2.5.67 kernel BUG at fs/buffer.c:2538

2.5.67 kernel BUG at fs/buffer.c:2538

[Steven Cole]

This may be a case of:

"It hurts when I do this:"
"Then don't do that!"

I started the latest security updated package of samba with
kernel 2.5.67, but I had forgotten to set CONFIG_SMB_FS.
The box became unresponsive, and then locked hard. Didn't respond
to pings, etc.  So I had to alt-sysrq-b. Now, samba is broken somehow, 
ERROR: Samba cannot create a SAM SID.
so I can't presently retest with SMB_FS=y and 2.5.67 or with
the vendor kernel (2.4.21-0.13mdk).  Unfortunately I had not yet 
tested the new security updated packages for samba with the vendor kernel.

However dumb it was to start samba without SMB_FS enabled,
the box shouldn't die like this.

I'm going to replace Bamboo with Shrike on that machine today,
so I won't be able to reproduce this anytime soon. (I hope).

Steven


Apr  8 07:30:34 spc1 smb: smbd startup succeeded
Apr  8 07:30:34 spc1 smb: nmbd startup succeeded
Apr  8 07:30:38 spc1 nmbd[8744]: [2003/04/08 07:30:38, 0] nmbd/nmbd_responserecordsdb.c:find_response_record(235) 
Apr  8 07:30:38 spc1 nmbd[8744]:   find_response_record: response packet id 19534 received with no matching record. 
Apr  8 07:30:38 spc1 nmbd[8744]: [2003/04/08 07:30:38, 0] nmbd/nmbd_responserecordsdb.c:find_response_record(235) 
Apr  8 07:30:38 spc1 nmbd[8744]:   find_response_record: response packet id 19535 received with no matching record. 
Apr  8 07:30:38 spc1 su(pam_unix)[8691]: session closed for user root
Apr  8 07:31:00 spc1 CROND[8752]: (mail) CMD (/usr/bin/python -S /var/lib/mailman/cron/qrunner) 
Apr  8 07:31:12 spc1 kernel: ------------[ cut here ]------------
Apr  8 07:31:12 spc1 kernel: kernel BUG at fs/buffer.c:2538!
Apr  8 07:31:12 spc1 kernel: invalid operand: 0000 [#1]
Apr  8 07:31:12 spc1 kernel: CPU:    0
Apr  8 07:31:12 spc1 kernel: EIP:    0060:[<c014b937>]    Not tainted
Apr  8 07:31:12 spc1 kernel: EFLAGS: 00010202
Apr  8 07:31:12 spc1 kernel: EIP is at submit_bh+0x1b7/0x1e0
Apr  8 07:31:12 spc1 kernel: eax: 00000010   ebx: c0b3f92c   ecx: 00000000   edx: cfcfd26c
Apr  8 07:31:12 spc1 kernel: esi: 00000002   edi: 00000001   ebp: c0b3f92c   esp: cfc39e10
Apr  8 07:31:12 spc1 kernel: ds: 007b   es: 007b   ss: 0068
Apr  8 07:31:12 spc1 kernel: Process kjournald (pid: 16, threadinfo=cfc38000 task=cfc378a0)
Apr  8 07:31:12 spc1 kernel: Stack: 00000010 00000001 000092b0 00000000 c69e8260 00000001 00000002 00000001 
Apr  8 07:31:12 spc1 kernel:        c0b3f92c 00000002 00000001 00000001 c014b9cc 00000001 c0b3f92c c0b3fc38 
Apr  8 07:31:12 spc1 kernel:        cfc39ea0 c38538e0 cea92dc0 c01906a6 00000001 00000004 cfc39ea0 00000cfc 
Apr  8 07:31:12 spc1 kernel: Call Trace:
Apr  8 07:31:12 spc1 kernel:  [<c014b9cc>] ll_rw_block+0x6c/0x90
Apr  8 07:31:12 spc1 kernel:  [<c01906a6>] journal_commit_transaction+0x1056/0x1208
Apr  8 07:31:12 spc1 kernel:  [<c01157c9>] schedule+0x1a9/0x3b0
Apr  8 07:31:12 spc1 kernel:  [<c010b77a>] do_IRQ+0xba/0x120
Apr  8 07:31:12 spc1 kernel:  [<c0115a20>] default_wake_function+0x0/0x20
Apr  8 07:31:12 spc1 kernel:  [<c01929a0>] kjournald+0x130/0x1e0
Apr  8 07:31:12 spc1 kernel:  [<c0109192>] ret_from_fork+0x6/0x14
Apr  8 07:31:12 spc1 kernel:  [<c0192850>] commit_timeout+0x0/0x10
Apr  8 07:31:12 spc1 kernel:  [<c0192870>] kjournald+0x0/0x1e0
Apr  8 07:31:12 spc1 kernel:  [<c01071dd>] kernel_thread_helper+0x5/0x18
Apr  8 07:31:12 spc1 kernel: 
Apr  8 07:31:12 spc1 kernel: Code: 0f 0b ea 09 26 27 38 c0 e9 6d fe ff ff 0f 0b e9 09 26 27 38 
Apr  8 07:31:33 spc1 smbd[8734]: [2003/04/08 07:31:33, 0] smbd/server.c:open_sockets(238) 
Apr  8 07:31:33 spc1 smbd[8734]:   Got SIGHUP 
Apr  8 07:31:33 spc1 smb: smbd -HUP succeeded

