| View previous topic :: View next topic |
| Author |
Message |
hpestilence
Tux's lil' helper
Joined: 16 Dec 2004
Posts: 82
|
 Posted: Tue Mar 22, 2005 9:28 pm Post subject: reiser4 2.6.11 out |
 |
|
i just noticed that reiser4 for kernel 2.6.11 is out
ftp://ftp.namesys.com/pub/reiser4-for-2.6/2.6.11
has anyone tried it out yet? any problems?
im thinking of upgrading my kernel to gentoo-dev-sources stable with that patch 
EDIT:
I just compiled gentoo-dev-sources 2.6.11-r4 with reiser4 and everything is running smoothly so far 
Modified ebuild from https://forums.gentoo.org/viewtopic-t-295168.html
reiser4-gentoo-dev-sources-2.6.11-r4.ebuild
| Code: | # Copyright 1999-2005 Gentoo Foundation
# Distributed under the terms of the GNU General Public License v2
ETYPE="sources"
IUSE="ultra1"
inherit kernel-2
detect_version
detect_arch
#version of gentoo patchset
GPV="11.04"
GPV_SRC="mirror://gentoo/genpatches-${KV_MAJOR}.${KV_MINOR}-${GPV}-base.tar.bz2
mirror://gentoo/genpatches-${KV_MAJOR}.${KV_MINOR}-${GPV}-extras.tar.bz2"
KEYWORDS="~x86 ~ppc ~amd64 ~ppc64"
HOMEPAGE="http://dev.gentoo.org/~dsd/gentoo-dev-sources"
UNIPATCH_LIST="${DISTDIR}/genpatches-${KV_MAJOR}.${KV_MINOR}-${GPV}-base.tar.bz2
${DISTDIR}/genpatches-${KV_MAJOR}.${KV_MINOR}-${GPV}-extras.tar.bz2
${DISTDIR}/reiser4-for-${KV/-reiser4*/-1}.gz"
UNIPATCH_DOCS="${WORKDIR}/patches/genpatches-${KV_MAJOR}.${KV_MINOR}-${GPV}/0000_README"
DESCRIPTION="Full sources including patchset for the ${KV_MAJOR}.${KV_MINOR} kernel tree and extra patches for reiser4 support"
SRC_URI="${KERNEL_URI} ${GPV_SRC} ${ARCH_URI}
ftp://ftp.namesys.com/pub/reiser4-for-${KV_MAJOR}.${KV_MINOR}/${KV%%-reiser4*}/reiser4-for-${KV/-reiser4*/-1}.gz"
pkg_setup() {
if use sparc; then
# hmu lockup hack on ultra1
use ultra1 || UNIPATCH_EXCLUDE="${UNIPATCH_EXCLUDE} 1399_sparc-U1-hme-lockup.patch"
fi
}
pkg_postinst() {
postinst_sources
echo
if [ "${ARCH}" = "sparc" ]; then
if [ x"`cat /proc/openprom/name 2>/dev/null`" \
= x"'SUNW,Ultra-1'" ]; then
einfo "For users with an Enterprise model Ultra 1 using the HME"
einfo "network interface, please emerge the kernel using the"
einfo "following command: USE=ultra1 emerge ${PN}"
fi
fi
einfo "For more info on this patchset, and how to report problems, see:"
einfo "${HOMEPAGE}"
echo
echo
ewarn "IMPORTANT:"
ewarn "The reiser4 filesystem is not considered stable in linux yet. The filesystem"
ewarn "is very much usable though."
ewarn "Always keep a second stable and bootable kernel apart in your boot manager."
ewarn "DO NOT use reiser4 for your /boot partition unless you really know what you"
ewarn "are doing... even then I do not recommend it."
echo
ewarn "Make sure that 4kb stacks are turned OFF in the kernel. Found under kernel hacking subsection in the kernel config"
echo
} |
_________________
When the plague descends from Heaven above only Darkness will be your one true ally.
Last edited by hpestilence on Tue Mar 22, 2005 11:16 pm; edited 1 time in total |
|
| Back to top |
|
 |
firephoto
Veteran
Joined: 29 Oct 2003
Posts: 1612
Location: +48° 5' 23.40", -119° 48' 30.00"
|
| Posted: Tue Mar 22, 2005 10:06 pm Post subject: |
|
|
Hey, I just went looking and found this out myself just now.
From the README:
| Quote: |
03/22/2005
reiser4 for 2.6.11
Changes: major change is that concurrent read/write performance may
increase in some cases, cleanups
|
Going to try it with 2.6.11-ck2 and make a partition of 2 on the extra 80g drive I have that's been spinning for a month doing next to nothing. |
|
| Back to top |
|
|
firephoto
Veteran
Joined: 29 Oct 2003
Posts: 1612
Location: +48° 5' 23.40", -119° 48' 30.00"
|
| Posted: Tue Mar 22, 2005 11:24 pm Post subject: |
|
|
I tried it with ck2 but it has a hunk failing when it patches and it errors out on the compile.
| Code: |
# patch -p1 < /home/tom/download/Kernels/reiser4-for-2.6.11-1
patching file fs/fs-writeback.c
patching file include/linux/fs.h
Hunk #1 succeeded at 1026 (offset 19 lines).
