Re: [ 102/127] iommu/amd: Workaround for ERBT1312

Re: [ 102/127] iommu/amd: Workaround for ERBT1312

[Andreas Hartmann]

Hello Joerg, hello Alex,

the subsequent patch and the patch "iommu/amd: Re-enable IOMMU event log
interrupt after handling." 925fe08bce38d1ff052fe2209b9e2b8d5fbb7f98
spread /var/log/messages with the following line (> 700 lines/second)
right after loading vfio:

AMD-Vi: Event logged [IO_PAGE_FAULT device=00:14.0 domain=0x0000 address=0x000000fdf9103300 flags=0x0600]

lspci -vvvs 0:14.0
00:14.0 SMBus: Advanced Micro Devices [AMD] nee ATI SBx00 SMBus Controller (rev 42)
        Control: I/O+ Mem+ BusMaster- SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx+
        Status: Cap- 66MHz+ UDF- FastB2B- ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-


Besides the enormous pollution I couldn't see any malfunction at all.
At first, I didn't realised it at all (-> the SSD was fast enough to
cover it silently). I saw it the first time I rebooted because X didn't start any more because
the /var partition was completely full. 

I removed the two mentioned patches and all is working
fine again as before.

Any idea?


Thanks,
kind regards,
Andreas


Greg Kroah-Hartman wrote:
> 3.9-stable review patch.  If anyone has any objections, please let me know.
> 
> ------------------
> 
> From: Joerg Roedel <joro@8bytes.org>
> 
> commit d3263bc29706e42f74d8800807c2dedf320d77f1 upstream.
> 
> Work around an IOMMU  hardware bug where clearing the
> EVT_INT or PPR_INT bit in the status register may race with
> the hardware trying to set it again. When not handled the
> bit might not be cleared and we lose all future event or ppr
> interrupts.
> 
> Reported-by: Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>
> Signed-off-by: Joerg Roedel <joro@8bytes.org>
> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
> 
> ---
>  drivers/iommu/amd_iommu.c |   34 ++++++++++++++++++++++++++--------
>  1 file changed, 26 insertions(+), 8 deletions(-)
> 
> --- a/drivers/iommu/amd_iommu.c
> +++ b/drivers/iommu/amd_iommu.c
> @@ -700,14 +700,23 @@ retry:
>  
>  static void iommu_poll_events(struct amd_iommu *iommu)
>  {
> -	u32 head, tail;
> +	u32 head, tail, status;
>  	unsigned long flags;
>  
> -	/* enable event interrupts again */
> -	writel(MMIO_STATUS_EVT_INT_MASK, iommu->mmio_base + MMIO_STATUS_OFFSET);
> -
>  	spin_lock_irqsave(&iommu->lock, flags);
>  
> +	/* enable event interrupts again */
> +	do {
> +		/*
> +		 * Workaround for Erratum ERBT1312
> +		 * Clearing the EVT_INT bit may race in the hardware, so read
> +		 * it again and make sure it was really cleared
> +		 */
> +		status = readl(iommu->mmio_base + MMIO_STATUS_OFFSET);
> +		writel(MMIO_STATUS_EVT_INT_MASK,
> +		       iommu->mmio_base + MMIO_STATUS_OFFSET);
> +	} while (status & MMIO_STATUS_EVT_INT_MASK);
> +
>  	head = readl(iommu->mmio_base + MMIO_EVT_HEAD_OFFSET);
>  	tail = readl(iommu->mmio_base + MMIO_EVT_TAIL_OFFSET);
>  
> @@ -744,16 +753,25 @@ static void iommu_handle_ppr_entry(struc
>  static void iommu_poll_ppr_log(struct amd_iommu *iommu)
>  {
>  	unsigned long flags;
> -	u32 head, tail;
> +	u32 head, tail, status;
>  
>  	if (iommu->ppr_log == NULL)
>  		return;
>  
> -	/* enable ppr interrupts again */
> -	writel(MMIO_STATUS_PPR_INT_MASK, iommu->mmio_base + MMIO_STATUS_OFFSET);
> -
>  	spin_lock_irqsave(&iommu->lock, flags);
>  
> +	/* enable ppr interrupts again */
> +	do {
> +		/*
> +		 * Workaround for Erratum ERBT1312
> +		 * Clearing the PPR_INT bit may race in the hardware, so read
> +		 * it again and make sure it was really cleared
> +		 */
> +		status = readl(iommu->mmio_base + MMIO_STATUS_OFFSET);
> +		writel(MMIO_STATUS_PPR_INT_MASK,
> +		       iommu->mmio_base + MMIO_STATUS_OFFSET);
> +	} while (status & MMIO_STATUS_PPR_INT_MASK);
> +
>  	head = readl(iommu->mmio_base + MMIO_PPR_HEAD_OFFSET);
>  	tail = readl(iommu->mmio_base + MMIO_PPR_TAIL_OFFSET);
>  

Re: [ 102/127] iommu/amd: Workaround for ERBT1312

[Alex Williamson]

On Fri, 2013-06-28 at 18:11 +0200, Andreas Hartmann wrote:
> Hello Joerg, hello Alex,
> 
> the subsequent patch and the patch "iommu/amd: Re-enable IOMMU event log
> interrupt after handling." 925fe08bce38d1ff052fe2209b9e2b8d5fbb7f98
> spread /var/log/messages with the following line (> 700 lines/second)
> right after loading vfio:
> 
> AMD-Vi: Event logged [IO_PAGE_FAULT device=00:14.0 domain=0x0000 address=0x000000fdf9103300 flags=0x0600]

That's interesting, I PXE boot my system from one NIC then use a
different NIC for the iSCSI root.  The PXE boot NIC now screams like
this, _until_ I attach it to vfio, then it quiets down.

> lspci -vvvs 0:14.0
> 00:14.0 SMBus: Advanced Micro Devices [AMD] nee ATI SBx00 SMBus Controller (rev 42)
>         Control: I/O+ Mem+ BusMaster- SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx+
>         Status: Cap- 66MHz+ UDF- FastB2B- ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
> 
> 
> Besides the enormous pollution I couldn't see any malfunction at all.
> At first, I didn't realised it at all (-> the SSD was fast enough to
> cover it silently). I saw it the first time I rebooted because X didn't start any more because
> the /var partition was completely full. 
> 
> I removed the two mentioned patches and all is working
> fine again as before.
> 
> Any idea?