Re: 2.5.67 kernel BUG at fs/buffer.c:2538

[Helge Hafting]

On Tue, Apr 08, 2003 at 09:27:33AM -0600, Steven Cole wrote:
> This may be a case of:
> 
> "It hurts when I do this:"
> "Then don't do that!"
> 
> I started the latest security updated package of samba with
> kernel 2.5.67, but I had forgotten to set CONFIG_SMB_FS.
> The box became unresponsive, and then locked hard. Didn't respond
> to pings, etc.  So I had to alt-sysrq-b. Now, samba is broken somehow, 
> ERROR: Samba cannot create a SAM SID.
> so I can't presently retest with SMB_FS=y and 2.5.67 or with
> the vendor kernel (2.4.21-0.13mdk). 

I tried starting samba on 2.5.66-mm3, and it died the same way,
even though I had smbfs compiled in.  I guess smbfs in 2.5 is a bit
shaky, at least in some configurations.  Do you too have
preempt?

 Unfortunately I had not yet 
> tested the new security updated packages for samba with the vendor kernel.
> 
> However dumb it was to start samba without SMB_FS enabled,
> the box shouldn't die like this.

Definitely not.  Samba software is supposed to simply fail when
the kernel don't support it.
 
Helge Hafting

Re: 2.5.67 kernel BUG at fs/buffer.c:2538

[Steven Cole]

On Tue, 2003-04-08 at 10:30, Helge Hafting wrote:
> On Tue, Apr 08, 2003 at 09:27:33AM -0600, Steven Cole wrote:
> > This may be a case of:
> > 
> > "It hurts when I do this:"
> > "Then don't do that!"
> > 
> > I started the latest security updated package of samba with
> > kernel 2.5.67, but I had forgotten to set CONFIG_SMB_FS.
> > The box became unresponsive, and then locked hard. Didn't respond
> > to pings, etc.  So I had to alt-sysrq-b. Now, samba is broken somehow, 
> > ERROR: Samba cannot create a SAM SID.
> > so I can't presently retest with SMB_FS=y and 2.5.67 or with
> > the vendor kernel (2.4.21-0.13mdk). 
> 
> I tried starting samba on 2.5.66-mm3, and it died the same way,
> even though I had smbfs compiled in.  I guess smbfs in 2.5 is a bit
> shaky, at least in some configurations.  Do you too have
> preempt?

