On Wed, Jun 21, 2017 at 12:25:51PM +0300, Ilya Maximets wrote:
> On 21.06.2017 11:58, Bruce Richardson wrote:
> > On Wed, Jun 21, 2017 at 10:51:58AM +0200, Thomas Monjalon wrote:
> >> 21/06/2017 10:04, Ilya Maximets:
> >>> +CONFIG_RTE_LIBRTE_EAL_NUMA_AWARE_HUGEPAGES=n
> >>
> >> We can stop insertin
21/06/2017 10:58, Bruce Richardson:
> On Wed, Jun 21, 2017 at 10:51:58AM +0200, Thomas Monjalon wrote:
> > 21/06/2017 10:04, Ilya Maximets:
> > > +CONFIG_RTE_LIBRTE_EAL_NUMA_AWARE_HUGEPAGES=n
> >
> > We can stop inserting LIBRTE in our config options.
> > CONFIG_RTE_EAL_ is long enough :)
> >
> C
On 21.06.2017 11:58, Bruce Richardson wrote:
> On Wed, Jun 21, 2017 at 10:51:58AM +0200, Thomas Monjalon wrote:
>> 21/06/2017 10:04, Ilya Maximets:
>>> +CONFIG_RTE_LIBRTE_EAL_NUMA_AWARE_HUGEPAGES=n
>>
>> We can stop inserting LIBRTE in our config options.
>> CONFIG_RTE_EAL_ is long enough :)
>>
> C
On Wed, Jun 21, 2017 at 10:51:58AM +0200, Thomas Monjalon wrote:
> 21/06/2017 10:04, Ilya Maximets:
> > +CONFIG_RTE_LIBRTE_EAL_NUMA_AWARE_HUGEPAGES=n
>
> We can stop inserting LIBRTE in our config options.
> CONFIG_RTE_EAL_ is long enough :)
>
Consistency. While I agree it's unneeded should have
21/06/2017 10:04, Ilya Maximets:
> +CONFIG_RTE_LIBRTE_EAL_NUMA_AWARE_HUGEPAGES=n
We can stop inserting LIBRTE in our config options.
CONFIG_RTE_EAL_ is long enough :)
Currently EAL allocates hugepages one by one not paying attention
from which NUMA node allocation was done.
Such behaviour leads to allocation failure if number of available
hugepages for application limited by cgroups or hugetlbfs and
memory requested not only from the first socket.
Example:
6 matches
Mail list logo