* [19.11][20.11][21.11] ppc64 fixes backport
@ 2022-06-29 13:40 Xueming(Steven) Li
2022-07-06 18:53 ` Luca Boccassi
0 siblings, 1 reply; 5+ messages in thread
From: Xueming(Steven) Li @ 2022-06-29 13:40 UTC (permalink / raw)
To: Kevin Traynor, Luca Boccassi, christian.ehrhardt
Cc: stable, NBU-Contact-Thomas Monjalon (EXTERNAL)
Hi LTS maintainers,
Confirmed with Thomas, mlx5 PMD needs the following patches to compile with the upstream rdma-core lib.
2022-05-25 2022-06-01 b251bb7630 eal/ppc: undefine AltiVec keyword vector @Thomas Monjalon
2022-05-03 2022-05-25 64fcadeac0 avoid AltiVec keyword vector @Thomas Monjalon
I tried to backport to 20.11.6 LTS, cherry-pick works without conflicts.
Best Regards,
Xueming Li
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [19.11][20.11][21.11] ppc64 fixes backport
2022-06-29 13:40 [19.11][20.11][21.11] ppc64 fixes backport Xueming(Steven) Li
@ 2022-07-06 18:53 ` Luca Boccassi
2022-07-07 6:51 ` Thomas Monjalon
0 siblings, 1 reply; 5+ messages in thread
From: Luca Boccassi @ 2022-07-06 18:53 UTC (permalink / raw)
To: Xueming(Steven) Li
Cc: Kevin Traynor, christian.ehrhardt, stable,
NBU-Contact-Thomas Monjalon (EXTERNAL)
On Wed, 29 Jun 2022 at 14:40, Xueming(Steven) Li <xuemingl@nvidia.com> wrote:
>
> Hi LTS maintainers,
>
> Confirmed with Thomas, mlx5 PMD needs the following patches to compile with the upstream rdma-core lib.
>
> 2022-05-25 2022-06-01 b251bb7630 eal/ppc: undefine AltiVec keyword vector @Thomas Monjalon
> 2022-05-03 2022-05-25 64fcadeac0 avoid AltiVec keyword vector @Thomas Monjalon
>
> I tried to backport to 20.11.6 LTS, cherry-pick works without conflicts.
>
> Best Regards,
> Xueming Li
Hi,
This patch:
2022-05-25 2022-06-01 b251bb7630 eal/ppc: undefine AltiVec keyword
vector @Thomas Monjalon
Looks like an incompatible change? Commit message says so too:
"This is a compatibility breakage for applications which were using
the keyword "vector" for its AltiVec meaning."
The first one looks like but I don't think we should pick up the
second one for the LTS branches?
Kind regards,
Luca Boccassi
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [19.11][20.11][21.11] ppc64 fixes backport
2022-07-06 18:53 ` Luca Boccassi
@ 2022-07-07 6:51 ` Thomas Monjalon
2022-07-07 7:44 ` Christian Ehrhardt
0 siblings, 1 reply; 5+ messages in thread
From: Thomas Monjalon @ 2022-07-07 6:51 UTC (permalink / raw)
To: Xueming(Steven) Li, Luca Boccassi
Cc: Kevin Traynor, christian.ehrhardt, stable
06/07/2022 20:53, Luca Boccassi:
> On Wed, 29 Jun 2022 at 14:40, Xueming(Steven) Li <xuemingl@nvidia.com> wrote:
> >
> > Hi LTS maintainers,
> >
> > Confirmed with Thomas, mlx5 PMD needs the following patches to compile with the upstream rdma-core lib.
> >
> > 2022-05-25 2022-06-01 b251bb7630 eal/ppc: undefine AltiVec keyword vector @Thomas Monjalon
> > 2022-05-03 2022-05-25 64fcadeac0 avoid AltiVec keyword vector @Thomas Monjalon
> >
> > I tried to backport to 20.11.6 LTS, cherry-pick works without conflicts.
> >
> > Best Regards,
> > Xueming Li
>
> Hi,
>
> This patch:
>
> 2022-05-25 2022-06-01 b251bb7630 eal/ppc: undefine AltiVec keyword
> vector @Thomas Monjalon
>
> Looks like an incompatible change? Commit message says so too:
>
> "This is a compatibility breakage for applications which were using
> the keyword "vector" for its AltiVec meaning."
>
> The first one looks like but I don't think we should pick up the
> second one for the LTS branches?
Yes if we undefine "vector", we break applications using this syntax.
That's a required change for recent rdma-core which uses the name "vector"
for another purpose.
I don't know what is best/worst.
