* [PATCH v3 1/8] iommu/amd: Refactor AMD IOMMU debugfs initial setup
2025-02-06 5:59 [PATCH v3 0/8] Introduce debugfs support in IOMMU Dheeraj Kumar Srivastava
@ 2025-02-06 5:59 ` Dheeraj Kumar Srivastava
2025-03-13 10:27 ` Vasant Hegde
2025-02-06 5:59 ` [PATCH v3 2/8] iommu/amd: Add debugfs support to dump IOMMU MMIO registers Dheeraj Kumar Srivastava
` (8 subsequent siblings)
9 siblings, 1 reply; 22+ messages in thread
From: Dheeraj Kumar Srivastava @ 2025-02-06 5:59 UTC (permalink / raw)
To: joro, suravee.suthikulpanit, will, robin.murphy, linux-kernel,
iommu
Cc: vasant.hegde
Rearrange initial setup of AMD IOMMU debugfs to segregate per IOMMU
setup and setup which is common for all IOMMUs. This ensures that common
debugfs paths (introduced in subsequent patches) are created only once
instead of being created for each IOMMU.
With the change, there is no need to use lock as amd_iommu_debugfs_setup()
will be called only once during AMD IOMMU initialization. So remove lock
acquisition in amd_iommu_debugfs_setup().
Signed-off-by: Dheeraj Kumar Srivastava <dheerajkumar.srivastava@amd.com>
Reviewed-by: Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>
---
drivers/iommu/amd/amd_iommu.h | 4 ++--
drivers/iommu/amd/debugfs.c | 16 +++++++---------
drivers/iommu/amd/init.c | 5 ++---
3 files changed, 11 insertions(+), 14 deletions(-)
diff --git a/drivers/iommu/amd/amd_iommu.h b/drivers/iommu/amd/amd_iommu.h
index 68debf5ee2d7..58308ed12d83 100644
--- a/drivers/iommu/amd/amd_iommu.h
+++ b/drivers/iommu/amd/amd_iommu.h
@@ -28,9 +28,9 @@ void *__init iommu_alloc_4k_pages(struct amd_iommu *iommu,
gfp_t gfp, size_t size);
#ifdef CONFIG_AMD_IOMMU_DEBUGFS
-void amd_iommu_debugfs_setup(struct amd_iommu *iommu);
+void amd_iommu_debugfs_setup(void);
#else
-static inline void amd_iommu_debugfs_setup(struct amd_iommu *iommu) {}
+static inline void amd_iommu_debugfs_setup(void) {}
#endif
/* Needed for interrupt remapping */
diff --git a/drivers/iommu/amd/debugfs.c b/drivers/iommu/amd/debugfs.c
index 545372fcc72f..ff9520e002be 100644
--- a/drivers/iommu/amd/debugfs.c
+++ b/drivers/iommu/amd/debugfs.c
@@ -13,20 +13,18 @@
#include "amd_iommu.h"
static struct dentry *amd_iommu_debugfs;
-static DEFINE_MUTEX(amd_iommu_debugfs_lock);
#define MAX_NAME_LEN 20
-void amd_iommu_debugfs_setup(struct amd_iommu *iommu)
+void amd_iommu_debugfs_setup(void)
{
+ struct amd_iommu *iommu;
char name[MAX_NAME_LEN + 1];
- mutex_lock(&amd_iommu_debugfs_lock);
- if (!amd_iommu_debugfs)
- amd_iommu_debugfs = debugfs_create_dir("amd",
- iommu_debugfs_dir);
- mutex_unlock(&amd_iommu_debugfs_lock);
+ amd_iommu_debugfs = debugfs_create_dir("amd", iommu_debugfs_dir);
- snprintf(name, MAX_NAME_LEN, "iommu%02d", iommu->index);
- iommu->debugfs = debugfs_create_dir(name, amd_iommu_debugfs);
+ for_each_iommu(iommu) {
+ snprintf(name, MAX_NAME_LEN, "iommu%02d", iommu->index);
+ iommu->debugfs = debugfs_create_dir(name, amd_iommu_debugfs);
+ }
}
diff --git a/drivers/iommu/amd/init.c b/drivers/iommu/amd/init.c
index c5cd92edada0..1f0d422bfcbf 100644
--- a/drivers/iommu/amd/init.c
+++ b/drivers/iommu/amd/init.c
@@ -3375,7 +3375,6 @@ int amd_iommu_enable_faulting(unsigned int cpu)
*/
static int __init amd_iommu_init(void)
{
- struct amd_iommu *iommu;
int ret;
ret = iommu_go_to_state(IOMMU_INITIALIZED);
@@ -3389,8 +3388,8 @@ static int __init amd_iommu_init(void)
}
#endif
- for_each_iommu(iommu)
- amd_iommu_debugfs_setup(iommu);
+ if (!ret)
+ amd_iommu_debugfs_setup();
return ret;
}
--
2.25.1
^ permalink raw reply related [flat|nested] 22+ messages in thread* Re: [PATCH v3 1/8] iommu/amd: Refactor AMD IOMMU debugfs initial setup
2025-02-06 5:59 ` [PATCH v3 1/8] iommu/amd: Refactor AMD IOMMU debugfs initial setup Dheeraj Kumar Srivastava
@ 2025-03-13 10:27 ` Vasant Hegde
0 siblings, 0 replies; 22+ messages in thread
From: Vasant Hegde @ 2025-03-13 10:27 UTC (permalink / raw)
To: Dheeraj Kumar Srivastava, joro, suravee.suthikulpanit, will,
robin.murphy, linux-kernel, iommu
On 2/6/2025 11:29 AM, Dheeraj Kumar Srivastava wrote:
> Rearrange initial setup of AMD IOMMU debugfs to segregate per IOMMU
> setup and setup which is common for all IOMMUs. This ensures that common
> debugfs paths (introduced in subsequent patches) are created only once
> instead of being created for each IOMMU.
>
> With the change, there is no need to use lock as amd_iommu_debugfs_setup()
> will be called only once during AMD IOMMU initialization. So remove lock
> acquisition in amd_iommu_debugfs_setup().
>
> Signed-off-by: Dheeraj Kumar Srivastava <dheerajkumar.srivastava@amd.com>
> Reviewed-by: Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>
Reviewed-by: Vasant Hegde <vasant.hegde@amd.com>
-Vasant
^ permalink raw reply [flat|nested] 22+ messages in thread
* [PATCH v3 2/8] iommu/amd: Add debugfs support to dump IOMMU MMIO registers
2025-02-06 5:59 [PATCH v3 0/8] Introduce debugfs support in IOMMU Dheeraj Kumar Srivastava
2025-02-06 5:59 ` [PATCH v3 1/8] iommu/amd: Refactor AMD IOMMU debugfs initial setup Dheeraj Kumar Srivastava
@ 2025-02-06 5:59 ` Dheeraj Kumar Srivastava
2025-03-13 10:28 ` Vasant Hegde
2025-02-06 5:59 ` [PATCH v3 3/8] iommu/amd: Add debugfs support to dump IOMMU Capability registers Dheeraj Kumar Srivastava
` (7 subsequent siblings)
9 siblings, 1 reply; 22+ messages in thread
From: Dheeraj Kumar Srivastava @ 2025-02-06 5:59 UTC (permalink / raw)
To: joro, suravee.suthikulpanit, will, robin.murphy, linux-kernel,
iommu
Cc: vasant.hegde
Analyzing IOMMU MMIO registers gives a view of what IOMMU is
configured with on the system and is helpful to debug issues
with IOMMU.
eg.
1. To get mmio registers value for iommu<x>
# echo "0x18" > /sys/kernel/debug/iommu/amd/iommu00/mmio
# cat /sys/kernel/debug/iommu/amd/iommu00/mmio
Signed-off-by: Dheeraj Kumar Srivastava <dheerajkumar.srivastava@amd.com>
---
drivers/iommu/amd/debugfs.c | 45 +++++++++++++++++++++++++++++++++++++
1 file changed, 45 insertions(+)
diff --git a/drivers/iommu/amd/debugfs.c b/drivers/iommu/amd/debugfs.c
index ff9520e002be..b16b62ae7111 100644
--- a/drivers/iommu/amd/debugfs.c
+++ b/drivers/iommu/amd/debugfs.c
@@ -15,6 +15,48 @@
static struct dentry *amd_iommu_debugfs;
#define MAX_NAME_LEN 20
+#define OFS_IN_SZ 8
+
+static int mmio_offset = -1;
+
+static ssize_t iommu_mmio_write(struct file *filp, const char __user *ubuf,
+ size_t cnt, loff_t *ppos)
+{
+ struct seq_file *m = filp->private_data;
+ struct amd_iommu *iommu = m->private;
+ int ret;
+
+ if (cnt > OFS_IN_SZ)
+ return -EINVAL;
+
+ ret = kstrtou32_from_user(ubuf, cnt, 0, &mmio_offset);
+ if (ret)
+ return ret;
+
+ if (mmio_offset > iommu->mmio_phys_end - 4) {
+ mmio_offset = -1;
+ return -EINVAL;
+ }
+
+ return cnt;
+}
+
+static int iommu_mmio_show(struct seq_file *m, void *unused)
+{
+ struct amd_iommu *iommu = m->private;
+ u32 value;
+
+ if (mmio_offset < 0) {
+ seq_puts(m, "Please provide mmio register's offset\n");
+ return 0;
+ }
+
+ value = readl(iommu->mmio_base + mmio_offset);
+ seq_printf(m, "Offset:0x%x Value:0x%08x\n", mmio_offset, value);
+
+ return 0;
+}
+DEFINE_SHOW_STORE_ATTRIBUTE(iommu_mmio);
void amd_iommu_debugfs_setup(void)
{
@@ -26,5 +68,8 @@ void amd_iommu_debugfs_setup(void)
for_each_iommu(iommu) {
snprintf(name, MAX_NAME_LEN, "iommu%02d", iommu->index);
iommu->debugfs = debugfs_create_dir(name, amd_iommu_debugfs);
+
+ debugfs_create_file("mmio", 0644, iommu->debugfs, iommu,
+ &iommu_mmio_fops);
}
}
--
2.25.1
^ permalink raw reply related [flat|nested] 22+ messages in thread* Re: [PATCH v3 2/8] iommu/amd: Add debugfs support to dump IOMMU MMIO registers
2025-02-06 5:59 ` [PATCH v3 2/8] iommu/amd: Add debugfs support to dump IOMMU MMIO registers Dheeraj Kumar Srivastava
@ 2025-03-13 10:28 ` Vasant Hegde
2025-03-17 11:22 ` Dheeraj Kumar Srivastava
0 siblings, 1 reply; 22+ messages in thread
From: Vasant Hegde @ 2025-03-13 10:28 UTC (permalink / raw)
To: Dheeraj Kumar Srivastava, joro, suravee.suthikulpanit, will,
robin.murphy, linux-kernel, iommu
On 2/6/2025 11:29 AM, Dheeraj Kumar Srivastava wrote:
> Analyzing IOMMU MMIO registers gives a view of what IOMMU is
> configured with on the system and is helpful to debug issues
> with IOMMU.