Hunk #2 succeeded at 1375 (offset 19 lines).
Hunk #3 succeeded at 1429 (offset 19 lines).
patching file fs/inode.c
patching file include/linux/mm.h
patching file mm/truncate.c
patching file mm/filemap.c
patching file mm/readahead.c
patching file mm/vmscan.c
Hunk #1 succeeded at 466 (offset -2 lines).
patching file include/linux/rcupdate.h
patching file kernel/rcupdate.c
patching file mm/swap.c
patching file lib/radix-tree.c
patching file include/linux/radix-tree.h
patching file include/linux/gfp.h
patching file include/linux/init_task.h
Hunk #1 FAILED at 112.
1 out of 1 hunk FAILED -- saving rejects to file include/linux/init_task.h.rej
patching file include/linux/sched.h
Hunk #1 succeeded at 486 with fuzz 2 (offset -199 lines).
patching file kernel/fork.c
patching file mm/page_alloc.c
Hunk #1 succeeded at 612 (offset 4 lines).
Hunk #2 succeeded at 810 (offset 4 lines).
patching file fs/Kconfig
patching file fs/Makefile
Hunk #1 succeeded at 49 (offset 1 line).
patching file Documentation/Changes
patching file Documentation/filesystems/reiser4.txt
patching file fs/reiser4/Kconfig
patching file fs/reiser4/Makefile
patching file fs/reiser4/README
patching file fs/reiser4/as_ops.c
patching file fs/reiser4/block_alloc.c
patching file fs/reiser4/block_alloc.h
patching file fs/reiser4/blocknrset.c
patching file fs/reiser4/carry.c
patching file fs/reiser4/carry.h
patching file fs/reiser4/carry_ops.c
patching file fs/reiser4/carry_ops.h
patching file fs/reiser4/cluster.c
patching file fs/reiser4/cluster.h
patching file fs/reiser4/context.c
patching file fs/reiser4/context.h
patching file fs/reiser4/coord.c
patching file fs/reiser4/coord.h
patching file fs/reiser4/crypt.c
patching file fs/reiser4/debug.c
patching file fs/reiser4/debug.h
patching file fs/reiser4/dformat.h
patching file fs/reiser4/dscale.c
patching file fs/reiser4/dscale.h
patching file fs/reiser4/emergency_flush.c
patching file fs/reiser4/emergency_flush.h
patching file fs/reiser4/entd.c
patching file fs/reiser4/entd.h
patching file fs/reiser4/eottl.c
patching file fs/reiser4/estimate.c
patching file fs/reiser4/file_ops.c
patching file fs/reiser4/flush.c
patching file fs/reiser4/flush.h
patching file fs/reiser4/flush_queue.c
patching file fs/reiser4/forward.h
patching file fs/reiser4/init_super.c
patching file fs/reiser4/init_super.h
patching file fs/reiser4/inode.c
patching file fs/reiser4/inode.h
patching file fs/reiser4/inode_ops.c
patching file fs/reiser4/ioctl.h
patching file fs/reiser4/jnode.c
patching file fs/reiser4/jnode.h
patching file fs/reiser4/kassign.c
patching file fs/reiser4/kassign.h
patching file fs/reiser4/kcond.c
patching file fs/reiser4/kcond.h
patching file fs/reiser4/key.c
patching file fs/reiser4/key.h
patching file fs/reiser4/ktxnmgrd.c
patching file fs/reiser4/ktxnmgrd.h
patching file fs/reiser4/lib.h
patching file fs/reiser4/lock.c
patching file fs/reiser4/lock.h
patching file fs/reiser4/oid.c
patching file fs/reiser4/page_cache.c
patching file fs/reiser4/page_cache.h
patching file fs/reiser4/plugin/compress/compress.c
patching file fs/reiser4/plugin/compress/compress.h
patching file fs/reiser4/plugin/compress/lzoconf.h
patching file fs/reiser4/plugin/compress/minilzo.c
patching file fs/reiser4/plugin/compress/minilzo.h
patching file fs/reiser4/plugin/cryptcompress.c
patching file fs/reiser4/plugin/cryptcompress.h
patching file fs/reiser4/plugin/digest.c
patching file fs/reiser4/plugin/dir/dir.c
patching file fs/reiser4/plugin/dir/dir.h
patching file fs/reiser4/plugin/dir/hashed_dir.c
patching file fs/reiser4/plugin/dir/hashed_dir.h
patching file fs/reiser4/plugin/dir/pseudo_dir.c
patching file fs/reiser4/plugin/dir/pseudo_dir.h
patching file fs/reiser4/plugin/disk_format/disk_format.c
patching file fs/reiser4/plugin/disk_format/disk_format.h
patching file fs/reiser4/plugin/disk_format/disk_format40.c
patching file fs/reiser4/plugin/disk_format/disk_format40.h
patching file fs/reiser4/plugin/fibration.c
patching file fs/reiser4/plugin/fibration.h
patching file fs/reiser4/plugin/file/file.c
patching file fs/reiser4/plugin/file/file.h
patching file fs/reiser4/plugin/file/funcs.h
patching file fs/reiser4/plugin/file/invert.c
patching file fs/reiser4/plugin/file/pseudo.c
patching file fs/reiser4/plugin/file/pseudo.h
patching file fs/reiser4/plugin/file/symfile.c
