linux-kernel - Re: [PATCH v4 2/2] perf record: Update docs regarding the maximum limitation of AUX area

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Fri, 4 Aug 2023 19:47:09 +0800
From:   Shuai Xue <xueshuai@...ux.alibaba.com>
To:     Leo Yan <leo.yan@...aro.org>, James Clark <james.clark@....com>
Cc:     alexander.shishkin@...ux.intel.com, peterz@...radead.org,
        mingo@...hat.com, baolin.wang@...ux.alibaba.com, acme@...nel.org,
        mark.rutland@....com, jolsa@...nel.org, namhyung@...nel.org,
        irogers@...gle.com, adrian.hunter@...el.com,
        linux-perf-users@...r.kernel.org, linux-kernel@...r.kernel.org,
        nathan@...nel.org, bpf@...r.kernel.org
Subject: Re: [PATCH v4 2/2] perf record: Update docs regarding the maximum
 limitation of AUX area



On 2023/8/4 16:46, Leo Yan wrote:
> On Fri, Aug 04, 2023 at 03:29:45PM +0800, Shuai Xue wrote:
>> The maximum AUX area is limited by the page size of the system. Update
>> the documentation to reflect this.
>>
>> Signed-off-by: Shuai Xue <xueshuai@...ux.alibaba.com>
>> ---
>>  tools/perf/Documentation/perf-record.txt | 3 +++
>>  1 file changed, 3 insertions(+)
>>
>> diff --git a/tools/perf/Documentation/perf-record.txt b/tools/perf/Documentation/perf-record.txt
>> index 680396c56bd1..b0ee7b63da0e 100644
>> --- a/tools/perf/Documentation/perf-record.txt
>> +++ b/tools/perf/Documentation/perf-record.txt
>> @@ -292,6 +292,9 @@ OPTIONS
>>  	Also, by adding a comma, the number of mmap pages for AUX
>>  	area tracing can be specified.
>>  
>> +	The maximum AUX area is limited by the page size of the system. For
>> +	example with 4K pages configured, the maximum is 2GiB.
>> +
> 
> This statement is incorrect as it fails to give out prerequisites.
> 
> E.g., on Arm64, for 4KiB, 16KiB or 64KiB base page size, different page
> size has different default values for MAX_ORDER.  Furthermore, MAX_ORDER
> can be set by config ARCH_FORCE_MAX_ORDER, thus we cannot arbitrarily
> say the maximum allocation size is 2GiB for 4KiB page size.
> 

Hi, Leo,

You are right, thank you for point this out.

Maybe we could consider to use a formula to present the avaliable
> maximum buffer size:
> 
>      PAGE_SIZE << MAX_ORDER
>    ( ---------------------- ) * PAGE_SIZE
>       sizeof(page_pointer)
> 
>    PAGE_SIZE << MAX_ORDER : the size of maximal physically
>    contiguous allocations, which is the maximum size can be
>    allocated by slab/slub.

I agree that a formula to present that limitation is more accurate. But as
@James commented in last v3, "Minor nit: I wouldn't expect a Perf tool user
to know what "MAX_ORDER" is", how about to keep both:

  The maximum AUX area is limited by the maximum physically contiguous memory
  allocated from slab/slub. It can be calculated with following formula:


     PAGE_SIZE << MAX_ORDER
    ( ---------------------- ) * PAGE_SIZE
       sizeof(page_pointer)

  For example with 4K pages and MAX_ORDER=10 configured, the maximum AUX area
  is 2GiB.

Thank you.

Best Regards,
Shuai