Skip to contentSkip to conversation

mm snapshot broken-out-2006-08-02-00-27.tar.gz uploaded

DISCUSSION Michal Piotrowski 3 participants 6 messages Aug 2, 2006 – Aug 3, 2006 1 message not in the archive

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

#1 lore raw

Hi,

Quoted 6 lines · On 02/08/06, [email protected] <[email protected]> wrote:

The mm snapshot broken-out-2006-08-02-00-27.tar.gz has been uploaded to

   ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/mm/broken-out-2006-08-02-00-27.tar.gz

It contains the following patches against 2.6.18-rc3:

This was reported, git-cpufreq.patch breaks this.

Aug  2 19:55:52 ltg01-fedora kernel: p4-clockmod: P4/Xeon(TM) CPU
On-Demand Clock Modulation available
Aug  2 19:55:52 ltg01-fedora kernel: ip_tables: (C) 2000-2006
Netfilter Core Team
Aug  2 19:55:52 ltg01-fedora kernel: BUG: warning at
/usr/src/linux-work4/kernel/cpu.c:56/unlock_cpu_hotplug()
Aug  2 19:55:52 ltg01-fedora kernel:  [<c010425e>] show_trace_log_lvl+0x58/0x152
Aug  2 19:55:52 ltg01-fedora kernel:  [<c01049a7>] show_trace+0xd/0x10
Aug  2 19:55:52 ltg01-fedora kernel:  [<c0104a6f>] dump_stack+0x19/0x1b
Aug  2 19:55:52 ltg01-fedora kernel:  [<c013ddca>] unlock_cpu_hotplug+0x43/0x7c
Aug  2 19:55:52 ltg01-fedora kernel:  [<fd94a1e8>]
store_speed+0x8f/0x9b [cpufreq_userspace]
Aug  2 19:55:52 ltg01-fedora kernel:  [<c029f46d>] store+0x37/0x48
Aug  2 19:55:52 ltg01-fedora kernel:  [<c01a87e7>] sysfs_write_file+0xaa/0xd3
Aug  2 19:55:52 ltg01-fedora kernel:  [<c0171e5f>] vfs_write+0xcd/0x179
Aug  2 19:55:52 ltg01-fedora kernel:  [<c0172508>] sys_write+0x3b/0x71
Aug  2 19:55:52 ltg01-fedora kernel:  [<c0103181>] sysenter_past_esp+0x56/0x8d
Aug  2 19:55:52 ltg01-fedora kernel: DWARF2 unwinder stuck at
sysenter_past_esp+0x56/0x8d
Aug  2 19:55:52 ltg01-fedora kernel: Leftover inexact backtrace:
Aug  2 19:55:52 ltg01-fedora kernel:  [<c01049a7>] show_trace+0xd/0x10
Aug  2 19:55:52 ltg01-fedora kernel:  [<c0104a6f>] dump_stack+0x19/0x1b
Aug  2 19:55:52 ltg01-fedora kernel:  [<c013ddca>] unlock_cpu_hotplug+0x43/0x7c
Aug  2 19:55:52 ltg01-fedora kernel:  [<fd94a1e8>]
store_speed+0x8f/0x9b [cpufreq_userspace]
Aug  2 19:55:52 ltg01-fedora kernel:  [<c029f46d>] store+0x37/0x48
Aug  2 19:55:52 ltg01-fedora kernel:  [<c01a87e7>] sysfs_write_file+0xaa/0xd3
Aug  2 19:55:52 ltg01-fedora kernel:  [<c0171e5f>] vfs_write+0xcd/0x179
Aug  2 19:55:52 ltg01-fedora kernel:  [<c0172508>] sys_write+0x3b/0x71
Aug  2 19:55:52 ltg01-fedora kernel:  [<c0103181>] sysenter_past_esp+0x56/0x8d
Aug  2 19:55:52 ltg01-fedora kernel: Netfilter messages via NETLINK v0.30.