patching file fs/reiser4/plugin/file/tail_conversion.c
patching file fs/reiser4/plugin/hash.c
patching file fs/reiser4/plugin/item/acl.h
patching file fs/reiser4/plugin/item/blackbox.c
patching file fs/reiser4/plugin/item/blackbox.h
patching file fs/reiser4/plugin/item/cde.c
patching file fs/reiser4/plugin/item/cde.h
patching file fs/reiser4/plugin/item/ctail.c
patching file fs/reiser4/plugin/item/ctail.h
patching file fs/reiser4/plugin/item/extent.c
patching file fs/reiser4/plugin/item/extent.h
patching file fs/reiser4/plugin/item/extent_file_ops.c
patching file fs/reiser4/plugin/item/extent_flush_ops.c
patching file fs/reiser4/plugin/item/extent_item_ops.c
patching file fs/reiser4/plugin/item/internal.c
patching file fs/reiser4/plugin/item/internal.h
patching file fs/reiser4/plugin/item/item.c
patching file fs/reiser4/plugin/item/item.h
patching file fs/reiser4/plugin/item/sde.c
patching file fs/reiser4/plugin/item/sde.h
patching file fs/reiser4/plugin/item/static_stat.c
patching file fs/reiser4/plugin/item/static_stat.h
patching file fs/reiser4/plugin/item/tail.c
patching file fs/reiser4/plugin/item/tail.h
patching file fs/reiser4/plugin/node/node.c
patching file fs/reiser4/plugin/node/node.h
patching file fs/reiser4/plugin/node/node40.c
patching file fs/reiser4/plugin/node/node40.h
patching file fs/reiser4/plugin/object.c
patching file fs/reiser4/plugin/object.h
patching file fs/reiser4/plugin/plugin.c
patching file fs/reiser4/plugin/plugin.h
patching file fs/reiser4/plugin/plugin_header.h
patching file fs/reiser4/plugin/plugin_set.c
patching file fs/reiser4/plugin/plugin_set.h
patching file fs/reiser4/plugin/pseudo/pseudo.c
patching file fs/reiser4/plugin/pseudo/pseudo.h
patching file fs/reiser4/plugin/security/perm.c
patching file fs/reiser4/plugin/security/perm.h
patching file fs/reiser4/plugin/space/bitmap.c
patching file fs/reiser4/plugin/space/bitmap.h
patching file fs/reiser4/plugin/space/space_allocator.h
patching file fs/reiser4/plugin/symlink.c
patching file fs/reiser4/plugin/symlink.h
patching file fs/reiser4/plugin/tail_policy.c
patching file fs/reiser4/pool.c
patching file fs/reiser4/pool.h
patching file fs/reiser4/readahead.c
patching file fs/reiser4/readahead.h
patching file fs/reiser4/reiser4.h
patching file fs/reiser4/safe_link.c
patching file fs/reiser4/safe_link.h
patching file fs/reiser4/seal.c
patching file fs/reiser4/seal.h
patching file fs/reiser4/search.c
patching file fs/reiser4/spin_macros.h
patching file fs/reiser4/status_flags.c
patching file fs/reiser4/status_flags.h
patching file fs/reiser4/super.c
patching file fs/reiser4/super.h
patching file fs/reiser4/tap.c
patching file fs/reiser4/tap.h
patching file fs/reiser4/tree.c
patching file fs/reiser4/tree.h
patching file fs/reiser4/tree_mod.c
patching file fs/reiser4/tree_mod.h
patching file fs/reiser4/tree_walk.c
patching file fs/reiser4/tree_walk.h
patching file fs/reiser4/txnmgr.c
patching file fs/reiser4/txnmgr.h
patching file fs/reiser4/type_safe_hash.h
patching file fs/reiser4/type_safe_list.h
patching file fs/reiser4/vfs_ops.c
patching file fs/reiser4/vfs_ops.h
patching file fs/reiser4/wander.c
patching file fs/reiser4/wander.h
patching file fs/reiser4/writeout.h
patching file fs/reiser4/znode.c
patching file fs/reiser4/znode.h
|
| Code: |
# cat include/linux/init_task.h.rej
***************
*** 112,117 ****
.proc_lock = SPIN_LOCK_UNLOCKED, \
.switch_lock = SPIN_LOCK_UNLOCKED, \
.journal_info = NULL, \
}
--- 112,119 ----
.proc_lock = SPIN_LOCK_UNLOCKED, \
.switch_lock = SPIN_LOCK_UNLOCKED, \
.journal_info = NULL, \
+ .private_pages = LIST_HEAD_INIT(tsk.private_pages), \
+ .private_pages_count = 0, \
}
|
| Code: |
kernel/fork.c: In function `dup_task_struct':
kernel/fork.c:157: error: structure has no member named `private_pages'
kernel/fork.c:157: error: structure has no member named `private_pages'
kernel/fork.c:157: error: structure has no member named `private_pages'
kernel/fork.c:157: error: structure has no member named `private_pages'
kernel/fork.c:158: error: structure has no member named `private_pages_count'
make[1]: *** [kernel/fork.o] Error 1
|
|
|
| Back to top |
|
|
Jake
Veteran
Joined: 31 Jul 2003
Posts: 1132
|
| Posted: Wed Mar 23, 2005 7:43 am Post subject: |
|
|
| firephoto wrote: | I tried it with ck2 but it has a hunk failing when it patches and it errors out on the compile.
