Re: [Qemu-devel] [PATCH v2] msix: don't mask already masked vectors on reset

[Michael S. Tsirkin]

On Fri, Dec 08, 2017 at 09:09:06AM -0500, Marcel Apfelbaum wrote:
> 
> 
> ----- Original Message -----
> > From: "Ladi Prosek" <address@hidden>
> > To: address@hidden
> > Cc: address@hidden, address@hidden, address@hidden, "alex williamson" 
> > <address@hidden>
> > Sent: Friday, December 8, 2017 8:31:13 AM
> > Subject: [PATCH v2] msix: don't mask already masked vectors on reset
> > 
> > msix_mask_all() is supposed to invoke the release vector notifier if the
> > state of the
> > respective vector changed from unmasked to masked. The way it's currently
> > called from
> > msix_reset(), though, may result in calling the release notifier even if the
> > vector
> > is already masked.
> > 
> > 1) msix_reset() clears out the msix_cap field and the msix_table.
> > 2) msix_mask_all() runs with was_masked=false for all vectors because of 1),
> > which
> >    results in calling the release notifier on all vectors.
> > 3) if msix_reset() is subsequently called again, it goes through the same
> > steps and
> >    calls the release notifier on all vectors again.
> > 
> > This commit moves msix_mask_all() up so it runs before the device state is
> > lost. And
> > it adds an assignment to msix_function_masked so that the device remembers
> > that
> > MSI-X is masked.
> > 
> > This is likely a low impact issue, found while debugging an already broken
> > device. It
> > is however easy to fix and the expectation that the use and release notifier
> > invocations
> > are always balanced is very natural.
> > 
> > Signed-off-by: Ladi Prosek <address@hidden>
> > ---
> > 
> > v1->v2:
> > * fixed typo in commit message "or" -> "to" (Marcel)
> > * directly set msix_function_masked to true instead of calling
> >   msix_update_function_masked() (Marcel)
> > 
> > 
> >  hw/pci/msix.c | 3 ++-
> >  1 file changed, 2 insertions(+), 1 deletion(-)
> > 
> > diff --git a/hw/pci/msix.c b/hw/pci/msix.c
> > index c944c02135..d6a4dbdb6b 100644
> > --- a/hw/pci/msix.c
> > +++ b/hw/pci/msix.c
> > @@ -500,11 +500,12 @@ void msix_reset(PCIDevice *dev)
> >          return;
> >      }
> >      msix_clear_all_vectors(dev);
> > +    msix_mask_all(dev, dev->msix_entries_nr);
> >      dev->config[dev->msix_cap + MSIX_CONTROL_OFFSET] &=
> >         ~dev->wmask[dev->msix_cap + MSIX_CONTROL_OFFSET];
> >      memset(dev->msix_table, 0, dev->msix_entries_nr * PCI_MSIX_ENTRY_SIZE);
> >      memset(dev->msix_pba, 0, QEMU_ALIGN_UP(dev->msix_entries_nr, 64) / 8);
> > -    msix_mask_all(dev, dev->msix_entries_nr);
> > +    dev->msix_function_masked = true;
> >  }
> >  
> >  /* PCI spec suggests that devices make it possible for software to 
> > configure
> > --
> > 2.13.6
> > 
> > 
> 
> Thanks Ladi!
> 
> Reviewed-by: Marcel Apfelbaum <address@hidden>

This breaks make check though.

Marcel - could you take a look please?

-- 
MST