>
> eg.
> 1. To get mmio registers value for iommu<x>
> # echo "0x18" > /sys/kernel/debug/iommu/amd/iommu00/mmio
> # cat /sys/kernel/debug/iommu/amd/iommu00/mmio
Many of MMIO registers are 8 bytes. By default can we print 8 bytes instead of 4
bytes?
-Vasant
^ permalink raw reply [flat|nested] 22+ messages in thread
* Re: [PATCH v3 2/8] iommu/amd: Add debugfs support to dump IOMMU MMIO registers
2025-03-13 10:28 ` Vasant Hegde
@ 2025-03-17 11:22 ` Dheeraj Kumar Srivastava
0 siblings, 0 replies; 22+ messages in thread
From: Dheeraj Kumar Srivastava @ 2025-03-17 11:22 UTC (permalink / raw)
To: Vasant Hegde, joro, suravee.suthikulpanit, will, robin.murphy,
linux-kernel, iommu
Hi Vasant,
On 3/13/2025 3:58 PM, Vasant Hegde wrote:
>
>
> On 2/6/2025 11:29 AM, Dheeraj Kumar Srivastava wrote:
>> Analyzing IOMMU MMIO registers gives a view of what IOMMU is
>> configured with on the system and is helpful to debug issues
>> with IOMMU.
>>
>> eg.
>> 1. To get mmio registers value for iommu<x>
>> # echo "0x18" > /sys/kernel/debug/iommu/amd/iommu00/mmio
>> # cat /sys/kernel/debug/iommu/amd/iommu00/mmio
>
> Many of MMIO registers are 8 bytes. By default can we print 8 bytes instead of 4
> bytes?
Sure will print 64 bits (8 bytes) for MMIO registers dump.
Thanks
Dheeraj
>
> -Vasant
>
^ permalink raw reply [flat|nested] 22+ messages in thread
* [PATCH v3 3/8] iommu/amd: Add debugfs support to dump IOMMU Capability registers
2025-02-06 5:59 [PATCH v3 0/8] Introduce debugfs support in IOMMU Dheeraj Kumar Srivastava
2025-02-06 5:59 ` [PATCH v3 1/8] iommu/amd: Refactor AMD IOMMU debugfs initial setup Dheeraj Kumar Srivastava
2025-02-06 5:59 ` [PATCH v3 2/8] iommu/amd: Add debugfs support to dump IOMMU MMIO registers Dheeraj Kumar Srivastava
@ 2025-02-06 5:59 ` Dheeraj Kumar Srivastava
2025-03-13 10:32 ` Vasant Hegde
2025-02-06 5:59 ` [PATCH v3 4/8] iommu/amd: Add debugfs support to dump IOMMU command buffer Dheeraj Kumar Srivastava
` (6 subsequent siblings)
9 siblings, 1 reply; 22+ messages in thread
From: Dheeraj Kumar Srivastava @ 2025-02-06 5:59 UTC (permalink / raw)
To: joro, suravee.suthikulpanit, will, robin.murphy, linux-kernel,
iommu
Cc: vasant.hegde
IOMMU Capability registers defines capabilities of IOMMU and information
needed for initialising MMIO registers and device table. This is useful
to dump these registers for debugging IOMMU related issues.
e.g.To get capability registers value for iommu<x>
# echo "0x10" > /sys/kernel/debug/iommu/amd/iommu00/capability
# cat /sys/kernel/debug/iommu/amd/iommu00/capability
Signed-off-by: Dheeraj Kumar Srivastava <dheerajkumar.srivastava@amd.com>
---
drivers/iommu/amd/debugfs.c | 47 +++++++++++++++++++++++++++++++++++++
1 file changed, 47 insertions(+)
diff --git a/drivers/iommu/amd/debugfs.c b/drivers/iommu/amd/debugfs.c
index b16b62ae7111..dc5fdc905754 100644
--- a/drivers/iommu/amd/debugfs.c
+++ b/drivers/iommu/amd/debugfs.c
@@ -18,6 +18,7 @@ static struct dentry *amd_iommu_debugfs;
#define OFS_IN_SZ 8
static int mmio_offset = -1;
+static int cap_offset = -1;
static ssize_t iommu_mmio_write(struct file *filp, const char __user *ubuf,
size_t cnt, loff_t *ppos)
@@ -58,6 +59,50 @@ static int iommu_mmio_show(struct seq_file *m, void *unused)
}
DEFINE_SHOW_STORE_ATTRIBUTE(iommu_mmio);
+static ssize_t iommu_capability_write(struct file *filp, const char __user *ubuf,
+ size_t cnt, loff_t *ppos)
+{
+ int ret;
+
+ if (cnt > OFS_IN_SZ)
+ return -EINVAL;
+
+ ret = kstrtou32_from_user(ubuf, cnt, 0, &cap_offset);
+ if (ret)
+ return ret;
+
+ /* Capability register at offset 0x14 is the last IOMMU capability register. */
+ if (cap_offset > 0x14) {
+ cap_offset = -1;
+ return -EINVAL;
+ }
+
+ return cnt;
+}
+
+static int iommu_capability_show(struct seq_file *m, void *unused)
+{
+ struct amd_iommu *iommu = m->private;
+ u32 value;
+ int err;
+
+ if (cap_offset < 0) {
+ seq_puts(m, "Please provide capability register's offset\n");
+ return 0;
+ }
+
+ err = pci_read_config_dword(iommu->dev, iommu->cap_ptr + cap_offset, &value);
+ if (err) {
+ seq_printf(m, "Not able to read capability register at 0x%x\n", cap_offset);
+ return 0;
+ }
+
+ seq_printf(m, "Offset:0x%x Value:0x%08x\n", cap_offset, value);
+
+ return 0;
+}
+DEFINE_SHOW_STORE_ATTRIBUTE(iommu_capability);
+
void amd_iommu_debugfs_setup(void)
{
struct amd_iommu *iommu;
@@ -71,5 +116,7 @@ void amd_iommu_debugfs_setup(void)
debugfs_create_file("mmio", 0644, iommu->debugfs, iommu,
&iommu_mmio_fops);
+ debugfs_create_file("capability", 0644, iommu->debugfs, iommu,
+ &iommu_capability_fops);
}
}
--
2.25.1
^ permalink raw reply related [flat|nested] 22+ messages in thread* Re: [PATCH v3 3/8] iommu/amd: Add debugfs support to dump IOMMU Capability registers
2025-02-06 5:59 ` [PATCH v3 3/8] iommu/amd: Add debugfs support to dump IOMMU Capability registers Dheeraj Kumar Srivastava
@ 2025-03-13 10:32 ` Vasant Hegde
2025-03-17 11:23 ` Dheeraj Kumar Srivastava
0 siblings, 1 reply; 22+ messages in thread
From: Vasant Hegde @ 2025-03-13 10:32 UTC (permalink / raw)
To: Dheeraj Kumar Srivastava, joro, suravee.suthikulpanit, will,
robin.murphy, linux-kernel, iommu
Dheeraj,
On 2/6/2025 11:29 AM, Dheeraj Kumar Srivastava wrote:
> IOMMU Capability registers defines capabilities of IOMMU and information
> needed for initialising MMIO registers and device table. This is useful
> to dump these registers for debugging IOMMU related issues.
