linux-kernel - Re: [PATCH v8 2/7] perf/x86/intel: Record branch type

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] [day] [month] [year] [list]

Message-ID: <c325bd11-81d2-9015-eebc-0fda83f00b58@linux.intel.com>
Date:   Thu, 13 Jul 2017 23:15:14 +0800
From:   "Jin, Yao" <yao.jin@...ux.intel.com>
To:     Peter Zijlstra <peterz@...radead.org>
Cc:     acme@...nel.org, jolsa@...nel.org, mingo@...hat.com,
        alexander.shishkin@...ux.intel.com, mpe@...erman.id.au,
        Linux-kernel@...r.kernel.org, ak@...ux.intel.com,
        kan.liang@...el.com, yao.jin@...el.com
Subject: Re: [PATCH v8 2/7] perf/x86/intel: Record branch type



On 7/13/2017 11:06 PM, Jin, Yao wrote:
>
> Sorry, please ignore my previous response.
>
>
> On 7/13/2017 10:31 PM, Peter Zijlstra wrote:
>> On Thu, Jul 13, 2017 at 08:04:14PM +0800, Jin Yao wrote:
>>> +#define X86_BR_TYPE_MAP_MAX    16
>>> +
>>> +static int
>>> +common_branch_type(int type)
>>> +{
>>> +    int i, mask;
>>> +    const int branch_map[X86_BR_TYPE_MAP_MAX] = {
>>> +        PERF_BR_CALL,        /* X86_BR_CALL */
>>> +        PERF_BR_RET,        /* X86_BR_RET */
>>> +        PERF_BR_SYSCALL,    /* X86_BR_SYSCALL */
>>> +        PERF_BR_SYSRET,        /* X86_BR_SYSRET */
>>> +        PERF_BR_UNKNOWN,    /* X86_BR_INT */
>>> +        PERF_BR_UNKNOWN,    /* X86_BR_IRET */
>>> +        PERF_BR_COND,        /* X86_BR_JCC */
>>> +        PERF_BR_UNCOND,        /* X86_BR_JMP */
>>> +        PERF_BR_UNKNOWN,    /* X86_BR_IRQ */
>>> +        PERF_BR_IND_CALL,    /* X86_BR_IND_CALL */
>>> +        PERF_BR_UNKNOWN,    /* X86_BR_ABORT */
>>> +        PERF_BR_UNKNOWN,    /* X86_BR_IN_TX */
>>> +        PERF_BR_UNKNOWN,    /* X86_BR_NO_TX */
>>> +        PERF_BR_CALL,        /* X86_BR_ZERO_CALL */
>>> +        PERF_BR_UNKNOWN,    /* X86_BR_CALL_STACK */
>>> +        PERF_BR_IND,        /* X86_BR_IND_JMP */
>>> +    };
>>> +
>>> +    type >>= 2; /* skip X86_BR_USER and X86_BR_KERNEL */
>>
>>> +    mask = ~(~0 << 1);
>> OCC worthy means of writing: 1
>>
>>> +
>>> +    for (i = 0; i < X86_BR_TYPE_MAP_MAX; i++) {
>>> +        if (type & mask)
>>> +            return branch_map[i];
>>> +
>>> +        type >>= 1;
>>> +    }
>> That is some of the more confused code I've seen in a while :/
>>
>>     if (type)
>>         return branch_map[__ffs(type)];
>>
>> is what you meant to write, no?
>
> Now I understand what you suggest. Yes, that's right.
>
> Do I need to update the patch?
>
> Thanks
> Jin Yao
>

Looks it should be:

if (type)
     return branch_map[__ffs(type) - 1];

>>> +
>>> +    return PERF_BR_UNKNOWN;
>>> +}
>