Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1219192 > unrolled thread

[PATCH v4 0/4] PCI: arm64/powerpc: Fix parsing of linux,pci-probe-only

Back to article view | Back to linux.kernel

Contents

  [PATCH v4 0/4] PCI: arm64/powerpc: Fix parsing of linux,pci-probe-only Marc Zyngier <marc.zyngier@arm.com> - 2015-09-04 19:00 +0200
    [PATCH v4 3/4] powerpc: PCI: Fix lookup of linux,pci-probe-only property Marc Zyngier <marc.zyngier@arm.com> - 2015-09-04 19:00 +0200
      Re: [PATCH v4 3/4] powerpc: PCI: Fix lookup of linux,pci-probe-only  property Michael Ellerman <mpe@ellerman.id.au> - 2015-09-07 11:00 +0200
    Re: [PATCH v4 0/4] PCI: arm64/powerpc: Fix parsing of linux,pci-probe-only Rob Herring <robherring2@gmail.com> - 2015-09-04 21:00 +0200
    Re: [PATCH v4 0/4] PCI: arm64/powerpc: Fix parsing of  linux,pci-probe-only Bjorn Helgaas <bhelgaas@google.com> - 2015-09-17 17:40 +0200
      Re: [PATCH v4 0/4] PCI: arm64/powerpc: Fix parsing of linux,pci-probe-only Marc Zyngier <marc.zyngier@arm.com> - 2015-09-17 19:20 +0200
        Re: [PATCH v4 0/4] PCI: arm64/powerpc: Fix parsing of linux,pci-probe-only Bjorn Helgaas <bhelgaas@google.com> - 2015-09-17 19:30 +0200

#1219192 — [PATCH v4 0/4] PCI: arm64/powerpc: Fix parsing of linux,pci-probe-only

The pci-host-generic driver parses the linux,pci-probe-only property,
and assumes that it will have a boolean parameter.

Turns out that the Seattle DTS file has a naked "linux,pci-probe-only"
property, which leads to the driver dereferencing some unsuspecting
memory location. Nothing really bad happens (we end up reading some
other bit of DT, fortunately), but that not a reason to keep it this
way. Turns out that the Pseries code (where this code was lifted from)
may suffer from the same issue.

The first patch introduces a common (and fixed) version of that check
that can be used by drivers and architectures that require it. The two
following patches change the pci-host-generic driver and the powerpc
code to use it.

Finally, the bad property is removed from the Seatle DTS, because it
is simply not necessary (it actually prevents me from using SR-IOV,
which otherwise runs fine without the probe-only thing).

This has been tested on the offending Seattle board.

* From v3:
  - Restrict the property lookup to /chosen (Rob)
  - Acked-by on patch #4 from Suravee
  - I swear this is the last time I rework these patches! ;-)

* From v2:
  - Use of_property_read_u32 to safely read the property (Rob)
  - Add a log message to indicate when we enable probe-only
    (probably quite useful for debugging)

* From v1:
  - Consolidate the parsing in of_pci.c (Bjorn)

Marc Zyngier (4):
  of/pci: Add of_pci_check_probe_only to parse "linux,pci-probe-only"
  PCI: pci-host-generic: Fix lookup of linux,pci-probe-only property
  powerpc: PCI: Fix lookup of linux,pci-probe-only property
  arm64: dts: Drop linux,pci-probe-only from the Seattle DTS

 arch/arm64/boot/dts/amd/amd-overdrive.dts |  1 -
 arch/powerpc/platforms/pseries/setup.c    | 14 ++------------
 drivers/of/of_pci.c                       | 28 ++++++++++++++++++++++++++++
 drivers/pci/host/pci-host-generic.c       |  9 +--------
 include/linux/of_pci.h                    |  3 +++
 5 files changed, 34 insertions(+), 21 deletions(-)

-- 
2.1.4

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [next] | [standalone]

#1219198 — [PATCH v4 3/4] powerpc: PCI: Fix lookup of linux,pci-probe-only property

When find_and_init_phbs() looks for the probe-only property, it seems
to trust the firmware to be correctly written, and assumes that there
is a parameter to the property.

It is conceivable that the firmware could not be that perfect, and it
could expose this property naked (at least one arm64 platform seems to
exhibit this exact behaviour). The setup code the ends up making
a decision based on whatever the property pointer points to, which
is likely to be junk.

Instead, switch to the common of_pci.c implementation that doesn't
suffer from this problem and ignore the property if the firmware
couldn't make up its mind.

Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
---
 arch/powerpc/platforms/pseries/setup.c | 14 ++------------
 1 file changed, 2 insertions(+), 12 deletions(-)

diff --git a/arch/powerpc/platforms/pseries/setup.c b/arch/powerpc/platforms/pseries/setup.c
index 39a74fa..6016709 100644
--- a/arch/powerpc/platforms/pseries/setup.c
+++ b/arch/powerpc/platforms/pseries/setup.c
@@ -40,6 +40,7 @@
 #include <linux/seq_file.h>
 #include <linux/root_dev.h>
 #include <linux/of.h>
+#include <linux/of_pci.h>
 #include <linux/kexec.h>
 
 #include <asm/mmu.h>
@@ -495,18 +496,7 @@ static void __init find_and_init_phbs(void)
 	 * PCI_PROBE_ONLY and PCI_REASSIGN_ALL_BUS can be set via properties
 	 * in chosen.
 	 */
-	if (of_chosen) {
-		const int *prop;
-
-		prop = of_get_property(of_chosen,
-				"linux,pci-probe-only", NULL);
-		if (prop) {
-			if (*prop)
-				pci_add_flags(PCI_PROBE_ONLY);
-			else
-				pci_clear_flags(PCI_PROBE_ONLY);
-		}
-	}
+	of_pci_check_probe_only();
 }
 
 static void __init pSeries_setup_arch(void)
-- 
2.1.4

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [next] | [standalone]

#1220006 — Re: [PATCH v4 3/4] powerpc: PCI: Fix lookup of linux,pci-probe-only property

On Fri, 2015-09-04 at 17:50 +0100, Marc Zyngier wrote:
> When find_and_init_phbs() looks for the probe-only property, it seems
> to trust the firmware to be correctly written, and assumes that there
> is a parameter to the property.
> 
> It is conceivable that the firmware could not be that perfect, and it
> could expose this property naked (at least one arm64 platform seems to
> exhibit this exact behaviour). The setup code the ends up making
> a decision based on whatever the property pointer points to, which
> is likely to be junk.
> 
> Instead, switch to the common of_pci.c implementation that doesn't
> suffer from this problem and ignore the property if the firmware
> couldn't make up its mind.
> 
> Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
> ---
>  arch/powerpc/platforms/pseries/setup.c | 14 ++------------

Thanks.

I can't find any powerpc machine that is currently using this, or much info on
what ever did use it, so I think it's pretty low impact, and the change looks
sane.

Acked-by: Michael Ellerman <mpe@ellerman.id.au>

cheers


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [next] | [standalone]

#1219256

On Fri, Sep 4, 2015 at 11:50 AM, Marc Zyngier <marc.zyngier@arm.com> wrote:
> The pci-host-generic driver parses the linux,pci-probe-only property,
> and assumes that it will have a boolean parameter.
>
> Turns out that the Seattle DTS file has a naked "linux,pci-probe-only"
> property, which leads to the driver dereferencing some unsuspecting
> memory location. Nothing really bad happens (we end up reading some
> other bit of DT, fortunately), but that not a reason to keep it this
> way. Turns out that the Pseries code (where this code was lifted from)
> may suffer from the same issue.
>
> The first patch introduces a common (and fixed) version of that check
> that can be used by drivers and architectures that require it. The two
> following patches change the pci-host-generic driver and the powerpc
> code to use it.
>
> Finally, the bad property is removed from the Seatle DTS, because it
> is simply not necessary (it actually prevents me from using SR-IOV,
> which otherwise runs fine without the probe-only thing).
>
> This has been tested on the offending Seattle board.
>
> * From v3:
>   - Restrict the property lookup to /chosen (Rob)
>   - Acked-by on patch #4 from Suravee
>   - I swear this is the last time I rework these patches! ;-)
>
> * From v2:
>   - Use of_property_read_u32 to safely read the property (Rob)
>   - Add a log message to indicate when we enable probe-only
>     (probably quite useful for debugging)
>
> * From v1:
>   - Consolidate the parsing in of_pci.c (Bjorn)
>
> Marc Zyngier (4):
>   of/pci: Add of_pci_check_probe_only to parse "linux,pci-probe-only"
>   PCI: pci-host-generic: Fix lookup of linux,pci-probe-only property
>   powerpc: PCI: Fix lookup of linux,pci-probe-only property
>   arm64: dts: Drop linux,pci-probe-only from the Seattle DTS
>
>  arch/arm64/boot/dts/amd/amd-overdrive.dts |  1 -
>  arch/powerpc/platforms/pseries/setup.c    | 14 ++------------
>  drivers/of/of_pci.c                       | 28 ++++++++++++++++++++++++++++
>  drivers/pci/host/pci-host-generic.c       |  9 +--------
>  include/linux/of_pci.h                    |  3 +++
>  5 files changed, 34 insertions(+), 21 deletions(-)

For the series:

Reviewed-by: Rob Herring <robh@kernel.org>

Rob
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [next] | [standalone]