>
> e.g.To get capability registers value for iommu<x>
> # echo "0x10" > /sys/kernel/debug/iommu/amd/iommu00/capability
> # cat /sys/kernel/debug/iommu/amd/iommu00/capability
>
> Signed-off-by: Dheeraj Kumar Srivastava <dheerajkumar.srivastava@amd.com>
> ---
> drivers/iommu/amd/debugfs.c | 47 +++++++++++++++++++++++++++++++++++++
> 1 file changed, 47 insertions(+)
>
> diff --git a/drivers/iommu/amd/debugfs.c b/drivers/iommu/amd/debugfs.c
> index b16b62ae7111..dc5fdc905754 100644
> --- a/drivers/iommu/amd/debugfs.c
> +++ b/drivers/iommu/amd/debugfs.c
> @@ -18,6 +18,7 @@ static struct dentry *amd_iommu_debugfs;
> #define OFS_IN_SZ 8
>
> static int mmio_offset = -1;
> +static int cap_offset = -1;
>
> static ssize_t iommu_mmio_write(struct file *filp, const char __user *ubuf,
> size_t cnt, loff_t *ppos)
> @@ -58,6 +59,50 @@ static int iommu_mmio_show(struct seq_file *m, void *unused)
> }
> DEFINE_SHOW_STORE_ATTRIBUTE(iommu_mmio);
>
> +static ssize_t iommu_capability_write(struct file *filp, const char __user *ubuf,
> + size_t cnt, loff_t *ppos)
> +{
> + int ret;
> +
> + if (cnt > OFS_IN_SZ)
> + return -EINVAL;
> +
> + ret = kstrtou32_from_user(ubuf, cnt, 0, &cap_offset);
> + if (ret)
> + return ret;
> +
> + /* Capability register at offset 0x14 is the last IOMMU capability register. */
> + if (cap_offset > 0x14) {
> + cap_offset = -1;
> + return -EINVAL;
> + }
> +
> + return cnt;
> +}
> +
> +static int iommu_capability_show(struct seq_file *m, void *unused)
> +{
> + struct amd_iommu *iommu = m->private;
> + u32 value;
> + int err;
> +
> + if (cap_offset < 0) {
> + seq_puts(m, "Please provide capability register's offset\n");
May be you can add last offset details... so that it becomes easy to find out
the offset range.
-Vasant
^ permalink raw reply [flat|nested] 22+ messages in thread* Re: [PATCH v3 3/8] iommu/amd: Add debugfs support to dump IOMMU Capability registers
2025-03-13 10:32 ` Vasant Hegde
@ 2025-03-17 11:23 ` Dheeraj Kumar Srivastava
0 siblings, 0 replies; 22+ messages in thread
From: Dheeraj Kumar Srivastava @ 2025-03-17 11:23 UTC (permalink / raw)
To: Vasant Hegde, joro, suravee.suthikulpanit, will, robin.murphy,
linux-kernel, iommu
Hi Vasant,
On 3/13/2025 4:02 PM, Vasant Hegde wrote:
> Dheeraj,
>
>
> On 2/6/2025 11:29 AM, Dheeraj Kumar Srivastava wrote:
>> IOMMU Capability registers defines capabilities of IOMMU and information
>> needed for initialising MMIO registers and device table. This is useful
>> to dump these registers for debugging IOMMU related issues.
>>
>> e.g.To get capability registers value for iommu<x>
>> # echo "0x10" > /sys/kernel/debug/iommu/amd/iommu00/capability
>> # cat /sys/kernel/debug/iommu/amd/iommu00/capability
>>
>> Signed-off-by: Dheeraj Kumar Srivastava <dheerajkumar.srivastava@amd.com>
>> ---
>> drivers/iommu/amd/debugfs.c | 47 +++++++++++++++++++++++++++++++++++++
>> 1 file changed, 47 insertions(+)
>>
>> diff --git a/drivers/iommu/amd/debugfs.c b/drivers/iommu/amd/debugfs.c
>> index b16b62ae7111..dc5fdc905754 100644
>> --- a/drivers/iommu/amd/debugfs.c
>> +++ b/drivers/iommu/amd/debugfs.c
>> @@ -18,6 +18,7 @@ static struct dentry *amd_iommu_debugfs;
>> #define OFS_IN_SZ 8
>>
>> static int mmio_offset = -1;
>> +static int cap_offset = -1;
>>
>> static ssize_t iommu_mmio_write(struct file *filp, const char __user *ubuf,
>> size_t cnt, loff_t *ppos)
>> @@ -58,6 +59,50 @@ static int iommu_mmio_show(struct seq_file *m, void *unused)
>> }
>> DEFINE_SHOW_STORE_ATTRIBUTE(iommu_mmio);
>>
>> +static ssize_t iommu_capability_write(struct file *filp, const char __user *ubuf,
>> + size_t cnt, loff_t *ppos)
>> +{
>> + int ret;
>> +
>> + if (cnt > OFS_IN_SZ)
>> + return -EINVAL;
>> +
>> + ret = kstrtou32_from_user(ubuf, cnt, 0, &cap_offset);
>> + if (ret)
>> + return ret;
>> +
>> + /* Capability register at offset 0x14 is the last IOMMU capability register. */
>> + if (cap_offset > 0x14) {
>> + cap_offset = -1;
>> + return -EINVAL;
>> + }
>> +
>> + return cnt;
>> +}
>> +
>> +static int iommu_capability_show(struct seq_file *m, void *unused)
>> +{
>> + struct amd_iommu *iommu = m->private;
>> + u32 value;
>> + int err;
>> +
>> + if (cap_offset < 0) {
>> + seq_puts(m, "Please provide capability register's offset\n");
>
> May be you can add last offset details... so that it becomes easy to find out
> the offset range.
Sure.
Thanks
Dheeraj
>
> -Vasant
>
>
^ permalink raw reply [flat|nested] 22+ messages in thread
* [PATCH v3 4/8] iommu/amd: Add debugfs support to dump IOMMU command buffer
2025-02-06 5:59 [PATCH v3 0/8] Introduce debugfs support in IOMMU Dheeraj Kumar Srivastava
` (2 preceding siblings ...)
2025-02-06 5:59 ` [PATCH v3 3/8] iommu/amd: Add debugfs support to dump IOMMU Capability registers Dheeraj Kumar Srivastava
@ 2025-02-06 5:59 ` Dheeraj Kumar Srivastava
2025-03-13 10:38 ` Vasant Hegde
2025-02-06 6:00 ` [PATCH v3 5/8] iommu/amd: Add support for device id user input Dheeraj Kumar Srivastava
` (5 subsequent siblings)
9 siblings, 1 reply; 22+ messages in thread
From: Dheeraj Kumar Srivastava @ 2025-02-06 5:59 UTC (permalink / raw)
To: joro, suravee.suthikulpanit, will, robin.murphy, linux-kernel,
iommu
Cc: vasant.hegde
IOMMU driver sends command to IOMMU hardware via command buffer. In cases
where IOMMU hardware fails to process commands in command buffer, dumping
it is a valuable input to debug the issue.
IOMMU hardware processes command buffer entry at offset equals to the head
pointer. Dumping just the entry at the head pointer may not always be
useful. The current head may not be pointing to the entry of the command
buffer which is causing the issue. IOMMU Hardware may have processed the
entry and updated the head pointer. So dumping the entire command buffer
gives a broad understanding of what hardware was/is doing. The command
buffer dump will have all entries from start to end of the command buffer.
Along with that, it will have a head and tail command buffer pointer
register dump to facilitate where the IOMMU driver and hardware are in
the command buffer for injecting and processing the entries respectively.
Command buffer is a per IOMMU data structure. So dumping on per IOMMU
basis.
eg. To get command buffer dump for iommu<x>
#cat /sys/kernel/debug/iommu/amd/iommu<x>/cmdbuf
Signed-off-by: Dheeraj Kumar Srivastava <dheerajkumar.srivastava@amd.com>
---
drivers/iommu/amd/amd_iommu_types.h | 7 +++++++
drivers/iommu/amd/debugfs.c | 26 ++++++++++++++++++++++++++
drivers/iommu/amd/iommu.c | 7 -------
3 files changed, 33 insertions(+), 7 deletions(-)
diff --git a/drivers/iommu/amd/amd_iommu_types.h b/drivers/iommu/amd/amd_iommu_types.h
index 0bbda60d3cdc..c4ca6d9c6b51 100644
--- a/drivers/iommu/amd/amd_iommu_types.h
+++ b/drivers/iommu/amd/amd_iommu_types.h
@@ -892,6 +892,13 @@ struct dev_table_entry {
};
};
+/*
+ * Structure defining one entry in the command buffer
+ */
+struct iommu_cmd {
+ u32 data[4];
+};
+
/*
* Structure to sture persistent DTE flags from IVHD
*/
diff --git a/drivers/iommu/amd/debugfs.c b/drivers/iommu/amd/debugfs.c
index dc5fdc905754..bf718e7258c5 100644
--- a/drivers/iommu/amd/debugfs.c
+++ b/drivers/iommu/amd/debugfs.c
@@ -103,6 +103,30 @@ static int iommu_capability_show(struct seq_file *m, void *unused)
}
DEFINE_SHOW_STORE_ATTRIBUTE(iommu_capability);
+static int iommu_cmdbuf_show(struct seq_file *m, void *unused)
+{
+ struct amd_iommu *iommu = m->private;
+ struct iommu_cmd *cmd;
+ unsigned long flag;
+ u32 head, tail;
+ int i;
+
+ raw_spin_lock_irqsave(&iommu->lock, flag);
+ head = readl(iommu->mmio_base + MMIO_CMD_HEAD_OFFSET);
+ tail = readl(iommu->mmio_base + MMIO_CMD_TAIL_OFFSET);
+ seq_printf(m, "CMD Buffer Head Offset:%d Tail Offset:%d\n",
+ (head >> 4) & 0x7fff, (tail >> 4) & 0x7fff);
+ for (i = 0; i < CMD_BUFFER_ENTRIES; i++) {
+ cmd = (struct iommu_cmd *)(iommu->cmd_buf + i * sizeof(*cmd));
+ seq_printf(m, "%3d: %08x%08x%08x%08x\n", i, cmd->data[0],
+ cmd->data[1], cmd->data[2], cmd->data[3]);
+ }
+ raw_spin_unlock_irqrestore(&iommu->lock, flag);
+
+ return 0;
+}
+DEFINE_SHOW_ATTRIBUTE(iommu_cmdbuf);
+
void amd_iommu_debugfs_setup(void)
{
struct amd_iommu *iommu;
@@ -118,5 +142,7 @@ void amd_iommu_debugfs_setup(void)
&iommu_mmio_fops);
debugfs_create_file("capability", 0644, iommu->debugfs, iommu,
&iommu_capability_fops);
+ debugfs_create_file("cmdbuf", 0444, iommu->debugfs, iommu,
+ &iommu_cmdbuf_fops);
}
}
diff --git a/drivers/iommu/amd/iommu.c b/drivers/iommu/amd/iommu.c
index b48a72bd7b23..b23fabea0e8a 100644
--- a/drivers/iommu/amd/iommu.c
+++ b/drivers/iommu/amd/iommu.c
@@ -62,13 +62,6 @@ static const struct iommu_dirty_ops amd_dirty_ops;
int amd_iommu_max_glx_val = -1;
-/*
- * general struct to manage commands send to an IOMMU
- */
-struct iommu_cmd {
- u32 data[4];
-};
-
/*
* AMD IOMMU allows up to 2^16 different protection domains. This is a bitmap
* to know which ones are already in use.