| Code: |
# cat include/linux/init_task.h.rej
***************
*** 112,117 ****
.proc_lock = SPIN_LOCK_UNLOCKED, \
.switch_lock = SPIN_LOCK_UNLOCKED, \
.journal_info = NULL, \
}
--- 112,119 ----
.proc_lock = SPIN_LOCK_UNLOCKED, \
.switch_lock = SPIN_LOCK_UNLOCKED, \
.journal_info = NULL, \
+ .private_pages = LIST_HEAD_INIT(tsk.private_pages), \
+ .private_pages_count = 0, \
}
|
|
Edit include/linux/init_task.h and insert the two rejected lines somewhere around line 115, after everything else. -ck probably added something after the ".journal_info = NULL, \" line, confusing patch. You don't really need to know any C; it's simple pattern recognition. |
|
| Back to top |
|
|
mpalencia
n00b
Joined: 16 Feb 2005
Posts: 51
Location: Colombia
|
| Posted: Wed Mar 23, 2005 7:46 am Post subject: Reiser4 |
|
|
I managed to patch and compile it without issues, but on a vanilla 2.6.11.1 not on ck2
I think its because ck2 is based on 2.6.11 and this version of reiser4 needs > 2.6.11.1
Im gonna try it with broken-out ck and see if it works.
Cheers:D
_________________
Lucky Livecd 32 and 64 bit versions (reiser4 suppport)
https://forums.gentoo.org/viewtopic-p-2986139.html#2986139
There are 2 things infinite: the universe and human stupidity, but I am not sure about the first one. (Einstein) |
|
| Back to top |
|
|
Jake
Veteran
Joined: 31 Jul 2003
Posts: 1132
|
| Posted: Wed Mar 23, 2005 8:50 am Post subject: Re: Reiser4 |
|
|
| mpalencia wrote: | I managed to patch and compile it without issues, but on a vanilla 2.6.11.1 not on ck2
I think its because ck2 is based on 2.6.11 and this version of reiser4 needs > 2.6.11.1
Im gonna try it with broken-out ck and see if it works.
Cheers:D |
The problem is cfq-ts-21.diff, which includes
| Code: | Index: linux-2.6.11-ck2/include/linux/init_task.h
===================================================================
--- linux-2.6.11-ck2.orig/include/linux/init_task.h 2005-03-09 20:07:30.000000000 +1100
+++ linux-2.6.11-ck2/include/linux/init_task.h 2005-03-09 20:07:31.000000000 +1100
@@ -80,6 +80,7 @@ extern struct group_info init_groups;
.mm = NULL, \
.active_mm = &init_mm, \
.run_list = LIST_HEAD_INIT(tsk.run_list), \
+ .ioprio = 0, \
.time_slice = HZ, \
.tasks = LIST_HEAD_INIT(tsk.tasks), \
.ptrace_children= LIST_HEAD_INIT(tsk.ptrace_children), \
@@ -112,6 +113,7 @@ extern struct group_info init_groups;
.proc_lock = SPIN_LOCK_UNLOCKED, \
.switch_lock = SPIN_LOCK_UNLOCKED, \
.journal_info = NULL, \
+ .fs_excl = ATOMIC_INIT(0), \
} |
Like I said in my reply to firephoto, all you need to do is add the two lines that fail. |
|
| Back to top |
|
|
firephoto
Veteran
Joined: 29 Oct 2003
Posts: 1612
Location: +48° 5' 23.40", -119° 48' 30.00"
|
| Posted: Wed Mar 23, 2005 3:40 pm Post subject: |
|
|
| I tried adding the lines after the journal_info entry and before the atomic one added by ck but that failed too. I just wiped ck2 and am going to try it from a clean ck2, patch, edit the file, then see if it works. I'll post results in a few. |
|
| Back to top |
|
|
firephoto
Veteran
Joined: 29 Oct 2003
Posts: 1612
Location: +48° 5' 23.40", -119° 48' 30.00"
|
| Posted: Wed Mar 23, 2005 3:59 pm Post subject: |
|
|
I edited include/linux/init_task.h to look like this.
