Greg, Jesse,
Is there any issue with the patch? Tomasz need kernel to have the patch to work
on his machine.
Thanks,
Yanmin
On Tue, 2008-09-30 at 12:29 +0800, Zhang, Yanmin wrote:
> Subject: pci-e: ignore unknown capability and continue searching
> From: "Zhang, Yanmin" <[email protected]>
>
>
> Tomasz reported AER driver couldn't work on his machine. With the output
> of "lcpsi -vvv", I found the root port's extended capabilities are
> Capabilities: [100] Unknown (11)
> Capabilities: [150] Advanced Error Reporting
> Capabilities: [190] Unknown (13)
> Such Unknown capability is not expected. During pci-e initialization,function
> get_port_device_capability just returns if it hits an unknown capability when
> searching the AER capability.
>
> I worked out a patch against 2.6.27-rc7. When hitting an unkown capability,
> function get_port_device_capability continues the searching.
>
> Tomasz tested it and the patch does work well.
>
> Signed-off-by Zhang Yanmin <[email protected]>
> Reported-by Tomasz Czernecki <[email protected]>
>
> ---
>
> --- linux-2.6.27-rc7/drivers/pci/pcie/portdrv_core.c 2008-09-27 09:35:32.000000000 +0800
> +++ linux-2.6.27-rc7_aer/drivers/pci/pcie/portdrv_core.c 2008-09-27 10:10:39.000000000 +0800
> @@ -195,23 +195,25 @@ static int get_port_device_capability(st
> /* PME Capable - root port capability */
> if (((reg16 >> 4) & PORT_TYPE_MASK) == PCIE_RC_PORT)
> services |= PCIE_PORT_SERVICE_PME;
> -
> +
> pos = PCI_CFG_SPACE_SIZE;
> while (pos) {
> - pci_read_config_dword(dev, pos, ®32);
> - switch (reg32 & 0xffff) {
> + if (pci_read_config_dword(dev, pos, ®32))
> + break;
> +
> + /* some broken boards return ~0 */
> + if (reg32 == 0xffffffff)
> + break;
> +
> + switch (PCI_EXT_CAP_ID(reg32)) {
> case PCI_EXT_CAP_ID_ERR:
> services |= PCIE_PORT_SERVICE_AER;
> - pos = reg32 >> 20;
> break;
> case PCI_EXT_CAP_ID_VC:
> services |= PCIE_PORT_SERVICE_VC;
> - pos = reg32 >> 20;
> - break;
> - default:
> - pos = 0;
> break;
> }
> + pos = PCI_EXT_CAP_NEXT(reg32);
> }
>
> return services;
>
>
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
I'd rather get rid of this function altogether. Can you check out my "use
pci_find_ext_capability everywhere" patch and see if you can do something
similar for the port driver? I mainly just want to avoid having duplicated
find_capability code everywhere...
Thanks,
Jesse
On Friday, October 3, 2008 1:21 am Zhang, Yanmin wrote:
> Greg, Jesse,
>
> Is there any issue with the patch? Tomasz need kernel to have the patch to
> work on his machine.
>
> Thanks,
> Yanmin
>
> On Tue, 2008-09-30 at 12:29 +0800, Zhang, Yanmin wrote:
> > Subject: pci-e: ignore unknown capability and continue searching
> > From: "Zhang, Yanmin" <[email protected]>
> >
> >
> > Tomasz reported AER driver couldn't work on his machine. With the output
> > of "lcpsi -vvv", I found the root port's extended capabilities are
> > Capabilities: [100] Unknown (11)
> > Capabilities: [150] Advanced Error Reporting
> > Capabilities: [190] Unknown (13)
> > Such Unknown capability is not expected. During pci-e
> > initialization,function get_port_device_capability just returns if it
> > hits an unknown capability when searching the AER capability.
> >
> > I worked out a patch against 2.6.27-rc7. When hitting an unkown
> > capability, function get_port_device_capability continues the searching.
> >
> > Tomasz tested it and the patch does work well.