--
2.25.1
^ permalink raw reply related [flat|nested] 22+ messages in thread* Re: [PATCH v3 4/8] iommu/amd: Add debugfs support to dump IOMMU command buffer
2025-02-06 5:59 ` [PATCH v3 4/8] iommu/amd: Add debugfs support to dump IOMMU command buffer Dheeraj Kumar Srivastava
@ 2025-03-13 10:38 ` Vasant Hegde
2025-03-17 11:24 ` Dheeraj Kumar Srivastava
0 siblings, 1 reply; 22+ messages in thread
From: Vasant Hegde @ 2025-03-13 10:38 UTC (permalink / raw)
To: Dheeraj Kumar Srivastava, joro, suravee.suthikulpanit, will,
robin.murphy, linux-kernel, iommu
Dheeraj,
On 2/6/2025 11:29 AM, Dheeraj Kumar Srivastava wrote:
> IOMMU driver sends command to IOMMU hardware via command buffer. In cases
> where IOMMU hardware fails to process commands in command buffer, dumping
> it is a valuable input to debug the issue.
>
> IOMMU hardware processes command buffer entry at offset equals to the head
> pointer. Dumping just the entry at the head pointer may not always be
> useful. The current head may not be pointing to the entry of the command
> buffer which is causing the issue. IOMMU Hardware may have processed the
> entry and updated the head pointer. So dumping the entire command buffer
> gives a broad understanding of what hardware was/is doing. The command
> buffer dump will have all entries from start to end of the command buffer.
> Along with that, it will have a head and tail command buffer pointer
> register dump to facilitate where the IOMMU driver and hardware are in
> the command buffer for injecting and processing the entries respectively.
>
> Command buffer is a per IOMMU data structure. So dumping on per IOMMU
> basis.
> eg. To get command buffer dump for iommu<x>
> #cat /sys/kernel/debug/iommu/amd/iommu<x>/cmdbuf
>
> Signed-off-by: Dheeraj Kumar Srivastava <dheerajkumar.srivastava@amd.com>
> ---
> drivers/iommu/amd/amd_iommu_types.h | 7 +++++++
> drivers/iommu/amd/debugfs.c | 26 ++++++++++++++++++++++++++
> drivers/iommu/amd/iommu.c | 7 -------
> 3 files changed, 33 insertions(+), 7 deletions(-)
>
> diff --git a/drivers/iommu/amd/amd_iommu_types.h b/drivers/iommu/amd/amd_iommu_types.h
> index 0bbda60d3cdc..c4ca6d9c6b51 100644
> --- a/drivers/iommu/amd/amd_iommu_types.h
> +++ b/drivers/iommu/amd/amd_iommu_types.h
> @@ -892,6 +892,13 @@ struct dev_table_entry {
> };
> };
>
> +/*
> + * Structure defining one entry in the command buffer
> + */
> +struct iommu_cmd {
> + u32 data[4];
> +};
> +
> /*
> * Structure to sture persistent DTE flags from IVHD
> */
> diff --git a/drivers/iommu/amd/debugfs.c b/drivers/iommu/amd/debugfs.c
> index dc5fdc905754..bf718e7258c5 100644
> --- a/drivers/iommu/amd/debugfs.c
> +++ b/drivers/iommu/amd/debugfs.c
> @@ -103,6 +103,30 @@ static int iommu_capability_show(struct seq_file *m, void *unused)
> }
> DEFINE_SHOW_STORE_ATTRIBUTE(iommu_capability);
>
> +static int iommu_cmdbuf_show(struct seq_file *m, void *unused)
> +{
> + struct amd_iommu *iommu = m->private;
> + struct iommu_cmd *cmd;
> + unsigned long flag;
> + u32 head, tail;
> + int i;
> +
> + raw_spin_lock_irqsave(&iommu->lock, flag);
> + head = readl(iommu->mmio_base + MMIO_CMD_HEAD_OFFSET);
> + tail = readl(iommu->mmio_base + MMIO_CMD_TAIL_OFFSET);
> + seq_printf(m, "CMD Buffer Head Offset:%d Tail Offset:%d\n",
> + (head >> 4) & 0x7fff, (tail >> 4) & 0x7fff);
> + for (i = 0; i < CMD_BUFFER_ENTRIES; i++) {
> + cmd = (struct iommu_cmd *)(iommu->cmd_buf + i * sizeof(*cmd));
> + seq_printf(m, "%3d: %08x%08x%08x%08x\n", i, cmd->data[0],
Can you give space between words so that its easy to read?
-Vasant
^ permalink raw reply [flat|nested] 22+ messages in thread* Re: [PATCH v3 4/8] iommu/amd: Add debugfs support to dump IOMMU command buffer
2025-03-13 10:38 ` Vasant Hegde
@ 2025-03-17 11:24 ` Dheeraj Kumar Srivastava
0 siblings, 0 replies; 22+ messages in thread
From: Dheeraj Kumar Srivastava @ 2025-03-17 11:24 UTC (permalink / raw)
To: Vasant Hegde, joro, suravee.suthikulpanit, will, robin.murphy,
linux-kernel, iommu
Hi Vasant,
On 3/13/2025 4:08 PM, Vasant Hegde wrote:
> Dheeraj,
>
>
> On 2/6/2025 11:29 AM, Dheeraj Kumar Srivastava wrote:
>> IOMMU driver sends command to IOMMU hardware via command buffer. In cases
>> where IOMMU hardware fails to process commands in command buffer, dumping
>> it is a valuable input to debug the issue.
>>
>> IOMMU hardware processes command buffer entry at offset equals to the head
>> pointer. Dumping just the entry at the head pointer may not always be
>> useful. The current head may not be pointing to the entry of the command
>> buffer which is causing the issue. IOMMU Hardware may have processed the
>> entry and updated the head pointer. So dumping the entire command buffer
>> gives a broad understanding of what hardware was/is doing. The command
>> buffer dump will have all entries from start to end of the command buffer.
>> Along with that, it will have a head and tail command buffer pointer
>> register dump to facilitate where the IOMMU driver and hardware are in
>> the command buffer for injecting and processing the entries respectively.
>>
>> Command buffer is a per IOMMU data structure. So dumping on per IOMMU
>> basis.
>> eg. To get command buffer dump for iommu<x>
>> #cat /sys/kernel/debug/iommu/amd/iommu<x>/cmdbuf
>>
>> Signed-off-by: Dheeraj Kumar Srivastava <dheerajkumar.srivastava@amd.com>
>> ---
>> drivers/iommu/amd/amd_iommu_types.h | 7 +++++++
>> drivers/iommu/amd/debugfs.c | 26 ++++++++++++++++++++++++++
>> drivers/iommu/amd/iommu.c | 7 -------
>> 3 files changed, 33 insertions(+), 7 deletions(-)
>>
>> diff --git a/drivers/iommu/amd/amd_iommu_types.h b/drivers/iommu/amd/amd_iommu_types.h
>> index 0bbda60d3cdc..c4ca6d9c6b51 100644
>> --- a/drivers/iommu/amd/amd_iommu_types.h
>> +++ b/drivers/iommu/amd/amd_iommu_types.h
>> @@ -892,6 +892,13 @@ struct dev_table_entry {
>> };
>> };
>>
>> +/*
>> + * Structure defining one entry in the command buffer
>> + */
>> +struct iommu_cmd {
>> + u32 data[4];
>> +};
>> +
>> /*
>> * Structure to sture persistent DTE flags from IVHD
>> */
>> diff --git a/drivers/iommu/amd/debugfs.c b/drivers/iommu/amd/debugfs.c
>> index dc5fdc905754..bf718e7258c5 100644
>> --- a/drivers/iommu/amd/debugfs.c
>> +++ b/drivers/iommu/amd/debugfs.c
>> @@ -103,6 +103,30 @@ static int iommu_capability_show(struct seq_file *m, void *unused)
>> }
>> DEFINE_SHOW_STORE_ATTRIBUTE(iommu_capability);
>>
>> +static int iommu_cmdbuf_show(struct seq_file *m, void *unused)
>> +{
>> + struct amd_iommu *iommu = m->private;
>> + struct iommu_cmd *cmd;
>> + unsigned long flag;
>> + u32 head, tail;
>> + int i;
>> +
>> + raw_spin_lock_irqsave(&iommu->lock, flag);
>> + head = readl(iommu->mmio_base + MMIO_CMD_HEAD_OFFSET);
>> + tail = readl(iommu->mmio_base + MMIO_CMD_TAIL_OFFSET);
>> + seq_printf(m, "CMD Buffer Head Offset:%d Tail Offset:%d\n",
>> + (head >> 4) & 0x7fff, (tail >> 4) & 0x7fff);
>> + for (i = 0; i < CMD_BUFFER_ENTRIES; i++) {
>> + cmd = (struct iommu_cmd *)(iommu->cmd_buf + i * sizeof(*cmd));
>> + seq_printf(m, "%3d: %08x%08x%08x%08x\n", i, cmd->data[0],
>
> Can you give space between words so that its easy to read?
>
Sure. Will update in v4.
Thanks
Dheeraj
> -Vasant
>
>
^ permalink raw reply [flat|nested] 22+ messages in thread
* [PATCH v3 5/8] iommu/amd: Add support for device id user input
2025-02-06 5:59 [PATCH v3 0/8] Introduce debugfs support in IOMMU Dheeraj Kumar Srivastava
` (3 preceding siblings ...)
2025-02-06 5:59 ` [PATCH v3 4/8] iommu/amd: Add debugfs support to dump IOMMU command buffer Dheeraj Kumar Srivastava
@ 2025-02-06 6:00 ` Dheeraj Kumar Srivastava
2025-03-13 10:52 ` Vasant Hegde
2025-02-06 6:00 ` [PATCH v3 6/8] iommu/amd: Add debugfs support to dump device table Dheeraj Kumar Srivastava
` (4 subsequent siblings)
9 siblings, 1 reply; 22+ messages in thread
From: Dheeraj Kumar Srivastava @ 2025-02-06 6:00 UTC (permalink / raw)
To: joro, suravee.suthikulpanit, will, robin.murphy, linux-kernel,
iommu
Cc: vasant.hegde
Device id user input is needed for dumping IOMMU data structures like
device table, IRT etc in AMD IOMMU debugfs.
eg.
1. # echo 0000:01:00.0 > /sys/kernel/debug/iommu/amd/devid
# cat /sys/kernel/debug/iommu/amd/devid
Output : 0000:01:00.0
2. # echo 01:00.0 > /sys/kernel/debug/iommu/amd/devid
# cat /sys/kernel/debug/iommu/amd/devid
Output : 0000:01:00.0
Signed-off-by: Dheeraj Kumar Srivastava <dheerajkumar.srivastava@amd.com>
---
drivers/iommu/amd/debugfs.c | 78 +++++++++++++++++++++++++++++++++++++
1 file changed, 78 insertions(+)
diff --git a/drivers/iommu/amd/debugfs.c b/drivers/iommu/amd/debugfs.c
index bf718e7258c5..f292b93f9318 100644
--- a/drivers/iommu/amd/debugfs.c
+++ b/drivers/iommu/amd/debugfs.c
@@ -16,9 +16,11 @@ static struct dentry *amd_iommu_debugfs;
#define MAX_NAME_LEN 20
#define OFS_IN_SZ 8
+#define DEVID_IN_SZ 16
static int mmio_offset = -1;
static int cap_offset = -1;
+static int sbdf = -1;
static ssize_t iommu_mmio_write(struct file *filp, const char __user *ubuf,
size_t cnt, loff_t *ppos)
@@ -127,6 +129,80 @@ static int iommu_cmdbuf_show(struct seq_file *m, void *unused)
}
DEFINE_SHOW_ATTRIBUTE(iommu_cmdbuf);
+static ssize_t devid_write(struct file *filp, const char __user *ubuf,
+ size_t cnt, loff_t *ppos)
+{
+ struct amd_iommu_pci_seg *pci_seg;
+ int seg, bus, slot, func;
+ struct amd_iommu *iommu;
+ char *srcid_ptr;
+ u16 devid;
+ int i;
+
+ sbdf = -1;
+
+ if (cnt >= DEVID_IN_SZ)
+ return -EINVAL;
+
+ srcid_ptr = memdup_user_nul(ubuf, cnt);
+ if (IS_ERR(srcid_ptr))
+ return PTR_ERR(srcid_ptr);
+
+ i = sscanf(srcid_ptr, "%x:%x:%x.%x", &seg, &bus, &slot, &func);
+ if (i != 4) {
+ i = sscanf(srcid_ptr, "%x:%x.%x", &bus, &slot, &func);
+ if (i != 3) {
+ kfree(srcid_ptr);
+ return -EINVAL;
+ }
+ seg = 0;
+ }
+
+ devid = PCI_DEVID(bus, PCI_DEVFN(slot, func));
+
+ /* Check if user device id input is a valid input */
+ for_each_pci_segment(pci_seg) {
+ if (pci_seg->id != seg)
+ continue;
+ if (devid > pci_seg->last_bdf) {
+ kfree(srcid_ptr);
+ return -EINVAL;
+ }
+ iommu = pci_seg->rlookup_table[devid];
+ if (!iommu) {
+ kfree(srcid_ptr);
+ return -ENODEV;
+ }
+ break;
+ }
+
+ if (pci_seg->id != seg) {
+ kfree(srcid_ptr);
+ return -EINVAL;
+ }
+
+ sbdf = PCI_SEG_DEVID_TO_SBDF(seg, devid);
+
+ kfree(srcid_ptr);
+
+ return cnt;
+}
+
+static int devid_show(struct seq_file *m, void *unused)
+{
+ u16 devid;
+
+ if (sbdf >= 0) {
+ devid = PCI_SBDF_TO_DEVID(sbdf);
+ seq_printf(m, "%04x:%02x:%02x:%x\n", PCI_SBDF_TO_SEGID(sbdf),
+ PCI_BUS_NUM(devid), PCI_SLOT(devid), PCI_FUNC(devid));
+ } else
+ seq_puts(m, "No or Invalid input provided\n");
+
+ return 0;
+}
+DEFINE_SHOW_STORE_ATTRIBUTE(devid);
+
void amd_iommu_debugfs_setup(void)
{
struct amd_iommu *iommu;
@@ -145,4 +221,6 @@ void amd_iommu_debugfs_setup(void)
debugfs_create_file("cmdbuf", 0444, iommu->debugfs, iommu,
&iommu_cmdbuf_fops);
}
+ debugfs_create_file("devid", 0644, amd_iommu_debugfs, NULL,
+ &devid_fops);
}
--
2.25.1
^ permalink raw reply related [flat|nested] 22+ messages in thread* Re: [PATCH v3 5/8] iommu/amd: Add support for device id user input
2025-02-06 6:00 ` [PATCH v3 5/8] iommu/amd: Add support for device id user input Dheeraj Kumar Srivastava
@ 2025-03-13 10:52 ` Vasant Hegde
2025-03-17 11:25 ` Dheeraj Kumar Srivastava
0 siblings, 1 reply; 22+ messages in thread
From: Vasant Hegde @ 2025-03-13 10:52 UTC (permalink / raw)
To: Dheeraj Kumar Srivastava, joro, suravee.suthikulpanit, will,
robin.murphy, linux-kernel, iommu
Dheeraj,
On 2/6/2025 11:30 AM, Dheeraj Kumar Srivastava wrote:
> Device id user input is needed for dumping IOMMU data structures like
> device table, IRT etc in AMD IOMMU debugfs.
You may want to rephrase the description with our approach of printing data per
device instead of printing everything.
-Vasant
^ permalink raw reply [flat|nested] 22+ messages in thread
* Re: [PATCH v3 5/8] iommu/amd: Add support for device id user input
2025-03-13 10:52 ` Vasant Hegde
@ 2025-03-17 11:25 ` Dheeraj Kumar Srivastava
0 siblings, 0 replies; 22+ messages in thread
From: Dheeraj Kumar Srivastava @ 2025-03-17 11:25 UTC (permalink / raw)
To: Vasant Hegde, joro, suravee.suthikulpanit, will, robin.murphy,
linux-kernel, iommu
Hi Vasant
On 3/13/2025 4:22 PM, Vasant Hegde wrote:
> Dheeraj,
>
> On 2/6/2025 11:30 AM, Dheeraj Kumar Srivastava wrote:
>> Device id user input is needed for dumping IOMMU data structures like
>> device table, IRT etc in AMD IOMMU debugfs.
>
> You may want to rephrase the description with our approach of printing data per
> device instead of printing everything.
>
Sure i will add the need to print data per device instead of everything.
Thanks
Dheeraj
>
> -Vasant
>
>
^ permalink raw reply [flat|nested] 22+ messages in thread
* [PATCH v3 6/8] iommu/amd: Add debugfs support to dump device table
2025-02-06 5:59 [PATCH v3 0/8] Introduce debugfs support in IOMMU Dheeraj Kumar Srivastava
` (4 preceding siblings ...)
2025-02-06 6:00 ` [PATCH v3 5/8] iommu/amd: Add support for device id user input Dheeraj Kumar Srivastava
@ 2025-02-06 6:00 ` Dheeraj Kumar Srivastava
2025-02-06 6:00 ` [PATCH v3 7/8] iommu/amd: Add debugfs support to dump IRT Table Dheeraj Kumar Srivastava
` (3 subsequent siblings)
9 siblings, 0 replies; 22+ messages in thread
From: Dheeraj Kumar Srivastava @ 2025-02-06 6:00 UTC (permalink / raw)
To: joro, suravee.suthikulpanit, will, robin.murphy, linux-kernel,
iommu
Cc: vasant.hegde
IOMMU uses device table data structure to get per-device information for
DMA remapping, interrupt remapping, and other functionalities. It's a
valuable data structure to visualize for debugging issues related to
IOMMU.
eg.