Not really without some digging.  I wonder if it's a new event each time
or if something is just not clearing a previous event.  ISTR that a boot
used to often, but not always, generate a couple faults between the
IOMMU being enabled and the NIC driver being loaded.  All the faults I
see are to the same address, so my guess is that it's getting replayed.
Thanks,

Alex

> Greg Kroah-Hartman wrote:
> > 3.9-stable review patch.  If anyone has any objections, please let me know.
> > 
> > ------------------
> > 
> > From: Joerg Roedel <joro@8bytes.org>
> > 
> > commit d3263bc29706e42f74d8800807c2dedf320d77f1 upstream.
> > 
> > Work around an IOMMU  hardware bug where clearing the
> > EVT_INT or PPR_INT bit in the status register may race with
> > the hardware trying to set it again. When not handled the
> > bit might not be cleared and we lose all future event or ppr
> > interrupts.
> > 
> > Reported-by: Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>
> > Signed-off-by: Joerg Roedel <joro@8bytes.org>
> > Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
> > 
> > ---
> >  drivers/iommu/amd_iommu.c |   34 ++++++++++++++++++++++++++--------
> >  1 file changed, 26 insertions(+), 8 deletions(-)
> > 
> > --- a/drivers/iommu/amd_iommu.c
> > +++ b/drivers/iommu/amd_iommu.c
> > @@ -700,14 +700,23 @@ retry:
> >  
> >  static void iommu_poll_events(struct amd_iommu *iommu)
> >  {
> > -	u32 head, tail;
> > +	u32 head, tail, status;
> >  	unsigned long flags;
> >  
> > -	/* enable event interrupts again */
> > -	writel(MMIO_STATUS_EVT_INT_MASK, iommu->mmio_base + MMIO_STATUS_OFFSET);
> > -
> >  	spin_lock_irqsave(&iommu->lock, flags);
> >  
> > +	/* enable event interrupts again */
> > +	do {
> > +		/*
> > +		 * Workaround for Erratum ERBT1312
> > +		 * Clearing the EVT_INT bit may race in the hardware, so read
> > +		 * it again and make sure it was really cleared
> > +		 */
> > +		status = readl(iommu->mmio_base + MMIO_STATUS_OFFSET);
> > +		writel(MMIO_STATUS_EVT_INT_MASK,
> > +		       iommu->mmio_base + MMIO_STATUS_OFFSET);
> > +	} while (status & MMIO_STATUS_EVT_INT_MASK);
> > +
> >  	head = readl(iommu->mmio_base + MMIO_EVT_HEAD_OFFSET);
> >  	tail = readl(iommu->mmio_base + MMIO_EVT_TAIL_OFFSET);
> >  
> > @@ -744,16 +753,25 @@ static void iommu_handle_ppr_entry(struc
> >  static void iommu_poll_ppr_log(struct amd_iommu *iommu)
> >  {
> >  	unsigned long flags;
> > -	u32 head, tail;
> > +	u32 head, tail, status;
> >  
> >  	if (iommu->ppr_log == NULL)
> >  		return;
> >  
> > -	/* enable ppr interrupts again */
> > -	writel(MMIO_STATUS_PPR_INT_MASK, iommu->mmio_base + MMIO_STATUS_OFFSET);
> > -
> >  	spin_lock_irqsave(&iommu->lock, flags);
> >  
> > +	/* enable ppr interrupts again */
> > +	do {
> > +		/*
> > +		 * Workaround for Erratum ERBT1312
> > +		 * Clearing the PPR_INT bit may race in the hardware, so read
> > +		 * it again and make sure it was really cleared
> > +		 */
> > +		status = readl(iommu->mmio_base + MMIO_STATUS_OFFSET);
> > +		writel(MMIO_STATUS_PPR_INT_MASK,
> > +		       iommu->mmio_base + MMIO_STATUS_OFFSET);
> > +	} while (status & MMIO_STATUS_PPR_INT_MASK);
> > +
> >  	head = readl(iommu->mmio_base + MMIO_PPR_HEAD_OFFSET);
> >  	tail = readl(iommu->mmio_base + MMIO_PPR_TAIL_OFFSET);
> >  

Re: [ 102/127] iommu/amd: Workaround for ERBT1312

[Joerg Roedel]

Hi Alex,

On Fri, Jun 28, 2013 at 11:49:04AM -0600, Alex Williamson wrote:
> That's interesting, I PXE boot my system from one NIC then use a
> different NIC for the iSCSI root.  The PXE boot NIC now screams like
> this, _until_ I attach it to vfio, then it quiets down.

Can you please send an example line of the reported fault? The addresses
it faults on would be interesting.

> > Any idea?
> 
> Not really without some digging.  I wonder if it's a new event each time
> or if something is just not clearing a previous event.  ISTR that a boot
> used to often, but not always, generate a couple faults between the
> IOMMU being enabled and the NIC driver being loaded.  All the faults I
> see are to the same address, so my guess is that it's getting replayed.

Well, I think it is a problem uncovered by the patch that re-enables the
event-log interrupt after it happened once. We need to find a strategy
to cope with those problems.

To my mind as a quick-fix comes rate-limiting for the printks. Or we use
the suppress-pf bit in the DTE to suppress all page-faults after the
first one.


	Joerg

Re: [ 102/127] iommu/amd: Workaround for ERBT1312

[Alex Williamson]

On Fri, 2013-06-28 at 20:29 +0200, Joerg Roedel wrote:
> Hi Alex,
> 
> On Fri, Jun 28, 2013 at 11:49:04AM -0600, Alex Williamson wrote:
> > That's interesting, I PXE boot my system from one NIC then use a
> > different NIC for the iSCSI root.  The PXE boot NIC now screams like
> > this, _until_ I attach it to vfio, then it quiets down.
> 
> Can you please send an example line of the reported fault? The addresses
> it faults on would be interesting.

[   99.613489] AMD-Vi: Event logged [IO_PAGE_FAULT device=02:00.0
domain=0x0000 address=0x000000000008f880 flags=0x0050]

> > > Any idea?
> > 
> > Not really without some digging.  I wonder if it's a new event each time
> > or if something is just not clearing a previous event.  ISTR that a boot
> > used to often, but not always, generate a couple faults between the
> > IOMMU being enabled and the NIC driver being loaded.  All the faults I
> > see are to the same address, so my guess is that it's getting replayed.
> 
> Well, I think it is a problem uncovered by the patch that re-enables the
> event-log interrupt after it happened once. We need to find a strategy
> to cope with those problems.
> 
> To my mind as a quick-fix comes rate-limiting for the printks. Or we use
> the suppress-pf bit in the DTE to suppress all page-faults after the
> first one.
> 
> 
> 	Joerg
> 
> 

Re: [ 102/127] iommu/amd: Workaround for ERBT1312

[Andreas Hartmann]

Alex Williamson wrote:
> On Fri, 2013-06-28 at 18:11 +0200, Andreas Hartmann wrote:
>> Hello Joerg, hello Alex,
>>
>> the subsequent patch and the patch "iommu/amd: Re-enable IOMMU event log
>> interrupt after handling." 925fe08bce38d1ff052fe2209b9e2b8d5fbb7f98
>> spread /var/log/messages with the following line (> 700 lines/second)
>> right after loading vfio:
>>
>> AMD-Vi: Event logged [IO_PAGE_FAULT device=00:14.0 domain=0x0000 address=0x000000fdf9103300 flags=0x0600]
> 
> That's interesting, I PXE boot my system from one NIC then use a
> different NIC for the iSCSI root.  The PXE boot NIC now screams like
> this, _until_ I attach it to vfio, then it quiets down.

Hmm, I just remembered an active workaround I implemented to "resolve"
an error like this when starting my VM to passthrough my intel pci
ethernet device since I applied a new kvm version:


qemu-kvm: -device vfio-pci,host=06:06.0: vfio: failed to set iommu for
container: Device or resource busy

qemu-kvm: -device vfio-pci,host=06:06.0: vfio: failed to setup container
for group 12

qemu-kvm: -device vfio-pci,host=06:06.0: vfio: failed to get group 12

qemu-kvm: -device vfio-pci,host=06:06.0: Device 'vfio-pci' could not be
initialized


The workaround was to bind the individual multifunction devices during
boot one time to vfio and release them after 2 seconds again and rebind
them to the original drivers as they where bound before (if it was bound
to any).

