On Sat, Apr 9, 2011 at 4:04 PM, Sergei Trofimovich <sly...@gmail.com> wrote:
> I've decided to use UML when caught some hard to
> debug OOpses on btrfs.
>
> The first attempt to use btrfs on UML gave me stable
> UML crash. I suspect it's a UML's problem.
>
> So I have some questions here (I'm on x86_64,
> 2.6.39-rc2):
>
> 1. (major one) BUG_ON trace in UML does not look
> as in real kernel. ud2 handler does not show us nice
> backtrace, but calls some suspicious handler.
>
> /mnt/btr # touch asd-`seq 1 10000`
> [ 95.540000] Kernel panic - not syncing: Kernel mode signal 4
> [ 95.540000] Call Trace:
> [ 95.540000] 602d9908: [<60227a48>] panic+0xea/0x1e0
> [ 95.540000] 602d99b8: [<600342ed>] do_softirq+0x4d/0x70
> [ 95.540000] 602d9a08: [<60016947>] relay_signal+0x87/0xa0
> [ 95.540000] 602d9a18: [<60021f70>] set_signals+0x30/0x40
> [ 95.540000] 602d9a38: [<60021df9>] sig_handler_common+0x59/0xd0
> [ 95.540000] 602d9a58: [<60021f2c>] real_alarm_handler+0x3c/0x50
> [ 95.540000] 602d9ae0: [<6018c0a0>] strncpy+0x20/0x30
> [ 95.540000] 602d9b78: [<6002200f>] sig_handler+0x3f/0x50
> [ 95.540000] 602d9b98: [<600222a1>] handle_signal+0x71/0xb0
> [ 95.540000] 602d9be8: [<60023630>] hard_handler+0x10/0x20
> [ 95.540000] 602d9ca8: [<6011263c>]
> lookup_inline_extent_backref+0x2dc/0x3f0
> [ 95.540000]
> [ 95.540000]
> [ 95.540000] Pid: 1, comm: sh Not tainted 2.6.39-rc2+
>
> Here we have BUG_ON in lookup_inline_extent_backref+0x2dc/0x3f0, but that
> pesky
> hard_handler. Is it an UML bug or known and expected behaviour?
Can you please test the attached patch?
The trace was broken long before I started working on UML.
> Now I use 'gdb -p $pid' to get nicer backtrace:
>
> Program received signal SIGILL, Illegal instruction.
> lookup_inline_extent_backref (trans=0x70c6e000, root=<value optimized out>,
> path=0x70c82000,
> ref_ret=<value optimized out>, bytenr=<value optimized out>,
> num_bytes=<value optimized out>, parent=0,
> root_objectid=5, owner=0, offset=0, insert=0) at
> /home/slyfox/linux-2.6/fs/btrfs/extent-tree.c:1399
> 1399 BUG_ON(ret);
>
> Can UML call the same ud2 handler as the real handler? We would see
> something like:
>
> [ 155.038388] kernel BUG at /mnt/archive/src/linux-2.6/fs/btrfs/inode.c:2967!
> [ 155.038643] invalid opcode: 0000 [#1] PREEMPT SMP
>
> How does UML implement running userspace processes and kernel threads in it?
> I guess it's a problem, why UML can't distinct where ud2 comes from and call
> proper handler.
>
> 2. btrfs in UML does not work for me at all. It corrupts data right off
> the start. I only need to format the filesystem, touch there ~20
> files and I get corrupted FS:
>
> I run the following tiny script in UML:
>
> /tools/mkfs.btrfs /dev/ubda
> mount /mnt/btr/
> cd /mnt/btr/
> touch `seq 1 11`
> ls
>
> The result is:
> / # ./kill_btr
>
> WARNING! - Btrfs v0.19-35-g1b444cd-dirty IS EXPERIMENTAL
> WARNING! - see http://btrfs.wiki.kernel.org before using
>
> fs created label (null) on /dev/ubda
> nodesize 4096 leafsize 4096 sectorsize 4096 size 1.91GB
> Btrfs v0.19-35-g1b444cd-dirty
> [ 2.770000] device fsid f2407093ae1f3a78-9e203194e541bfbd devid 1 transid
> 7 /dev/ubda
> [ 2.770000] btrfS: invalid dir item name len: 12594
> [ 2.770000] btrfS: invalid dir item name len: 0
> [ 2.770000] btrfS: invalid dir item name len: 0
> 11
>
> Corruption. Only last file is seen. Later, after unmount I get an OOps.
>
> It does not happen in real OS. I've enabled all the corruption
> detecting tehniques in the kernel config and it didn't change picture.
> I can share .config if you like. kernel builds only 4 minutes on my slow
> laptop.
>
> Is there serious problems in UML block device handling?
>
> I run UML this way:
> ./vmlinux \
> ubd0=$(pwd)/btr.img \
> root=/dev/root \
> rootflags="$(pwd)/root" rw rootfstype=hostfs \
> mem=256M init=/init \
>
> btr.img is a 2048000000 bytes sized file. rootfs contains statically linked
> busybox and latest statically linked btrfs-progs. I can share my root if
> you are interested. It's tiny (btrfs-progs tree is mounted from hostfs):
> $ find root/ -type f
> root/bin/busybox
> root/bin/strace
> root/kill_btr
> root/etc/fstab
> root/init
>
> Thank you!
>
> --
>
> Sergei
>
> ------------------------------------------------------------------------------
> Xperia(TM) PLAY
> It's a major breakthrough. An authentic gaming
> smartphone on the nation's most reliable network.
> And it wants your games.
> http://p.sf.net/sfu/verizon-sfdev
> _______________________________________________
> User-mode-linux-devel mailing list
> User-mode-linux-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel
>
>
--
Thanks,
//richard
From d3808c146d987e0add9e793f56f1d58883eb5c7a Mon Sep 17 00:00:00 2001
From: Richard Weinberger <rich...@nod.at>
Date: Tue, 12 Apr 2011 00:53:32 +0200
Subject: [PATCH] um: Fix call tracer and bug handler
Commit 1de1502c (x86, um: now we can get rid of trivial uml headers)
removed accidentally bug.h which broke UML's call tracer and
bug handler.
Since v2.6.28-rc1 the call trace was incomplete and nobody noticed. >:)
Signed-off-by: Richard Weinberger <rich...@nod.at>
---
arch/um/include/asm/bug.h | 6 ++++++
1 files changed, 6 insertions(+), 0 deletions(-)
create mode 100644 arch/um/include/asm/bug.h
diff --git a/arch/um/include/asm/bug.h b/arch/um/include/asm/bug.h
new file mode 100644
index 0000000..9e33b86
--- /dev/null
+++ b/arch/um/include/asm/bug.h
@@ -0,0 +1,6 @@
+#ifndef __UM_BUG_H
+#define __UM_BUG_H
+
+#include <asm-generic/bug.h>
+
+#endif
--
1.6.6.1
------------------------------------------------------------------------------
Forrester Wave Report - Recovery time is now measured in hours and minutes
not days. Key insights are discussed in the 2010 Forrester Wave Report as
part of an in-depth evaluation of disaster recovery service providers.
Forrester found the best-in-class provider in terms of services and vision.
Read this report now! http://p.sf.net/sfu/ibm-webcastpromo
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel
