Coverity CID 142811: pci_set_vga_state() operands don't affect result

Coverity CID 142811: pci_set_vga_state() operands don't affect result

[Bjorn Helgaas]

Coverity complains about this in drivers/pci/pci.c:

CID 142811 (#1 of 1): Operands don't affect result (CONSTANT_EXPRESSION_RESULT)
result_independent_of_operands: flags & (2U /* 1 << 1 */) &
(command_bits & 4294967292U /* ~(1 | 2) */) is always 0 regardless of
the values of its operands. This occurs as the logical operand of if.

4128        WARN_ON((flags & PCI_VGA_STATE_CHANGE_DECODES) &
(command_bits & ~(PCI_COMMAND_IO|PCI_COMMAND_MEMORY)));

This is a result of 3448a19da479 "vgaarb: use bridges to control VGA
routing where possible."

I wonder if that middle "&" was intended to be "&&"?

Bjorn

Re: Coverity CID 142811: pci_set_vga_state() operands don't affect result

[Yinghai Lu]

On Thu, Apr 3, 2014 at 1:43 PM, Bjorn Helgaas <bhelgaas@google.com> wrote:
> Coverity complains about this in drivers/pci/pci.c:
>
> CID 142811 (#1 of 1): Operands don't affect result (CONSTANT_EXPRESSION_RESULT)
> result_independent_of_operands: flags & (2U /* 1 << 1 */) &
> (command_bits & 4294967292U /* ~(1 | 2) */) is always 0 regardless of
> the values of its operands. This occurs as the logical operand of if.
>
> 4128        WARN_ON((flags & PCI_VGA_STATE_CHANGE_DECODES) &
> (command_bits & ~(PCI_COMMAND_IO|PCI_COMMAND_MEMORY)));
>
> This is a result of 3448a19da479 "vgaarb: use bridges to control VGA
> routing where possible."
>
> I wonder if that middle "&" was intended to be "&&"?

must be. otherwise always get 0.

Re: Coverity CID 142811: pci_set_vga_state() operands don't affect result

[David Airlie]

----- Original Message -----
> From: "Yinghai Lu" <yinghai@kernel.org>
> To: "Bjorn Helgaas" <bhelgaas@google.com>
> Cc: "Dave Airlie" <airlied@redhat.com>, linux-pci@vger.kernel.org
> Sent: Friday, 4 April, 2014 10:06:32 AM
> Subject: Re: Coverity CID 142811: pci_set_vga_state() operands don't affect result
> 
> On Thu, Apr 3, 2014 at 1:43 PM, Bjorn Helgaas <bhelgaas@google.com> wrote:
> > Coverity complains about this in drivers/pci/pci.c:
> >
> > CID 142811 (#1 of 1): Operands don't affect result
> > (CONSTANT_EXPRESSION_RESULT)
> > result_independent_of_operands: flags & (2U /* 1 << 1 */) &
> > (command_bits & 4294967292U /* ~(1 | 2) */) is always 0 regardless of
> > the values of its operands. This occurs as the logical operand of if.
> >
> > 4128        WARN_ON((flags & PCI_VGA_STATE_CHANGE_DECODES) &
> > (command_bits & ~(PCI_COMMAND_IO|PCI_COMMAND_MEMORY)));
> >
> > This is a result of 3448a19da479 "vgaarb: use bridges to control VGA
> > routing where possible."
> >
> > I wonder if that middle "&" was intended to be "&&"?
> 
> must be. otherwise always get 0.

Yeah I'd agree :-)

Dave.

Re: Coverity CID 142811: pci_set_vga_state() operands don't affect result

[Bjorn Helgaas]

On Thu, Apr 03, 2014 at 02:43:05PM -0600, Bjorn Helgaas wrote:
> Coverity complains about this in drivers/pci/pci.c:
> 
> CID 142811 (#1 of 1): Operands don't affect result (CONSTANT_EXPRESSION_RESULT)
> result_independent_of_operands: flags & (2U /* 1 << 1 */) &
> (command_bits & 4294967292U /* ~(1 | 2) */) is always 0 regardless of
> the values of its operands. This occurs as the logical operand of if.
> 
> 4128        WARN_ON((flags & PCI_VGA_STATE_CHANGE_DECODES) &
> (command_bits & ~(PCI_COMMAND_IO|PCI_COMMAND_MEMORY)));
> 
> This is a result of 3448a19da479 "vgaarb: use bridges to control VGA
> routing where possible."
> 
> I wonder if that middle "&" was intended to be "&&"?