| Code: |
.proc_lock = SPIN_LOCK_UNLOCKED, \
.switch_lock = SPIN_LOCK_UNLOCKED, \
.journal_info = NULL, \
.fs_excl = ATOMIC_INIT(0), \
.private_pages = LIST_HEAD_INIT(tsk.private_pages), \
.private_pages_count = 0, \
}
#endif
|
It failed and gave this.
| Code: |
arch/i386/kernel/init_task.c:37: error: unknown field `private_pages' specified in initializer
arch/i386/kernel/init_task.c:37: error: extra brace group at end of initializer
arch/i386/kernel/init_task.c:37: error: (near initialization for `init_task')
arch/i386/kernel/init_task.c:37: error: structure has no member named `private_pages'
arch/i386/kernel/init_task.c:37: error: structure has no member named `private_pages'
arch/i386/kernel/init_task.c:37: warning: excess elements in struct initializer
arch/i386/kernel/init_task.c:37: warning: (near initialization for `init_task')
arch/i386/kernel/init_task.c:37: error: unknown field `private_pages_count' specified in initializer
arch/i386/kernel/init_task.c:37: warning: excess elements in struct initializer
arch/i386/kernel/init_task.c:37: warning: (near initialization for `init_task')
make[1]: *** [arch/i386/kernel/init_task.o] Error 1
|
|
|
| Back to top |
|
|
Jake
Veteran
Joined: 31 Jul 2003
Posts: 1132
|
| Posted: Wed Mar 23, 2005 6:50 pm Post subject: |
|
|
| Rather than trying to fix the -ck conflicts, I recommend avoiding this reiser4 patch. I get oopses whenever I try to unmount a reiser4 filesystem while using this patch, and someone else on the mailing list is also having umount problems. |
|
| Back to top |
|
|
fallow
Bodhisattva
Joined: 08 Jan 2004
Posts: 2208
Location: Poland
|
| Posted: Wed Mar 23, 2005 7:00 pm Post subject: |
|
|
the same - oopses with remounting here.
I tried it with upcoming love release , but in this case ( oopses ) I returend to good working version from 2.6.12-rc1-mm1.
cheers
_________________
"Time is a companion that goes with us on a journey. It reminds us to cherish each moment, because it will never come again. What we leave behind is not as important as how we have lived" J-L. Picard  |
|
| Back to top |
|
|
firephoto
Veteran
Joined: 29 Oct 2003
Posts: 1612
Location: +48° 5' 23.40", -119° 48' 30.00"
|
| Posted: Wed Mar 23, 2005 7:23 pm Post subject: |
|
|
I have the patch working with 2.6.11.4 and I made a new reiser4 partition last night.
I could have sworn I unmounted and remounted the reiser4 partition last night but after reading your guy's oops report I unmounted hdb1 and it gave me a "segmentation fault" then I went to mount it again and it just sat there and I couldn't ctrl-c it or anything so I closed the konsole. I have ktxnmgrd:hdb1:w and ent:hdb1. processes running that I can't kill now.
Here's the dmesg.
| Code: |
Unable to handle kernel paging request at virtual address 4347326f
printing eip:
c01577b1
*pde = 00000000
Oops: 0000 [#1]
PREEMPT
Modules linked in: it87 i2c_sensor i2c_isa nvidia
CPU: 0
EIP: 0060:[<c01577b1>] Tainted: P VLI
EFLAGS: 00210217 (2.6.11.4)
EIP is at invalidate_inode_buffers+0x31/0x70
eax: 4347326b ebx: d4f6e5cc ecx: 4347326b edx: d4f6e4f0
esi: d4f6e4f0 edi: 65735f74 ebp: dc5ddf10 esp: dc5dded8
ds: 007b es: 007b ss: 0068
Process umount (pid: 8199, threadinfo=dc5dc000 task=e5d8e020)
Stack: d4f6e4f0 d4f6e4f0 d4f6e500 c0170196 d4f6e4f0 00000006 00000000 dc5dc000
dc5ddf10 c6c3a1f4 dc5dc000 c0170294 e8c0fa74 dc5ddf10 e36fc6f8 d462fc78
e31215a4 e8c0fa00 c6c3a1f4 c015bd19 e8c0fa00 e8c0fa00 c18e3900 00000000
Call Trace:
[<c0170196>] invalidate_list+0x46/0x100
[<c0170294>] invalidate_inodes+0x44/0x80
[<c015bd19>] generic_shutdown_super+0x59/0x140
[<c015c9bd>] kill_block_super+0x2d/0x50
[<c015bbaa>] deactivate_super+0x5a/0x90
[<c017353f>] sys_umount+0x3f/0x90
[<c0148377>] do_munmap+0x147/0x1c0
[<c0148434>] sys_munmap+0x44/0x70
[<c01735a7>] sys_oldumount+0x17/0x20
[<c0102719>] sysenter_past_esp+0x52/0x75
Code: 74 24 10 89 34 24 e8 cf fa ff ff 85 c0 74 42 8d 9e dc 00 00 00 b8 00 e0 ff ff 21 e0 ff 40 14 8b 86 dc 00 00 00 39 d8 74 17 89 c1 <8b> 51 04 8b 01 89 02 89 50 04 89 09 89 49 04 8b 0b 39 d9 75 eb
<6>note: umount[8199] exited with preempt_count 2
|
|
|
| Back to top |
|
|
superstoned
Guru
Joined: 17 Dec 2004
Posts: 432
|
| Posted: Wed Mar 23, 2005 8:05 pm Post subject: |
|
|
| firephoto wrote: | Hey, I just went looking and found this out myself just now.
