linux-kernel - Re: [PATCH] perf: Fix intel shared extra msr allocation

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]
Message-ID: <CABPqkBRA0Z-oZfSgdyurtW0=KUbuGJ+uFqiVKiwbt9Wpk1Ph7w@mail.gmail.com>
Date:	Mon, 4 Jun 2012 15:12:49 +0200
From:	Stephane Eranian <eranian@...gle.com>
To:	"Yan, Zheng" <zheng.z.yan@...ux.intel.com>
Cc:	"Yan, Zheng" <zheng.z.yan@...el.com>, a.p.zijlstra@...llo.nl,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH] perf: Fix intel shared extra msr allocation

On Fri, Jun 1, 2012 at 4:11 PM, Yan, Zheng <zheng.z.yan@...ux.intel.com> wrote:
> On Fri, Jun 1, 2012 at 5:35 PM, Stephane Eranian <eranian@...gle.com> wrote:
>> On Fri, Jun 1, 2012 at 5:20 AM, Yan, Zheng <zheng.z.yan@...el.com> wrote:
>>> From: "Yan, Zheng" <zheng.z.yan@...el.com>
>>>
>>> intel_shared_reg_get/put_constraints() can be indirectly called
>>> by validate_group(). In that case, they should avoid modifying
>>> the perf_event date structure because the event can be already
>>> in active state. Otherwise the shared extra msr's reference
>>> count will be left in inconsistent state.
>>>
>> I understand the problem but I am wondering if you actually saw
>> it in real life. The reason I am asking is because  of the way
>> validate_group() collects the events and how they are added
>> to sibling_list. The new event is added at the tail. Thus it will
>> come last, and will get to __intel_shared_reg_get_constraints()
>> last, thus I am wondering if it can really modify the programming
>> on the existing events.
>
> The real problem is from __intel_shared_reg_put_constraints(). it set
> reg->alloc to 0 and decreases fake_cpuc->shared_regs->regs[reg->idx]'s
> reference count. Later when deleting the event,  put_constraints() will find
> reg->alloc is 0 and it won't decrease the shared msr's reference count.
>
> Run 'perf stat --group -a -C 0 -e LLC-loads -e LLC-stores sleep 1" on
> Nehalem can trigger the bug.
>
And what do you see in this particular example?

I'd like to see the results via libpfm4 and /perf_examples/syst_count:
$ sudo ./syst_count -e offcore_response_0:dmnd_data_rd
-eoffcore_response_0:dmnd_rfo  -p -d 10 -c 0

Should see 50% for each event group.

>>
>> See more comments inline.
>>
>>> Signed-off-by: Zheng Yan <zheng.z.yan@...el.com>
>>> ---
>>>  arch/x86/kernel/cpu/perf_event_intel.c |   31 +++++++++++++++++++++++--------
>>>  1 files changed, 23 insertions(+), 8 deletions(-)
>>>
>>> diff --git a/arch/x86/kernel/cpu/perf_event_intel.c b/arch/x86/kernel/cpu/perf_event_intel.c
>>> index 166546e..10840d0 100644
>>> --- a/arch/x86/kernel/cpu/perf_event_intel.c
>>> +++ b/arch/x86/kernel/cpu/perf_event_intel.c
>>> @@ -1119,11 +1119,21 @@ intel_bts_constraints(struct perf_event *event)
>>>        return NULL;
>>>  }
>>>
>>> -static bool intel_try_alt_er(struct perf_event *event, int orig_idx)
>>> +static bool intel_try_alt_er(struct perf_event *event, int *idx,
>>> +                            int orig_idx, bool fake_cpuc)
>>>  {
>>> -       if (!(x86_pmu.er_flags & ERF_HAS_RSP_1))
>>> +       if (!(x86_pmu.er_flags & ERF_HAS_RSP_1) || *idx != orig_idx)
>>>                return false;
>>>
>>> +       /* don't modify the event structure if the cpuc is faked */
>>> +       if (fake_cpuc) {
>>> +               if (*idx == EXTRA_REG_RSP_0)
>>> +                       *idx = EXTRA_REG_RSP_1;
>>> +               else if (*idx == EXTRA_REG_RSP_1)
>>> +                       *idx = EXTRA_REG_RSP_0;
>>> +               return (*idx != orig_idx);
>>> +       }
>>> +
>> I understand that.
>>
>>>        if (event->hw.extra_reg.idx == EXTRA_REG_RSP_0) {
>>>                event->hw.config &= ~INTEL_ARCH_EVENT_MASK;
>>>                event->hw.config |= 0x01bb;
>>> @@ -1139,6 +1149,7 @@ static bool intel_try_alt_er(struct perf_event *event, int orig_idx)
>>>        if (event->hw.extra_reg.idx == orig_idx)
>>>                return false;
>>>
>>> +       *idx = event->hw.extra_reg.idx;
>>>        return true;
>>>  }
>>>
>>> @@ -1155,16 +1166,18 @@ __intel_shared_reg_get_constraints(struct cpu_hw_events *cpuc,
>>>                                   struct hw_perf_event_extra *reg)
>>>  {
>>>        struct event_constraint *c = &emptyconstraint;
>>> +       struct intel_shared_regs *shared_regs = cpuc->shared_regs;
>>>        struct er_account *era;
>>>        unsigned long flags;
>>>        int orig_idx = reg->idx;
>>> +       int idx = orig_idx;
>>>
>>> -       /* already allocated shared msr */
>>> -       if (reg->alloc)
>>> +       /* shared msr is already allocated and cpuc is not faked */
>>> +       if (reg->alloc && shared_regs->core_id != -1)
>>>                return NULL; /* call x86_get_event_constraint() */
>>>
>> I don't understand what you need this stuff. Shared_regs is faked as well.
>
> The event can be in active state, we should avoid clearing reg->alloc.
>
> Regards
> Yan, Zheng
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/