Re: [PATCH] async: Suppress GCC13 false positive in aio_bh_poll()

["Daniel P. Berrangé" <berrange@redhat.com>]

On Tue, Apr 25, 2023 at 02:24:59PM +0100, Daniel P. Berrangé wrote:
> On Tue, Apr 25, 2023 at 03:22:15PM +0200, Juan Quintela wrote:
> > Cédric Le Goater <clg@kaod.org> wrote:
> > > From: Cédric Le Goater <clg@redhat.com>
> > >
> > > GCC13 reports an error :
> > >
> > > ../util/async.c: In function ‘aio_bh_poll’:
> > > include/qemu/queue.h:303:22: error: storing the address of local
> > > variable ‘slice’ in ‘*ctx.bh_slice_list.sqh_last’
> > > [-Werror=dangling-pointer=]
> > >   303 |     (head)->sqh_last = &(elm)->field.sqe_next;                          \
> > >       |     ~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~
> > > ../util/async.c:169:5: note: in expansion of macro ‘QSIMPLEQ_INSERT_TAIL’
> > >   169 |     QSIMPLEQ_INSERT_TAIL(&ctx->bh_slice_list, &slice, next);
> > >       |     ^~~~~~~~~~~~~~~~~~~~
> > > ../util/async.c:161:17: note: ‘slice’ declared here
> > >   161 |     BHListSlice slice;
> > >       |                 ^~~~~
> > > ../util/async.c:161:17: note: ‘ctx’ declared here
> > >
> > > But the local variable 'slice' is removed from the global context list
> > > in following loop of the same routine. Add a pragma to silent GCC.
> > >
> > > Cc: Stefan Hajnoczi <stefanha@redhat.com>
> > > Cc: Paolo Bonzini <pbonzini@redhat.com>
> > > Cc: Daniel P. Berrangé <berrange@redhat.com>
> > > Signed-off-by: Cédric Le Goater <clg@redhat.com>
> > > ---
> > >  util/async.c | 14 ++++++++++++++
> > >  1 file changed, 14 insertions(+)
> > >
> > > diff --git a/util/async.c b/util/async.c
> > > index 21016a1ac7..856e1a8a33 100644
> > > --- a/util/async.c
> > > +++ b/util/async.c
> > > @@ -164,7 +164,21 @@ int aio_bh_poll(AioContext *ctx)
> > >  
> > >      /* Synchronizes with QSLIST_INSERT_HEAD_ATOMIC in aio_bh_enqueue().  */
> > >      QSLIST_MOVE_ATOMIC(&slice.bh_list, &ctx->bh_list);
> > > +
> > > +    /*
> > > +     * GCC13 [-Werror=dangling-pointer=] complains that the local variable
> > > +     * 'slice' is being stored in the global 'ctx->bh_slice_list' but the
> > > +     * list is emptied before this function returns.
> > > +     */
> > > +#if !defined(__clang__)
> > > +#pragma GCC diagnostic push
> > > +#pragma GCC diagnostic ignored "-Wpragmas"
> > > +#pragma GCC diagnostic ignored "-Wdangling-pointer="
> > > +#endif
> > >      QSIMPLEQ_INSERT_TAIL(&ctx->bh_slice_list, &slice, next);
> > > +#if !defined(__clang__)
> > > +#pragma GCC diagnostic pop
> > > +#endif
> > >  
> > >      while ((s = QSIMPLEQ_FIRST(&ctx->bh_slice_list))) {
> > >          QEMUBH *bh;
> > 
> > I know, I know.
> > 
> > I like to make fun of the compiler as the next guy.  But it is not
> > simpler this other change, just put the variable in the heap?
> > 
> > Later, Juan.
> > 
> > 
> > From bb5792a6763a451c72ef5cfd78b09032689f54e5 Mon Sep 17 00:00:00 2001
> > From: Juan Quintela <quintela@redhat.com>
> > Date: Tue, 25 Apr 2023 15:19:11 +0200
> > Subject: [PATCH] Silent GCC13 warning
> > 
> > Gcc complains about putting a local variable on a global list, not
> > noticing that we remove the whole list before leaving the function.
> > 
> > Signed-off-by: Juan Quintela <quintela@redhat.com>
> > ---
> >  util/async.c | 10 +++++++---
> >  1 file changed, 7 insertions(+), 3 deletions(-)
> > 
> > diff --git a/util/async.c b/util/async.c
> > index 21016a1ac7..7a8432e9e9 100644
> > --- a/util/async.c
> > +++ b/util/async.c
> > @@ -158,13 +158,17 @@ void aio_bh_call(QEMUBH *bh)
> >  /* Multiple occurrences of aio_bh_poll cannot be called concurrently. */
> >  int aio_bh_poll(AioContext *ctx)
> >  {
> > -    BHListSlice slice;
> > +    /*
> > +     * gcc13 complains about putting a local variable
> > +     * in a global list, so put it on the heap.
> > +     */
> > +    g_autofree BHListSlice *slice = g_new(BHListSlice, 1);
> >      BHListSlice *s;
> >      int ret = 0;
> >  
> >      /* Synchronizes with QSLIST_INSERT_HEAD_ATOMIC in aio_bh_enqueue().  */
> > -    QSLIST_MOVE_ATOMIC(&slice.bh_list, &ctx->bh_list);
> > -    QSIMPLEQ_INSERT_TAIL(&ctx->bh_slice_list, &slice, next);
> > +    QSLIST_MOVE_ATOMIC(&slice->bh_list, &ctx->bh_list);
> > +    QSIMPLEQ_INSERT_TAIL(&ctx->bh_slice_list, slice, next);
> >  
> >      while ((s = QSIMPLEQ_FIRST(&ctx->bh_slice_list))) {
> >          QEMUBH *bh;
> 
> This must be a memory leak since you're adding a g_new but not
> adding any g_free

Sorry, I'm failing to read properly today. It uses g_autofree
so there is no leak.

With regards,
Daniel
-- 
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|