lkml.org 
[lkml]   [2022]   [Oct]   [21]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH v7 18/43] instrumented.h: add KMSAN support
On October 21, 2022 10:02:05 AM PDT, Alexander Potapenko <glider@google.com> wrote:
>On Fri, Oct 21, 2022 at 8:19 AM youling 257 <youling257@gmail.com> wrote:
>
>> CONFIG_DEBUG_INFO=y
>> CONFIG_AS_HAS_NON_CONST_LEB128=y
>> # CONFIG_DEBUG_INFO_NONE is not set
>> CONFIG_DEBUG_INFO_DWARF_TOOLCHAIN_DEFAULT=y
>> # CONFIG_DEBUG_INFO_DWARF4 is not set
>> # CONFIG_DEBUG_INFO_DWARF5 is not set
>> # CONFIG_DEBUG_INFO_REDUCED is not set
>> # CONFIG_DEBUG_INFO_COMPRESSED is not set
>> # CONFIG_DEBUG_INFO_SPLIT is not set
>> # CONFIG_DEBUG_INFO_BTF is not set
>> # CONFIG_GDB_SCRIPTS is not set
>>
>> perf top still no function name.
>>
>Will it help if you disable CONFIG_RANDOMIZE_BASE?
>(if it doesn't show the symbols, at least we'll be able to figure out the
>offending function by running nm)

Is KALLSYMS needed?

>
>
>>
>> 12.90% [kernel] [k] 0xffffffff833dfa64
>> 3.78% [kernel] [k] 0xffffffff8285b439
>> 3.61% [kernel] [k] 0xffffffff83370254
>> 2.32% [kernel] [k] 0xffffffff8337025b
>> 1.88% bluetooth.default.so [.] 0x000000000000d09d
>>
>> 2022-10-21 15:37 GMT+08:00, Marco Elver <elver@google.com>:
>> > On Thu, 20 Oct 2022 at 23:39, youling 257 <youling257@gmail.com> wrote:
>> >>
>> >> PerfTop: 8253 irqs/sec kernel:75.3% exact: 100.0% lost: 0/0 drop:
>> >> 0/17899 [4000Hz cycles], (all, 8 CPUs)
>> >>
>> ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
>> >>
>> >> 14.87% [kernel] [k] 0xffffffff941d1f37
>> >> 6.71% [kernel] [k] 0xffffffff942016cf
>> >>
>> >> what is 0xffffffff941d1f37?
>> >
>> > You need to build with debug symbols:
>> > CONFIG_DEBUG_INFO_DWARF_TOOLCHAIN_DEFAULT=y
>> >
>> > Then it'll show function names.
>> >
>> >> 2022-10-21 14:16 GMT+08:00, Marco Elver <elver@google.com>:
>> >> > On Thu, 20 Oct 2022 at 22:55, youling 257 <youling257@gmail.com>
>> wrote:
>> >> >>
>> >> >> How to use perf tool?
>> >> >
>> >> > The simplest would be to try just "perf top" - and see which kernel
>> >> > functions consume most CPU cycles. I would suggest you compare both
>> >> > kernels, and see if you can spot a function which uses more cycles% in
>> >> > the problematic kernel.
>> >> >
>> >
>>
>
>


--
Kees Cook

\
 
 \ /
  Last update: 2022-10-21 19:21    [W:0.083 / U:0.716 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site