Here is something new, previous mm snapshot was fine

Aug  2 19:56:08 ltg01-fedora kernel: ------------[ cut here ]------------
Aug  2 19:56:08 ltg01-fedora kernel: kernel BUG at
/usr/src/linux-work4/mm/shmem.c:1228!
Aug  2 19:56:08 ltg01-fedora kernel: invalid opcode: 0000 [#1]
Aug  2 19:56:08 ltg01-fedora kernel: 4K_STACKS PREEMPT SMP
Aug  2 19:56:08 ltg01-fedora kernel: last sysfs file:
/devices/system/cpu/cpu0/cpufreq/scaling_governor
Aug  2 19:56:08 ltg01-fedora kernel: Modules linked in: ipv6 w83627hf

hwmon_vid hwmon i2c_isa af_packet ip_conntrack_netbios _ns ipt_REJECT xt_state ip_conntrack nfnetlink xt_tcpudp iptable_filter ip_tables x_tables cpufreq_userspace p4_clockmod spe edstep_lib binfmt_misc thermal processor fan container evdev snd_intel8x0 snd_ac97_codec snd_ac97_bus snd_seq_dummy snd_seq_ oss snd_seq_midi_event snd_seq snd_seq_device snd_pcm_oss sk98lin snd_mixer_oss snd_pcm skge snd_timer snd soundcore snd_pag e_alloc i2c_i801 intel_agp ide_cd agpgart cdrom rtc unix

Aug  2 19:56:08 ltg01-fedora kernel: CPU:    1
Aug  2 19:56:08 ltg01-fedora kernel: EIP:    0060:[<c016afed>]    Not
tainted VLI
Aug  2 19:56:08 ltg01-fedora kernel: EFLAGS: 00013202   (2.6.18-rc3-mm1 #99)
Aug  2 19:56:08 ltg01-fedora kernel: EIP is at shmem_nopage+0x40/0xd8
Aug  2 19:56:08 ltg01-fedora kdm[1962]: Display :0 cannot be opened
Aug  2 19:56:08 ltg01-fedora kdm[1962]: Unable to fire up local
display :0; disabling.
Aug  2 19:56:08 ltg01-fedora kernel: eax: 00000001   ebx: 00000000
ecx: c0369848   edx: 00000001
Aug  2 19:56:09 ltg01-fedora kernel: esi: f7125120   edi: f7125120
ebp: f5f2ef28   esp: f5f2ef08
Aug  2 19:56:09 ltg01-fedora kernel: ds: 007b   es: 007b   ss: 0068
Aug  2 19:56:10 ltg01-fedora kernel: Process X (pid: 1964, ti=f5f2e000
task=f5accfa0 task.ti=f5f2e000)
Aug  2 19:56:10 ltg01-fedora kernel: Stack: f5f2ef78 00000000 f5f2ef4c
f5af46e0 00000000 c0372660 f7125120 f7125120
Aug  2 19:56:10 ltg01-fedora kernel:        f5f2ef88 c015f5af 00000000
00000001 00000000 f7125120 c767d774 00000000
Aug  2 19:56:10 ltg01-fedora kernel:        f5b3e000 f5f2ef74 00003246
00000001 00000002 00000000 c0116d05 c767d7d4
Aug  2 19:56:11 ltg01-fedora kernel: Call Trace:
Aug  2 19:56:12 ltg01-fedora kernel:  [<c015f5af>] __handle_mm_fault+0x2fe/0x963
Aug  2 19:56:12 ltg01-fedora kernel:  [<c0116db6>] do_page_fault+0x1be/0x4f0
Aug  2 19:56:13 ltg01-fedora kernel:  [<c0103d85>] error_code+0x39/0x40
Aug  2 19:56:14 ltg01-fedora kernel: DWARF2 unwinder stuck at
error_code+0x39/0x40
Aug  2 19:56:14 ltg01-fedora kernel: Leftover inexact backtrace:
Aug  2 19:56:14 ltg01-fedora kernel:  [<c01043e4>] show_stack_log_lvl+0x8c/0x97
Aug  2 19:56:14 ltg01-fedora kernel:  [<c010456c>] show_registers+0x17d/0x211
Aug  2 19:56:14 ltg01-fedora kernel:  [<c01047c2>] die+0x1c2/0x2d8
Aug  2 19:56:14 ltg01-fedora kernel:  [<c0104954>] do_trap+0x7c/0x96
Aug  2 19:56:14 ltg01-fedora kernel:  [<c01051a5>] do_invalid_op+0x89/0x93
Aug  2 19:56:14 ltg01-fedora kernel:  [<c0103d85>] error_code+0x39/0x40
Aug  2 19:56:15 ltg01-fedora kernel:  [<c015f5af>] __handle_mm_fault+0x2fe/0x963
Aug  2 19:56:15 ltg01-fedora kernel:  [<c0116db6>] do_page_fault+0x1be/0x4f0
Aug  2 19:56:15 ltg01-fedora kernel:  [<c0103d85>] error_code+0x39/0x40
Aug  2 19:56:15 ltg01-fedora kernel: Code: 08 8b 40 28 89 45 ec c7 45
f0 00 00 00 00 8b 56 14 81 f2 00 00 00 04 c1 ea 1a 83
e2 01 b8 48 98 36 c0 e8 1b 29 09 00 85 c0 74 08 <0f> 0b cc 04 89 da 31
c0 2b 5e 04 c1 eb 0c 03 5e 44 89 5d e4 c7
Aug  2 19:56:15 ltg01-fedora kernel: EIP: [<c016afed>]