I did this with a script beginning like this:

#!/bin/sh
modprobe vfio-pci

echo "1002 4385" > /sys/bus/pci/drivers/vfio-pci/new_id
echo 0000:00:14.0 > /sys/bus/pci/devices/0000:00:14.0/driver/unbind
echo 0000:00:14.0 > /sys/bus/pci/drivers/vfio-pci/bind
...

sleep 2

echo 0000:00:14.0 > /sys/bus/pci/drivers/vfio-pci/unbind
echo "1002 4385" > /sys/bus/pci/drivers/vfio-pci/remove_id
...

The logs in messages:

Jun 28 15:54:12 . kernel: [   48.860147] VFIO - User Level meta-driver version: 0.3
Jun 28 15:54:12 . kernel: [   48.875243] AMD-Vi: Event logged [IO_PAGE_FAULT device=00:14.0 domain=0x0000 address=0x000000fdf9103300 flags=0x0600]
...

Therefore, the logoutput most probably started after device 14.0 was
bound to vfio. If it would have started after removing vfio, I would
have expected 2 seconds between the start messages of vfio and the first
occurrence of the IO_PAGE_FAULT.

Today, I'm using kvm 1.3.1 and it isn't necessary to use the complete
workaround anymore. It is enough to bind / unbind the pci bridge
as described above before starting the VM with the passed through pci
ethernet device.
Because I now don't touch the 14.0 device any more, the IO_PAGE_FAULT
messages disappeared completely.

@Joerg:
Anyway, I'm going to test your provided patch tomorrow!

BTW: what does it mean: IO_PAGE_FAULT - what do I have to expect if I
see this message?



Thanks,
regards,
Andreas

Re: [ 102/127] iommu/amd: Workaround for ERBT1312

[Joerg Roedel]

Hi Andreas,

On Fri, Jun 28, 2013 at 06:11:36PM +0200, Andreas Hartmann wrote:
> Hello Joerg, hello Alex,
> 
> the subsequent patch and the patch "iommu/amd: Re-enable IOMMU event log
> interrupt after handling." 925fe08bce38d1ff052fe2209b9e2b8d5fbb7f98
> spread /var/log/messages with the following line (> 700 lines/second)
> right after loading vfio:
> 
> AMD-Vi: Event logged [IO_PAGE_FAULT device=00:14.0 domain=0x0000 address=0x000000fdf9103300 flags=0x0600]
> 
> lspci -vvvs 0:14.0
> 00:14.0 SMBus: Advanced Micro Devices [AMD] nee ATI SBx00 SMBus Controller (rev 42)
>         Control: I/O+ Mem+ BusMaster- SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx+
>         Status: Cap- 66MHz+ UDF- FastB2B- ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-

Most likely a BIOS issue that is uncovered by re-enabling the event-log
interrupt patch. The device itself is only used by the BIOS and not by
the Linux kernel

> Besides the enormous pollution I couldn't see any malfunction at all.
> At first, I didn't realised it at all (-> the SSD was fast enough to
> cover it silently). I saw it the first time I rebooted because X didn't start any more because
> the /var partition was completely full. 
> 
> I removed the two mentioned patches and all is working
> fine again as before.

Without these two patches, can you check dmesg after boot if there are
other lines which report IO_PAGE_FAULTs?


	Joerg

Re: [ 102/127] iommu/amd: Workaround for ERBT1312

[Andreas Hartmann]

Hello Joerg,

Joerg Roedel wrote:
> Hi Andreas,
> 
> On Fri, Jun 28, 2013 at 06:11:36PM +0200, Andreas Hartmann wrote:
>> Hello Joerg, hello Alex,
>>
>> the subsequent patch and the patch "iommu/amd: Re-enable IOMMU event log
>> interrupt after handling." 925fe08bce38d1ff052fe2209b9e2b8d5fbb7f98
>> spread /var/log/messages with the following line (> 700 lines/second)
>> right after loading vfio:
>>
>> AMD-Vi: Event logged [IO_PAGE_FAULT device=00:14.0 domain=0x0000 address=0x000000fdf9103300 flags=0x0600]
>>
>> lspci -vvvs 0:14.0
>> 00:14.0 SMBus: Advanced Micro Devices [AMD] nee ATI SBx00 SMBus Controller (rev 42)
>>         Control: I/O+ Mem+ BusMaster- SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx+
>>         Status: Cap- 66MHz+ UDF- FastB2B- ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
> 
> Most likely a BIOS issue that is uncovered by re-enabling the event-log
> interrupt patch. The device itself is only used by the BIOS and not by
> the Linux kernel

Thanks for this info! Good to know.

[...]

>> I removed the two mentioned patches and all is working
>> fine again as before.
> 
> Without these two patches, can you check dmesg after boot if there are
> other lines which report IO_PAGE_FAULTs?

You're right, there is exactly one entry directly after loading of vfio.
I can see this message, too, with linux 3.4.43.


Regards,
Andreas

Re: [ 102/127] iommu/amd: Workaround for ERBT1312

[Joerg Roedel]

Alex, Andreas,

On Fri, Jun 28, 2013 at 08:42:05PM +0200, Andreas Hartmann wrote:
> You're right, there is exactly one entry directly after loading of vfio.
> I can see this message, too, with linux 3.4.43.

Can you please test this patch? It should reduce the noise
significantly, but a few of those error messages are still expected.

diff --git a/drivers/iommu/amd_iommu_init.c b/drivers/iommu/amd_iommu_init.c
index e3c2d74..74e1d1c 100644
--- a/drivers/iommu/amd_iommu_init.c
+++ b/drivers/iommu/amd_iommu_init.c
@@ -1431,6 +1431,7 @@ static void init_device_table_dma(void)
 	for (devid = 0; devid <= amd_iommu_last_bdf; ++devid) {
 		set_dev_entry_bit(devid, DEV_ENTRY_VALID);
 		set_dev_entry_bit(devid, DEV_ENTRY_TRANSLATION);
+		set_dev_entry_bit(devid, DEV_ENTRY_ONE_FAULT);
 	}
 }
 
diff --git a/drivers/iommu/amd_iommu_types.h b/drivers/iommu/amd_iommu_types.h
index 083f98c..4872b68 100644
--- a/drivers/iommu/amd_iommu_types.h
+++ b/drivers/iommu/amd_iommu_types.h
@@ -173,6 +173,7 @@
 #define DEV_ENTRY_TRANSLATION   0x01
 #define DEV_ENTRY_IR            0x3d
 #define DEV_ENTRY_IW            0x3e
+#define DEV_ENTRY_ONE_FAULT	0x61
 #define DEV_ENTRY_NO_PAGE_FAULT	0x62
 #define DEV_ENTRY_EX            0x67
 #define DEV_ENTRY_SYSMGT1       0x68

Re: [ 102/127] iommu/amd: Workaround for ERBT1312

[Alex Williamson]

On Fri, 2013-06-28 at 21:23 +0200, Joerg Roedel wrote:
> Alex, Andreas,
> 
> On Fri, Jun 28, 2013 at 08:42:05PM +0200, Andreas Hartmann wrote:
> > You're right, there is exactly one entry directly after loading of vfio.
> > I can see this message, too, with linux 3.4.43.
> 
> Can you please test this patch? It should reduce the noise
> significantly, but a few of those error messages are still expected.