To dump device table entry for a particular device
#echo 0000:c4:00.0 > /sys/kernel/debug/iommu/amd/devid
#cat /sys/kernel/debug/iommu/amd/devtbl
or
#echo c4:00.0 > /sys/kernel/debug/iommu/amd/devid
#cat /sys/kernel/debug/iommu/amd/devtbl
Signed-off-by: Dheeraj Kumar Srivastava <dheerajkumar.srivastava@amd.com>
---
drivers/iommu/amd/debugfs.c | 49 +++++++++++++++++++++++++++++++++++++
1 file changed, 49 insertions(+)
diff --git a/drivers/iommu/amd/debugfs.c b/drivers/iommu/amd/debugfs.c
index f292b93f9318..1377795814d1 100644
--- a/drivers/iommu/amd/debugfs.c
+++ b/drivers/iommu/amd/debugfs.c
@@ -203,6 +203,53 @@ static int devid_show(struct seq_file *m, void *unused)
}
DEFINE_SHOW_STORE_ATTRIBUTE(devid);
+static void dump_dte(struct seq_file *m, struct amd_iommu_pci_seg *pci_seg, u16 devid)
+{
+ struct dev_table_entry *dev_table;
+ struct amd_iommu *iommu;
+
+ iommu = pci_seg->rlookup_table[devid];
+ if (!iommu)
+ return;
+
+ dev_table = get_dev_table(iommu);
+ if (!dev_table) {
+ seq_puts(m, "Device table not found");
+ return;
+ }
+
+ seq_printf(m, "%-12s %16s %16s %16s %16s iommu\n", "DeviceId",
+ "QWORD[3]", "QWORD[2]", "QWORD[1]", "QWORD[0]");
+ seq_printf(m, "%04x:%02x:%02x:%x ", pci_seg->id, PCI_BUS_NUM(devid),
+ PCI_SLOT(devid), PCI_FUNC(devid));
+ for (int i = 3; i >= 0; --i)
+ seq_printf(m, "%016llx ", dev_table[devid].data[i]);
+ seq_printf(m, "iommu%d\n", iommu->index);
+}
+
+static int iommu_devtbl_show(struct seq_file *m, void *unused)
+{
+ struct amd_iommu_pci_seg *pci_seg;
+ u16 seg, devid;
+
+ if (sbdf < 0) {
+ seq_puts(m, "Please provide valid device id input\n");
+ return 0;
+ }
+ seg = PCI_SBDF_TO_SEGID(sbdf);
+ devid = PCI_SBDF_TO_DEVID(sbdf);
+
+ for_each_pci_segment(pci_seg) {
+ if (pci_seg->id != seg)
+ continue;
+ dump_dte(m, pci_seg, devid);
+ break;
+ }
+
+ return 0;
+}
+DEFINE_SHOW_ATTRIBUTE(iommu_devtbl);
+
void amd_iommu_debugfs_setup(void)
{
struct amd_iommu *iommu;
@@ -223,4 +270,6 @@ void amd_iommu_debugfs_setup(void)
}
debugfs_create_file("devid", 0644, amd_iommu_debugfs, NULL,
&devid_fops);
+ debugfs_create_file("devtbl", 0444, amd_iommu_debugfs, NULL,
+ &iommu_devtbl_fops);
}
--
2.25.1
^ permalink raw reply related [flat|nested] 22+ messages in thread* [PATCH v3 7/8] iommu/amd: Add debugfs support to dump IRT Table
2025-02-06 5:59 [PATCH v3 0/8] Introduce debugfs support in IOMMU Dheeraj Kumar Srivastava
` (5 preceding siblings ...)
2025-02-06 6:00 ` [PATCH v3 6/8] iommu/amd: Add debugfs support to dump device table Dheeraj Kumar Srivastava
@ 2025-02-06 6:00 ` Dheeraj Kumar Srivastava
2025-03-13 11:03 ` Vasant Hegde
2025-02-06 6:00 ` [PATCH v3 8/8] iommu/amd: Add documentation for AMD IOMMU debugfs support Dheeraj Kumar Srivastava
` (2 subsequent siblings)
9 siblings, 1 reply; 22+ messages in thread
From: Dheeraj Kumar Srivastava @ 2025-02-06 6:00 UTC (permalink / raw)
To: joro, suravee.suthikulpanit, will, robin.murphy, linux-kernel,
iommu
Cc: vasant.hegde
In cases where we have an issue in the device interrupt path with IOMMU
interrupt remapping enabled, dumping valid IRT table entries for the device
is very useful and good input for debugging the issue.
eg.
To dump irte entries for a particular device
#echo "c4:00.0" > /sys/kernel/debug/iommu/amd/devid
#cat /sys/kernel/debug/iommu/amd/irqtbl | less
or
#echo "0000:c4:00.0" > /sys/kernel/debug/iommu/amd/devid
#cat /sys/kernel/debug/iommu/amd/irqtbl | less
Signed-off-by: Dheeraj Kumar Srivastava <dheerajkumar.srivastava@amd.com>
---
drivers/iommu/amd/debugfs.c | 89 +++++++++++++++++++++++++++++++++++++
1 file changed, 89 insertions(+)
diff --git a/drivers/iommu/amd/debugfs.c b/drivers/iommu/amd/debugfs.c
index 1377795814d1..0aff1d616108 100644
--- a/drivers/iommu/amd/debugfs.c
+++ b/drivers/iommu/amd/debugfs.c
@@ -11,6 +11,7 @@
#include <linux/pci.h>
#include "amd_iommu.h"
+#include "../irq_remapping.h"
static struct dentry *amd_iommu_debugfs;
@@ -250,6 +251,92 @@ static int iommu_devtbl_show(struct seq_file *m, void *unused)
}
DEFINE_SHOW_ATTRIBUTE(iommu_devtbl);
+static void dump_128_irte(struct seq_file *m, struct irq_remap_table *table)
+{
+ struct irte_ga *ptr, *irte;
+ int index;
+
+ for (index = 0; index < MAX_IRQS_PER_TABLE; index++) {
+ ptr = (struct irte_ga *)table->table;
+ irte = &ptr[index];
+
+ if (AMD_IOMMU_GUEST_IR_VAPIC(amd_iommu_guest_ir) &&
+ !irte->lo.fields_vapic.valid)
+ continue;
+ else if (!irte->lo.fields_remap.valid)
+ continue;
+ seq_printf(m, "IRT[%04d] %016llx%016llx\n", index, irte->hi.val, irte->lo.val);
+ }
+}
+
+static void dump_32_irte(struct seq_file *m, struct irq_remap_table *table)
+{
+ union irte *ptr, *irte;
+ int index;
+
+ for (index = 0; index < MAX_IRQS_PER_TABLE; index++) {
+ ptr = (union irte *)table->table;
+ irte = &ptr[index];
+
+ if (!irte->fields.valid)
+ continue;
+ seq_printf(m, "IRT[%04d] %08x\n", index, irte->val);
+ }
+}
+
+static void dump_irte(struct seq_file *m, u16 devid, struct amd_iommu_pci_seg *pci_seg)
+{
+ struct irq_remap_table *table;
+ unsigned long flags;
+
+ table = pci_seg->irq_lookup_table[devid];
+ if (!table) {
+ seq_printf(m, "IRQ lookup table not set for %04x:%02x:%02x:%x\n",
+ pci_seg->id, PCI_BUS_NUM(devid), PCI_SLOT(devid), PCI_FUNC(devid));
+ return;
+ }
+
+ seq_printf(m, "DeviceId %04x:%02x:%02x:%x\n", pci_seg->id, PCI_BUS_NUM(devid),
+ PCI_SLOT(devid), PCI_FUNC(devid));
+
+ raw_spin_lock_irqsave(&table->lock, flags);
+ if (AMD_IOMMU_GUEST_IR_GA(amd_iommu_guest_ir))
+ dump_128_irte(m, table);
+ else
+ dump_32_irte(m, table);
+ seq_puts(m, "\n");
+ raw_spin_unlock_irqrestore(&table->lock, flags);
+}
+
+static int iommu_irqtbl_show(struct seq_file *m, void *unused)
+{
+ struct amd_iommu_pci_seg *pci_seg;
+ u16 devid, seg;
+
+ if (!irq_remapping_enabled) {
+ seq_puts(m, "Interrupt remapping is disabled\n");
+ return 0;
+ }
+
+ if (sbdf < 0) {
+ seq_puts(m, "Please provide valid device id input\n");
+ return 0;
+ }
+
+ seg = PCI_SBDF_TO_SEGID(sbdf);
+ devid = PCI_SBDF_TO_DEVID(sbdf);
+
+ for_each_pci_segment(pci_seg) {
+ if (pci_seg->id != seg)
+ continue;
+ dump_irte(m, devid, pci_seg);
+ break;
+ }
+
+ return 0;
+}
+DEFINE_SHOW_ATTRIBUTE(iommu_irqtbl);
+
void amd_iommu_debugfs_setup(void)
{
struct amd_iommu *iommu;
@@ -272,4 +359,6 @@ void amd_iommu_debugfs_setup(void)
&devid_fops);
debugfs_create_file("devtbl", 0444, amd_iommu_debugfs, NULL,
&iommu_devtbl_fops);
+ debugfs_create_file("irqtbl", 0444, amd_iommu_debugfs, NULL,
+ &iommu_irqtbl_fops);
}
--
2.25.1
^ permalink raw reply related [flat|nested] 22+ messages in thread* Re: [PATCH v3 7/8] iommu/amd: Add debugfs support to dump IRT Table
2025-02-06 6:00 ` [PATCH v3 7/8] iommu/amd: Add debugfs support to dump IRT Table Dheeraj Kumar Srivastava
@ 2025-03-13 11:03 ` Vasant Hegde
2025-03-17 11:26 ` Dheeraj Kumar Srivastava
0 siblings, 1 reply; 22+ messages in thread
From: Vasant Hegde @ 2025-03-13 11:03 UTC (permalink / raw)
To: Dheeraj Kumar Srivastava, joro, suravee.suthikulpanit, will,
robin.murphy, linux-kernel, iommu
Hi,
On 2/6/2025 11:30 AM, Dheeraj Kumar Srivastava wrote:
> In cases where we have an issue in the device interrupt path with IOMMU
> interrupt remapping enabled, dumping valid IRT table entries for the device
> is very useful and good input for debugging the issue.
