* Re: [PATCH] iommu/vt-d: improve ITE fault handling if device was released
2024-02-29 4:07 [PATCH] iommu/vt-d: improve ITE fault handling if device was released Ethan Zhao
@ 2024-02-29 4:08 ` Ethan Zhao
2024-02-29 21:07 ` Bjorn Helgaas
2024-03-01 7:05 ` Ethan Zhao
2 siblings, 0 replies; 5+ messages in thread
From: Ethan Zhao @ 2024-02-29 4:08 UTC (permalink / raw)
To: baolu.lu, bhelgaas, robin.murphy, jgg
Cc: kevin.tian, dwmw2, will, lukas, yi.l.liu, dan.carpenter, iommu,
linux-kernel
On 2/29/2024 12:07 PM, Ethan Zhao wrote:
> Break the loop to blindly retry the timeout ATS invalidation request
> after ITE fault hit if device was released or isn't present anymore.
>
> This is part of the followup of prior proposed patchset
>
> https://do-db2.lkml.org/lkml/2024/2/22/350
>
> Fixes: 6ba6c3a4cacf ("VT-d: add device IOTLB invalidation support")
> Signed-off-by: Ethan Zhao <haifeng.zhao@linux.intel.com>
> ---
> drivers/iommu/intel/dmar.c | 25 +++++++++++++++++++++++++
> 1 file changed, 25 insertions(+)
>
> diff --git a/drivers/iommu/intel/dmar.c b/drivers/iommu/intel/dmar.c
> index d14797aabb7a..d01d68205557 100644
> --- a/drivers/iommu/intel/dmar.c
> +++ b/drivers/iommu/intel/dmar.c
> @@ -1273,6 +1273,9 @@ static int qi_check_fault(struct intel_iommu *iommu, int index, int wait_index)
> {
> u32 fault;
> int head, tail;
> + u64 iqe_err, ite_sid;
> + struct device *dev = NULL;
> + struct pci_dev *pdev = NULL;
> struct q_inval *qi = iommu->qi;
> int shift = qi_shift(iommu);
>
> @@ -1317,6 +1320,13 @@ static int qi_check_fault(struct intel_iommu *iommu, int index, int wait_index)
> tail = readl(iommu->reg + DMAR_IQT_REG);
> tail = ((tail >> shift) - 1 + QI_LENGTH) % QI_LENGTH;
>
> + /*
> + * SID field is valid only when the ITE field is Set in FSTS_REG
> + * see Intel VT-d spec r4.1, section 11.4.9.9
> + */
> + iqe_err = dmar_readq(iommu->reg + DMAR_IQER_REG);
> + ite_sid = DMAR_IQER_REG_ITESID(iqe_err);
> +
> writel(DMA_FSTS_ITE, iommu->reg + DMAR_FSTS_REG);
> pr_info("Invalidation Time-out Error (ITE) cleared\n");
>
> @@ -1326,6 +1336,21 @@ static int qi_check_fault(struct intel_iommu *iommu, int index, int wait_index)
> head = (head - 2 + QI_LENGTH) % QI_LENGTH;
> } while (head != tail);
>
> + /*
> + * If got ITE, we need to check if the sid of ITE is one of the
> + * current valid ATS invalidation target devices, if no, or the
> + * target device isn't presnet, don't try this request anymore.
> + * 0 value of ite_sid means old VT-d device, no ite_sid value.
