Patchwork PCI/MSI: don't disable AMD IOMMU MSI on Xen dom0 (was: Re: [PATCH V2] amd iommu: re-enable iommu msi if dom0 disabled it)

login
register
mail settings
Submitter Jan Beulich
Date June 21, 2012, 9:59 a.m.
Message ID <4FE30CBB020000780008B06B@nat28.tlf.novell.com>
Download mbox | patch
Permalink /patch/166291/
State Rejected
Headers show

Comments

Jan Beulich - June 21, 2012, 9:59 a.m.
>>> On 14.06.12 at 17:15, Wei Wang <wei.wang2@amd.com> wrote:
> Am 14.06.2012 16:18, schrieb Jan Beulich:
>> Have you at all considered getting this fixed on the kernel side?
>> As I don't have direct access to any AMD IOMMU capable
>> system - can one, other than by enumerating the respective
>> PCI IDs or reading ACPI tables, reasonably easily identify the
>> devices in question (e.g. via vendor/class/sub-class or some
>> such)? That might permit skipping those in the offending kernel
>> code...
> 
> AMD IOMMUs (both v1 and v2) uses class id 08 (System Base Peripheral) 
> and sub class id 06 (IOMMU). Combined with PCI_VENDEOR_ID_AMD, this 
> should be enough to identify amd iommu device. I could send you a kernel 
> patch for review using this approach. I would believe that fixing this 
> issue in 4.2, no matter how, is really important for amd iommu.

As you didn't come forward with anything, here's my first
take on this:

Commit d5dea7d95c48d7bc951cee4910a7fd9c0cd26fb0 ("PCI: msi: Disable msi
interrupts when we initialize a pci device") caused MSI to get disabled
on Xen Dom0 despite it having got turned on purposefully by the
hypervisor. As an immediate band aid, suppress the disabling in this
specific case.

I don't think, however, that either the original change or this fix are
really appropriate. The original fix, besides leaving aside the
presence of a hypervisor like Xen, doesn't deal with all possible
cases (in particular it has no effect if the secondary kernel was built
with CONFIG_PCI_MSI unset). And taking into account a hypervisor like
Xen, the logic like this should probably be skipped altogether (i.e. it
should be entirely left to the hypervisor, being the entity in control
of IRQs).

Signed-off-by: Jan Beulich <jbeulich@suse.com>
Cc: stable@kernel.org 

---
 drivers/pci/msi.c       |    7 +++++++
 include/linux/pci_ids.h |    1 +
 2 files changed, 8 insertions(+)




--
To unsubscribe from this list: send the line "unsubscribe linux-pci" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Eric W. Biederman - June 21, 2012, 11:08 a.m.
"Jan Beulich" <JBeulich@suse.com> writes:

>>>> On 14.06.12 at 17:15, Wei Wang <wei.wang2@amd.com> wrote:
>> Am 14.06.2012 16:18, schrieb Jan Beulich:
>>> Have you at all considered getting this fixed on the kernel side?
>>> As I don't have direct access to any AMD IOMMU capable
>>> system - can one, other than by enumerating the respective
>>> PCI IDs or reading ACPI tables, reasonably easily identify the
>>> devices in question (e.g. via vendor/class/sub-class or some
>>> such)? That might permit skipping those in the offending kernel
>>> code...
>> 
>> AMD IOMMUs (both v1 and v2) uses class id 08 (System Base Peripheral) 
>> and sub class id 06 (IOMMU). Combined with PCI_VENDEOR_ID_AMD, this 
>> should be enough to identify amd iommu device. I could send you a kernel 
>> patch for review using this approach. I would believe that fixing this 
>> issue in 4.2, no matter how, is really important for amd iommu.
>
> As you didn't come forward with anything, here's my first
> take on this:
>
> Commit d5dea7d95c48d7bc951cee4910a7fd9c0cd26fb0 ("PCI: msi: Disable msi
> interrupts when we initialize a pci device") caused MSI to get disabled
> on Xen Dom0 despite it having got turned on purposefully by the
> hypervisor. As an immediate band aid, suppress the disabling in this
> specific case.
>
> I don't think, however, that either the original change or this fix are
> really appropriate. The original fix, besides leaving aside the
> presence of a hypervisor like Xen, doesn't deal with all possible
> cases (in particular it has no effect if the secondary kernel was built
> with CONFIG_PCI_MSI unset). And taking into account a hypervisor like
> Xen, the logic like this should probably be skipped altogether (i.e. it
> should be entirely left to the hypervisor, being the entity in control
> of IRQs).