Seems to work for me.  I generated a few errors, but each one seems
unique and is more like what I would expect if the PXE ROM left the NIC
running:

[    2.574155] AMD-Vi: Event logged [IO_PAGE_FAULT device=02:00.0 domain=0x001b address=0x00000000000946b0 flags=0x0020]                                        
[    2.574157] AMD-Vi: Event logged [IO_PAGE_FAULT device=02:00.0 domain=0x001b address=0x000000000008f8d0 flags=0x0020]                                        
[    4.992074] AMD-Vi: Event logged [IO_PAGE_FAULT device=02:00.0 domain=0x001b address=0x000000000008f880 flags=0x0000] 

I booted the system four times and twice got just three log messages,
like above, and twice got no faults.  Thanks,

Alex

> diff --git a/drivers/iommu/amd_iommu_init.c b/drivers/iommu/amd_iommu_init.c
> index e3c2d74..74e1d1c 100644
> --- a/drivers/iommu/amd_iommu_init.c
> +++ b/drivers/iommu/amd_iommu_init.c
> @@ -1431,6 +1431,7 @@ static void init_device_table_dma(void)
>  	for (devid = 0; devid <= amd_iommu_last_bdf; ++devid) {
>  		set_dev_entry_bit(devid, DEV_ENTRY_VALID);
>  		set_dev_entry_bit(devid, DEV_ENTRY_TRANSLATION);
> +		set_dev_entry_bit(devid, DEV_ENTRY_ONE_FAULT);
>  	}
>  }
>  
> diff --git a/drivers/iommu/amd_iommu_types.h b/drivers/iommu/amd_iommu_types.h
> index 083f98c..4872b68 100644
> --- a/drivers/iommu/amd_iommu_types.h
> +++ b/drivers/iommu/amd_iommu_types.h
> @@ -173,6 +173,7 @@
>  #define DEV_ENTRY_TRANSLATION   0x01
>  #define DEV_ENTRY_IR            0x3d
>  #define DEV_ENTRY_IW            0x3e
> +#define DEV_ENTRY_ONE_FAULT	0x61
>  #define DEV_ENTRY_NO_PAGE_FAULT	0x62
>  #define DEV_ENTRY_EX            0x67
>  #define DEV_ENTRY_SYSMGT1       0x68
> 

Re: [ 102/127] iommu/amd: Workaround for ERBT1312

[Andreas Hartmann]

Joerg Roedel wrote:
> Alex, Andreas,
> 
> On Fri, Jun 28, 2013 at 08:42:05PM +0200, Andreas Hartmann wrote:
>> You're right, there is exactly one entry directly after loading of vfio.
>> I can see this message, too, with linux 3.4.43.
> 
> Can you please test this patch? It should reduce the noise
> significantly, but a few of those error messages are still expected.

Sorry, but it doesn't work for me at all :-(. Behaviour is unchanged. It
is exactly as described in the other mail: at the moment of binding vfio
to 14.0, the fire begins.

echo "1002 4385" > /sys/bus/pci/drivers/vfio-pci/new_id
echo 0000:00:14.0 > /sys/bus/pci/devices/0000:00:14.0/driver/unbind
echo 0000:00:14.0 > /sys/bus/pci/drivers/vfio-pci/bind


Regards,
Andreas

Re: [ 102/127] iommu/amd: Workaround for ERBT1312

[Joerg Roedel]

On Sat, Jun 29, 2013 at 07:54:20AM +0200, Andreas Hartmann wrote:
> Sorry, but it doesn't work for me at all :-(. Behaviour is unchanged. It
> is exactly as described in the other mail: at the moment of binding vfio
> to 14.0, the fire begins.

Hmm, VFIO attaches the device to a new domain. That clears the bit, how
about this patch:

diff --git a/drivers/iommu/amd_iommu.c b/drivers/iommu/amd_iommu.c
index 1a5285b..cdf346f 100644
--- a/drivers/iommu/amd_iommu.c
+++ b/drivers/iommu/amd_iommu.c
@@ -2111,6 +2111,8 @@ static void set_dte_entry(u16 devid, struct protection_domain *domain, bool ats)
 
 		tmp = DTE_GCR3_VAL_C(gcr3) << DTE_GCR3_SHIFT_C;
 		flags    |= tmp;
+
+		flags	 |= DTE_FLAG_SE;
 	}
 
 	flags &= ~(0xffffUL);
diff --git a/drivers/iommu/amd_iommu_init.c b/drivers/iommu/amd_iommu_init.c
index e3c2d74..74e1d1c 100644
--- a/drivers/iommu/amd_iommu_init.c
+++ b/drivers/iommu/amd_iommu_init.c
@@ -1431,6 +1431,7 @@ static void init_device_table_dma(void)
 	for (devid = 0; devid <= amd_iommu_last_bdf; ++devid) {
 		set_dev_entry_bit(devid, DEV_ENTRY_VALID);
 		set_dev_entry_bit(devid, DEV_ENTRY_TRANSLATION);
+		set_dev_entry_bit(devid, DEV_ENTRY_ONE_FAULT);
 	}
 }
 
diff --git a/drivers/iommu/amd_iommu_types.h b/drivers/iommu/amd_iommu_types.h
index 083f98c..2104ca4 100644
--- a/drivers/iommu/amd_iommu_types.h
+++ b/drivers/iommu/amd_iommu_types.h
@@ -173,6 +173,7 @@
 #define DEV_ENTRY_TRANSLATION   0x01
 #define DEV_ENTRY_IR            0x3d
 #define DEV_ENTRY_IW            0x3e
+#define DEV_ENTRY_ONE_FAULT	0x61
 #define DEV_ENTRY_NO_PAGE_FAULT	0x62
 #define DEV_ENTRY_EX            0x67
 #define DEV_ENTRY_SYSMGT1       0x68
@@ -282,6 +283,7 @@
 #define IOMMU_PTE_IW (1ULL << 62)
 
 #define DTE_FLAG_IOTLB	(0x01UL << 32)
+#define DTE_FLAG_SE	(0x01UL << 33)
 #define DTE_FLAG_GV	(0x01ULL << 55)
 #define DTE_GLX_SHIFT	(56)
 #define DTE_GLX_MASK	(3)

Re: [ 102/127] iommu/amd: Workaround for ERBT1312

[Andreas Hartmann]

Joerg Roedel schrieb:
> On Sat, Jun 29, 2013 at 07:54:20AM +0200, Andreas Hartmann wrote:
>> Sorry, but it doesn't work for me at all :-(. Behaviour is unchanged. It
>> is exactly as described in the other mail: at the moment of binding vfio
>> to 14.0, the fire begins.
>
> Hmm, VFIO attaches the device to a new domain. That clears the bit, how
> about this patch:

Didn't help, too :-(


Regards,
Andreas

[ 000/127] 3.9.5-stable review

[Greg Kroah-Hartman]

This is the start of the stable review cycle for the 3.9.5 release.
There are 127 patches in this series, all will be posted as a response
to this one.  If anyone has any issues with these being applied, please
let me know.