> + */
> + if (ite_sid) {
> + dev = device_rbtree_find(iommu, ite_sid);
> + if (!dev || !dev_is_pci(dev))
> + return -ETIMEDOUT;
> + pdev = to_pci_dev(dev);
> + if (!pci_device_is_present(pdev) &&
> + ite_sid == pci_dev_id(pci_physfn(pdev)))
> + return -ETIMEDOUT;
> + }
> if (qi->desc_status[wait_index] == QI_ABORT)
> return -EAGAIN;
> }
This patch based on Baolu's per iommu device rbtree patchset
https://github.com/LuBaolu/intel-iommu/commits/rbtree-for-device-info-v2
Thanks,
Ethan
^ permalink raw reply [flat|nested] 5+ messages in thread* Re: [PATCH] iommu/vt-d: improve ITE fault handling if device was released
2024-02-29 4:07 [PATCH] iommu/vt-d: improve ITE fault handling if device was released Ethan Zhao
2024-02-29 4:08 ` Ethan Zhao
@ 2024-02-29 21:07 ` Bjorn Helgaas
2024-03-01 1:51 ` Ethan Zhao
2024-03-01 7:05 ` Ethan Zhao
2 siblings, 1 reply; 5+ messages in thread
From: Bjorn Helgaas @ 2024-02-29 21:07 UTC (permalink / raw)
To: Ethan Zhao
Cc: baolu.lu, bhelgaas, robin.murphy, jgg, kevin.tian, dwmw2, will,
lukas, yi.l.liu, dan.carpenter, iommu, linux-kernel
On Wed, Feb 28, 2024 at 11:07:24PM -0500, Ethan Zhao wrote:
> Break the loop to blindly retry the timeout ATS invalidation request
> after ITE fault hit if device was released or isn't present anymore.
>
> This is part of the followup of prior proposed patchset
>
> https://do-db2.lkml.org/lkml/2024/2/22/350
Use lore URL, please.
> Fixes: 6ba6c3a4cacf ("VT-d: add device IOTLB invalidation support")
> Signed-off-by: Ethan Zhao <haifeng.zhao@linux.intel.com>
> ---
> drivers/iommu/intel/dmar.c | 25 +++++++++++++++++++++++++
> 1 file changed, 25 insertions(+)
>
> diff --git a/drivers/iommu/intel/dmar.c b/drivers/iommu/intel/dmar.c
> index d14797aabb7a..d01d68205557 100644
> --- a/drivers/iommu/intel/dmar.c
> +++ b/drivers/iommu/intel/dmar.c
> @@ -1273,6 +1273,9 @@ static int qi_check_fault(struct intel_iommu *iommu, int index, int wait_index)
> {
> u32 fault;
> int head, tail;
> + u64 iqe_err, ite_sid;
> + struct device *dev = NULL;
> + struct pci_dev *pdev = NULL;
> struct q_inval *qi = iommu->qi;
> int shift = qi_shift(iommu);
>
> @@ -1317,6 +1320,13 @@ static int qi_check_fault(struct intel_iommu *iommu, int index, int wait_index)
> tail = readl(iommu->reg + DMAR_IQT_REG);
> tail = ((tail >> shift) - 1 + QI_LENGTH) % QI_LENGTH;
>
> + /*
> + * SID field is valid only when the ITE field is Set in FSTS_REG
> + * see Intel VT-d spec r4.1, section 11.4.9.9
> + */
> + iqe_err = dmar_readq(iommu->reg + DMAR_IQER_REG);
> + ite_sid = DMAR_IQER_REG_ITESID(iqe_err);
> +
> writel(DMA_FSTS_ITE, iommu->reg + DMAR_FSTS_REG);
> pr_info("Invalidation Time-out Error (ITE) cleared\n");
>
> @@ -1326,6 +1336,21 @@ static int qi_check_fault(struct intel_iommu *iommu, int index, int wait_index)
> head = (head - 2 + QI_LENGTH) % QI_LENGTH;
> } while (head != tail);
>
> + /*
> + * If got ITE, we need to check if the sid of ITE is one of the
> + * current valid ATS invalidation target devices, if no, or the
> + * target device isn't presnet, don't try this request anymore.
> + * 0 value of ite_sid means old VT-d device, no ite_sid value.