>
> eg.
> To dump irte entries for a particular device
> #echo "c4:00.0" > /sys/kernel/debug/iommu/amd/devid
> #cat /sys/kernel/debug/iommu/amd/irqtbl | less
>
> or
>
> #echo "0000:c4:00.0" > /sys/kernel/debug/iommu/amd/devid
> #cat /sys/kernel/debug/iommu/amd/irqtbl | less
>
> Signed-off-by: Dheeraj Kumar Srivastava <dheerajkumar.srivastava@amd.com>
> ---
> drivers/iommu/amd/debugfs.c | 89 +++++++++++++++++++++++++++++++++++++
> 1 file changed, 89 insertions(+)
>
> diff --git a/drivers/iommu/amd/debugfs.c b/drivers/iommu/amd/debugfs.c
> index 1377795814d1..0aff1d616108 100644
> --- a/drivers/iommu/amd/debugfs.c
> +++ b/drivers/iommu/amd/debugfs.c
> @@ -11,6 +11,7 @@
> #include <linux/pci.h>
>
> #include "amd_iommu.h"
> +#include "../irq_remapping.h"
>
> static struct dentry *amd_iommu_debugfs;
>
> @@ -250,6 +251,92 @@ static int iommu_devtbl_show(struct seq_file *m, void *unused)
> }
> DEFINE_SHOW_ATTRIBUTE(iommu_devtbl);
>
> +static void dump_128_irte(struct seq_file *m, struct irq_remap_table *table)
> +{
> + struct irte_ga *ptr, *irte;
> + int index;
> +
> + for (index = 0; index < MAX_IRQS_PER_TABLE; index++) {
> + ptr = (struct irte_ga *)table->table;
> + irte = &ptr[index];
> +
> + if (AMD_IOMMU_GUEST_IR_VAPIC(amd_iommu_guest_ir) &&
> + !irte->lo.fields_vapic.valid)
> + continue;
> + else if (!irte->lo.fields_remap.valid)
> + continue;
> + seq_printf(m, "IRT[%04d] %016llx%016llx\n", index, irte->hi.val, irte->lo.val);
Can you please space between words so that its easy to read?
-Vasant
^ permalink raw reply [flat|nested] 22+ messages in thread* Re: [PATCH v3 7/8] iommu/amd: Add debugfs support to dump IRT Table
2025-03-13 11:03 ` Vasant Hegde
@ 2025-03-17 11:26 ` Dheeraj Kumar Srivastava
0 siblings, 0 replies; 22+ messages in thread
From: Dheeraj Kumar Srivastava @ 2025-03-17 11:26 UTC (permalink / raw)
To: Vasant Hegde, joro, suravee.suthikulpanit, will, robin.murphy,
linux-kernel, iommu
Hi Vasant,
On 3/13/2025 4:33 PM, Vasant Hegde wrote:
> Hi,
>
>
> On 2/6/2025 11:30 AM, Dheeraj Kumar Srivastava wrote:
>> In cases where we have an issue in the device interrupt path with IOMMU
>> interrupt remapping enabled, dumping valid IRT table entries for the device
>> is very useful and good input for debugging the issue.
>>
>> eg.
>> To dump irte entries for a particular device
>> #echo "c4:00.0" > /sys/kernel/debug/iommu/amd/devid
>> #cat /sys/kernel/debug/iommu/amd/irqtbl | less
>>
>> or
>>
>> #echo "0000:c4:00.0" > /sys/kernel/debug/iommu/amd/devid
>> #cat /sys/kernel/debug/iommu/amd/irqtbl | less
>>
>> Signed-off-by: Dheeraj Kumar Srivastava <dheerajkumar.srivastava@amd.com>
>> ---
>> drivers/iommu/amd/debugfs.c | 89 +++++++++++++++++++++++++++++++++++++
>> 1 file changed, 89 insertions(+)
>>
>> diff --git a/drivers/iommu/amd/debugfs.c b/drivers/iommu/amd/debugfs.c
>> index 1377795814d1..0aff1d616108 100644
>> --- a/drivers/iommu/amd/debugfs.c
>> +++ b/drivers/iommu/amd/debugfs.c
>> @@ -11,6 +11,7 @@
>> #include <linux/pci.h>
>>
>> #include "amd_iommu.h"
>> +#include "../irq_remapping.h"
>>
>> static struct dentry *amd_iommu_debugfs;
>>
>> @@ -250,6 +251,92 @@ static int iommu_devtbl_show(struct seq_file *m, void *unused)
>> }
>> DEFINE_SHOW_ATTRIBUTE(iommu_devtbl);
>>
>> +static void dump_128_irte(struct seq_file *m, struct irq_remap_table *table)
>> +{
>> + struct irte_ga *ptr, *irte;
>> + int index;
>> +
>> + for (index = 0; index < MAX_IRQS_PER_TABLE; index++) {
>> + ptr = (struct irte_ga *)table->table;
>> + irte = &ptr[index];
>> +
>> + if (AMD_IOMMU_GUEST_IR_VAPIC(amd_iommu_guest_ir) &&
>> + !irte->lo.fields_vapic.valid)
>> + continue;
>> + else if (!irte->lo.fields_remap.valid)
>> + continue;
>> + seq_printf(m, "IRT[%04d] %016llx%016llx\n", index, irte->hi.val, irte->lo.val);
>
> Can you please space between words so that its easy to read?
Sure. Will update in v4.
Thanks
Dheeraj
>
> -Vasant
>
>
^ permalink raw reply [flat|nested] 22+ messages in thread
* [PATCH v3 8/8] iommu/amd: Add documentation for AMD IOMMU debugfs support
2025-02-06 5:59 [PATCH v3 0/8] Introduce debugfs support in IOMMU Dheeraj Kumar Srivastava
` (6 preceding siblings ...)
2025-02-06 6:00 ` [PATCH v3 7/8] iommu/amd: Add debugfs support to dump IRT Table Dheeraj Kumar Srivastava
@ 2025-02-06 6:00 ` Dheeraj Kumar Srivastava
2025-02-06 8:56 ` [PATCH v3 0/8] Introduce debugfs support in IOMMU Suthikulpanit, Suravee
2025-03-07 8:17 ` Dheeraj Kumar Srivastava
9 siblings, 0 replies; 22+ messages in thread
From: Dheeraj Kumar Srivastava @ 2025-02-06 6:00 UTC (permalink / raw)
To: joro, suravee.suthikulpanit, will, robin.murphy, linux-kernel,
iommu
Cc: vasant.hegde
Add documentation describing how to use AMD IOMMU debugfs support to
dump IOMMU data structures - IRT table, Device table, Registers (MMIO and
Capability) and command buffer.
Signed-off-by: Dheeraj Kumar Srivastava <dheerajkumar.srivastava@amd.com>
---
Documentation/ABI/testing/debugfs-amd-iommu | 114 ++++++++++++++++++++
1 file changed, 114 insertions(+)
create mode 100644 Documentation/ABI/testing/debugfs-amd-iommu
diff --git a/Documentation/ABI/testing/debugfs-amd-iommu b/Documentation/ABI/testing/debugfs-amd-iommu
new file mode 100644
index 000000000000..f318e3418435
--- /dev/null
+++ b/Documentation/ABI/testing/debugfs-amd-iommu
@@ -0,0 +1,114 @@
+What: /sys/kernel/debug/iommu/amd/iommu<x>/mmio
+Date: September 2024
+Contact: Dheeraj Kumar Srivastava <dheerajkumar.srivastava@amd.com>
+Description:
+ This is an input read/write access file. In this file, the user input
+ mmio register offset for iommu<x> to print corresponding mmio register
+ of iommu<x>.
+
+ Example:
+ $ echo "0x18" > /sys/kernel/debug/iommu/amd/iommu00/mmio
+ $ cat /sys/kernel/debug/iommu/amd/iommu00/mmio
+
+ Output:
+ Offset:0x18 Value:0x0003f48d
+
+What: /sys/kernel/debug/iommu/amd/iommu<x>/capability
+Date: September 2024
+Contact: Dheeraj Kumar Srivastava <dheerajkumar.srivastava@amd.com>
+Description:
+ This is an input read/write access file. In this file, the user input
+ capability register offset for iommu<x> to print corresponding capability
+ register of iommu<x>.
+
+ Example:
+ $ echo "0x10" > /sys/kernel/debug/iommu/amd/iommu00/capability
+ $ cat /sys/kernel/debug/iommu/amd/iommu00/capability
+
+ Output:
+ Offset:0x10 Value:0x00203040
+
+What: /sys/kernel/debug/iommu/amd/iommu<x>/cmdbuf
+Date: September 2024
+Contact: Dheeraj Kumar Srivastava <dheerajkumar.srivastava@amd.com>
+Description:
+ This file is an output read only file that contains iommu<x> command
+ buffer entries.