The original fix is fine. Xen dom0 remains insane.  Although perhaps
some better than Xen dom0 once was.

Why does the dom0 kernel even get any access to pci devices that
Xen doesn't want it to touch?

As far as I can tell my patch has revealed a design bug in Xen.  If you
don't want to be messed up by the kernel don't let the kernel touch
things.  At the very least we need an abstraction much cleaner
than the patch below.

Eric


> Signed-off-by: Jan Beulich <jbeulich@suse.com>
> Cc: stable@kernel.org 
>
> ---
>  drivers/pci/msi.c       |    7 +++++++
>  include/linux/pci_ids.h |    1 +
>  2 files changed, 8 insertions(+)
>
> --- 3.5-rc3/drivers/pci/msi.c
> +++ 3.5-rc3-xen-pci-msi-no-iommu-disable/drivers/pci/msi.c
> @@ -20,6 +20,7 @@
>  #include <linux/errno.h>
>  #include <linux/io.h>
>  #include <linux/slab.h>
> +#include <xen/xen.h>
>  
>  #include "pci.h"
>  #include "msi.h"
> @@ -1022,7 +1023,13 @@ void pci_msi_init_pci_dev(struct pci_dev
>  	/* Disable the msi hardware to avoid screaming interrupts
>  	 * during boot.  This is the power on reset default so
>  	 * usually this should be a noop.
> +	 * But on a Xen host don't do this for IOMMUs which the hypervisor
> +	 * is in control of (and hence has already enabled on purpose).
>  	 */
> +	if (xen_initial_domain()
> +	    && (dev->class >> 8) == PCI_CLASS_SYSTEM_IOMMU
> +	    && dev->vendor == PCI_VENDOR_ID_AMD)
> +		return;
>  	pos = pci_find_capability(dev, PCI_CAP_ID_MSI);
>  	if (pos)
>  		msi_set_enable(dev, pos, 0);
> --- 3.5-rc3/include/linux/pci_ids.h
> +++ 3.5-rc3-xen-pci-msi-no-iommu-disable/include/linux/pci_ids.h
> @@ -75,6 +75,7 @@
>  #define PCI_CLASS_SYSTEM_RTC		0x0803
>  #define PCI_CLASS_SYSTEM_PCI_HOTPLUG	0x0804
>  #define PCI_CLASS_SYSTEM_SDHCI		0x0805
> +#define PCI_CLASS_SYSTEM_IOMMU		0x0806
>  #define PCI_CLASS_SYSTEM_OTHER		0x0880
>  
>  #define PCI_BASE_CLASS_INPUT		0x09
--
To unsubscribe from this list: send the line "unsubscribe linux-pci" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Jan Beulich - June 21, 2012, 12:28 p.m.
>>> On 21.06.12 at 13:08, Eric W. Biederman <ebiederm@xmission.com> wrote:
> "Jan Beulich" <JBeulich@suse.com> writes:
> 
>>>>> On 14.06.12 at 17:15, Wei Wang <wei.wang2@amd.com> wrote:
>>> Am 14.06.2012 16:18, schrieb Jan Beulich:
>>>> Have you at all considered getting this fixed on the kernel side?
>>>> As I don't have direct access to any AMD IOMMU capable
>>>> system - can one, other than by enumerating the respective
>>>> PCI IDs or reading ACPI tables, reasonably easily identify the
>>>> devices in question (e.g. via vendor/class/sub-class or some
>>>> such)? That might permit skipping those in the offending kernel
>>>> code...
>>> 
>>> AMD IOMMUs (both v1 and v2) uses class id 08 (System Base Peripheral) 
>>> and sub class id 06 (IOMMU). Combined with PCI_VENDEOR_ID_AMD, this 
>>> should be enough to identify amd iommu device. I could send you a kernel 
>>> patch for review using this approach. I would believe that fixing this 
>>> issue in 4.2, no matter how, is really important for amd iommu.
>>
>> As you didn't come forward with anything, here's my first
>> take on this:
>>
>> Commit d5dea7d95c48d7bc951cee4910a7fd9c0cd26fb0 ("PCI: msi: Disable msi
>> interrupts when we initialize a pci device") caused MSI to get disabled
>> on Xen Dom0 despite it having got turned on purposefully by the
>> hypervisor. As an immediate band aid, suppress the disabling in this
>> specific case.
>>
>> I don't think, however, that either the original change or this fix are
>> really appropriate. The original fix, besides leaving aside the
>> presence of a hypervisor like Xen, doesn't deal with all possible
>> cases (in particular it has no effect if the secondary kernel was built
>> with CONFIG_PCI_MSI unset). And taking into account a hypervisor like
>> Xen, the logic like this should probably be skipped altogether (i.e. it
>> should be entirely left to the hypervisor, being the entity in control
>> of IRQs).
> 
> The original fix is fine. Xen dom0 remains insane.  Although perhaps
> some better than Xen dom0 once was.
> 
> Why does the dom0 kernel even get any access to pci devices that
> Xen doesn't want it to touch?
> 
> As far as I can tell my patch has revealed a design bug in Xen.  If you
> don't want to be messed up by the kernel don't let the kernel touch
> things.