> + */
> + if (ite_sid) {
> + dev = device_rbtree_find(iommu, ite_sid);
> + if (!dev || !dev_is_pci(dev))
> + return -ETIMEDOUT;
> + pdev = to_pci_dev(dev);
> + if (!pci_device_is_present(pdev) &&
> + ite_sid == pci_dev_id(pci_physfn(pdev)))
> + return -ETIMEDOUT;
> + }
> if (qi->desc_status[wait_index] == QI_ABORT)
> return -EAGAIN;
> }
> --
> 2.31.1
>
^ permalink raw reply [flat|nested] 5+ messages in thread* Re: [PATCH] iommu/vt-d: improve ITE fault handling if device was released
2024-02-29 21:07 ` Bjorn Helgaas
@ 2024-03-01 1:51 ` Ethan Zhao
0 siblings, 0 replies; 5+ messages in thread
From: Ethan Zhao @ 2024-03-01 1:51 UTC (permalink / raw)
To: Bjorn Helgaas
Cc: baolu.lu, bhelgaas, robin.murphy, jgg, kevin.tian, dwmw2, will,
lukas, yi.l.liu, dan.carpenter, iommu, linux-kernel
On 3/1/2024 5:07 AM, Bjorn Helgaas wrote:
> On Wed, Feb 28, 2024 at 11:07:24PM -0500, Ethan Zhao wrote:
>> Break the loop to blindly retry the timeout ATS invalidation request
>> after ITE fault hit if device was released or isn't present anymore.
>>
>> This is part of the followup of prior proposed patchset
>>
>> https://do-db2.lkml.org/lkml/2024/2/22/350
> Use lore URL, please.
Sure !
>
>> Fixes: 6ba6c3a4cacf ("VT-d: add device IOTLB invalidation support")
>> Signed-off-by: Ethan Zhao <haifeng.zhao@linux.intel.com>
>> ---
>> drivers/iommu/intel/dmar.c | 25 +++++++++++++++++++++++++
>> 1 file changed, 25 insertions(+)
>>
>> diff --git a/drivers/iommu/intel/dmar.c b/drivers/iommu/intel/dmar.c
>> index d14797aabb7a..d01d68205557 100644
>> --- a/drivers/iommu/intel/dmar.c
>> +++ b/drivers/iommu/intel/dmar.c
>> @@ -1273,6 +1273,9 @@ static int qi_check_fault(struct intel_iommu *iommu, int index, int wait_index)
>> {
>> u32 fault;
>> int head, tail;
>> + u64 iqe_err, ite_sid;
>> + struct device *dev = NULL;
>> + struct pci_dev *pdev = NULL;
>> struct q_inval *qi = iommu->qi;
>> int shift = qi_shift(iommu);
>>
>> @@ -1317,6 +1320,13 @@ static int qi_check_fault(struct intel_iommu *iommu, int index, int wait_index)
>> tail = readl(iommu->reg + DMAR_IQT_REG);
>> tail = ((tail >> shift) - 1 + QI_LENGTH) % QI_LENGTH;
>>
>> + /*
>> + * SID field is valid only when the ITE field is Set in FSTS_REG
>> + * see Intel VT-d spec r4.1, section 11.4.9.9
>> + */
>> + iqe_err = dmar_readq(iommu->reg + DMAR_IQER_REG);
>> + ite_sid = DMAR_IQER_REG_ITESID(iqe_err);
>> +
>> writel(DMA_FSTS_ITE, iommu->reg + DMAR_FSTS_REG);
>> pr_info("Invalidation Time-out Error (ITE) cleared\n");
>>
>> @@ -1326,6 +1336,21 @@ static int qi_check_fault(struct intel_iommu *iommu, int index, int wait_index)
>> head = (head - 2 + QI_LENGTH) % QI_LENGTH;
>> } while (head != tail);
>>
>> + /*
>> + * If got ITE, we need to check if the sid of ITE is one of the
>> + * current valid ATS invalidation target devices, if no, or the
>> + * target device isn't presnet, don't try this request anymore.
>> + * 0 value of ite_sid means old VT-d device, no ite_sid value.
