[LU-16157] KASAN reports an read out of bound during lnet-selftest.

• Type: Bug
• Resolution: Fixed
• Priority: Critical
• Fix Version/s: Lustre 2.16.0
• Severity: 3
• Assignee: Alexey Lyashkov
• Reporter: Alexey Lyashkov

Description

easy to replicate - just build a lustre with with 2.15.0 (without my patches applied).
sysctl -w kernel.panic_on_warn=1
will helps with crash dump obtain.

[ 5715.612789] LNet: 6072:0:(rpc.c:616:srpc_service_add_buffers()) waiting for adding buffer
[ 5716.149797] LNet: 6072:0:(rpc.c:616:srpc_service_add_buffers()) waiting for adding buffer
[ 5716.683091] LNet: 6072:0:(rpc.c:616:srpc_service_add_buffers()) waiting for adding buffer
[ 5716.952287] ==================================================================
[ 5716.958820] BUG: KASAN: slab-out-of-bounds in lstcon_testrpc_prep+0x19e7/0x1bb0 [lnet_selftest]
[ 5716.960248] Read of size 4 at addr ffff8888bbaa866c by task lt-lst/6371
[ 5716.961381]
[ 5716.961667] CPU: 3 PID: 6371 Comm: lt-lst Tainted: G           OE    ---------r-  - 4.18.0-305.25.1.el8_4.x86_64+debug #1
[ 5716.963474] Hardware name: Red Hat KVM, BIOS 1.15.0-2.module_el8.6.0+2880+7d9e3703 04/01/2014
[ 5716.964554] Call Trace:
[ 5716.964970]  dump_stack+0x8e/0xd0
[ 5716.965562]  ? lstcon_testrpc_prep+0x19e7/0x1bb0 [lnet_selftest]
[ 5716.966596]  print_address_description.constprop.5+0x1e/0x230
[ 5716.967577]  ? kmsg_dump_rewind_nolock+0xd9/0xd9
[ 5716.968385]  ? lstcon_testrpc_prep+0x19e7/0x1bb0 [lnet_selftest]
[ 5716.969420]  ? lstcon_testrpc_prep+0x19e7/0x1bb0 [lnet_selftest]
[ 5716.970411]  ? lstcon_testrpc_prep+0x19e7/0x1bb0 [lnet_selftest]
[ 5716.971462]  __kasan_report.cold.7+0x37/0x86
[ 5716.972244]  ? lstcon_testrpc_prep+0x19e7/0x1bb0 [lnet_selftest]
[ 5716.973050]  kasan_report+0x37/0x50
[ 5716.973682]  lstcon_testrpc_prep+0x19e7/0x1bb0 [lnet_selftest]