[PATCH] iommu: amd: Fix IOMMU perf counter clobbering during init

[PATCH] iommu: amd: Fix IOMMU perf counter clobbering during init

[Shuah Khan]

init_iommu_perf_ctr() clobbers the register when it checks write access
to IOMMU perf counters and fails to restore when they are writable.

Add save and restore to fix it.

Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
---
 drivers/iommu/amd_iommu_init.c | 22 ++++++++++++++++------
 1 file changed, 16 insertions(+), 6 deletions(-)

diff --git a/drivers/iommu/amd_iommu_init.c b/drivers/iommu/amd_iommu_init.c
index 568c52317757..c0ad4f293522 100644
--- a/drivers/iommu/amd_iommu_init.c
+++ b/drivers/iommu/amd_iommu_init.c
@@ -1655,27 +1655,37 @@ static int iommu_pc_get_set_reg(struct amd_iommu *iommu, u8 bank, u8 cntr,
 static void init_iommu_perf_ctr(struct amd_iommu *iommu)
 {
 	struct pci_dev *pdev = iommu->dev;
-	u64 val = 0xabcd, val2 = 0;
+	u64 val = 0xabcd, val2 = 0, save_reg = 0;
 
 	if (!iommu_feature(iommu, FEATURE_PC))
 		return;
 
 	amd_iommu_pc_present = true;
 
+	/* save the value to restore, if writable */
+	if (iommu_pc_get_set_reg(iommu, 0, 0, 0, &save_reg, false))
+		goto pc_false;
+
 	/* Check if the performance counters can be written to */
 	if ((iommu_pc_get_set_reg(iommu, 0, 0, 0, &val, true)) ||
 	    (iommu_pc_get_set_reg(iommu, 0, 0, 0, &val2, false)) ||
-	    (val != val2)) {
-		pci_err(pdev, "Unable to write to IOMMU perf counter.\n");
-		amd_iommu_pc_present = false;
-		return;
-	}
+	    (val != val2))
+		goto pc_false;
+
+	/* restore */
+	if (iommu_pc_get_set_reg(iommu, 0, 0, 0, &save_reg, true))
+		goto pc_false;
 
 	pci_info(pdev, "IOMMU performance counters supported\n");
 
 	val = readl(iommu->mmio_base + MMIO_CNTR_CONF_OFFSET);
 	iommu->max_banks = (u8) ((val >> 12) & 0x3f);
 	iommu->max_counters = (u8) ((val >> 7) & 0xf);
+
+pc_false:
+	pci_err(pdev, "Unable to read/write to IOMMU perf counter.\n");
+	amd_iommu_pc_present = false;
+	return;
 }
 
 static ssize_t amd_iommu_show_cap(struct device *dev,
-- 
2.20.1

Re: [PATCH] iommu: amd: Fix IOMMU perf counter clobbering during init

[Joerg Roedel]

Adding Suravee, who wrote the IOMMU Perf Counter code.

On Tue, Jan 14, 2020 at 08:12:20AM -0700, Shuah Khan wrote:
> init_iommu_perf_ctr() clobbers the register when it checks write access
> to IOMMU perf counters and fails to restore when they are writable.
> 
> Add save and restore to fix it.
> 
> Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
> ---
>  drivers/iommu/amd_iommu_init.c | 22 ++++++++++++++++------
>  1 file changed, 16 insertions(+), 6 deletions(-)

Suravee, can you please review this patch?

> 
> diff --git a/drivers/iommu/amd_iommu_init.c b/drivers/iommu/amd_iommu_init.c
> index 568c52317757..c0ad4f293522 100644
> --- a/drivers/iommu/amd_iommu_init.c
> +++ b/drivers/iommu/amd_iommu_init.c
> @@ -1655,27 +1655,37 @@ static int iommu_pc_get_set_reg(struct amd_iommu *iommu, u8 bank, u8 cntr,
>  static void init_iommu_perf_ctr(struct amd_iommu *iommu)
>  {
>  	struct pci_dev *pdev = iommu->dev;
> -	u64 val = 0xabcd, val2 = 0;
> +	u64 val = 0xabcd, val2 = 0, save_reg = 0;
>  
>  	if (!iommu_feature(iommu, FEATURE_PC))
>  		return;
>  
>  	amd_iommu_pc_present = true;
>  
> +	/* save the value to restore, if writable */
> +	if (iommu_pc_get_set_reg(iommu, 0, 0, 0, &save_reg, false))
> +		goto pc_false;
> +
>  	/* Check if the performance counters can be written to */
>  	if ((iommu_pc_get_set_reg(iommu, 0, 0, 0, &val, true)) ||
>  	    (iommu_pc_get_set_reg(iommu, 0, 0, 0, &val2, false)) ||
> -	    (val != val2)) {
> -		pci_err(pdev, "Unable to write to IOMMU perf counter.\n");
> -		amd_iommu_pc_present = false;
> -		return;
> -	}
> +	    (val != val2))
> +		goto pc_false;
> +
> +	/* restore */
> +	if (iommu_pc_get_set_reg(iommu, 0, 0, 0, &save_reg, true))
> +		goto pc_false;
>  
>  	pci_info(pdev, "IOMMU performance counters supported\n");
>  
>  	val = readl(iommu->mmio_base + MMIO_CNTR_CONF_OFFSET);
>  	iommu->max_banks = (u8) ((val >> 12) & 0x3f);
>  	iommu->max_counters = (u8) ((val >> 7) & 0xf);
> +
> +pc_false:
> +	pci_err(pdev, "Unable to read/write to IOMMU perf counter.\n");
> +	amd_iommu_pc_present = false;
> +	return;
>  }
>  
>  static ssize_t amd_iommu_show_cap(struct device *dev,
> -- 
> 2.20.1

