Futex BUG in 2.6.18rc2-git7
One of my test machines (single socket core2 duo) running 2.6.18rc2-git7 over night under moderate load threw this, followed by an endless loop of soft lockup timeouts (one exemplar appended)
I assume it is related to the new PI mutexes.
-Andi
----------- [cut here ] --------- [please bite here ] ---------
Kernel BUG at ...v2.6/linux-2.6.18-rc2-git7/kernel/rtmutex_common.h:74 invalid opcode: 0000 [1] SMP CPU 0 Modules linked in: Pid: 23036, comm: ld-linux.so.2 Not tainted 2.6.18-rc2-git7 #7 RIP: 0010:[<ffffffff80247b36>] [<ffffffff80247b36>] rt_mutex_next_owner+0x1a/0x2c RSP: 0000:ffff8100033f5d70 EFLAGS: 00010207 RAX: ffff81003dc712d0 RBX: ffff81003dc712c0 RCX: 0000000000000469 RDX: 0000000000000000 RSI: ffff810031f907e0 RDI: ffff81003dc712d0 RBP: ffff810003cabc48 R08: ffffffff807a4e60 R09: 0000000000000000 R10: ffff8100033f4000 R11: 0000000000000002 R12: 00000000800059fc R13: 000000004013d468 R14: 000000004013d7ac R15: ffff81003dc712d0 FS: 0000000000000000(0000) GS:ffffffff807d5000(0063) knlGS:000000004053dba0 CS: 0010 DS: 002b ES: 002b CR0: 000000008005003b CR2: 000000004000ce80 CR3: 000000002e970000 CR4: 00000000000006e0 Process ld-linux.so.2 (pid: 23036, threadinfo ffff8100033f4000, task ffff810033cf9590) Stack: ffffffff80247064 0000000000000009 0000000000000009 7fffffffffffffff ffff810040012ff4 00000000033f5ef8 ffffffff807a4e58 0000000000000000 0000000000000000 0000000000000000 0000000000000000 0000000000000000 Call Trace: [<ffffffff80247064>] do_futex+0x95a/0xbf5 [<ffffffff8024787d>] compat_sys_futex+0xfd/0x11b [<ffffffff80220136>] ia32_sysret+0x0/0xa DWARF2 unwinder stuck at ia32_sysret+0x0/0xa
Leftover inexact backtrace:
Code: 0f 0b 68 9b b8 54 80 c2 4a 00 48 8b 50 50 48 89 d0 c3 48 83 RIP [<ffffffff80247b36>] rt_mutex_next_owner+0x1a/0x2c RSP <ffff8100033f5d70> <3>BUG: soft lockup detected on CPU#1! Call Trace: [<ffffffff8020ae03>] dump_stack+0x12/0x17 [<ffffffff802520b3>] softlockup_tick+0xdb/0xed [<ffffffff802397f1>] update_process_times+0x42/0x68 [<ffffffff80217e3b>] smp_local_timer_interrupt+0x23/0x47 [<ffffffff80218522>] smp_apic_timer_interrupt+0x41/0x47 [<ffffffff8020a215>] apic_timer_interrupt+0x65/0x6c DWARF2 unwinder stuck at apic_timer_interrupt+0x65/0x6c
Leftover inexact backtrace:
... same soft lockup follows forever...
Quoted 6 lines · Andi Kleen
One of my test machines (single socket core2 duo) running 2.6.18rc2-git7 over night under moderate load threw this, followed by an endless loop of soft lockup timeouts (one exemplar appended)
I assume it is related to the new PI mutexes.
Maybe triggered by this, if it was from wagner.suse.de:
(glibc mainline make check):
GCONV_PATH=/usr/src/packages/BUILD/glibc-2.4/cc-nptl/iconvdata LC_ALL=C /usr/src/packages/BUILD/glibc-2.4/cc-nptl/elf/ld-linux.so.2 --library-path /usr/src/packages/BUILD/glibc-2.4/cc-nptl:/usr/src/packages/BUILD/glibc-2.4/cc-nptl/math:/usr/src/packages/BUILD/glibc-2.4/cc-nptl/elf:/usr/src/packages/BUILD/glibc-2.4/cc-nptl/dlfcn:/usr/src/packages/BUILD/glibc-2.4/cc-nptl/nss:/usr/src/packages/BUILD/glibc-2.4/cc-nptl/nis:/usr/src/packages/BUILD/glibc-2.4/cc-nptl/rt:/usr/src/packages/BUILD/glibc-2.4/cc-nptl/resolv:/usr/src/packages/BUILD/glibc-2.4/cc-nptl/crypt:/usr/src/packages/BUILD/glibc-2.4/cc-nptl/nptl /usr/src/packages/BUILD/glibc-2.4/cc-nptl/nptl/tst-robustpi8 > /usr/src/packages/BUILD/glibc-2.4/cc-nptl/nptl/tst-robustpi8.out
Read from remote host wagner: Connection reset by peer
Quoted 9 lines · Olaf Hering
Quoted 6 lines · Andi Kleen
One of my test machines (single socket core2 duo) running 2.6.18rc2-git7 over night under moderate load threw this, followed by an endless loop of soft lockup timeouts (one exemplar appended)
I assume it is related to the new PI mutexes.
Maybe triggered by this, if it was from wagner.suse.de:
Yes it was that box. So it looks like the new mutex code cannot run the glibc test suite.
-Andi
Quoted 4 lines
Quoted 1 line
/usr/src/packages/BUILD/glibc-2.4/cc-nptl/nptl/tst-robustpi8.outRead from remote host wagner: Connection reset by peer
Quoted 7 lines · Andi Kleen
Quoted 6 lines · Olaf Hering
Quoted 5 lines · Andi Kleen
One of my test machines (single socket core2 duo) running 2.6.18rc2-git7 over night under moderate load threw this, followed by an endless loop of soft lockup timeouts (one exemplar appended)
Quoted 6 lines
Quoted 3 lines
Quoted 1 line
I assume it is related to the new PI mutexes.
Maybe triggered by this, if it was from wagner.suse.de:
Yes it was that box. So it looks like the new mutex code cannot run the glibc test suite.
Can you retest against -rc3-current + the compat fix I sent out earlier today (see also below) ?
Is the glibc the latest CVS version ?
tglx
kernel/futex_compat.c +3 -3
@@ -39,7 +39,7 @@ void compat_exit_robust_list(struct task {struct compat_robust_list_head __user *head = curr->compat_robust_list;struct robust_list __user *entry, *pending; -unsigned int limit =ROBUST_LIST_LIMIT , pi; +unsigned int limit =ROBUST_LIST_LIMIT , pi, pip;compat_uptr_t uentry, upending;compat_long_t futex_offset;
@@ -59,10 +59,10 @@ void compat_exit_robust_list(struct task* if it exists: */ if (fetch_robust_entry (&upending, &pending, - &head->list_op_pending, &pi)) + &head->list_op_pending, &pip))return ;if (upending) -handle_futex_death ((void *)pending + futex_offset, curr, pi); +handle_futex_death ((void *)pending + futex_offset, curr, pip);while (compat_ptr (uentry) != &head->list) {/*
Quoted 1 line · Thomas Gleixner
Is the glibc the latest CVS version ?
Its a snapshot from 2006073023.