Responses should be made by Fri Jun  7 21:30:02 UTC 2013.
Anything received after that time might be too late.

The whole patch series can be found in one patch at:
	kernel.org/pub/linux/kernel/v3.0/stable-review/patch-3.9.5-rc1.gz
and the diffstat can be found below.

thanks,

greg k-h

-------------
Pseudo-Shortlog of commits:

Greg Kroah-Hartman <gregkh@linuxfoundation.org>
    Linux 3.9.5-rc1

Wei Liu <wei.liu2@citrix.com>
    xen-netback: better names for thresholds

Johannes Berg <johannes.berg@intel.com>
    mac80211_hwsim: remove P2P_DEVICE support

Johannes Berg <johannes.berg@intel.com>
    iwlwifi: mvm: remove P2P_DEVICE support

Wei Liu <wei.liu2@citrix.com>
    xen-netfront: reduce gso_max_size to account for max TCP header

Wei Liu <wei.liu2@citrix.com>
    xen-netback: avoid allocating variable size array on stack

Wei Liu <wei.liu2@citrix.com>
    xen-netback: remove redundent parameter in netbk_count_requests

Wei Liu <wei.liu2@citrix.com>
    xen-netback: don't disconnect frontend when seeing oversize packet

Wei Liu <wei.liu2@citrix.com>
    xen-netback: coalesce slots in TX path and fix regressions

Timo Teräs <timo.teras@iki.f>
    leds: leds-gpio: reserve gpio before using it

Nicholas Bellinger <nab@linux-iscsi.org>
    target/file: Fix off-by-one READ_CAPACITY bug for !S_ISBLK export

Nicholas Bellinger <nab@linux-iscsi.org>
    target: Re-instate sess_wait_list for target_wait_for_sess_cmds

Miklos Szeredi <mszeredi@suse.cz>
    fuse: fix readdirplus Oops in fuse_dentry_revalidate

Daniel Mack <zonque@gmail.com>
    ASoC: davinci: fix sample rotation

Or Gerlitz <ogerlitz@mellanox.com>
    IB/iser: Return error to upper layers on EAGAIN registration failures

Jonas Peterson <jonas.peterson@gmail.com>
    net: can: kvaser_usb: fix reception on "USBcan Pro" and "USBcan R" type hardware.

Nithin Sujir <nsujir@broadcom.com>
    tg3: Add read dma workaround for 5720

Gabor Juhos <juhosg@openwrt.org>
    ath9k: use correct OTP register offsets for AR9550

Johannes Berg <johannes.berg@intel.com>
    cfg80211: check wdev->netdev in connection work

Emmanuel Grumbach <emmanuel.grumbach@intel.com>
    iwlwifi: mvm: tell firmware to let multicast frames in

Emmanuel Grumbach <emmanuel.grumbach@intel.com>
    iwlwifi: dvm: fix zero LQ CMD sending avoidance

Johannes Berg <johannes.berg@intel.com>
    mac80211: close AP_VLAN interfaces before unregistering all

Stanislaw Gruszka <sgruszka@redhat.com>
    mac80211: fix direct probe auth

Ben Skeggs <bskeggs@redhat.com>
    drm/nvc0/ce: disable ce1 on a number of chipsets

Bastian Triller <bastian.triller@gmail.com>
    ACPI / video: Add "Asus UL30A" to ACPI video detect blacklist

Rafael J. Wysocki <rafael.j.wysocki@intel.com>
    ACPI / PM: Allow device power states to be used for CONFIG_PM unset

Joerg Roedel <joro@8bytes.org>
    iommu/amd: Workaround for ERBT1312

Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>
    iommu/amd: Re-enable IOMMU event log interrupt after handling.

Emmanuel Grumbach <emmanuel.grumbach@intel.com>
    iwlwifi: mvm: fix aggregation drain flow

Wei Liu <wei.liu2@citrix.com>
    xen-netback: remove skb in xen_netbk_alloc_page

Dave Kleikamp <dave.kleikamp@oracle.com>
    jfs: fix a couple races

Jeff Mahoney <jeffm@jeffreymahoney.com>
    reiserfs: fix spurious multiple-fill in reiserfs_readdir_dentry

Jeff Mahoney <jeffm@suse.com>
    reiserfs: fix problems with chowning setuid file w/ xattrs

Jeff Mahoney <jeffm@suse.com>
    reiserfs: fix deadlock with nfs racing on create/lookup

Srivatsa S. Bhat <srivatsa.bhat@linux.vnet.ibm.com>
    powerpc/pseries: Always enable CONFIG_HOTPLUG_CPU on PSERIES SMP

Kishon Vijay Abraham I <kishon@ti.com>
    regulator: palmas: Fix "enable_reg" to point to the correct reg for SMPS10

Paolo Bonzini <pbonzini@redhat.com>
    KVM: fix sil/dil/bpl/spl in the mod/rm fields

Paolo Bonzini <pbonzini@redhat.com>
    KVM: Emulate multibyte NOP

Trond Myklebust <Trond.Myklebust@netapp.com>
    NFSv4: Fix a thinko in nfs4_try_open_cached

Nicholas Bellinger <nab@linux-iscsi.org>
    ib_srpt: Call target_sess_cmd_list_set_waiting during shutdown_session

Kees Cook <keescook@chromium.org>
    iscsi-target: fix heap buffer overflow on error

Michael Neuling <mikey@neuling.org>
    powerpc/tm: Move TM abort cause codes to uapi

Michael Neuling <mikey@neuling.org>
    powerpc/tm: Abort on emulation and alignment faults

Michael Neuling <mikey@neuling.org>
    powerpc/tm: Fix userspace stack corruption on signal delivery for active transactions

Michael Neuling <mikey@neuling.org>
    powerpc/tm: Update cause codes documentation

Michael Neuling <mikey@neuling.org>
    powerpc/tm: Make room for hypervisor in abort cause codes

Priyanka Jain <Priyanka.Jain@freescale.com>
    powerpc/32bit:Store temporary result in r0 instead of r8

Tejun Heo <tj@kernel.org>
    cgroup: fix a subtle bug in descendant pre-order walk

Li Zefan <lizefan@huawei.com>
    cgroup: initialize xattr before calling d_instantiate()

Christian Borntraeger <borntraeger@de.ibm.com>
    s390/pgtable: Fix check for pgste/storage key handling

Finn Thain <fthain@telegraphics.com.au>
    m68k/mac: Fix unexpected interrupt with CONFIG_EARLY_PRINTK

Vincent Pelletier <plr.vincent@gmail.com>
    libata: make ata_exec_internal_sg honor DMADIR

Chew, Chiau Ee <chiau.ee.chew@intel.com>
    ata_piix: add PCI IDs for Intel BayTail

Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>
    sata_rcar: fix interrupt handling

Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>
    sata_rcar: clear STOP bit in bmdma_start() method

Jeff Layton <jlayton@redhat.com>
    cifs: fix potential buffer overrun when composing a new options string

Horia Geanta <horia.geanta@freescale.com>
    crypto: caam - fix inconsistent assoc dma mapping direction

Alex Deucher <alexander.deucher@amd.com>
    drm/radeon: fix card_posted check for newer asics

Alex Deucher <alexander.deucher@amd.com>
    drm/radeon: fix typo in cu_per_sh on verde

Rodrigo Vivi <rodrigo.vivi@gmail.com>
    drm/i915: Adding more reserved PCI IDs for Haswell.

