Re: [PATCH 2/2] mempool: use rte constant macro instead of GCC builtin

[David Marchand <david.marchand@redhat.com>]

On Wed, Jul 3, 2024 at 3:49 PM Morten Brørup <mb@smartsharesystems.com> wrote:
>
> > From: Thomas Monjalon [mailto:thomas@monjalon.net]
> > Sent: Wednesday, 3 July 2024 15.17
> >
> > 14/06/2024 16:32, David Marchand:
> > > On Wed, May 29, 2024 at 4:51 PM Thomas Monjalon <thomas@monjalon.net> wrote:
> > > >
> > > > 20/03/2024 22:33, Tyler Retzlaff:
> > > > > Use newly introduced __rte_constant(e) macro instead of directly using
> > > > > __builtin_constant_p() allowing mempool to be built by MSVC.
> > > >
> > > > Does it mean we should enable mempool build?
> > > > If yes, please send a v2.
> > >
> > > I guess now it is possible, as I merged some other patches on mempool
> > > from Stephen that were for MSVC.
> > > Tyler, can you send a v2 so it passes through the CI?
> >
> > I tried a retest last week and there is this failure on Ubuntu 24.04
> > that I don't manage to reproduce locally:
> >
> > /usr/include/x86_64-linux-gnu/bits/string_fortified.h:29:10: error: argument 2
> > null where non-null expected [-Werror=nonnull]
> > 29 |   return __builtin___memcpy_chk (__dest, __src, __len,
> > |          ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > 30 |                                  __glibc_objsize0 (__dest));
> > |                                  ~~~~~~~~~~~~~~~~~~~~~~~~~~
> > /usr/include/x86_64-linux-gnu/bits/string_fortified.h:29:10: note: in a call
> > to built-in function '__builtin___memcpy_chk'
> > In function 'memcpy',
> > inlined from 'pcapng_add_option' at ../lib/pcapng/rte_pcapng.c:131:2,
>
> pcapng_add_option() in rte_pcapng.c has memcpy() on line 132 [1] (and has a fix for this error, by comparing len > 0 before calling memcpy()); older versions had memcpy() on line 131, so the CI must be building with an older version of rte_pcapng.c.
>
> [1]: https://elixir.bootlin.com/dpdk/v24.07-rc1/source/lib/pcapng/rte_pcapng.c#L132
> [2]: https://elixir.bootlin.com/dpdk/v24.03/source/lib/pcapng/rte_pcapng.c#L131

This is likely the reason.
Looking at the report:

http://mails.dpdk.org/archives/test-report/2024-June/717951.html

_Testing issues RETEST #1_

Submitter: Tyler Retzlaff <roretzla at linux.microsoft.com>
Date: Wednesday, March 20 2024 21:33:36
DPDK git baseline: Repo:dpdk
  Branch: master
  CommitID:80ecef6d1f71fcebc0a51d7cabc51f73ee142ff2


$ git describe --contains 80ecef6d1f71fcebc0a51d7cabc51f73ee142ff2
v24.03-rc3^0


From the discussions on the retest mechanism, I understand we need to
ask for a rebase.
I sent a new retest. Let's see...


-- 
David Marchand