From the README:
| Quote: |
03/22/2005
reiser4 for 2.6.11
Changes: major change is that concurrent read/write performance may
increase in some cases, cleanups
|
Going to try it with 2.6.11-ck2 and make a partition of 2 on the extra 80g drive I have that's been spinning for a month doing next to nothing. |
so probably this release is supposed to increase the interactivity a bit? hmmmm... pity its giving problems to some. alas, it'll get fixed soon |
|
| Back to top |
|
|
mpalencia
n00b
Joined: 16 Feb 2005
Posts: 51
Location: Colombia
|
| Posted: Thu Mar 24, 2005 12:07 am Post subject: It Works |
|
|
It Works perfectly here over 13 hours uptime and no problems, i tried unmounting other reiser4 partitions i´ve got and no oopses.
I had the problem earlier, but not on vanilla 2.6.11.5
Right now Im using vanilla 2.6.11.5 and these patches::
Staircase 10.5 (couldnt compile with 10.6)
cddvdm filter patch
defaut-deadline
nvidia compat
No problems at all with these patches, If someone else could give it a shot...
_________________
Lucky Livecd 32 and 64 bit versions (reiser4 suppport)
https://forums.gentoo.org/viewtopic-p-2986139.html#2986139
There are 2 things infinite: the universe and human stupidity, but I am not sure about the first one. (Einstein) |
|
| Back to top |
|
|
chjiang
n00b
Joined: 24 Mar 2005
Posts: 5
|
| Posted: Thu Mar 24, 2005 12:29 am Post subject: |
|
|
| firephoto wrote: | I edited include/linux/init_task.h to look like this.
| Code: |
.proc_lock = SPIN_LOCK_UNLOCKED, \
.switch_lock = SPIN_LOCK_UNLOCKED, \
.journal_info = NULL, \
.fs_excl = ATOMIC_INIT(0), \
.private_pages = LIST_HEAD_INIT(tsk.private_pages), \
.private_pages_count = 0, \
}
#endif
|
It failed and gave this.
| Code: |
arch/i386/kernel/init_task.c:37: error: unknown field `private_pages' specified in initializer
arch/i386/kernel/init_task.c:37: error: extra brace group at end of initializer
arch/i386/kernel/init_task.c:37: error: (near initialization for `init_task')
arch/i386/kernel/init_task.c:37: error: structure has no member named `private_pages'
arch/i386/kernel/init_task.c:37: error: structure has no member named `private_pages'
arch/i386/kernel/init_task.c:37: warning: excess elements in struct initializer
arch/i386/kernel/init_task.c:37: warning: (near initialization for `init_task')
arch/i386/kernel/init_task.c:37: error: unknown field `private_pages_count' specified in initializer
arch/i386/kernel/init_task.c:37: warning: excess elements in struct initializer
arch/i386/kernel/init_task.c:37: warning: (near initialization for `init_task')
make[1]: *** [arch/i386/kernel/init_task.o] Error 1
|
|
The problem is due to the fuzz in sched.h
| Code: | patching file include/linux/sched.h
Hunk #1 succeeded at 486 with fuzz 2 (offset -199 lines). |
You may need to apply this patch
| Code: | --- linux-2.6.11/include/linux/sched.h.orig 2005-03-23 16:25:25.678971656 -0800
+++ linux-2.6.11/include/linux/sched.h 2005-03-23 16:25:59.410843632 -0800
@@ -486,9 +486,6 @@ struct sched_domain {
unsigned long ttwu_wake_affine;
unsigned long ttwu_wake_balance;
#endif
-
- struct list_head private_pages; /* per-process private pages */
- int private_pages_count;
};
#ifdef ARCH_HAS_SCHED_DOMAIN
@@ -702,6 +699,9 @@ struct task_struct {
short il_next;
#endif
atomic_t fs_excl; /* holding fs exclusive resources */
+
+ struct list_head private_pages; /* per-process private pages */
+ int private_pages_count;
};
static inline pid_t process_group(struct task_struct *tsk) |
|
|
| Back to top |
|
|
firephoto
Veteran
Joined: 29 Oct 2003
Posts: 1612
Location: +48° 5' 23.40", -119° 48' 30.00"
|
| Posted: Thu Mar 24, 2005 1:34 am Post subject: |
|
|
| chjiang wrote: |
You may need to apply this patch
| Code: | --- linux-2.6.11/include/linux/sched.h.orig 2005-03-23 16:25:25.678971656 -0800
+++ linux-2.6.11/include/linux/sched.h 2005-03-23 16:25:59.410843632 -0800
@@ -486,9 +486,6 @@ struct sched_domain {
unsigned long ttwu_wake_affine;
unsigned long ttwu_wake_balance;
#endif
-
- struct list_head private_pages; /* per-process private pages */
- int private_pages_count;
};
#ifdef ARCH_HAS_SCHED_DOMAIN
@@ -702,6 +699,9 @@ struct task_struct {
short il_next;
#endif
atomic_t fs_excl; /* holding fs exclusive resources */
+
+ struct list_head private_pages; /* per-process private pages */
+ int private_pages_count;
};
static inline pid_t process_group(struct task_struct *tsk) |
|
The patch wouldn't apply but I made the change by hand and ck2 is compiling now.
