[PATCH 0/1] x86: fix a bug of wrong return erorr.

[PATCH 0/1] x86: fix a bug of wrong return erorr.

[Liming Wang]

Fix a bug to make a ltp test work on qemux86-64:
bug 900: [LTP] clock_nanosleep01 fails on qemux86-64

Liming Wang (1):
      x86: fix a bug of wrong return erorr.

 arch/x86/vdso/vclock_gettime.c |    2 --
 1 files changed, 0 insertions(+), 2 deletions(-)

[PATCH 1/1] x86: fix a bug of wrong return erorr.

[Liming Wang]

__vdso_clock_gettime should fall back to call vdso_fallback_gettime function
if no clockid is selected, not just return error.

Signed-off-by: Liming Wang <liming.wang@windriver.com>
---
 arch/x86/vdso/vclock_gettime.c |    2 --
 1 files changed, 0 insertions(+), 2 deletions(-)

diff --git a/arch/x86/vdso/vclock_gettime.c b/arch/x86/vdso/vclock_gettime.c
index 7bc4815..2365a5b 100644
--- a/arch/x86/vdso/vclock_gettime.c
+++ b/arch/x86/vdso/vclock_gettime.c
@@ -173,8 +173,6 @@ notrace int __vdso_clock_gettime(clockid_t clock, struct timespec *ts)
 			return do_trace_clock(ts);
 		case CLOCK_TRACE_FREQ:
 			return do_trace_clock_freq(ts);
-		default:
-			return -EINVAL;
 		}
 	return vdso_fallback_gettime(clock, ts);
 }
-- 
1.7.0.4

Re: [PATCH 1/1] x86: fix a bug of wrong return erorr.

[Bruce Ashfield]

On 11-08-23 10:45 PM, Liming Wang wrote:
> __vdso_clock_gettime should fall back to call vdso_fallback_gettime function
> if no clockid is selected, not just return error.

At a glance, this seems reasonable to me .. and then I
looked a bit more. This is in fact introduced by lttng and
the ENIVAL does look wrong. It's worth checking out the
latest lttng to make sure that this error isn't there as
well. Our 3.0 kernel won't have this bug yet, but I'll
keep an eye out for this during any lttng work.

I'll merge this into the 2.6.37 tree shortly.

Cheers,

Bruce

>
> Signed-off-by: Liming Wang<liming.wang@windriver.com>
> ---
>   arch/x86/vdso/vclock_gettime.c |    2 --
>   1 files changed, 0 insertions(+), 2 deletions(-)
>
> diff --git a/arch/x86/vdso/vclock_gettime.c b/arch/x86/vdso/vclock_gettime.c
> index 7bc4815..2365a5b 100644
> --- a/arch/x86/vdso/vclock_gettime.c
> +++ b/arch/x86/vdso/vclock_gettime.c
> @@ -173,8 +173,6 @@ notrace int __vdso_clock_gettime(clockid_t clock, struct timespec *ts)
>   			return do_trace_clock(ts);
>   		case CLOCK_TRACE_FREQ:
>   			return do_trace_clock_freq(ts);
> -		default:
> -			return -EINVAL;
>   		}
>   	return vdso_fallback_gettime(clock, ts);
>   }

Re: [PATCH 1/1] x86: fix a bug of wrong return erorr.

[Wang Liming]

On 08/24/11 21:40, Bruce Ashfield wrote:
> On 11-08-23 10:45 PM, Liming Wang wrote:
>> __vdso_clock_gettime should fall back to call vdso_fallback_gettime
>> function
>> if no clockid is selected, not just return error.
>
> At a glance, this seems reasonable to me .. and then I
> looked a bit more. This is in fact introduced by lttng and
> the ENIVAL does look wrong. It's worth checking out the
> latest lttng to make sure that this error isn't there as
It's still in the latest lttng patch:
http://git.kernel.org/?p=linux/kernel/git/compudj/linux-2.6-lttng.git;a=blobdiff;f=arch/x86/vdso/vclock_gettime.c;h=7bc481508d004c4e8dd0f5cff51aeeac8bfd0766;hp=ee55754cc3c5ff378b76f2065a610b72e757f088;hb=98052998fe2aee4423dc24fccfe991b305969656;hpb=b6c4d0eaca66305984cf1ce6bc9d49a3244b412b
> well. Our 3.0 kernel won't have this bug yet, but I'll
Yes, our 3.0 kernel hasn't this bug.

> keep an eye out for this during any lttng work.
>
> I'll merge this into the 2.6.37 tree shortly.
Please replace subject "erorr" with "error" for my fault.

Liming Wang
>
> Cheers,
>
> Bruce
>
>>
>> Signed-off-by: Liming Wang<liming.wang@windriver.com>
>> ---
>> arch/x86/vdso/vclock_gettime.c | 2 --
>> 1 files changed, 0 insertions(+), 2 deletions(-)
>>
>> diff --git a/arch/x86/vdso/vclock_gettime.c
>> b/arch/x86/vdso/vclock_gettime.c
>> index 7bc4815..2365a5b 100644
>> --- a/arch/x86/vdso/vclock_gettime.c
>> +++ b/arch/x86/vdso/vclock_gettime.c
>> @@ -173,8 +173,6 @@ notrace int __vdso_clock_gettime(clockid_t clock,
>> struct timespec *ts)
>> return do_trace_clock(ts);
>> case CLOCK_TRACE_FREQ:
>> return do_trace_clock_freq(ts);
>> - default:
>> - return -EINVAL;
>> }
>> return vdso_fallback_gettime(clock, ts);
>> }
>
>