On Wed, Apr 19, 2023 at 4:30 PM Luca Boccassi <[email protected]> wrote: > > On Wed, 19 Apr 2023 at 15:26, Kevin Traynor <[email protected]> wrote: > > > > On 19/04/2023 15:23, Ali Alnubani wrote: > > >> -----Original Message----- > > >> From: Luca Boccassi <[email protected]> > > >> Sent: Wednesday, April 19, 2023 5:15 PM > > >> To: David Marchand <[email protected]> > > >> Cc: Ali Alnubani <[email protected]>; [email protected]; [email protected]; > > >> Abhishek Marathe <[email protected]>; > > >> [email protected]; David Christensen <[email protected]>; > > >> Hemant Agrawal <[email protected]>; Ian Stokes > > >> <[email protected]>; Jerin Jacob <[email protected]>; John McNamara > > >> <[email protected]>; Ju-Hyoung Lee <[email protected]>; > > >> Kevin Traynor <[email protected]>; Pei Zhang <[email protected]>; > > >> [email protected]; Raslan Darawsheh <[email protected]>; NBU- > > >> Contact-Thomas Monjalon (EXTERNAL) <[email protected]>; Yanghang > > >> Liu <[email protected]>; [email protected]; > > >> [email protected] > > >> Subject: Re: 20.11.8 patches review and test > > >> > > >> On Wed, 19 Apr 2023 at 14:25, David Marchand > > >> <[email protected]> wrote: > > >>> > > >>> On Wed, Apr 19, 2023 at 3:14 PM Ali Alnubani <[email protected]> wrote: > > >>>>> > > >>>>> I don't think this function prototype is really different in the 20.11 > > >>>>> stable branch (putting the __rte_internal tag aside). > > >>>>> Can you trigger this issue on the main branch? > > >>>>> > > >>>> > > >>>> The main branch doesn't reproduce on the same environment. > > >>> > > >>> Ok, same for me, I reproduced for 20.11 but not for main. > > >>> > > >>> > > >>> I think I have the reason. > > >>> We filter headers to be processed by doxygen via a name filter: > > >>> > > >>> FILE_PATTERNS = rte_*.h \ > > >>> > > >>> But this rte_cryptodev_pmd.h driver header has been renamed as > > >>> cryptodev_pmd.h in 21.11. > > >>> So the issue you reported only affects 20.11. > > >> > > >> I can reproduce in debian testing, after backporting the mentioned > > >> commit from main it seems fine, can you confirm as well that it works > > >> in fedora? > > >> > > >> https://github.com/bluca/dpdk- > > >> stable/commit/77f65b61f955fd817b3791049dda3032c459ac59 > > > > > > Now I see a different error: > > > > > > [2/4] Generating doc/api/doxygen with a custom command > > > FAILED: doc/api/html > > > /tmp/dpdk/doc/api/generate_doxygen.sh doc/api/doxy-api.conf doc/api/html > > > /tmp/dpdk/doc/api/doxy-html-custom.sh > > > /tmp/dpdk/lib/librte_vhost/rte_vhost_async.h:49: error: > > > rte_vhost_async_channel_ops::transfer_data has @param documentation > > > sections but no arguments (warning treated as error, aborting now) > > > > See > > https://git.dpdk.org/dpdk-stable/commit/?h=21.11&id=adeaf361f0a5bb3992756fd4e727dd11ecca0943 > > Thanks, any more errors with > https://github.com/bluca/dpdk-stable/commit/b75ebf643de0204d0092a56fbec63f0bcd10e89f
With those two backports, it works for me. b75ebf643d - (HEAD -> 20.11, bluca/20.11) vhost: fix doxygen warnings (6 minutes ago) <Kevin Traynor> 77f65b61f9 - lib: remove empty return types from doxygen comments (23 minutes ago) <Ali Alnubani> a75247cbdc - (stable/20.11) crypto/snow3g: fix snow3g enqueue stat increment (24 hours ago) <Saoirse O'Donovan> -- David Marchand
