[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <ZrV6TcK7jFTXR0cz@rli9-mobl>
Date: Fri, 9 Aug 2024 10:09:17 +0800
From: Philip Li <philip.li@...el.com>
To: Nathan Chancellor <nathan@...nel.org>
CC: kernel test robot <lkp@...el.com>, Kent Overstreet
<kent.overstreet@...ux.dev>, <llvm@...ts.linux.dev>,
<oe-kbuild-all@...ts.linux.dev>, <linux-kernel@...r.kernel.org>
Subject: Re: ERROR: modpost: "__popcountsi2" [fs/bcachefs/bcachefs.ko]
undefined!
On Thu, Aug 08, 2024 at 05:39:01PM -0700, Nathan Chancellor wrote:
> On Thu, Aug 08, 2024 at 11:28:22AM +0800, Philip Li wrote:
> > On Wed, Aug 07, 2024 at 08:36:16AM -0700, Nathan Chancellor wrote:
> > > On Wed, Aug 07, 2024 at 04:43:13AM +0800, kernel test robot wrote:
> > > > Hi Kent,
> > > >
> > > > First bad commit (maybe != root cause):
> > > >
> > > > tree: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master
> > > > head: eb5e56d1491297e0881c95824e2050b7c205f0d4
> > > > commit: 9ae82fe6ace1b267005758ccfb2347a4a6aa4398 bcachefs: Inline make_bfloat() into __build_ro_aux_tree()
> > > > date: 10 months ago
> > > > config: arm-randconfig-002-20240805 (https://download.01.org/0day-ci/archive/20240807/202408070432.X6n56VaY-lkp@intel.com/config)
> > > > compiler: clang version 20.0.0git (https://github.com/llvm/llvm-project 423aec6573df4424f90555468128e17073ddc69e)
> > > > reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20240807/202408070432.X6n56VaY-lkp@intel.com/reproduce)
> > > >
> > > > If you fix the issue in a separate patch/commit (i.e. not just a new version of
> > > > the same patch/commit), kindly add following tags
> > > > | Reported-by: kernel test robot <lkp@...el.com>
> > > > | Closes: https://lore.kernel.org/oe-kbuild-all/202408070432.X6n56VaY-lkp@intel.com/
> > > >
> > > > All errors (new ones prefixed by >>, old ones prefixed by <<):
> > > ...
> > > > ERROR: modpost: "__popcountsi2" [fs/ext4/ext4.ko] undefined!
> > > > ERROR: modpost: "__popcountsi2" [fs/fat/fat.ko] undefined!
> > > > ERROR: modpost: "__popcountsi2" [fs/hfsplus/hfsplus.ko] undefined!
> > > > ERROR: modpost: "__popcountsi2" [fs/xfs/xfs.ko] undefined!
> > > > ERROR: modpost: "__popcountsi2" [fs/gfs2/gfs2.ko] undefined!
> > > > >> ERROR: modpost: "__popcountsi2" [fs/bcachefs/bcachefs.ko] undefined!
> > > > ERROR: modpost: "__aeabi_uldivmod" [fs/bcachefs/bcachefs.ko] undefined!
> > > > ERROR: modpost: "__popcountsi2" [drivers/block/virtio_blk.ko] undefined!
> > > > ERROR: modpost: "__popcountsi2" [drivers/net/ipa/ipa.ko] undefined!
> > > > ERROR: modpost: "__popcountsi2" [drivers/memory/emif.ko] undefined!
> > > > WARNING: modpost: suppressed 5 unresolved symbol warnings because there were too many)
> > >
> > > Intel folks, can you upgrade your build of Clang main to
> > > https://github.com/llvm/llvm-project/commit/4527fba9ad6bc682eceda603150bfaec65ec6916?
> >
> > Got it, we will update clang to this version 4527fba9.
>
> Thanks a lot! Is there a way to notice when a compiler upgrade causes a
> lot of similar reports to be generated? If so, would it be possible to
Sorry, this is not supported yet as an automatic approach. So far, we add
pattern to the configuration to indicate low confidence issue after getting
feedback or noticing abnormal pattern. Such issues will be only sent to
llvm@...ts.linux.dev.
> send the reports that would be generated to only me and
> llvm@...ts.linux.dev? We want to get notified when there would be a lot
> of reports generated after a compiler upgrade (because that usually
> points to a compiler regression) but kernel developers shouldn't get
> notified (since there is not anything for them to do).
Got it, we will think of this to reduce the impact.
>
> > > The current revision is broken and unrelated changes are getting
> > > notified. There might be one more regression that I am looking into, so
> > > consider not going further than that revision.
> >
> > Ok, we will keep at 4527fba9, and kindly let us know if we can move to
> > latest head.
>
> You should be able to move to the latest whenever now, the regression I
> noticed was fixed with a revert (it has since been reapplied with a
> fix).
Thanks for the info.
>
> Cheers,
> Nathan
Powered by blists - more mailing lists