eal: fix check when retrieving current cpu affinity
Checks
Commit Message
pthread_getaffinity_np returns a >0 value when failing.
This is mainly for the sake of correctness.
The only case where it could fail is when passing an incorrect cpuset
size wrt to the kernel.
Fixes: 2eba8d21f3c9 ("eal: restrict cores auto detection")
Signed-off-by: David Marchand <david.marchand@redhat.com>
---
lib/librte_eal/common/eal_common_options.c | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
Comments
On 14-Feb-19 1:27 PM, David Marchand wrote:
> pthread_getaffinity_np returns a >0 value when failing.
>
> This is mainly for the sake of correctness.
> The only case where it could fail is when passing an incorrect cpuset
> size wrt to the kernel.
>
> Fixes: 2eba8d21f3c9 ("eal: restrict cores auto detection")
> Signed-off-by: David Marchand <david.marchand@redhat.com>
> ---
> lib/librte_eal/common/eal_common_options.c | 5 ++---
> 1 file changed, 2 insertions(+), 3 deletions(-)
>
> diff --git a/lib/librte_eal/common/eal_common_options.c b/lib/librte_eal/common/eal_common_options.c
> index 6c96f45..1f45f82 100644
> --- a/lib/librte_eal/common/eal_common_options.c
> +++ b/lib/librte_eal/common/eal_common_options.c
> @@ -1343,10 +1343,9 @@ static int xdigit2val(unsigned char c)
> unsigned int lcore_id;
> unsigned int removed = 0;
> rte_cpuset_t affinity_set;
> - pthread_t tid = pthread_self();
>
> - if (pthread_getaffinity_np(tid, sizeof(rte_cpuset_t),
> - &affinity_set) < 0)
> + if (pthread_getaffinity_np(pthread_self(), sizeof(rte_cpuset_t),
> + &affinity_set))
> CPU_ZERO(&affinity_set);
>
> for (lcore_id = 0; lcore_id < RTE_MAX_LCORE; lcore_id++) {
>
CC: stable?
On Thu, Feb 14, 2019 at 5:44 PM Burakov, Anatoly <anatoly.burakov@intel.com>
wrote:
> On 14-Feb-19 1:27 PM, David Marchand wrote:
> > pthread_getaffinity_np returns a >0 value when failing.
> >
> > This is mainly for the sake of correctness.
> > The only case where it could fail is when passing an incorrect cpuset
> > size wrt to the kernel.
> >
> > Fixes: 2eba8d21f3c9 ("eal: restrict cores auto detection")
> > Signed-off-by: David Marchand <david.marchand@redhat.com>
> > ---
> > lib/librte_eal/common/eal_common_options.c | 5 ++---
> > 1 file changed, 2 insertions(+), 3 deletions(-)
> >
> > diff --git a/lib/librte_eal/common/eal_common_options.c
> b/lib/librte_eal/common/eal_common_options.c
> > index 6c96f45..1f45f82 100644
> > --- a/lib/librte_eal/common/eal_common_options.c
> > +++ b/lib/librte_eal/common/eal_common_options.c
> > @@ -1343,10 +1343,9 @@ static int xdigit2val(unsigned char c)
> > unsigned int lcore_id;
> > unsigned int removed = 0;
> > rte_cpuset_t affinity_set;
> > - pthread_t tid = pthread_self();
> >
> > - if (pthread_getaffinity_np(tid, sizeof(rte_cpuset_t),
> > - &affinity_set) < 0)
> > + if (pthread_getaffinity_np(pthread_self(), sizeof(rte_cpuset_t),
> > + &affinity_set))
> > CPU_ZERO(&affinity_set);
> >
> > for (lcore_id = 0; lcore_id < RTE_MAX_LCORE; lcore_id++) {
> >
> CC: stable?
>
Not really sure about CCing stable for this.
I did not get errors with pthread_getaffinity_np so far.
Afaiu, it would need the kernel and libc to have different cpuset sizes.
I did not investigate in which situations it could happen.
On 14-Feb-19 5:41 PM, David Marchand wrote:
>
>
> On Thu, Feb 14, 2019 at 5:44 PM Burakov, Anatoly
> <anatoly.burakov@intel.com <mailto:anatoly.burakov@intel.com>> wrote:
>
> On 14-Feb-19 1:27 PM, David Marchand wrote:
> > pthread_getaffinity_np returns a >0 value when failing.
> >
> > This is mainly for the sake of correctness.
> > The only case where it could fail is when passing an incorrect cpuset
> > size wrt to the kernel.
> >
> > Fixes: 2eba8d21f3c9 ("eal: restrict cores auto detection")
> > Signed-off-by: David Marchand <david.marchand@redhat.com
> <mailto:david.marchand@redhat.com>>
> > ---
> > lib/librte_eal/common/eal_common_options.c | 5 ++---
> > 1 file changed, 2 insertions(+), 3 deletions(-)
> >
> > diff --git a/lib/librte_eal/common/eal_common_options.c
> b/lib/librte_eal/common/eal_common_options.c
> > index 6c96f45..1f45f82 100644
> > --- a/lib/librte_eal/common/eal_common_options.c
> > +++ b/lib/librte_eal/common/eal_common_options.c
> > @@ -1343,10 +1343,9 @@ static int xdigit2val(unsigned char c)
> > unsigned int lcore_id;
> > unsigned int removed = 0;
> > rte_cpuset_t affinity_set;
> > - pthread_t tid = pthread_self();
> >
> > - if (pthread_getaffinity_np(tid, sizeof(rte_cpuset_t),
> > - &affinity_set) < 0)
> > + if (pthread_getaffinity_np(pthread_self(),
> sizeof(rte_cpuset_t),
> > + &affinity_set))
> > CPU_ZERO(&affinity_set);
> >
> > for (lcore_id = 0; lcore_id < RTE_MAX_LCORE; lcore_id++) {
> >
> CC: stable?
>
>
> Not really sure about CCing stable for this.
>
> I did not get errors with pthread_getaffinity_np so far.
> Afaiu, it would need the kernel and libc to have different cpuset sizes.
> I did not investigate in which situations it could happen.
>
You're incorrectly handling error condition - that's a bug regardless of
how likely it is to occur :)
On Tue, Feb 19, 2019 at 12:40 PM Burakov, Anatoly <anatoly.burakov@intel.com>
wrote:
> On 14-Feb-19 5:41 PM, David Marchand wrote:
> > On Thu, Feb 14, 2019 at 5:44 PM Burakov, Anatoly
> > <anatoly.burakov@intel.com <mailto:anatoly.burakov@intel.com>> wrote:
> > CC: stable?
> >
> > Not really sure about CCing stable for this.
> >
> > I did not get errors with pthread_getaffinity_np so far.
> > Afaiu, it would need the kernel and libc to have different cpuset sizes.
> > I did not investigate in which situations it could happen.
> >
>
> You're incorrectly handling error condition - that's a bug regardless of
> how likely it is to occur :)
>
Ok will resend with Cc:
@@ -1343,10 +1343,9 @@ static int xdigit2val(unsigned char c)
unsigned int lcore_id;
unsigned int removed = 0;
rte_cpuset_t affinity_set;
- pthread_t tid = pthread_self();
- if (pthread_getaffinity_np(tid, sizeof(rte_cpuset_t),
- &affinity_set) < 0)
+ if (pthread_getaffinity_np(pthread_self(), sizeof(rte_cpuset_t),
+ &affinity_set))
CPU_ZERO(&affinity_set);
for (lcore_id = 0; lcore_id < RTE_MAX_LCORE; lcore_id++) {