So I emerged 2.6.11-ck2, applied the reiser4-for-2.6.11-1 patch, patched sched.h like above, and patched init_task.h like previously mentioned. "Kernel: arch/i386/boot/bzImage is ready", so it seems to have worked, thanks!
The unmount/mount issue I had earlier eventually ended up freezing the whole system and I had to alt+SysRq+SUB it. I had just told portage to use /mnt/reiser4 for portdir so it wasn't a big issue except for the whole system stopping. |
|
| Back to top |
|
|
hpestilence
Tux's lil' helper
Joined: 16 Dec 2004
Posts: 82
|
| Posted: Thu Mar 24, 2005 3:49 am Post subject: |
|
|
My filesystem got corrupted today while updating mozilla. I tried fixing it but the error kept reappearing.
I don't think it was the kernel, I think it was from playing foobillard while emerge -uDa --newuse world. 
But I'm not really sure. 
Luckily, I just reinstalled gentoo a couple of weeks ago to try out reiser4 so no harm done. I'll stick with reiserfs when 2005.0 comes out then make the switch over once some more of the bugs are ironed out in reiser4.
_________________
When the plague descends from Heaven above only Darkness will be your one true ally. |
|
| Back to top |
|
|
superstoned
Guru
Joined: 17 Dec 2004
Posts: 432
|
| Posted: Thu Mar 24, 2005 10:48 am Post subject: |
|
|
| hpestilence wrote: | My filesystem got corrupted today while updating mozilla. I tried fixing it but the error kept reappearing.
I don't think it was the kernel, I think it was from playing foobillard while emerge -uDa --newuse world.
But I'm not really sure.
Luckily, I just reinstalled gentoo a couple of weeks ago to try out reiser4 so no harm done. I'll stick with reiserfs when 2005.0 comes out then make the switch over once some more of the bugs are ironed out in reiser4. |
hmm. maybe you are willing to try to help find the reason why your filesystem got corrupted, and try to get it fixed? I think it is important to use the opportunity to found out what is wrong if a bug in reiser4 is found... maybe someone can tell what to do, to help finding the bug? or just mail to the reiser4 people at namesys. |
|
| Back to top |
|
|
Jake
Veteran
Joined: 31 Jul 2003
Posts: 1132
|
| Posted: Thu Mar 24, 2005 5:42 pm Post subject: |
|
|
The following patch from Vladimir Saveliev seems to fix the umount problem.
| Code: | --- object.c~ 2005-03-23 12:30:52.000000000 +0300
+++ object.c 2005-03-23 13:31:39.000000000 +0300
@@ -976,6 +976,7 @@
* FIXME: this resembles generic_delete_inode
*/
list_del_init(&object->i_list);
+ list_del_init(&object->i_sb_list);
object->i_state |= I_FREEING;
inodes_stat.nr_inodes--;
spin_unlock(&inode_lock); |
|
|
| Back to top |
|
|
fallow
Bodhisattva
Joined: 08 Jan 2004
Posts: 2208
Location: Poland
|
| Posted: Fri Mar 25, 2005 12:53 pm Post subject: |
|
|
| Jake wrote: | The following patch from Vladimir Saveliev seems to fix the umount problem.
| Code: | --- object.c~ 2005-03-23 12:30:52.000000000 +0300
+++ object.c 2005-03-23 13:31:39.000000000 +0300
@@ -976,6 +976,7 @@
* FIXME: this resembles generic_delete_inode
*/
list_del_init(&object->i_list);
+ list_del_init(&object->i_sb_list);
object->i_state |= I_FREEING;
inodes_stat.nr_inodes--;
spin_unlock(&inode_lock); |
|
nice - it's time to test it . thanks.
cheers.