Charles Keepax <ckeepax@opensource.wolfsonmicro.com>
    ASoC: wm5110: Correct DSP4R Mixer control name

Nicolas Schichan <nschichan@freebox.fr>
    ASoC: cs42l52: fix default value for MASTERA_VOL.

J. Bruce Fields <bfields@redhat.com>
    svcrpc: fix failures to handle -1 uid's and gid's

Trond Myklebust <Trond.Myklebust@netapp.com>
    SUNRPC: Prevent an rpc_task wakeup race

Sylwester Nawrocki <s.nawrocki@samsung.com>
    ARM: SAMSUNG: Export MIPI CSIS/DSIM PHY control functions

Mark Rutland <mark.rutland@arm.com>
    arm64: don't kill the kernel on a bad esr from el0

Dave Chinner <dchinner@redhat.com>
    xfs: kill suid/sgid through the truncate path.

Rusty Russell <rusty@rustcorp.com.au>
    module: don't unlink the module until we've removed all exposure.

Jan Beulich <JBeulich@suse.com>
    x86, crc32-pclmul: Fix build with older binutils

Yinghai Lu <yinghai@kernel.org>
    x86, range: fix missing merge during add range

Zhang Yanfei <zhangyanfei@cn.fujitsu.com>
    x86-64, init: Fix a possible wraparound bug in switchover in head_64.S

Pekka Riikonen <priikone@iki.fi>
    x86: Allow FPU to be used at interrupt time even with eagerfpu

wenxiong@linux.vnet.ibm.com <wenxiong@linux.vnet.ibm.com>
    SCSI: ipr: Need to reset adapter after the 6th EEH error

Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
    mm/THP: use pmd_populate() to update the pmd with pgtable_t pointer

Cliff Wickman <cpw@sgi.com>
    mm/pagewalk.c: walk_page_range should avoid VM_PFNMAP areas

Brian Behlendorf <behlendorf1@llnl.gov>
    drivers/block/brd.c: fix brd_lookup_page() race

Johannes Weiner <hannes@cmpxchg.org>
    mm: memcg: remove incorrect VM_BUG_ON for swap cache pages in uncharge

Xiao Guangrong <xiaoguangrong@linux.vnet.ibm.com>
    mm: mmu_notifier: re-fix freed page still mapped in secondary MMU

Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>
    nilfs2: fix issue of nilfs_set_page_dirty() for page at EOF boundary

Imre Deak <imre.deak@intel.com>
    wait: fix false timeouts when using wait_event_timeout()

wang, biao <biao.wang@intel.com>
    klist: del waiter from klist_remove_waiters before wakeup waitting process

Leonid Yegoshin <Leonid.Yegoshin@imgtec.com>
    mm compaction: fix of improper cache flush in migration code

Alexandre Bounine <alexandre.bounine@idt.com>
    rapidio/tsi721: fix bug in MSI interrupt handling

Christian Gmeiner <christian.gmeiner@gmail.com>
    drivers/leds/leds-ot200.c: fix error caused by shifted mask

Martin Michlmayr <tbm@cyrius.com>
    Kirkwood: Enable PCIe port 1 on QNAP TS-11x/TS-21x

Joseph Qi <joseph.qi@huawei.com>
    ocfs2: goto out_unlock if ocfs2_get_clusters_nocache() failed in ocfs2_fiemap()

Sujith Manoharan <c_manoha@qca.qualcomm.com>
    ath9k_hw: Enable manual peak calibration for AR9485

Sujith Manoharan <c_manoha@qca.qualcomm.com>
    ath9k: Fix crash on module unload

Ben Hutchings <ben@decadent.org.uk>
    perf: net_dropmonitor: Fix symbol-relative addresses

Ben Hutchings <ben@decadent.org.uk>
    perf: net_dropmonitor: Fix trace parameter order

Michael Chan <mchan@broadcom.com>
    tg3: Fix data corruption on 5725 with TSO

Nithin Sujir <nsujir@broadcom.com>
    tg3: Skip powering down function 0 on certain serdes devices

OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
    fat: fix possible overflow for fat_clusters

Jiri Kosina <jkosina@suse.cz>
    random: fix accounting race condition with lockless irq entropy_count update

Jarod Wilson <jarod@redhat.com>
    drivers/char/random.c: fix priming of last_data

Vineet Gupta <vgupta@synopsys.com>
    ARC: copy_(to|from)_user() to honor usermode-access permissions

Jeff Layton <jlayton@redhat.com>
    cifs: only set ops for inodes in I_NEW state

K. Y. Srinivasan <kys@microsoft.com>
    Drivers: hv: Fix a bug in get_vp_index()

Heiko Abraham <abrahamh@web.de>
    Input: egalax_ts - ABS_MT_POSITION_Y not reported well

Alan Stern <stern@rowland.harvard.edu>
    USB: UHCI: fix for suspend of virtual HP controller

Alan Stern <stern@rowland.harvard.edu>
    USB: xHCI: override bogus bulk wMaxPacketSize values

Oliver Neukum <oliver@neukum.org>
    USB: reset resume quirk needed by a hub

Schemmel Hans-Christoph <Hans-Christoph.Schemmel@gemalto.com>
    USB: Blacklisted Cinterion's PLxx WWAN Interface

Wang YanQing <udknight@gmail.com>
    TTY: Fix tty miss restart after we turn off flow-control

Dan Carpenter <dan.carpenter@oracle.com>
    USB: cxacru: potential underflow in cxacru_cm_get_array()

Gomella, Andrew (NIH/NHLBI) [F] <andrew.gomella@nih.gov>
    USB: ftdi_sio: Add support for Newport CONEX motor drivers

Dan Williams <dcbw@redhat.com>
    USB: option: add device IDs for Dell 5804 (Novatel E371) WWAN card

Teppo Kotilainen <qubit303@gmail.com>
    usb: option: Add Telewell TW-LTE 4G

David Rientjes <rientjes@google.com>
    usb, chipidea: fix link error when USB_EHCI_HCD is a module

Hema Prathaban <hemaklnce@gmail.com>
    staging: vt6656: use free_netdev instead of kfree

Malcolm Priestley <tvboxspy@gmail.com>
    staging: vt6656: [bug] Fix missing spin lock in iwctl_siwpower.

