Comment 0 for bug 2048404

Revision history for this message
Tim Gardner (timg-tpi) wrote :

SRU Justification

[Impacy]

  perf/x86/uncore: Don't WARN_ON_ONCE() for a broken discovery table

    The kernel warning message is triggered, when SPR MCC is used.

    [ 17.945331] ------------[ cut here ]------------
    [ 17.946305] WARNING: CPU: 65 PID: 1 at
    arch/x86/events/intel/uncore_discovery.c:184
    intel_uncore_has_discovery_tables+0x4c0/0x65c
    [ 17.946305] Modules linked in:
    [ 17.946305] CPU: 65 PID: 1 Comm: swapper/0 Not tainted
    5.4.17-2136.313.1-X10-2c+ #4

    It's caused by the broken discovery table of UPI.

    The discovery tables are from hardware. Except for dropping the broken
    information, there is nothing Linux can do. Using WARN_ON_ONCE() is
    overkilled.

    Use the pr_info() to replace WARN_ON_ONCE(), and specify what uncore unit
    is dropped and the reason.

[Fix]

Commit 5d515ee40cb57ea5331998f27df7946a69f14dc3 ('perf/x86/uncore: Don't WARN_ON_ONCE() for a broken discovery table')

[Test]

AWS tested

[Regression potential]

Low regression potential. Only the code that prints the warning has been changed.

[Other Info]

SF: #00347830