#1227140 — Re: [PATCH v4 0/4] PCI: arm64/powerpc: Fix parsing of linux,pci-probe-only

On Fri, Sep 04, 2015 at 05:50:07PM +0100, Marc Zyngier wrote:
> The pci-host-generic driver parses the linux,pci-probe-only property,
> and assumes that it will have a boolean parameter.
> 
> Turns out that the Seattle DTS file has a naked "linux,pci-probe-only"
> property, which leads to the driver dereferencing some unsuspecting
> memory location. Nothing really bad happens (we end up reading some
> other bit of DT, fortunately), but that not a reason to keep it this
> way. Turns out that the Pseries code (where this code was lifted from)
> may suffer from the same issue.
> 
> The first patch introduces a common (and fixed) version of that check
> that can be used by drivers and architectures that require it. The two
> following patches change the pci-host-generic driver and the powerpc
> code to use it.
> 
> Finally, the bad property is removed from the Seatle DTS, because it
> is simply not necessary (it actually prevents me from using SR-IOV,
> which otherwise runs fine without the probe-only thing).
> 
> This has been tested on the offending Seattle board.
> 
> * From v3:
>   - Restrict the property lookup to /chosen (Rob)
>   - Acked-by on patch #4 from Suravee
>   - I swear this is the last time I rework these patches! ;-)
> 
> * From v2:
>   - Use of_property_read_u32 to safely read the property (Rob)
>   - Add a log message to indicate when we enable probe-only
>     (probably quite useful for debugging)
> 
> * From v1:
>   - Consolidate the parsing in of_pci.c (Bjorn)
> 
> Marc Zyngier (4):
>   of/pci: Add of_pci_check_probe_only to parse "linux,pci-probe-only"
>   PCI: pci-host-generic: Fix lookup of linux,pci-probe-only property
>   powerpc: PCI: Fix lookup of linux,pci-probe-only property
>   arm64: dts: Drop linux,pci-probe-only from the Seattle DTS
> 
>  arch/arm64/boot/dts/amd/amd-overdrive.dts |  1 -
>  arch/powerpc/platforms/pseries/setup.c    | 14 ++------------
>  drivers/of/of_pci.c                       | 28 ++++++++++++++++++++++++++++
>  drivers/pci/host/pci-host-generic.c       |  9 +--------
>  include/linux/of_pci.h                    |  3 +++
>  5 files changed, 34 insertions(+), 21 deletions(-)

Applied with the comment tweak and acks to pci/host-generic for v4.4,
thanks!
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [next] | [standalone]

#1227229

On 17/09/15 16:30, Bjorn Helgaas wrote:
> On Fri, Sep 04, 2015 at 05:50:07PM +0100, Marc Zyngier wrote:
>> The pci-host-generic driver parses the linux,pci-probe-only property,
>> and assumes that it will have a boolean parameter.
>>
>> Turns out that the Seattle DTS file has a naked "linux,pci-probe-only"
>> property, which leads to the driver dereferencing some unsuspecting
>> memory location. Nothing really bad happens (we end up reading some
>> other bit of DT, fortunately), but that not a reason to keep it this
>> way. Turns out that the Pseries code (where this code was lifted from)
>> may suffer from the same issue.
>>
>> The first patch introduces a common (and fixed) version of that check
>> that can be used by drivers and architectures that require it. The two
>> following patches change the pci-host-generic driver and the powerpc
>> code to use it.
>>
>> Finally, the bad property is removed from the Seatle DTS, because it
>> is simply not necessary (it actually prevents me from using SR-IOV,
>> which otherwise runs fine without the probe-only thing).
>>
>> This has been tested on the offending Seattle board.
>>
>> * From v3:
>>   - Restrict the property lookup to /chosen (Rob)
>>   - Acked-by on patch #4 from Suravee
>>   - I swear this is the last time I rework these patches! ;-)
>>
>> * From v2:
>>   - Use of_property_read_u32 to safely read the property (Rob)
>>   - Add a log message to indicate when we enable probe-only
>>     (probably quite useful for debugging)
>>
>> * From v1:
>>   - Consolidate the parsing in of_pci.c (Bjorn)
>>
>> Marc Zyngier (4):
>>   of/pci: Add of_pci_check_probe_only to parse "linux,pci-probe-only"
>>   PCI: pci-host-generic: Fix lookup of linux,pci-probe-only property
>>   powerpc: PCI: Fix lookup of linux,pci-probe-only property
>>   arm64: dts: Drop linux,pci-probe-only from the Seattle DTS
>>
>>  arch/arm64/boot/dts/amd/amd-overdrive.dts |  1 -
>>  arch/powerpc/platforms/pseries/setup.c    | 14 ++------------
>>  drivers/of/of_pci.c                       | 28 ++++++++++++++++++++++++++++
>>  drivers/pci/host/pci-host-generic.c       |  9 +--------
>>  include/linux/of_pci.h                    |  3 +++
>>  5 files changed, 34 insertions(+), 21 deletions(-)
> 
> Applied with the comment tweak and acks to pci/host-generic for v4.4,
> thanks!

