORUN: err = 0; - autostart_arrays(); + autostart_arrays(arg); goto done; #endif default:; @@ -3577,7 +3582,7 @@ void md_autodetect_dev(dev_t dev) } -static void autostart_arrays(void) +static void autostart_arrays(int part) { char b[BDEVNAME_SIZE]; mdk_rdev_t *rdev; @@ -3602,7 +3607,7 @@ static void autostart_arrays(void) } dev_cnt = 0; - autorun_devices(); + autorun_devices(part); } #endif diff ./init/do_mounts_md.c~current~ ./init/do_mounts_md.c --- ./init/do_mounts_md.c~current~ 2004-03-05 16:43:23.000000000 +1100 +++ ./init/do_mounts_md.c 2004-03-05 16:43:24.000000000 +1100 @@ -12,14 +12,17 @@ * The code for that is here. */ -static int __initdata raid_noautodetect; +static int __initdata raid_noautodetect, raid_autopart; static struct { - char device_set [MAX_MD_DEVS]; - int pers[MAX_MD_DEVS]; - int chunk[MAX_MD_DEVS]; - char *device_names[MAX_MD_DEVS]; -} md_setup_args __initdata; + int minor; + int partitioned; + int pers; + int chunk; + char *device_names; +} md_setup_args[MAX_MD_DEVS] __initdata; + +static int md_setup_ents __initdata; /* * Parse the command-line parameters given our kernel, but do not @@ -43,21 +46,37 @@ static struct { */ static int __init md_setup(char *str) { - int minor, level, factor, fault, pers; + int minor, level, factor, fault, pers, partitioned = 0; char *pername = ""; - char *str1 = str; + char *str1; + int ent; + if (*str == 'd') { + partitioned = 1; + str++; + } if (get_option(&str, &minor) != 2) { /* MD Number */ printk(KERN_WARNING "md: Too few arguments supplied to md=.\n"); return 0; } + str1 = str; if (minor >= MAX_MD_DEVS) { printk(KERN_WARNING "md: md=%d, Minor device number too high.\n", minor); return 0; - } else if (md_setup_args.device_names[minor]) { - printk(KERN_WARNING "md: md=%d, Specified more than once. " - "Replacing previous definition.\n", minor); } + for (ent=0 ; ent< md_setup_ents ; ent++) + if (md_setup_args[ent].minor == minor && + md_setup_args[ent].partitioned == partitioned) { + printk(KERN_WARNING "md: md=%s%d, Specified more than once. " + "Replacing previous definition.\n", partitioned?"d":"", minor); + break; + } + if (ent >= MAX_MD_DEVS) { + printk(KERN_WARNING "md: md=%s%d - too many md initialisations\n", partitioned?"d":"", minor); + return 0; + } + if (ent >= md_setup_ents) + md_setup_ents++; switch (get_option(&str, &level)) { /* RAID Personality */ case 2: /* could be 0 or -1.. */ if (level == 0 || level == LEVEL_LINEAR) { @@ -66,24 +85,16 @@ static int __init md_setup(char *str) printk(KERN_WARNING "md: Too few arguments supplied to md=.\n"); return 0; } - md_setup_args.pers[minor] = level; - md_setup_args.chunk[minor] = 1 << (factor+12); - switch(level) { - case LEVEL_LINEAR: + md_setup_args[ent].pers = level; + md_setup_args[ent].chunk = 1 << (factor+12); + if (level == LEVEL_LINEAR) { pers = LINEAR; pername = "linear"; - break; - case 0: + } else { pers = RAID0; pername = "raid0"; - break; - default: - printk(KERN_WARNING - "md: The kernel has not been configured for raid%d support!\n", - level); - return 0; } - md_setup_args.pers[minor] = pers; + md_setup_args[ent].pers = pers; break; } /* FALL THROUGH */ @@ -91,35 +102,38 @@ static int __init md_setup(char *str) str = str1; /* FALL THROUGH */ case 0: - md_setup_args.pers[minor] = 0; + md_setup_args[ent].pers = 0; pername="super-block"; } printk(KERN_INFO "md: Will configure md%d (%s) from %s, below.\n", minor, pername, str); - md_setup_args.device_names[minor] = str; + md_setup_args[ent].device_names = str; + md_setup_args[ent].partitioned = partitioned; + md_setup_args[ent].minor = minor; return 1; } static void __init md_setup_drive(void) { - int minor, i; + int minor, i, ent, partitioned; dev_t dev; dev_t devices[MD_SB_DISKS+1]; - for (minor = 0; minor < MAX_MD_DEVS; minor++) { + for (ent = 0; ent < md_setup_ents ; ent++) { int fd; int err = 0; char *devname; mdu_disk_info_t dinfo; char name[16], devfs_name[16]; - if (!(devname = md_setup_args.device_names[minor])) - continue; - - sprintf(name, "/dev/md%d", minor); - sprintf(devfs_name, "/dev/md/%d", minor); + minor = md_setup_args[ent].minor; + partitioned = md_setup_args[ent].partitioned; + devname = md_setup_args[ent].device_names; + + sprintf(name, "/dev/md%s%d", partitioned?"_d":"", minor); + sprintf(devfs_name, "/dev/md/%s%d", partitioned?"d":"", minor); create_dev(name, MKDEV(MD_MAJOR, minor), devfs_name); for (i = 0; i < MD_SB_DISKS && devname != 0; i++) { char *p; @@ -143,20 +157,19 @@ static void __init md_setup_drive(void) } devices[i] = dev; - md_setup_args.device_set[minor] = 1; devname = p; } devices[i] = 0; - if (!md_setup_args.device_set[minor]) + if (!i) continue; - printk(KERN_INFO "md: Loading md%d: %s\n", minor, md_setup_args.device_names[minor]); + printk(KERN_INFO "md: Loading md%s%d: %s\n", partitioned?"