> Roland Dreier <[email protected]> wrote:
>> Can you be explicit about the race you're worried about?
Roland,
This patch was made after we got the below report from the field.
Dotan, can we get Roland access to the full vmcore image?
Here's the report I got:
"[...] box panic'ed when trying to find a completion queue. There is
no corruption but there is a possible race which could result in
mlx4_cq_completion getting wrong height of the radix tree and
following a bit too deep into the chains. In the other code which uses
this radix tree the access is protected by the lock but
mlx4_cq_completion is running win the interrupt context and cannot
take locks, so instead it run without any protection whatsoever."
The below is the stack trace, the kernel is an RHEL5
2.6.18-274.something, so in that respect, maybe your comment on the
radix tree RCU doesn't take effect there? you can see the problem was
somewhere into the radix tree code.
crash> bt
PID: 8178 TASK: ffff810b91a52400 CPU: 11 COMMAND: "universal_consu"
#0 [ffff81182fe3bc70] crash_kexec at ffffffff800ba60c
#1 [ffff81182fe3bd30] __die at ffffffff80069047
#2 [ffff81182fe3bd70] die at ffffffff8007075b
#3 [ffff81182fe3bda0] do_general_protection at ffffffff80069375
#4 [ffff81182fe3bde0] error_exit at ffffffff80060e9d
[exception RIP: radix_tree_lookup+38]
RIP: ffffffff8016405f RSP: ffff81182fe3be90 RFLAGS: 00210002
RAX: 6b6b6b6b6b6b6b8b RBX: ffff810c2fb90000 RCX: 0000000000000006
RDX: 0000000000000001 RSI: 00000000000000c0 RDI: 6b6b6b6b6b6b6b6b
RBP: 00000000000000c0 R8: 0000000000010000 R9: 000000000920ea94
R10: 00000000000000b3 R11: ffffffff800c7ea5 R12: 00000000000000b3
R13: ffff810c2b15a280 R14: ffff810b7a98ff58 R15: 0000000000000000
ORIG_RAX: ffffffffffffffff CS: 0010 SS: 0000
#5 [ffff81182fe3be90] mlx4_cq_completion at ffffffff886f2367
#6 [ffff81182fe3beb0] mlx4_eq_int at ffffffff886f2bd3
#7 [ffff81182fe3bf10] mlx4_msi_x_interrupt at ffffffff886f3078
#8 [ffff81182fe3bf20] handle_IRQ_event at ffffffff8001167c
#9 [ffff81182fe3bf50] __do_IRQ at ffffffff800c7f40
#10 [ffff81182fe3bf90] do_IRQ at ffffffff80071659
--- <IRQ stack> ---
#11 [ffff810b7a98ff58] ret_from_intr at ffffffff80060652
RIP: 00000000f5cd8859 RSP: 00000000f3eafa28 RFLAGS: 00200202
RAX: 0000000000000000 RBX: 00000000f3eafaf8 RCX: 000000000b6d4f00
RDX: 000000000000001c RSI: 000000000000001c RDI: 0000000000000000
RBP: ffff810bf1a0a120 R8: 0000000000000000 R9: 0000000000000000
R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000000001
R13: 0000000000000000 R14: ffff810b9e7d5bf0 R15: 0000000000000000
ORIG_RAX: ffffffffffffff4c CS: 0023 SS: 002b
crash>
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