Yes, PREEMPT=y.  For what it's worth, here is the
output of grep ^CONFIG .config

CONFIG_X86=y
CONFIG_MMU=y
CONFIG_UID16=y
CONFIG_GENERIC_ISA_DMA=y
CONFIG_EXPERIMENTAL=y
CONFIG_SWAP=y
CONFIG_SYSVIPC=y
CONFIG_SYSCTL=y
CONFIG_LOG_BUF_SHIFT=14
CONFIG_X86_PC=y
CONFIG_MPENTIUMIII=y
CONFIG_X86_CMPXCHG=y
CONFIG_X86_XADD=y
CONFIG_X86_L1_CACHE_SHIFT=5
CONFIG_RWSEM_XCHGADD_ALGORITHM=y
CONFIG_X86_WP_WORKS_OK=y
CONFIG_X86_INVLPG=y
CONFIG_X86_BSWAP=y
CONFIG_X86_POPAD_OK=y
CONFIG_X86_GOOD_APIC=y
CONFIG_X86_INTEL_USERCOPY=y
CONFIG_X86_USE_PPRO_CHECKSUM=y
CONFIG_X86_PREFETCH=y
CONFIG_PREEMPT=y
CONFIG_X86_UP_APIC=y
CONFIG_X86_UP_IOAPIC=y
CONFIG_X86_LOCAL_APIC=y
CONFIG_X86_IO_APIC=y
CONFIG_X86_TSC=y
CONFIG_X86_MCE=y
CONFIG_NOHIGHMEM=y
CONFIG_HAVE_DEC_LOCK=y
CONFIG_PM=y
CONFIG_PCI=y
CONFIG_PCI_GOANY=y
CONFIG_PCI_BIOS=y
CONFIG_PCI_DIRECT=y
CONFIG_PCI_NAMES=y
CONFIG_ISA=y
CONFIG_KCORE_ELF=y
CONFIG_BINFMT_AOUT=y
CONFIG_BINFMT_ELF=y
CONFIG_BINFMT_MISC=y
CONFIG_PARPORT=y
CONFIG_PARPORT_PC=y
CONFIG_PNP=y
CONFIG_ISAPNP=y
CONFIG_BLK_DEV_FD=y
CONFIG_LBD=y
CONFIG_IDE=y
CONFIG_BLK_DEV_IDE=y
CONFIG_BLK_DEV_IDEDISK=y
CONFIG_IDEDISK_MULTI_MODE=y
CONFIG_BLK_DEV_IDECD=y
CONFIG_BLK_DEV_CMD640=y
CONFIG_BLK_DEV_IDEPCI=y
CONFIG_BLK_DEV_GENERIC=y
CONFIG_IDEPCI_SHARE_IRQ=y
CONFIG_BLK_DEV_IDEDMA_PCI=y
CONFIG_IDEDMA_PCI_AUTO=y
CONFIG_BLK_DEV_IDEDMA=y
CONFIG_BLK_DEV_ADMA=y
CONFIG_BLK_DEV_PIIX=y
CONFIG_BLK_DEV_RZ1000=y
CONFIG_IDEDMA_AUTO=y
CONFIG_BLK_DEV_IDE_MODES=y
CONFIG_SCSI=y
CONFIG_CHR_DEV_SG=y
CONFIG_SCSI_MULTI_LUN=y
CONFIG_SCSI_CONSTANTS=y
CONFIG_SCSI_SYM53C8XX=y
CONFIG_SCSI_NCR53C8XX_DEFAULT_TAGS=4
CONFIG_SCSI_NCR53C8XX_MAX_TAGS=32
CONFIG_SCSI_NCR53C8XX_SYNC=20
CONFIG_NET=y
CONFIG_PACKET=y
CONFIG_UNIX=y
CONFIG_INET=y
CONFIG_IP_MULTICAST=y
CONFIG_IPV6_SCTP__=y
CONFIG_NETDEVICES=y
CONFIG_DUMMY=y
CONFIG_NET_ETHERNET=y
CONFIG_NET_VENDOR_3COM=y
CONFIG_VORTEX=y
CONFIG_NET_PCI=y
CONFIG_EEPRO100=y
CONFIG_INPUT=y
CONFIG_INPUT_MOUSEDEV=y
CONFIG_INPUT_MOUSEDEV_PSAUX=y
CONFIG_INPUT_MOUSEDEV_SCREEN_X=1024
CONFIG_INPUT_MOUSEDEV_SCREEN_Y=768
CONFIG_SOUND_GAMEPORT=y
CONFIG_SERIO=y
CONFIG_SERIO_I8042=y
CONFIG_INPUT_KEYBOARD=y
CONFIG_KEYBOARD_ATKBD=y
CONFIG_KEYBOARD_XTKBD=y
CONFIG_INPUT_MOUSE=y
CONFIG_MOUSE_PS2=y
CONFIG_VT=y
CONFIG_VT_CONSOLE=y
CONFIG_HW_CONSOLE=y
CONFIG_UNIX98_PTYS=y
CONFIG_UNIX98_PTY_COUNT=256
CONFIG_PRINTER=y
CONFIG_AGP=y
CONFIG_AGP_INTEL=y
CONFIG_AGP_VIA=y
CONFIG_AGP_AMD=y
CONFIG_AGP_SIS=y
CONFIG_AGP_ALI=y
CONFIG_DRM=y
CONFIG_HANGCHECK_TIMER=y
CONFIG_EXT2_FS=y
CONFIG_EXT3_FS=y
CONFIG_EXT3_FS_XATTR=y
CONFIG_EXT3_FS_POSIX_ACL=y
CONFIG_JBD=y
CONFIG_FS_MBCACHE=y
CONFIG_REISERFS_FS=y
CONFIG_JFS_FS=y
CONFIG_FS_POSIX_ACL=y
CONFIG_XFS_FS=y
CONFIG_AUTOFS4_FS=y
CONFIG_ISO9660_FS=y
CONFIG_PROC_FS=y
CONFIG_DEVFS_FS=y
CONFIG_DEVFS_MOUNT=y
CONFIG_DEVPTS_FS=y
CONFIG_TMPFS=y
CONFIG_RAMFS=y
CONFIG_NFS_FS=y
CONFIG_NFSD=y
CONFIG_LOCKD=y
CONFIG_EXPORTFS=y
CONFIG_SUNRPC=y
CONFIG_SMB_FS=y
CONFIG_SMB_NLS_DEFAULT=y
CONFIG_SMB_NLS_REMOTE="n"
CONFIG_MSDOS_PARTITION=y
CONFIG_SMB_NLS=y
CONFIG_NLS=y
CONFIG_NLS_DEFAULT="iso8859-1"
CONFIG_NLS_CODEPAGE_850=y
CONFIG_NLS_ISO8859_1=y
CONFIG_FB=y
CONFIG_VGA_CONSOLE=y
CONFIG_DUMMY_CONSOLE=y
CONFIG_LOGO=y
CONFIG_LOGO_LINUX_MONO=y
CONFIG_LOGO_LINUX_VGA16=y
CONFIG_LOGO_LINUX_CLUT224=y
CONFIG_USB=y
CONFIG_USB_STORAGE=y
CONFIG_DEBUG_KERNEL=y
CONFIG_MAGIC_SYSRQ=y
CONFIG_KALLSYMS=y
CONFIG_X86_EXTRA_IRQS=y
CONFIG_X86_FIND_SMP_CONFIG=y
CONFIG_X86_MPPARSE=y
CONFIG_X86_BIOS_REBOOT=y