Re: [PATCH] iommu: amd: Fix IOMMU perf counter clobbering during init

[Suravee Suthikulpanit]

On 1/17/2020 5:08 PM, Joerg Roedel wrote:
> Adding Suravee, who wrote the IOMMU Perf Counter code.
> 
> On Tue, Jan 14, 2020 at 08:12:20AM -0700, Shuah Khan wrote:
>> init_iommu_perf_ctr() clobbers the register when it checks write access
>> to IOMMU perf counters and fails to restore when they are writable.
>>
>> Add save and restore to fix it.
>>
>> Signed-off-by: Shuah Khan<skhan@linuxfoundation.org>
>> ---
>>   drivers/iommu/amd_iommu_init.c | 22 ++++++++++++++++------
>>   1 file changed, 16 insertions(+), 6 deletions(-)
> Suravee, can you please review this patch?
> 

This looks ok. Does this fix certain issues? Or is this just for sanity.

Reviewed-by: Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>

Thanks,
Suravee

Re: [PATCH] iommu: amd: Fix IOMMU perf counter clobbering during init

[Shuah Khan]

On 1/20/20 7:10 PM, Suravee Suthikulpanit wrote:
> On 1/17/2020 5:08 PM, Joerg Roedel wrote:
>> Adding Suravee, who wrote the IOMMU Perf Counter code.
>>
>> On Tue, Jan 14, 2020 at 08:12:20AM -0700, Shuah Khan wrote:
>>> init_iommu_perf_ctr() clobbers the register when it checks write access
>>> to IOMMU perf counters and fails to restore when they are writable.
>>>
>>> Add save and restore to fix it.
>>>
>>> Signed-off-by: Shuah Khan<skhan@linuxfoundation.org>
>>> ---
>>>   drivers/iommu/amd_iommu_init.c | 22 ++++++++++++++++------
>>>   1 file changed, 16 insertions(+), 6 deletions(-)
>> Suravee, can you please review this patch?
>>
> 
> This looks ok. Does this fix certain issues? Or is this just for sanity.

I didn't notice any problems. Counters aren't writable on my system.
However, it certainly looks like a bog since registers aren't restored
like in other places in this file where such checks are done on other
registers.

I see 2 banks and 4 counters on my system. Is it sufficient to check
the first bank and first counter? In other words, if the first one
isn't writable, are all counters non-writable?

Should we read the config first and then, try to see if any of the
counters are writable? I have a patch that does that, I can send it
out for review.

> 
> Reviewed-by: Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>
Thanks for the review.

thanks,
-- Shuah

Re: [PATCH] iommu: amd: Fix IOMMU perf counter clobbering during init

[Shuah Khan]

On 1/21/20 8:32 AM, Shuah Khan wrote:
> On 1/20/20 7:10 PM, Suravee Suthikulpanit wrote:
>> On 1/17/2020 5:08 PM, Joerg Roedel wrote:
>>> Adding Suravee, who wrote the IOMMU Perf Counter code.
>>>
>>> On Tue, Jan 14, 2020 at 08:12:20AM -0700, Shuah Khan wrote:
>>>> init_iommu_perf_ctr() clobbers the register when it checks write access
>>>> to IOMMU perf counters and fails to restore when they are writable.
>>>>
>>>> Add save and restore to fix it.
>>>>
>>>> Signed-off-by: Shuah Khan<skhan@linuxfoundation.org>
>>>> ---
>>>>   drivers/iommu/amd_iommu_init.c | 22 ++++++++++++++++------
>>>>   1 file changed, 16 insertions(+), 6 deletions(-)
>>> Suravee, can you please review this patch?
>>>
>>
>> This looks ok. Does this fix certain issues? Or is this just for sanity.
> 
> I didn't notice any problems. Counters aren't writable on my system.
> However, it certainly looks like a bog since registers aren't restored
> like in other places in this file where such checks are done on other
> registers.
> 
> I see 2 banks and 4 counters on my system. Is it sufficient to check
> the first bank and first counter? In other words, if the first one
> isn't writable, are all counters non-writable?
> 
> Should we read the config first and then, try to see if any of the
> counters are writable? I have a patch that does that, I can send it
> out for review.
> 
>>
>> Reviewed-by: Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>

Joerg,

Please don't pull this in. I introduced a bug in this patch. It always
returns amd_iommu_pc_present false even when it can write to the
counters. My bad.

I will send v2.

thanks,
-- Shuah