Alexander Bondar <alexander.bondar@intel.com>
    iwlwifi: mvm: Prevent setting assoc flag in MAC_CONTEXT_CMD

Ilan Peer <ilan.peer@intel.com>
    iwlwifi: mvm: Always use SCAN_TYPE_FORCED

Johannes Berg <johannes.berg@intel.com>
    mac80211: fix AP-mode frame matching

Johannes Berg <johannes.berg@intel.com>
    mac80211: use just spin_lock() in ieee80211_get_tkip_p2k()

Johannes Berg <johannes.berg@intel.com>
    cfg80211: fix sending WoWLAN TCP wakeup settings

Johannes Berg <johannes.berg@intel.com>
    cfg80211: fix wiphy_register error path

Johannes Berg <johannes.berg@intel.com>
    cfg80211: fix WoWLAN wakeup tracing

Boris BREZILLON <b.brezillon@overkiz.com>
    ARM: at91/dt: fix macb pinctrl_macb_rmii_mii_alt definition

Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
    ARM: at91: rm9200 fix time support

Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
    ARM: at91/trivial: fix model name for SAM9X25-EK

Ard Biesheuvel <ard.biesheuvel@gmail.com>
    ARM: 7723/1: crypto: sha1-armv4-large.S: fix SP handling

Vaibhav Hiremath <hvaibhav@ti.com>
    ARM: AM33XX: Add missing .clkdm_name to clkdiv32k_ick clock

jean-philippe francois <jp.francois@cynove.com>
    ARM: OMAP: fix __init section mismatch for _enable_preprogram

Gregory CLEMENT <gregory.clement@free-electrons.com>
    ARM: plat-orion: Fix num_resources and id for ge10 and ge11

Alan Stern <stern@rowland.harvard.edu>
    USB: fix latency in uhci-hcd and ohci-hcd

Alan Stern <stern@rowland.harvard.edu>
    USB: OHCI: fix logic for scheduling isochronous URBs

Hans-Christian Egtvedt <egtvedt@samfundet.no>
    avr32: fix relocation check for signed 18-bit offset


-------------

Diffstat:

 Documentation/powerpc/transactional_memory.txt    |  27 +-
 Makefile                                          |   4 +-
 arch/arc/include/asm/pgtable.h                    |  26 +-
 arch/arc/include/asm/tlb.h                        |   2 +-
 arch/arc/mm/tlbex.S                               |   6 +-
 arch/arm/boot/dts/at91sam9260.dtsi                |   2 +-
 arch/arm/boot/dts/at91sam9x25ek.dts               |   2 +-
 arch/arm/crypto/sha1-armv4-large.S                |   2 +-
 arch/arm/mach-at91/at91rm9200_time.c              |   7 +-
 arch/arm/mach-kirkwood/ts219-setup.c              |   2 +-
 arch/arm/mach-omap2/cclock33xx_data.c             |  26 +-
 arch/arm/mach-omap2/omap_hwmod.c                  |   2 +-
 arch/arm/plat-orion/common.c                      |  12 +-
 arch/arm/plat-samsung/setup-mipiphy.c             |   3 +
 arch/arm64/kernel/traps.c                         |  12 +-
 arch/avr32/kernel/module.c                        |   2 +-
 arch/m68k/kernel/head.S                           |  29 +-
 arch/powerpc/include/asm/processor.h              |  13 +-
 arch/powerpc/include/asm/reg.h                    |  11 -
 arch/powerpc/include/asm/signal.h                 |   3 +
 arch/powerpc/include/asm/tm.h                     |   2 +
 arch/powerpc/include/uapi/asm/Kbuild              |   1 +
 arch/powerpc/include/uapi/asm/tm.h                |  18 ++
 arch/powerpc/kernel/entry_32.S                    |   2 +-
 arch/powerpc/kernel/signal.c                      |  40 ++-
 arch/powerpc/kernel/signal.h                      |   2 +-
 arch/powerpc/kernel/signal_32.c                   |  10 +-
 arch/powerpc/kernel/signal_64.c                   |  23 +-
 arch/powerpc/kernel/traps.c                       |  29 ++
 arch/powerpc/platforms/pseries/Kconfig            |   2 +
 arch/s390/include/asm/pgtable.h                   |  15 +-
 arch/x86/crypto/crc32-pclmul_asm.S                |   2 +-
 arch/x86/include/asm/inst.h                       |  74 ++++-
 arch/x86/kernel/head_64.S                         |   6 +-
 arch/x86/kernel/i387.c                            |  14 +-
 arch/x86/kvm/emulate.c                            |   9 +-
 drivers/acpi/Makefile                             |   2 +-
 drivers/acpi/device_pm.c                          | 126 ++++-----
 drivers/acpi/video_detect.c                       |   8 +
 drivers/ata/ata_piix.c                            |  15 ++
 drivers/ata/libata-core.c                         |   6 +
 drivers/ata/sata_rcar.c                           |  24 +-
 drivers/block/brd.c                               |   4 +-
 drivers/char/random.c                             |  54 ++--
 drivers/crypto/caam/caamalg.c                     |   4 +-
 drivers/gpu/drm/i915/i915_drv.c                   |  46 +++-
 drivers/gpu/drm/nouveau/core/subdev/device/nvc0.c |   2 -
 drivers/gpu/drm/radeon/radeon_device.c            |  19 +-
 drivers/gpu/drm/radeon/si.c                       |   2 +-
 drivers/hv/channel_mgmt.c                         |   2 +-
 drivers/infiniband/ulp/iser/iser_memory.c         |   3 +-
 drivers/infiniband/ulp/srpt/ib_srpt.c             |  32 ++-
 drivers/infiniband/ulp/srpt/ib_srpt.h             |   1 +
 drivers/input/touchscreen/egalax_ts.c             |   2 +-
 drivers/iommu/amd_iommu.c                         |  31 ++-
 drivers/iommu/amd_iommu_types.h                   |   1 +
 drivers/leds/leds-gpio.c                          |   9 +-
 drivers/leds/leds-ot200.c                         |  14 +-
 drivers/net/can/usb/kvaser_usb.c                  |  64 +++--
 drivers/net/ethernet/broadcom/tg3.c               |  70 ++++-
 drivers/net/ethernet/broadcom/tg3.h               |   5 +-
 drivers/net/wireless/ath/ath9k/ar9003_calib.c     |   2 +-
 drivers/net/wireless/ath/ath9k/ar9003_eeprom.h    |   9 +-
 drivers/net/wireless/ath/ath9k/ar9485_initvals.h  |   2 +-
 drivers/net/wireless/ath/ath9k/debug.c            |   8 +
 drivers/net/wireless/ath/ath9k/debug.h            |   5 +
 drivers/net/wireless/ath/ath9k/init.c             |  10 +-
 drivers/net/wireless/iwlwifi/dvm/sta.c            |   2 +-
 drivers/net/wireless/iwlwifi/mvm/fw-api.h         |  27 ++
 drivers/net/wireless/iwlwifi/mvm/mac-ctxt.c       |  12 +-
 drivers/net/wireless/iwlwifi/mvm/mac80211.c       |  31 ++-
 drivers/net/wireless/iwlwifi/mvm/mvm.h            |   1 +
 drivers/net/wireless/iwlwifi/mvm/ops.c            |   1 +
 drivers/net/wireless/iwlwifi/mvm/scan.c           |   6 +
 drivers/net/wireless/iwlwifi/mvm/sta.c            |  13 +-
 drivers/net/wireless/iwlwifi/mvm/sta.h            |   2 -
 drivers/net/wireless/iwlwifi/mvm/tx.c             |  48 +++-
 drivers/net/wireless/mac80211_hwsim.c             |   4 +-
 drivers/net/xen-netback/netback.c                 | 311 ++++++++++++++++++----
 drivers/net/xen-netfront.c                        |  17 +-
 drivers/rapidio/devices/tsi721.c                  |  12 +
 drivers/regulator/palmas-regulator.c              |   2 +-
 drivers/scsi/ipr.c                                |  12 +-
 drivers/staging/vt6656/hostap.c                   |   2 +-
 drivers/staging/vt6656/iwctl.c                    |   6 +
 drivers/target/iscsi/iscsi_target_parameters.c    |   8 +-
 drivers/target/iscsi/iscsi_target_parameters.h    |   4 +-
 drivers/target/target_core_file.c                 |  12 +-
 drivers/target/target_core_transport.c            |  18 +-
 drivers/tty/n_tty.c                               |   8 +
 drivers/usb/atm/cxacru.c                          |   3 +-
 drivers/usb/chipidea/Kconfig                      |   2 +-
 drivers/usb/core/quirks.c                         |   3 +
 drivers/usb/host/ohci-hcd.c                       |  34 +--
 drivers/usb/host/uhci-hub.c                       |   3 +-
 drivers/usb/host/uhci-q.c                         |   2 +-
 drivers/usb/host/xhci-mem.c                       |  17 +-
 drivers/usb/serial/ftdi_sio.c                     |   2 +
 drivers/usb/serial/ftdi_sio_ids.h                 |   2 +
 drivers/usb/serial/option.c                       |  13 +-
 fs/cifs/cifs_dfs_ref.c                            |   4 +-
 fs/cifs/inode.c                                   |   3 +-
 fs/fat/inode.c                                    |  15 +-
 fs/fuse/dir.c                                     |  12 +-
 fs/fuse/inode.c                                   |   7 +-
 fs/jfs/inode.c                                    |   2 +-
 fs/jfs/jfs_logmgr.c                               |   3 +-
 fs/nfs/nfs4proc.c                                 |   2 +-
 fs/nilfs2/inode.c                                 |  27 +-
 fs/ocfs2/extent_map.c                             |   2 +-
 fs/reiserfs/dir.c                                 |   2 +
 fs/reiserfs/inode.c                               |   9 +-
 fs/reiserfs/xattr.c                               |  14 +-
 fs/reiserfs/xattr_acl.c                           |   3 +
 fs/xfs/xfs_iops.c                                 |  47 ++--
 include/acpi/acpi_bus.h                           |  40 +--
 include/linux/cgroup.h                            |   2 +-
 include/linux/wait.h                              |  16 +-
 include/target/target_core_base.h                 |   1 +
 include/xen/interface/io/netif.h                  |  19 ++
 kernel/cgroup.c                                   |  18 +-
 kernel/module.c                                   |  13 +-
 kernel/range.c                                    |   8 +-
 lib/klist.c                                       |   2 +-
 mm/huge_memory.c                                  |   7 +-
 mm/memcontrol.c                                   |  14 +-
 mm/migrate.c                                      |   2 +-
 mm/mmu_notifier.c                                 |  79 +++---
 mm/pagewalk.c                                     |  70 ++---
 net/mac80211/iface.c                              |   9 +
 net/mac80211/mlme.c                               |  12 +-
 net/mac80211/rx.c                                 |   3 +
 net/mac80211/tkip.c                               |   4 +-
 net/sunrpc/sched.c                                |   8 +-
 net/sunrpc/svcauth_unix.c                         |  12 +-
 net/wireless/core.c                               |  16 +-
 net/wireless/nl80211.c                            |   2 +
 net/wireless/sme.c                                |   3 +
 net/wireless/trace.h                              |  23 +-
 sound/soc/codecs/cs42l52.c                        |   2 +-
 sound/soc/codecs/wm5110.c                         |   2 +-
 sound/soc/davinci/davinci-mcasp.c                 |   7 +-
 tools/perf/scripts/python/net_dropmonitor.py      |   8 +-
 143 files changed, 1545 insertions(+), 673 deletions(-)