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [19.11][20.11][21.11] ppc64 fixes backport
2022-07-07 6:51 ` Thomas Monjalon
@ 2022-07-07 7:44 ` Christian Ehrhardt
2022-07-07 8:21 ` Thomas Monjalon
0 siblings, 1 reply; 5+ messages in thread
From: Christian Ehrhardt @ 2022-07-07 7:44 UTC (permalink / raw)
To: Thomas Monjalon; +Cc: Xueming(Steven) Li, Luca Boccassi, Kevin Traynor, stable
On Thu, Jul 7, 2022 at 8:51 AM Thomas Monjalon <thomas@monjalon.net> wrote:
>
> 06/07/2022 20:53, Luca Boccassi:
> > On Wed, 29 Jun 2022 at 14:40, Xueming(Steven) Li <xuemingl@nvidia.com> wrote:
> > >
> > > Hi LTS maintainers,
> > >
> > > Confirmed with Thomas, mlx5 PMD needs the following patches to compile with the upstream rdma-core lib.
> > >
> > > 2022-05-25 2022-06-01 b251bb7630 eal/ppc: undefine AltiVec keyword vector @Thomas Monjalon
> > > 2022-05-03 2022-05-25 64fcadeac0 avoid AltiVec keyword vector @Thomas Monjalon
> > >
> > > I tried to backport to 20.11.6 LTS, cherry-pick works without conflicts.
> > >
> > > Best Regards,
> > > Xueming Li
> >
> > Hi,
> >
> > This patch:
> >
> > 2022-05-25 2022-06-01 b251bb7630 eal/ppc: undefine AltiVec keyword
> > vector @Thomas Monjalon
> >
> > Looks like an incompatible change? Commit message says so too:
> >
> > "This is a compatibility breakage for applications which were using
> > the keyword "vector" for its AltiVec meaning."
> >
> > The first one looks like but I don't think we should pick up the
> > second one for the LTS branches?
>
> Yes if we undefine "vector", we break applications using this syntax.
> That's a required change for recent rdma-core which uses the name "vector"
> for another purpose.
> I don't know what is best/worst.
It is ok for new DPDK releases to add this, but not for LTS backports.
In general LTS releases are usually updated and rebuild against their
own older environment.
So if "new rdma-core" is a problem it only is a problem for "new DPDK".
IMHO LTS versions of DPDK should not take such a change.
>
--
Christian Ehrhardt
Senior Staff Engineer, Ubuntu Server
Canonical Ltd
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [19.11][20.11][21.11] ppc64 fixes backport
2022-07-07 7:44 ` Christian Ehrhardt
@ 2022-07-07 8:21 ` Thomas Monjalon
0 siblings, 0 replies; 5+ messages in thread
From: Thomas Monjalon @ 2022-07-07 8:21 UTC (permalink / raw)
To: Christian Ehrhardt
Cc: Xueming(Steven) Li, Luca Boccassi, Kevin Traynor, stable
07/07/2022 09:44, Christian Ehrhardt:
> On Thu, Jul 7, 2022 at 8:51 AM Thomas Monjalon <thomas@monjalon.net> wrote:
> >
> > 06/07/2022 20:53, Luca Boccassi:
> > > On Wed, 29 Jun 2022 at 14:40, Xueming(Steven) Li <xuemingl@nvidia.com> wrote:
> > > >
> > > > Hi LTS maintainers,
> > > >
> > > > Confirmed with Thomas, mlx5 PMD needs the following patches to compile with the upstream rdma-core lib.
> > > >
> > > > 2022-05-25 2022-06-01 b251bb7630 eal/ppc: undefine AltiVec keyword vector @Thomas Monjalon
> > > > 2022-05-03 2022-05-25 64fcadeac0 avoid AltiVec keyword vector @Thomas Monjalon
> > > >
> > > > I tried to backport to 20.11.6 LTS, cherry-pick works without conflicts.
> > > >
> > > > Best Regards,
> > > > Xueming Li
> > >
> > > Hi,
> > >
> > > This patch:
> > >
> > > 2022-05-25 2022-06-01 b251bb7630 eal/ppc: undefine AltiVec keyword
> > > vector @Thomas Monjalon
> > >
> > > Looks like an incompatible change? Commit message says so too:
> > >
> > > "This is a compatibility breakage for applications which were using
> > > the keyword "vector" for its AltiVec meaning."
> > >
> > > The first one looks like but I don't think we should pick up the
> > > second one for the LTS branches?
> >
> > Yes if we undefine "vector", we break applications using this syntax.
> > That's a required change for recent rdma-core which uses the name "vector"
> > for another purpose.
> > I don't know what is best/worst.
>
> It is ok for new DPDK releases to add this, but not for LTS backports.
>
> In general LTS releases are usually updated and rebuild against their
> own older environment.
> So if "new rdma-core" is a problem it only is a problem for "new DPDK".
>
> IMHO LTS versions of DPDK should not take such a change.
OK
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2022-07-07 8:21 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-06-29 13:40 [19.11][20.11][21.11] ppc64 fixes backport Xueming(Steven) Li
2022-07-06 18:53 ` Luca Boccassi
2022-07-07 6:51 ` Thomas Monjalon
2022-07-07 7:44 ` Christian Ehrhardt
2022-07-07 8:21 ` Thomas Monjalon
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).