I propose the following patch for this.  Unless there's objection, I'll
queue this for v3.16.  I took the liberty of interpreting your email
responses as acks.


PCI: Fix incorrect vgaarb conditional in WARN_ON()

From: Bjorn Helgaas <bhelgaas@google.com>

3448a19da479 "vgaarb: use bridges to control VGA routing where possible"
added the "flags & PCI_VGA_STATE_CHANGE_DECODES" condition to an existing
WARN_ON(), but used bitwise AND (&) instead of logical AND (&&), so the
condition is never true.  Replace with logical AND.

Found by Coverity (CID 142811).

Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Yinghai Lu <yinghai@kernel.org>
Acked-by: David Airlie <airlied@redhat.com>
---
 drivers/pci/pci.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c
index 7325d43bf030..39012831867e 100644
--- a/drivers/pci/pci.c
+++ b/drivers/pci/pci.c
@@ -4125,7 +4125,7 @@ int pci_set_vga_state(struct pci_dev *dev, bool decode,
 	u16 cmd;
 	int rc;
 
-	WARN_ON((flags & PCI_VGA_STATE_CHANGE_DECODES) & (command_bits & ~(PCI_COMMAND_IO|PCI_COMMAND_MEMORY)));
+	WARN_ON((flags & PCI_VGA_STATE_CHANGE_DECODES) && (command_bits & ~(PCI_COMMAND_IO|PCI_COMMAND_MEMORY)));
 
 	/* ARCH specific VGA enables */
 	rc = pci_set_vga_state_arch(dev, decode, command_bits, flags);

Re: Coverity CID 142811: pci_set_vga_state() operands don't affect result

[Bjorn Helgaas]

On Fri, Apr 04, 2014 at 09:47:44AM -0600, Bjorn Helgaas wrote:
> On Thu, Apr 03, 2014 at 02:43:05PM -0600, Bjorn Helgaas wrote:
> > Coverity complains about this in drivers/pci/pci.c:
> > 
> > CID 142811 (#1 of 1): Operands don't affect result (CONSTANT_EXPRESSION_RESULT)
> > result_independent_of_operands: flags & (2U /* 1 << 1 */) &
> > (command_bits & 4294967292U /* ~(1 | 2) */) is always 0 regardless of
> > the values of its operands. This occurs as the logical operand of if.
> > 
> > 4128        WARN_ON((flags & PCI_VGA_STATE_CHANGE_DECODES) &
> > (command_bits & ~(PCI_COMMAND_IO|PCI_COMMAND_MEMORY)));
> > 
> > This is a result of 3448a19da479 "vgaarb: use bridges to control VGA
> > routing where possible."
> > 
> > I wonder if that middle "&" was intended to be "&&"?
> 
> I propose the following patch for this.  Unless there's objection, I'll
> queue this for v3.16.  I took the liberty of interpreting your email
> responses as acks.
> 
> 
> PCI: Fix incorrect vgaarb conditional in WARN_ON()
> 
> From: Bjorn Helgaas <bhelgaas@google.com>
> 
> 3448a19da479 "vgaarb: use bridges to control VGA routing where possible"
> added the "flags & PCI_VGA_STATE_CHANGE_DECODES" condition to an existing
> WARN_ON(), but used bitwise AND (&) instead of logical AND (&&), so the
> condition is never true.  Replace with logical AND.
> 
> Found by Coverity (CID 142811).
> 
> Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
> Acked-by: Yinghai Lu <yinghai@kernel.org>
> Acked-by: David Airlie <airlied@redhat.com>

I applied this to pci/misc for v3.16.

I didn't mark it for stable because the only effect of backporting it would
be to cause warnings in some new cases, without actually fixing any
problems.

> ---
>  drivers/pci/pci.c |    2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c
> index 7325d43bf030..39012831867e 100644
> --- a/drivers/pci/pci.c
> +++ b/drivers/pci/pci.c
> @@ -4125,7 +4125,7 @@ int pci_set_vga_state(struct pci_dev *dev, bool decode,
>  	u16 cmd;
>  	int rc;
>  
> -	WARN_ON((flags & PCI_VGA_STATE_CHANGE_DECODES) & (command_bits & ~(PCI_COMMAND_IO|PCI_COMMAND_MEMORY)));
> +	WARN_ON((flags & PCI_VGA_STATE_CHANGE_DECODES) && (command_bits & ~(PCI_COMMAND_IO|PCI_COMMAND_MEMORY)));
>  
>  	/* ARCH specific VGA enables */
>  	rc = pci_set_vga_state_arch(dev, decode, command_bits, flags);