shmem_nopage+0x40/0xd8 SS:ESP 0068:f5f2ef08

(gdb) list *0xc016afed
0xc016afed is in shmem_nopage (/usr/src/linux-work4/mm/shmem.c:1228).

1223            struct inode *inode = vma->vm_file->f_dentry->d_inode;
1224            struct page *page = NULL;
1225            unsigned long idx;
1226            int error;
1227
1228            BUG_ON(!(vma->vm_flags & VM_CAN_INVALIDATE));
1229
1230            idx = (address - vma->vm_start) >> PAGE_SHIFT;
1231            idx += vma->vm_pgoff;
1232            idx >>= PAGE_CACHE_SHIFT - PAGE_SHIFT;

I will revert fix-the-race-between-invalidate_inode_pages-and-do_no_page.patch fix-the-race-between-invalidate_inode_pages-and-do_no_page-tidy.patch and see what will happen.

http://www.stardust.webpages.pl/files/mm/2.6.18-rc3-mm1/mm-config

Regards,
Michal

Signature · 3 lines
Michal K. K. Piotrowski
LTG - Linux Testers Group
(http://www.stardust.webpages.pl/ltg/wiki/)
#2 lore raw
Quoted 6 lines · Michal Piotrowski
Quoted 5 lines · On 02/08/06, [email protected] <[email protected]> wrote:

The mm snapshot broken-out-2006-08-02-00-27.tar.gz has been uploaded to

   ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/mm/broken-out-2006-08-02-00-27.tar.gz

It contains the following patches against 2.6.18-rc3:

....

Quoted 5 lines

Here is something new, previous mm snapshot was fine

Aug  2 19:56:08 ltg01-fedora kernel: ------------[ cut here ]------------
Aug  2 19:56:08 ltg01-fedora kernel: kernel BUG at
Quoted 7 lines

/usr/src/linux-work4/mm/shmem.c:1228!
1228 BUG_ON(!(vma->vm_flags & VM_CAN_INVALIDATE));

I will revert fix-the-race-between-invalidate_inode_pages-and-do_no_page.patch fix-the-race-between-invalidate_inode_pages-and-do_no_page-tidy.patch and see what will happen.

Rather than reverting those, please add something like (I've not seen this particular tree) the fixup below - it's easily missed that ipc/shm.c has its own reference to shmem_nopage. (Last I heard, the flag was called VM_CAN_INVLD, but it looks like I'm not the only one averse to unpronounceables.)

Hugh

ipc/shm.c +1 -0
@@ -264,6 +264,7 @@ static struct vm_operations_struct shm_v
 	.set_policy = shmem_set_policy,
 	.get_policy = shmem_get_policy,
 #endif
+	.vm_flags = VM_CAN_INVALIDATE,
 };
 
 static int newseg (struct ipc_namespace *ns, key_t key, int shmflg, size_t size)
#3 lore raw

Hi Hugh,

Quoted 21 lines · Hugh Dickins
Quoted 6 lines · Michal Piotrowski
Quoted 5 lines · On 02/08/06, [email protected] <[email protected]> wrote:

The mm snapshot broken-out-2006-08-02-00-27.tar.gz has been uploaded to

   ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/mm/broken-out-2006-08-02-00-27.tar.gz

It contains the following patches against 2.6.18-rc3:

....