_d":"", minor, md_setup_args[ent].device_names); fd = open(name, 0, 0); if (fd < 0) { - printk(KERN_ERR "md: open failed - cannot start array %d\n", minor); + printk(KERN_ERR "md: open failed - cannot start array %s\n", name); continue; } if (sys_ioctl(fd, SET_ARRAY_INFO, 0) == -EBUSY) { @@ -167,10 +180,10 @@ static void __init md_setup_drive(void) continue; } - if (md_setup_args.pers[minor]) { + if (md_setup_args[ent].pers) { /* non-persistent */ mdu_array_info_t ainfo; - ainfo.level = pers_to_level(md_setup_args.pers[minor]); + ainfo.level = pers_to_level(md_setup_args[ent].pers); ainfo.size = 0; ainfo.nr_disks =0; ainfo.raid_disks =0; @@ -181,7 +194,7 @@ static void __init md_setup_drive(void) ainfo.state = (1 << MD_SB_CLEAN); ainfo.layout = 0; - ainfo.chunk_size = md_setup_args.chunk[minor]; + ainfo.chunk_size = md_setup_args[ent].chunk; err = sys_ioctl(fd, SET_ARRAY_INFO, (long)&ainfo); for (i = 0; !err && i <= MD_SB_DISKS; i++) { dev = devices[i]; @@ -229,6 +242,10 @@ static int __init raid_setup(char *str) if (!strncmp(str, "noautodetect", wlen)) raid_noautodetect = 1; + if (strncmp(str, "partitionable", wlen)==0) + raid_autopart = 1; + if (strncmp(str, "part", wlen)==0) + raid_autopart = 1; pos += wlen+1; } return 1; @@ -245,7 +262,7 @@ void __init md_run_setup(void) else { int fd = open("/dev/md0", 0, 0); if (fd >= 0) { - sys_ioctl(fd, RAID_AUTORUN, 0); + sys_ioctl(fd, RAID_AUTORUN, raid_autopart); close(fd); } } ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH] md - 0 of 2 - Introduction 2004-03-05 5:45 [PATCH] md - 0 of 2 - Introduction NeilBrown 2004-03-05 5:45 ` [PATCH] md - 1 of 2 - Use "shedule_timeout(2)" instead of yield() as it seems to wait for less time NeilBrown 2004-03-05 5:45 ` [PATCH] md - 2 of 2 - Allow assembling of partitioned arrays at boot time NeilBrown @ 2004-03-05 8:42 ` Clemens Schwaighofer 2004-03-05 10:01 ` Michael Tokarev 2 siblings, 1 reply; 7+ messages in thread From: Clemens Schwaighofer @ 2004-03-05 8:42 UTC (permalink / raw) To: NeilBrown; +Cc: linux-raid -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 NeilBrown wrote: | Here are two patches for md that I believe are suitable | for 2.6.4. | | 2/ Support assembling partitionable raid at boot-time so that it is | possible to boot from a partition of a raid array without needing | an initramfs. I have a RAID 1 here and I boot with grub and I don't need ana initramfs (initrd) ... - -- Clemens Schwaighofer - IT Engineer & System Administration ========================================================== Tequila Japan, 6-17-2 Ginza Chuo-ku, Tokyo 104-8167, JAPAN Tel: +81-(0)3-3545-7703 Fax: +81-(0)3-3545-7343 http://www.tequila.jp ========================================================== -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.4 (GNU/Linux) iD8DBQFASD1ojBz/yQjBxz8RAq6lAKDdN8hKgdBocQZOqC0yGoYd/Z6MCQCfdaL/ BFv6yTZznpg7wIEWxlQELak= =Zht7 -----END PGP SIGNATURE----- ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH] md - 0 of 2 - Introduction 2004-03-05 8:42 ` [PATCH] md - 0 of 2 - Introduction Clemens Schwaighofer @ 2004-03-05 10:01 ` Michael Tokarev 0 siblings, 0 replies; 7+ messages in thread From: Michael Tokarev @ 2004-03-05 10:01 UTC (permalink / raw) Cc: linux-raid Clemens Schwaighofer wrote: > > NeilBrown wrote: > | Here are two patches for md that I believe are suitable > | for 2.6.4. > | > | 2/ Support assembling partitionable raid at boot-time so that it is > | possible to boot from a partition of a raid array without needing > | an initramfs. > > I have a RAID 1 here and I boot with grub and I don't need ana initramfs > (initrd) ... The keyword here is "partitionable". /mjt ^ permalink raw reply [flat|nested] 7+ messages in thread
end of thread, other threads:[~2004-03-10 2:26 UTC | newest] Thread overview: 7+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2004-03-05 5:45 [PATCH] md - 0 of 2 - Introduction NeilBrown 2004-03-05 5:45 ` [PATCH] md - 1 of 2 - Use "shedule_timeout(2)" instead of yield() as it seems to wait for less time NeilBrown 2004-03-05 6:16 ` Andrew Morton 2004-03-10 2:26 ` Neil Brown 2004-03-05 5:45 ` [PATCH] md - 2 of 2 - Allow assembling of partitioned arrays at boot time NeilBrown 2004-03-05 8:42 ` [PATCH] md - 0 of 2 - Introduction Clemens Schwaighofer 2004-03-05 10:01 ` Michael Tokarev
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).