On 17/09/13 23:23, Vince Weaver wrote:
>
> This patch somehow breaks the perf-ABI.
>
> If I take a program that reads "mmap->cap_usr_rdpmc" and compile it
> against the new header with this change (say from 3.12-rc1)
> and then run it on an old kernel (say 3.11) then I get "0" for
> cap_usr_rdpmc.
>
> If I take the same program and recompile against the old (without this
> patch) header and run it on 3.11, I get the expected "1" value.
>
> So something about this changed the bit pattern in an incompatible
> fashion.
cap_usr_time and cap_usr_rdpmc were occupying the same bit position i.e. bit 0
That means that cap_usr_time and cap_usr_rdpmc were both unreliable.
If you look at the logic:
void arch_perf_update_userpage(struct perf_event_mmap_page *userpg, u64 now)
{
userpg->cap_usr_time = 0;
userpg->cap_usr_time_zero = 0;
userpg->cap_usr_rdpmc = x86_pmu.attr_rdpmc;
userpg->pmc_width = x86_pmu.cntval_bits;
if (!boot_cpu_has(X86_FEATURE_CONSTANT_TSC))
return;
if (!boot_cpu_has(X86_FEATURE_NONSTOP_TSC))
return;
userpg->cap_usr_time = 1;
userpg->time_mult = this_cpu_read(cyc2ns);
userpg->time_shift = CYC2NS_SCALE_FACTOR;
userpg->time_offset = this_cpu_read(cyc2ns_offset) - now;
if (sched_clock_stable && !check_tsc_disabled()) {
userpg->cap_usr_time_zero = 1;
userpg->time_zero = this_cpu_read(cyc2ns_offset);
}
}
The incorrect union caused 2 bugs:
1. On hardware with constant, non-stop TSC cap_usr_rdpmc was always 1.
2. On hardware without constant, non-stop TSC cap_usr_time was still 1 if
rdpmc was allowed in userspace.
Possible improvements are one or both of:
1. Add cap_usr_fixed to identify kernels that have the capabilities bits fixed
2. Swap the positions of cap_usr_time and cap_usr_rdpmc so that
cap_usr_rdpmc remains in bit 0
>
> Vince
>
>
>
>
> On Tue, 23 Jul 2013, tip-bot for Adrian Hunter wrote:
>
>> Commit-ID: 860f085b74e9f0075de8140ed3a1e5b5e3e39aa8
>> Gitweb:
>> http://git.kernel.org/tip/860f085b74e9f0075de8140ed3a1e5b5e3e39aa8
>> Author: Adrian Hunter <[email protected]>
>> AuthorDate: Fri, 28 Jun 2013 16:22:17 +0300
>> Committer: Ingo Molnar <[email protected]>
>> CommitDate: Tue, 23 Jul 2013 12:17:10 +0200
>>
>> perf: Fix broken union in 'struct perf_event_mmap_page'
>>
>> The capabilities bits must not be "union'ed" together.
>> Put them in a separate struct.
>>
>> Signed-off-by: Adrian Hunter <[email protected]>
>> Signed-off-by: Peter Zijlstra <[email protected]>
>> Link:
>> http://lkml.kernel.org/r/[email protected]
>> Signed-off-by: Ingo Molnar <[email protected]>
>> ---
>> include/uapi/linux/perf_event.h | 8 +++++---
>> 1 file changed, 5 insertions(+), 3 deletions(-)
>>
>> diff --git a/include/uapi/linux/perf_event.h
>> b/include/uapi/linux/perf_event.h
>> index 00d8274..0041aed 100644
>> --- a/include/uapi/linux/perf_event.h
>> +++ b/include/uapi/linux/perf_event.h
>> @@ -375,9 +375,11 @@ struct perf_event_mmap_page {
>> __u64 time_running; /* time event on cpu */
>> union {
>> __u64 capabilities;
>> - __u64 cap_usr_time : 1,
>> - cap_usr_rdpmc : 1,
>> - cap_____res : 62;
>> + struct {
>> + __u64 cap_usr_time : 1,
>> + cap_usr_rdpmc : 1,
>> + cap_____res : 62;
>> + };
>> };
>>
>> /*
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
>> the body of a message to [email protected]
>> More majordomo info at http://vger.kernel.org/majordomo-info.html
>> Please read the FAQ at http://www.tux.org/lkml/
>>
> k
>
>
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