Quoted 5 lines

Here is something new, previous mm snapshot was fine

Aug  2 19:56:08 ltg01-fedora kernel: ------------[ cut here ]------------
Aug  2 19:56:08 ltg01-fedora kernel: kernel BUG at
Quoted 7 lines

/usr/src/linux-work4/mm/shmem.c:1228!
1228 BUG_ON(!(vma->vm_flags & VM_CAN_INVALIDATE));

I will revert fix-the-race-between-invalidate_inode_pages-and-do_no_page.patch fix-the-race-between-invalidate_inode_pages-and-do_no_page-tidy.patch and see what will happen.

Quoted 5 lines

Rather than reverting those, please add something like (I've not seen this particular tree) the fixup below - it's easily missed that ipc/shm.c has its own reference to shmem_nopage. (Last I heard, the flag was called VM_CAN_INVLD, but it looks like I'm not the only one averse to unpronounceables.)

Oops is gone. Thanks!

Andrew, can you add this patch to -mm?

Quoted 3 lines

Hugh

Regards,
Michal

Signature · 3 lines
Michal K. K. Piotrowski
LTG - Linux Testers Group
(http://www.stardust.webpages.pl/ltg/wiki/)
#4 lore raw
In reply to a message not in the archive
Quoted 4 lines · On 02/08/06, [email protected] <[email protected]> wrote:

The mm snapshot broken-out-2006-08-02-00-27.tar.gz has been uploaded to

   ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/mm/broken-out-2006-08-02-00-27.tar.gz

There is something wrong with this kernel. I have noticed, that after 1,5 hour some of the keys on my keyboard doesn't work... amarok doesn't want to play music (30 sec gaps between songs etc.), switching between firefox/openoffice takes 1 min. I don't see nothing special in the logs. It is a CPU scheduler problem?

Regards,
Michal

Signature · 3 lines
Michal K. K. Piotrowski
LTG - Linux Testers Group
(http://www.stardust.webpages.pl/ltg/wiki/)
#5 lore raw
Quoted 12 lines · On Thu, 3 Aug 2006 02:00:42 +0200 "Michal Piotrowski"…
Quoted 4 lines · On 02/08/06, [email protected] <[email protected]> wrote:

The mm snapshot broken-out-2006-08-02-00-27.tar.gz has been uploaded to

   ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/mm/broken-out-2006-08-02-00-27.tar.gz

There is something wrong with this kernel. I have noticed, that after 1,5 hour some of the keys on my keyboard doesn't work... amarok doesn't want to play music (30 sec gaps between songs etc.), switching between firefox/openoffice takes 1 min. I don't see nothing special in the logs. It is a CPU scheduler problem?

Could be a timekeeping problem, perhaps. Is it SMP? Is the time-of-day increasing at the right speed? Does `sleep 5' do the right thing?

#6 lore raw
Quoted 9 lines · Andrew Morton
Quoted 6 lines · On Thu, 3 Aug 2006 02:00:42 +0200 "Michal Piotrowski"…
Quoted 4 lines · On 02/08/06, [email protected] <[email protected]> wrote:

The mm snapshot broken-out-2006-08-02-00-27.tar.gz has been uploaded to

   ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/mm/broken-out-2006-08-02-00-27.tar.gz
Quoted 8 lines
Quoted 6 lines

There is something wrong with this kernel. I have noticed, that after 1,5 hour some of the keys on my keyboard doesn't work... amarok doesn't want to play music (30 sec gaps between songs etc.), switching between firefox/openoffice takes 1 min. I don't see nothing special in the logs. It is a CPU scheduler problem?

Could be a timekeeping problem, perhaps. Is it SMP?

Yes, it is.

Quoted 2 lines

Is the time-of-day
increasing at the right speed?

Yes.

Quoted 1 line

Does `sleep 5' do the right thing?

Yes, it does.

time sleep 5

real    0m5.016s
user    0m0.000s
sys     0m0.014s

Problems starts after stress testing (LTP).

Regards,
Michal

Signature · 3 lines
Michal K. K. Piotrowski
LTG - Linux Testers Group
(http://www.stardust.webpages.pl/ltg/wiki/)