Re: [Bulk] Re: [PATCH 1/2] [PCI] Check that MWI bit really did get set

[Andrew Morton <akpm@osdl.org>]

On Sun, 15 Oct 2006 07:57:56 -0600
Matthew Wilcox <matthew@wil.cx> wrote:

> On Sun, Oct 15, 2006 at 03:21:22PM +0100, Alan Cox wrote:
> > Ar Sul, 2006-10-15 am 00:08 -0700, ysgrifennodd David Brownell:
> > > Since it's not an error, there should be no such printk ... which
> > > is exactly how it's coded above.
> > 
> > The underlying bug is that someone marked pci_set_mwi must-check, that's
> > wrong for most of the drivers that use it. If you remove the must check
> > annotation from it then the problem and a thousand other spurious
> > warnings go away.
> 
> There's only about 20 users of pci_set_mwi ... about 12 of them seem to
> check it, one of them uses a variable called
> compiler_warning_pointless_fix which leaves about 7 warnings to be
> removed by removing the __must_check.
> 
> However, I do believe the __must_check should be removed.  For example,
> the LSI 53c1030 has *nothing* to be done if setting MWI fails.  It just
> doesn't work, and the device copes.

If the drivers doesn't care and if it makes no difference to performance
then just delete the call to pci_set_mwi().

But if MWI _does_ make a difference to performance then we should tell
someone that it isn't working rather than silently misbehaving?