I rather think this is a design bug of the AMD IOMMU - it should
never have got surfaced as a PCI device.

Furthermore, fully hiding _any_ PCI device from Dom0 has its
own set of problems - depending on the nature of the device,
full emulation of all devices may become necessary to get this
implemented (because this can implicitly hide other devices, or
the Dom0 kernel re-assigning BARs could get in conflict with
what the hidden devices use).

Xen's model has always been to allow Dom0 full control over all
peripherals and bridges, so if all of the sudden PCI devices of
other kinds show up, we can't simply re-model everything.

>  At the very least we need an abstraction much cleaner
> than the patch below.

Probably - any suggestion? As said in the mail I sent, this is
meant to overcome the immediate problem, and a more
permanent fix would be desirable (ideally suppressing the
playing with the MSI related data altogether, following the
rest of the MSI support in Xen/Dom0).

Jan

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

Patch

--- 3.5-rc3/drivers/pci/msi.c
+++ 3.5-rc3-xen-pci-msi-no-iommu-disable/drivers/pci/msi.c
@@ -20,6 +20,7 @@ 
 #include <linux/errno.h>
 #include <linux/io.h>
 #include <linux/slab.h>
+#include <xen/xen.h>
 
 #include "pci.h"
 #include "msi.h"
@@ -1022,7 +1023,13 @@  void pci_msi_init_pci_dev(struct pci_dev
 	/* Disable the msi hardware to avoid screaming interrupts
 	 * during boot.  This is the power on reset default so
 	 * usually this should be a noop.
+	 * But on a Xen host don't do this for IOMMUs which the hypervisor
+	 * is in control of (and hence has already enabled on purpose).
 	 */
+	if (xen_initial_domain()
+	    && (dev->class >> 8) == PCI_CLASS_SYSTEM_IOMMU
+	    && dev->vendor == PCI_VENDOR_ID_AMD)
+		return;
 	pos = pci_find_capability(dev, PCI_CAP_ID_MSI);
 	if (pos)
 		msi_set_enable(dev, pos, 0);
--- 3.5-rc3/include/linux/pci_ids.h
+++ 3.5-rc3-xen-pci-msi-no-iommu-disable/include/linux/pci_ids.h
@@ -75,6 +75,7 @@ 
 #define PCI_CLASS_SYSTEM_RTC		0x0803
 #define PCI_CLASS_SYSTEM_PCI_HOTPLUG	0x0804
 #define PCI_CLASS_SYSTEM_SDHCI		0x0805
+#define PCI_CLASS_SYSTEM_IOMMU		0x0806
 #define PCI_CLASS_SYSTEM_OTHER		0x0880
 
 #define PCI_BASE_CLASS_INPUT		0x09