[PATCH] PCI: Fix BUILD_BUG_ON usage for old gcc

[PATCH] PCI: Fix BUILD_BUG_ON usage for old gcc

[Alex Williamson]

As reported in the below link, it seems older versions of gcc cannot
determine that the howmany variable is known for all callers.  Include
a test so that newer compilers can still enforce this sanity check and
older compilers can still work.

Fixes: 4453f360862e ("PCI: Batch BAR sizing operations")
Link: https://lore.kernel.org/all/20250209154512.GA18688@redhat.com
Reported-by: Oleg Nesterov <oleg@redhat.com>
Suggested-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
---

Verified against gcc 14.2.1 to still trigger a build error if called
with a constant value greater than 6, Oleg to confirm build issue is
resolved for gcc 5.3.1.

 drivers/pci/probe.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c
index b6536ed599c3..1bde89d0dc0d 100644
--- a/drivers/pci/probe.c
+++ b/drivers/pci/probe.c
@@ -345,7 +345,8 @@ static void pci_read_bases(struct pci_dev *dev, unsigned int howmany, int rom)
 	unsigned int pos, reg;
 	u16 orig_cmd;
 
-	BUILD_BUG_ON(howmany > PCI_STD_NUM_BARS);
+	BUILD_BUG_ON(__builtin_constant_p(howmany) &&
+		     howmany > PCI_STD_NUM_BARS);
 
 	if (dev->non_compliant_bars)
 		return;
-- 
2.47.1

Re: [PATCH] PCI: Fix BUILD_BUG_ON usage for old gcc

[Oleg Nesterov]

On 02/10, Alex Williamson wrote:
>
> As reported in the below link, it seems older versions of gcc cannot
> determine that the howmany variable is known for all callers.  Include
> a test so that newer compilers can still enforce this sanity check and
> older compilers can still work.
>
> Fixes: 4453f360862e ("PCI: Batch BAR sizing operations")
> Link: https://lore.kernel.org/all/20250209154512.GA18688@redhat.com
> Reported-by: Oleg Nesterov <oleg@redhat.com>
> Suggested-by: Oleg Nesterov <oleg@redhat.com>
  ^^^^^^^^^^^^

Well, thanks, but I didn't ;)

> @@ -345,7 +345,8 @@ static void pci_read_bases(struct pci_dev *dev, unsigned int howmany, int rom)
>  	unsigned int pos, reg;
>  	u16 orig_cmd;
>  
> -	BUILD_BUG_ON(howmany > PCI_STD_NUM_BARS);
> +	BUILD_BUG_ON(__builtin_constant_p(howmany) &&
> +		     howmany > PCI_STD_NUM_BARS);

Thanks!

Tested-by: Oleg Nesterov <oleg@redhat.com>

Re: [PATCH] PCI: Fix BUILD_BUG_ON usage for old gcc

[Oleg Nesterov]

On 02/10, Oleg Nesterov wrote:
>
> On 02/10, Alex Williamson wrote:
> >
> > @@ -345,7 +345,8 @@ static void pci_read_bases(struct pci_dev *dev, unsigned int howmany, int rom)
> >  	unsigned int pos, reg;
> >  	u16 orig_cmd;
> >
> > -	BUILD_BUG_ON(howmany > PCI_STD_NUM_BARS);
> > +	BUILD_BUG_ON(__builtin_constant_p(howmany) &&
> > +		     howmany > PCI_STD_NUM_BARS);
>
> Thanks!
>
> Tested-by: Oleg Nesterov <oleg@redhat.com>

Just in case... I agree with David, statically_true() looks a bit
better and

	BUILD_BUG_ON(statically_true(howmany > PCI_STD_NUM_BARS));

also works for me, so if you decide to update this patch feel free
to keep my Tested-by.

Oleg.

Re: [PATCH] PCI: Fix BUILD_BUG_ON usage for old gcc

[Mitchell Augustin]

On Ubuntu's (modern) GCC 13.3.0-6ubuntu2~24.04, both proposed options
also build as expected for me when howmany is set correctly, and fail
as expected when I force howmany = 7.

Tested-by: Mitchell Augustin <mitchell.augustin@canonical.com>

On Tue, Feb 11, 2025 at 4:05 AM Oleg Nesterov <oleg@redhat.com> wrote:
>
> On 02/10, Oleg Nesterov wrote:
> >
> > On 02/10, Alex Williamson wrote:
> > >
> > > @@ -345,7 +345,8 @@ static void pci_read_bases(struct pci_dev *dev, unsigned int howmany, int rom)
> > >     unsigned int pos, reg;
> > >     u16 orig_cmd;
> > >
> > > -   BUILD_BUG_ON(howmany > PCI_STD_NUM_BARS);
> > > +   BUILD_BUG_ON(__builtin_constant_p(howmany) &&
> > > +                howmany > PCI_STD_NUM_BARS);
> >
> > Thanks!
> >
> > Tested-by: Oleg Nesterov <oleg@redhat.com>
>
> Just in case... I agree with David, statically_true() looks a bit
> better and
>
>         BUILD_BUG_ON(statically_true(howmany > PCI_STD_NUM_BARS));
>
> also works for me, so if you decide to update this patch feel free
> to keep my Tested-by.
>
> Oleg.
>


-- 
Mitchell Augustin
Software Engineer - Ubuntu Partner Engineering