Turns out that the 01.org infrastructure has picked up on a compilation
bug with randconfig. The following patch seems to fix it and should be
applied on the first patch:

diff --git a/drivers/of/of_pci.c b/drivers/of/of_pci.c
index 485d625..2da5abc 100644
--- a/drivers/of/of_pci.c
+++ b/drivers/of/of_pci.c
@@ -6,6 +6,8 @@
 #include <linux/of_pci.h>
 #include <linux/slab.h>
 
+#include <asm-generic/pci-bridge.h>
+
 static inline int __of_pci_pci_compare(struct device_node *node,
 				       unsigned int data)
 {

Sorry for the annoyance.

	M.
-- 
Jazz is not dead. It just smells funny...
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [next] | [standalone]

#1227234

On Thu, Sep 17, 2015 at 12:17 PM, Marc Zyngier <marc.zyngier@arm.com> wrote:
> On 17/09/15 16:30, Bjorn Helgaas wrote:
>> On Fri, Sep 04, 2015 at 05:50:07PM +0100, Marc Zyngier wrote:
>>> The pci-host-generic driver parses the linux,pci-probe-only property,
>>> and assumes that it will have a boolean parameter.
>>>
>>> Turns out that the Seattle DTS file has a naked "linux,pci-probe-only"
>>> property, which leads to the driver dereferencing some unsuspecting
>>> memory location. Nothing really bad happens (we end up reading some
>>> other bit of DT, fortunately), but that not a reason to keep it this
>>> way. Turns out that the Pseries code (where this code was lifted from)
>>> may suffer from the same issue.
>>>
>>> The first patch introduces a common (and fixed) version of that check
>>> that can be used by drivers and architectures that require it. The two
>>> following patches change the pci-host-generic driver and the powerpc
>>> code to use it.
>>>
>>> Finally, the bad property is removed from the Seatle DTS, because it
>>> is simply not necessary (it actually prevents me from using SR-IOV,
>>> which otherwise runs fine without the probe-only thing).
>>>
>>> This has been tested on the offending Seattle board.
>>>
>>> * From v3:
>>>   - Restrict the property lookup to /chosen (Rob)
>>>   - Acked-by on patch #4 from Suravee
>>>   - I swear this is the last time I rework these patches! ;-)
>>>
>>> * From v2:
>>>   - Use of_property_read_u32 to safely read the property (Rob)
>>>   - Add a log message to indicate when we enable probe-only
>>>     (probably quite useful for debugging)
>>>
>>> * From v1:
>>>   - Consolidate the parsing in of_pci.c (Bjorn)
>>>
>>> Marc Zyngier (4):
>>>   of/pci: Add of_pci_check_probe_only to parse "linux,pci-probe-only"
>>>   PCI: pci-host-generic: Fix lookup of linux,pci-probe-only property
>>>   powerpc: PCI: Fix lookup of linux,pci-probe-only property
>>>   arm64: dts: Drop linux,pci-probe-only from the Seattle DTS
>>>
>>>  arch/arm64/boot/dts/amd/amd-overdrive.dts |  1 -
>>>  arch/powerpc/platforms/pseries/setup.c    | 14 ++------------
>>>  drivers/of/of_pci.c                       | 28 ++++++++++++++++++++++++++++
>>>  drivers/pci/host/pci-host-generic.c       |  9 +--------
>>>  include/linux/of_pci.h                    |  3 +++
>>>  5 files changed, 34 insertions(+), 21 deletions(-)
>>
>> Applied with the comment tweak and acks to pci/host-generic for v4.4,
>> thanks!
>
> Turns out that the 01.org infrastructure has picked up on a compilation
> bug with randconfig. The following patch seems to fix it and should be
> applied on the first patch:
>
> diff --git a/drivers/of/of_pci.c b/drivers/of/of_pci.c
> index 485d625..2da5abc 100644
> --- a/drivers/of/of_pci.c
> +++ b/drivers/of/of_pci.c
> @@ -6,6 +6,8 @@
>  #include <linux/of_pci.h>
>  #include <linux/slab.h>
>
> +#include <asm-generic/pci-bridge.h>
> +
>  static inline int __of_pci_pci_compare(struct device_node *node,
>                                        unsigned int data)
>  {
>
> Sorry for the annoyance.

Folded in and re-pushed, thanks!
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web