Re: 2.5.67 kernel BUG at fs/buffer.c:2538

[Steven Cole]

On Tue, 2003-04-08 at 09:27, Steven Cole wrote:
> This may be a case of:
> 
> "It hurts when I do this:"
> "Then don't do that!"
> 
> I started the latest security updated package of samba with
> kernel 2.5.67, but I had forgotten to set CONFIG_SMB_FS.
> The box became unresponsive, and then locked hard. Didn't respond
> to pings, etc.  So I had to alt-sysrq-b. Now, samba is broken somehow, 
> ERROR: Samba cannot create a SAM SID.
> so I can't presently retest with SMB_FS=y and 2.5.67 or with
> the vendor kernel (2.4.21-0.13mdk).  Unfortunately I had not yet 
> tested the new security updated packages for samba with the vendor kernel.
> 
> However dumb it was to start samba without SMB_FS enabled,
> the box shouldn't die like this.
> 
> I'm going to replace Bamboo with Shrike on that machine today,
> so I won't be able to reproduce this anytime soon. (I hope).
> 
Replying to myself with an update.  After reading the fs/Kconfig help
for SMB_FS I now realize that having that compiled in or not was 
orthogonal to starting the samba server.

I installed Redhat 9 (Shrike) on the test box, and tested samba with
both the vendor kernel and 2.5.67 (with and without SMB_FS before I
realized that was not the issue).  It all worked.  Then I upgraded
the samba packages (security updates) and tested everything again, and
it still all works. 

I cannot reproduce the kernel BUG at fs/buffer.c:2538, at least not yet.

Steven