>> + */
>> + if (ite_sid) {
>> + dev = device_rbtree_find(iommu, ite_sid);
>> + if (!dev || !dev_is_pci(dev))
>> + return -ETIMEDOUT;
>> + pdev = to_pci_dev(dev);
>> + if (!pci_device_is_present(pdev) &&
>> + ite_sid == pci_dev_id(pci_physfn(pdev)))
>> + return -ETIMEDOUT;
>> + }
>> if (qi->desc_status[wait_index] == QI_ABORT)
>> return -EAGAIN;
>> }
>> --
>> 2.31.1
>>
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] iommu/vt-d: improve ITE fault handling if device was released
2024-02-29 4:07 [PATCH] iommu/vt-d: improve ITE fault handling if device was released Ethan Zhao
2024-02-29 4:08 ` Ethan Zhao
2024-02-29 21:07 ` Bjorn Helgaas
@ 2024-03-01 7:05 ` Ethan Zhao
2 siblings, 0 replies; 5+ messages in thread
From: Ethan Zhao @ 2024-03-01 7:05 UTC (permalink / raw)
To: baolu.lu, bhelgaas, robin.murphy, jgg
Cc: kevin.tian, dwmw2, will, lukas, yi.l.liu, dan.carpenter, iommu,
linux-kernel
On 2/29/2024 12:07 PM, Ethan Zhao wrote:
> Break the loop to blindly retry the timeout ATS invalidation request
> after ITE fault hit if device was released or isn't present anymore.
>
> This is part of the followup of prior proposed patchset
>
> https://do-db2.lkml.org/lkml/2024/2/22/350
>
> Fixes: 6ba6c3a4cacf ("VT-d: add device IOTLB invalidation support")
> Signed-off-by: Ethan Zhao <haifeng.zhao@linux.intel.com>
> ---
> drivers/iommu/intel/dmar.c | 25 +++++++++++++++++++++++++
> 1 file changed, 25 insertions(+)
>
> diff --git a/drivers/iommu/intel/dmar.c b/drivers/iommu/intel/dmar.c
> index d14797aabb7a..d01d68205557 100644
> --- a/drivers/iommu/intel/dmar.c
> +++ b/drivers/iommu/intel/dmar.c
> @@ -1273,6 +1273,9 @@ static int qi_check_fault(struct intel_iommu *iommu, int index, int wait_index)
> {
> u32 fault;
> int head, tail;
> + u64 iqe_err, ite_sid;
> + struct device *dev = NULL;
> + struct pci_dev *pdev = NULL;
> struct q_inval *qi = iommu->qi;
> int shift = qi_shift(iommu);
>
> @@ -1317,6 +1320,13 @@ static int qi_check_fault(struct intel_iommu *iommu, int index, int wait_index)
> tail = readl(iommu->reg + DMAR_IQT_REG);
> tail = ((tail >> shift) - 1 + QI_LENGTH) % QI_LENGTH;
>
> + /*
> + * SID field is valid only when the ITE field is Set in FSTS_REG
> + * see Intel VT-d spec r4.1, section 11.4.9.9
> + */
> + iqe_err = dmar_readq(iommu->reg + DMAR_IQER_REG);
> + ite_sid = DMAR_IQER_REG_ITESID(iqe_err);
> +
> writel(DMA_FSTS_ITE, iommu->reg + DMAR_FSTS_REG);
> pr_info("Invalidation Time-out Error (ITE) cleared\n");
>
> @@ -1326,6 +1336,21 @@ static int qi_check_fault(struct intel_iommu *iommu, int index, int wait_index)
> head = (head - 2 + QI_LENGTH) % QI_LENGTH;
> } while (head != tail);
>
> + /*
> + * If got ITE, we need to check if the sid of ITE is one of the
> + * current valid ATS invalidation target devices, if no, or the
> + * target device isn't presnet, don't try this request anymore.
> + * 0 value of ite_sid means old VT-d device, no ite_sid value.
> + */
> + if (ite_sid) {
> + dev = device_rbtree_find(iommu, ite_sid);
> + if (!dev || !dev_is_pci(dev))
> + return -ETIMEDOUT;
> + pdev = to_pci_dev(dev);
> + if (!pci_device_is_present(pdev) &&
> + ite_sid == pci_dev_id(pci_physfn(pdev)))
> + return -ETIMEDOUT;
> + }
> if (qi->desc_status[wait_index] == QI_ABORT)
> return -EAGAIN;
> }
Given maintainer is going to pick up patchset
https://lore.kernel.org/lkml/2d1788da-521c-4531-a159-81d2fb801d6c@linux.intel.com/T/
and this one is mutually exclusive with it, suspend.
Thanks,
Ethan
^ permalink raw reply [flat|nested] 5+ messages in thread