[ 102/127] iommu/amd: Workaround for ERBT1312

[Greg Kroah-Hartman]

3.9-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Joerg Roedel <joro@8bytes.org>

commit d3263bc29706e42f74d8800807c2dedf320d77f1 upstream.

Work around an IOMMU  hardware bug where clearing the
EVT_INT or PPR_INT bit in the status register may race with
the hardware trying to set it again. When not handled the
bit might not be cleared and we lose all future event or ppr
interrupts.

Reported-by: Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>
Signed-off-by: Joerg Roedel <joro@8bytes.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

---
 drivers/iommu/amd_iommu.c |   34 ++++++++++++++++++++++++++--------
 1 file changed, 26 insertions(+), 8 deletions(-)

--- a/drivers/iommu/amd_iommu.c
+++ b/drivers/iommu/amd_iommu.c
@@ -700,14 +700,23 @@ retry:
 
 static void iommu_poll_events(struct amd_iommu *iommu)
 {
-	u32 head, tail;
+	u32 head, tail, status;
 	unsigned long flags;
 
-	/* enable event interrupts again */
-	writel(MMIO_STATUS_EVT_INT_MASK, iommu->mmio_base + MMIO_STATUS_OFFSET);
-
 	spin_lock_irqsave(&iommu->lock, flags);
 
+	/* enable event interrupts again */
+	do {
+		/*
+		 * Workaround for Erratum ERBT1312
+		 * Clearing the EVT_INT bit may race in the hardware, so read
+		 * it again and make sure it was really cleared
+		 */
+		status = readl(iommu->mmio_base + MMIO_STATUS_OFFSET);
+		writel(MMIO_STATUS_EVT_INT_MASK,
+		       iommu->mmio_base + MMIO_STATUS_OFFSET);
+	} while (status & MMIO_STATUS_EVT_INT_MASK);
+
 	head = readl(iommu->mmio_base + MMIO_EVT_HEAD_OFFSET);
 	tail = readl(iommu->mmio_base + MMIO_EVT_TAIL_OFFSET);
 
@@ -744,16 +753,25 @@ static void iommu_handle_ppr_entry(struc
 static void iommu_poll_ppr_log(struct amd_iommu *iommu)
 {
 	unsigned long flags;
-	u32 head, tail;
+	u32 head, tail, status;
 
 	if (iommu->ppr_log == NULL)
 		return;
 
-	/* enable ppr interrupts again */
-	writel(MMIO_STATUS_PPR_INT_MASK, iommu->mmio_base + MMIO_STATUS_OFFSET);
-
 	spin_lock_irqsave(&iommu->lock, flags);
 
+	/* enable ppr interrupts again */
+	do {
+		/*
+		 * Workaround for Erratum ERBT1312
+		 * Clearing the PPR_INT bit may race in the hardware, so read
+		 * it again and make sure it was really cleared
+		 */
+		status = readl(iommu->mmio_base + MMIO_STATUS_OFFSET);
+		writel(MMIO_STATUS_PPR_INT_MASK,
+		       iommu->mmio_base + MMIO_STATUS_OFFSET);
+	} while (status & MMIO_STATUS_PPR_INT_MASK);
+
 	head = readl(iommu->mmio_base + MMIO_PPR_HEAD_OFFSET);
 	tail = readl(iommu->mmio_base + MMIO_PPR_TAIL_OFFSET);