> >
> > Signed-off-by Zhang Yanmin <[email protected]>
> > Reported-by Tomasz Czernecki <[email protected]>
> >
> > ---
> >
> > --- linux-2.6.27-rc7/drivers/pci/pcie/portdrv_core.c 2008-09-27
> > 09:35:32.000000000 +0800 +++
> > linux-2.6.27-rc7_aer/drivers/pci/pcie/portdrv_core.c 2008-09-27
> > 10:10:39.000000000 +0800 @@ -195,23 +195,25 @@ static int
> > get_port_device_capability(st
> > /* PME Capable - root port capability */
> > if (((reg16 >> 4) & PORT_TYPE_MASK) == PCIE_RC_PORT)
> > services |= PCIE_PORT_SERVICE_PME;
> > -
> > +
> > pos = PCI_CFG_SPACE_SIZE;
> > while (pos) {
> > - pci_read_config_dword(dev, pos, ®32);
> > - switch (reg32 & 0xffff) {
> > + if (pci_read_config_dword(dev, pos, ®32))
> > + break;
> > +
> > + /* some broken boards return ~0 */
> > + if (reg32 == 0xffffffff)
> > + break;
> > +
> > + switch (PCI_EXT_CAP_ID(reg32)) {
> > case PCI_EXT_CAP_ID_ERR:
> > services |= PCIE_PORT_SERVICE_AER;
> > - pos = reg32 >> 20;
> > break;
> > case PCI_EXT_CAP_ID_VC:
> > services |= PCIE_PORT_SERVICE_VC;
> > - pos = reg32 >> 20;
> > - break;
> > - default:
> > - pos = 0;
> > break;
> > }
> > + pos = PCI_EXT_CAP_NEXT(reg32);
> > }
> >
> > return services;
> >
> >
> >
> > --
> > To unsubscribe from this list: send the line "unsubscribe linux-kernel"
> > in the body of a message to [email protected]
> > More majordomo info at http://vger.kernel.org/majordomo-info.html
> > Please read the FAQ at http://www.tux.org/lkml/
--
Jesse Barnes, Intel Open Source Technology Center
On Fri, 2008-10-03 at 08:34 -0700, Jesse Barnes wrote:
> I'd rather get rid of this function altogether. Can you check out my "use
> pci_find_ext_capability everywhere" patch and see if you can do something
> similar for the port driver?
Jesse,
I did a quick check. If applying the patches from you and Matthew, port driver
has no its own version of find capability.
Yanmin
> I mainly just want to avoid having duplicated
> find_capability code everywhere...
>
> Thanks,
> Jesse
>
> On Friday, October 3, 2008 1:21 am Zhang, Yanmin wrote:
> > Greg, Jesse,
> >
> > Is there any issue with the patch? Tomasz need kernel to have the patch to
> > work on his machine.
> >
> > Thanks,
> > Yanmin
> >
> > On Tue, 2008-09-30 at 12:29 +0800, Zhang, Yanmin wrote:
> > > Subject: pci-e: ignore unknown capability and continue searching
> > > From: "Zhang, Yanmin" <[email protected]>
> > >
> > >
> > > Tomasz reported AER driver couldn't work on his machine. With the output
> > > of "lcpsi -vvv", I found the root port's extended capabilities are
> > > Capabilities: [100] Unknown (11)
> > > Capabilities: [150] Advanced Error Reporting
> > > Capabilities: [190] Unknown (13)
> > > Such Unknown capability is not expected. During pci-e
> > > initialization,function get_port_device_capability just returns if it
> > > hits an unknown capability when searching the AER capability.
> > >
> > > I worked out a patch against 2.6.27-rc7. When hitting an unkown
> > > capability, function get_port_device_capability continues the searching.
> > >
> > > Tomasz tested it and the patch does work well.
> > >
> > > Signed-off-by Zhang Yanmin <[email protected]>
> > > Reported-by Tomasz Czernecki <[email protected]>
> > >
> > > ---
> > >
> > > --- linux-2.6.27-rc7/drivers/pci/pcie/portdrv_core.c 2008-09-27
> > > 09:35:32.000000000 +0800 +++
> > > linux-2.6.27-rc7_aer/drivers/pci/pcie/portdrv_core.c 2008-09-27
> > > 10:10:39.000000000 +0800 @@ -195,23 +195,25 @@ static int
> > > get_port_device_capability(st
> > > /* PME Capable - root port capability */
> > > if (((reg16 >> 4) & PORT_TYPE_MASK) == PCIE_RC_PORT)
> > > services |= PCIE_PORT_SERVICE_PME;
> > > -
> > > +
> > > pos = PCI_CFG_SPACE_SIZE;
> > > while (pos) {
> > > - pci_read_config_dword(dev, pos, ®32);
> > > - switch (reg32 & 0xffff) {
> > > + if (pci_read_config_dword(dev, pos, ®32))
> > > + break;
> > > +
> > > + /* some broken boards return ~0 */
> > > + if (reg32 == 0xffffffff)
> > > + break;
> > > +
> > > + switch (PCI_EXT_CAP_ID(reg32)) {
> > > case PCI_EXT_CAP_ID_ERR:
> > > services |= PCIE_PORT_SERVICE_AER;
> > > - pos = reg32 >> 20;
> > > break;
> > > case PCI_EXT_CAP_ID_VC:
> > > services |= PCIE_PORT_SERVICE_VC;
> > > - pos = reg32 >> 20;
> > > - break;
> > > - default:
> > > - pos = 0;
> > > break;
> > > }
> > > + pos = PCI_EXT_CAP_NEXT(reg32);
> > > }
> > >
> > > return services;
> > >
> > >
> > >
> > > --
> > > To unsubscribe from this list: send the line "unsubscribe linux-kernel"
> > > in the body of a message to [email protected]
> > > More majordomo info at http://vger.kernel.org/majordomo-info.html
> > > Please read the FAQ at http://www.tux.org/lkml/
>
>
>