_________________
"Time is a companion that goes with us on a journey. It reminds us to cherish each moment, because it will never come again. What we leave behind is not as important as how we have lived" J-L. Picard |
|
| Back to top |
|
|
firephoto
Veteran
Joined: 29 Oct 2003
Posts: 1612
Location: +48° 5' 23.40", -119° 48' 30.00"
|
| Posted: Fri Mar 25, 2005 7:04 pm Post subject: |
|
|
Looks like the patches are updated and split out now, not sure what's changed besides the object fix.
I'm going to patch on a clean ck2 again and see what the results are, hopefully try to correct the ones that are broke for ck. |
|
| Back to top |
|
|
fallow
Bodhisattva
Joined: 08 Jan 2004
Posts: 2208
Location: Poland
|
|
| Back to top |
|
|
ollonois
Tux's lil' helper
Joined: 01 Sep 2004
Posts: 130
Location: Germany
|
| Posted: Fri Mar 25, 2005 7:57 pm Post subject: |
|
|
I also have this unmountig problem.
Now there are a lot of files yxz.patch
can someone tell me how can I patch the kernel using this files? |
|
| Back to top |
|
|
firephoto
Veteran
Joined: 29 Oct 2003
Posts: 1612
Location: +48° 5' 23.40", -119° 48' 30.00"
|
| Posted: Fri Mar 25, 2005 9:12 pm Post subject: |
|
|
Ok here's a diff on reiser4-perthread-pages.patch so it will apply to 2.6.11-ck2. All the other patches applied ok.
| Code: |
--- reiser4-perthread-pages.patch 2005-03-25 12:57:23.000000000 -0800
+++ reiser4-perthread-pages.patch.ck2 2005-03-25 12:45:34.000000000 -0800
@@ -46,10 +46,10 @@
diff -puN include/linux/init_task.h~reiser4-perthread-pages include/linux/init_task.h
--- linux-2.6.11/include/linux/init_task.h~reiser4-perthread-pages 2005-03-25 18:01:49.000000000 +0300
+++ linux-2.6.11-vs/include/linux/init_task.h 2005-03-25 18:03:22.000000000 +0300
-@@ -112,6 +112,8 @@ extern struct group_info init_groups;
- .proc_lock = SPIN_LOCK_UNLOCKED, \
+@@ -114,6 +114,8 @@ extern struct group_info init_groups;
.switch_lock = SPIN_LOCK_UNLOCKED, \
.journal_info = NULL, \
+ .fs_excl = ATOMIC_INIT(0), \
+ .private_pages = LIST_HEAD_INIT(tsk.private_pages), \
+ .private_pages_count = 0, \
}
@@ -58,7 +58,7 @@
diff -puN include/linux/sched.h~reiser4-perthread-pages include/linux/sched.h
--- linux-2.6.11/include/linux/sched.h~reiser4-perthread-pages 2005-03-25 18:01:49.000000000 +0300
+++ linux-2.6.11-vs/include/linux/sched.h 2005-03-25 18:05:00.000000000 +0300
-@@ -474,6 +474,9 @@ struct sched_domain {
+@@ -486,6 +486,9 @@ struct sched_domain {
unsigned long ttwu_wake_affine;
unsigned long ttwu_wake_balance;
#endif
@@ -68,14 +68,14 @@
};
#ifdef ARCH_HAS_SCHED_DOMAIN
-@@ -685,6 +688,8 @@ struct task_struct {
- struct mempolicy *mempolicy;
+@@ -699,6 +702,8 @@ struct task_struct {
short il_next;
#endif
+ atomic_t fs_excl; /* holding fs exclusive resources */
+ struct list_head private_pages; /* per-process private pages */
+ int private_pages_count;
};
-
+ p
static inline pid_t process_group(struct task_struct *tsk)
diff -puN kernel/fork.c~reiser4-perthread-pages kernel/fork.c
--- linux-2.6.11/kernel/fork.c~reiser4-perthread-pages 2005-03-25 18:01:49.000000000 +0300
|
It has a little fuzz but nothing critical it seems and the kernel compiled ok.
It's from the new patches here.
@ollonois
Just look at the "serie" file and patch in that order. |
|
| Back to top |
|
|
StifflerStealth
Retired Dev
Joined: 03 Jul 2002
Posts: 968
|
| Posted: Sat Mar 26, 2005 6:46 pm Post subject: |
|
|
In the new broken out patches for 2.6.11, was the unmounting bug fix, or do we still need that small fix posted a few posts up?
Thanks.
_________________
Nothing to read in this sig. Move along. |
|
| Back to top |
|
|
Jake
Veteran
Joined: 31 Jul 2003
Posts: 1132
|
| Posted: Sat Mar 26, 2005 9:31 pm Post subject: |
|
|
| StifflerStealth wrote: | In the new broken out patches for 2.6.11, was the unmounting bug fix, or do we still need that small fix posted a few posts up?
Thanks. |
The new reiser4-only patch has the fix integrated. |
|
| Back to top |
|
|
|
Unsupported Software