+
+ Examples:
+ $ cat /sys/kernel/debug/iommu/amd/iommu<x>/cmdbuf
+
+ Output:
+ CMD Buffer Head Offset:339 Tail Offset:339
+ 0: 008350011000000100003c0000000000
+ 1: 0000000030000005fffff0037fffffff
+ 2: 008350011000000100003c0100000000
+ 3: 0000000030000005fffff0037fffffff
+ 4: 008350011000000100003c0200000000
+ 5: 0000000030000005fffff0037fffffff
+ 6: 008350011000000100003c0300000000
+ 7: 0000000030000005fffff0037fffffff
+ 8: 008350011000000100003c0400000000
+ 9: 0000000030000005fffff0037fffffff
+ 10: 008350011000000100003c0500000000
+ 11: 0000000030000005fffff0037fffffff
+ [...]
+
+What: /sys/kernel/debug/iommu/amd/devid
+Date: September 2024
+Contact: Dheeraj Kumar Srivastava <dheerajkumar.srivastava@amd.com>
+Description:
+ This is an input read/write file that takes device id user input.
+ This input can be used for dumping iommu data structures like
+ interrupt remapping table, device table etc.
+
+ Example:
+ 1.
+ $ echo 0000:01:00.0 > /sys/kernel/debug/iommu/amd/devid
+ $ cat /sys/kernel/debug/iommu/amd/devid
+
+ Output:
+ 0000:01:00.0
+
+ 2.
+ $ echo 01:00.0 > /sys/kernel/debug/iommu/amd/devid
+ $ cat /sys/kernel/debug/iommu/amd/devid
+
+ Output:
+ 0000:01:00.0
+
+What: /sys/kernel/debug/iommu/amd/devtbl
+Date: September 2024
+Contact: Dheeraj Kumar Srivastava <dheerajkumar.srivastava@amd.com>
+Description:
+ This is an output read only file that contains device table entry for
+ the device id input given in /sys/kernel/debug/iommu/amd/devid.
+
+ Example:
+ $ cat /sys/kernel/debug/iommu/amd/devtbl
+
+ Output:
+ DeviceId QWORD[3] QWORD[2] QWORD[1] QWORD[0] iommu
+ 0000:01:00.0 0000000000000000 20000001373b8013 0000000000000038 6000000114d7b603 iommu3
+
+What: /sys/kernel/debug/iommu/amd/irqtbl
+Date: September 2024
+Contact: Dheeraj Kumar Srivastava <dheerajkumar.srivastava@amd.com>
+Description:
+ This is an output read only file that contains IRT table valid entries
+ for the device id input given in /sys/kernel/debug/iommu/amd/devid.
+
+ Example:
+ $ cat /sys/kernel/debug/iommu/amd/irqtbl
+
+ Output:
+ DeviceId 0000:01:00.0
+ IRT[0000] 00000000000000200000000000000241
+ IRT[0001] 00000000000000200000000000000841
+ IRT[0002] 00000000000000200000000000002041
+ IRT[0003] 00000000000000200000000000008041
+ IRT[0004] 00000000000000200000000000020041
+ IRT[0005] 00000000000000200000000000080041
+ IRT[0006] 00000000000000200000000000200041
+ IRT[0007] 00000000000000200000000000800041
+ [...]
--
2.25.1
^ permalink raw reply related [flat|nested] 22+ messages in thread* Re: [PATCH v3 0/8] Introduce debugfs support in IOMMU
2025-02-06 5:59 [PATCH v3 0/8] Introduce debugfs support in IOMMU Dheeraj Kumar Srivastava
` (7 preceding siblings ...)
2025-02-06 6:00 ` [PATCH v3 8/8] iommu/amd: Add documentation for AMD IOMMU debugfs support Dheeraj Kumar Srivastava
@ 2025-02-06 8:56 ` Suthikulpanit, Suravee
2025-03-07 8:17 ` Dheeraj Kumar Srivastava
9 siblings, 0 replies; 22+ messages in thread
From: Suthikulpanit, Suravee @ 2025-02-06 8:56 UTC (permalink / raw)
To: Dheeraj Kumar Srivastava, joro, will, robin.murphy, linux-kernel,
iommu
Cc: vasant.hegde
On 2/6/2025 12:59 PM, Dheeraj Kumar Srivastava wrote:
> From: root <root@BLRDHSRIVAS.amd.com>
>
> Introducing debugfs support in AMD/IOMMU driver that will allow
> userspace to dump below IOMMU information
> 1) MMIO and Capability register per IOMMU
> 2) Command buffer
> 3) Device table entry
> 4) Interrupt remapping table entry
>
> Analyzing contents of IOMMU data structures helps in understanding IOMMU
> capabilities and behavior and debug issues faster.
>
> 1. MMIO and Capability registers - Add support to dump MMIO and Capability
> registers per IOMMU.
>
> Example:
> a. Write MMIO register offset to dump it
> $ echo 0x18 > /sys/kernel/debug/iommu/amd/iommu00/mmio
> $ cat /sys/kernel/debug/iommu/amd/iommu00/mmio
>
> Output:
> $ 0x0003f48d
Correction: Based on the documentation in patch 8/8, the output would be
Offset:0x18 Value:0x0003f48d
> b. Write capability register offset to dump it
> $ echo 0x10 > /sys/kernel/debug/iommu/amd/iommu00/capability
> $ cat /sys/kernel/debug/iommu/amd/iommu00/capability
>
> Output:
> $ 0x00203040
Correction: Based on the documentation in patch 8/8, the output would be
Offset:0x10 Value:0x00203040
Suravee
^ permalink raw reply [flat|nested] 22+ messages in thread* Re: [PATCH v3 0/8] Introduce debugfs support in IOMMU
2025-02-06 5:59 [PATCH v3 0/8] Introduce debugfs support in IOMMU Dheeraj Kumar Srivastava
` (8 preceding siblings ...)
2025-02-06 8:56 ` [PATCH v3 0/8] Introduce debugfs support in IOMMU Suthikulpanit, Suravee
@ 2025-03-07 8:17 ` Dheeraj Kumar Srivastava
9 siblings, 0 replies; 22+ messages in thread
From: Dheeraj Kumar Srivastava @ 2025-03-07 8:17 UTC (permalink / raw)
To: joro, suravee.suthikulpanit, will, robin.murphy, linux-kernel,
iommu
Cc: vasant.hegde
Hi,
I hope you're doing well. Just a friendly reminder to kindly review the
patches at your earliest convenience. Your feedback is greatly appreciated.
Thanks
Dheeraj
On 2/6/2025 11:29 AM, Dheeraj Kumar Srivastava wrote:
> From: root <root@BLRDHSRIVAS.amd.com>
>
> Introducing debugfs support in AMD/IOMMU driver that will allow
> userspace to dump below IOMMU information
> 1) MMIO and Capability register per IOMMU
> 2) Command buffer
> 3) Device table entry
> 4) Interrupt remapping table entry
>
> Analyzing contents of IOMMU data structures helps in understanding IOMMU
> capabilities and behavior and debug issues faster.
>
> 1. MMIO and Capability registers - Add support to dump MMIO and Capability
> registers per IOMMU.
>
> Example:
> a. Write MMIO register offset to dump it
> $ echo 0x18 > /sys/kernel/debug/iommu/amd/iommu00/mmio
> $ cat /sys/kernel/debug/iommu/amd/iommu00/mmio
>
> Output:
> $ 0x0003f48d
>
> b. Write capability register offset to dump it
> $ echo 0x10 > /sys/kernel/debug/iommu/amd/iommu00/capability
> $ cat /sys/kernel/debug/iommu/amd/iommu00/capability
>
> Output:
> $ 0x00203040
>
> 2. Command buffer - Add support to dump per IOMMU command buffer.
>
> Example:
> a. cat /sys/kernel/debug/iommu/amd/iommu00/cmdbuf
>
> Output:
> CMD Buffer Head Offset:339 Tail Offset:339
> 0: 00835001100000010000990000000000
> 1: 0000000030000005fffff0037fffffff
> 2: 00835001100000010000990100000000
> 3: 0000000030000005fffff0037fffffff
> 4: 00835001100000010000990200000000
> ...................................
> ...................................
> ...................................
>
> 3. Device table - Add support to dump device table per IOMMU.
>
> Example:
> a. Write device id to dump device table entry for that device
> $ echo 0000:01:00.0 > /sys/kernel/debug/iommu/amd/devid
> $ cat /sys/kernel/debug/iommu/amd/devid
>
> Output:
> 0000:01:00.0
>
> Dump the device table entry for the input given
> $ cat /sys/kernel/debug/iommu/amd/devtbl
>
> Output:
> DeviceId QWORD[3] QWORD[2] QWORD[1] QWORD[0] iommu
> 0000:01:00.0 0000000000000000 20000001373b8013 0000000000000038 6000000114d7b603 iommu3
>
> b. Write device id to dump device table entry for that device
> $ echo 01:00.0 > /sys/kernel/debug/iommu/amd/devid
> $ cat /sys/kernel/debug/iommu/amd/devid
>
> Output:
> 0000:01:00.0
>
> Dump the device table entry for the input given
> $ cat /sys/kernel/debug/iommu/amd/devtbl
>
> Output:
> DeviceId QWORD[3] QWORD[2] QWORD[1] QWORD[0] iommu
> 0000:01:00.0 0000000000000000 20000001373b8013 0000000000000038 6000000114d7b603 iommu3
>
> 4. Interrupt remapping table - Add support to dump IRT table valid entries in
> "iommu_irqtbl" file. This supports user input to dump IRT entry for a
> specific pci device.
>
> Example:
> a. Write device id to dump device table entry for that device
> $ echo 0000:01:00.0 > /sys/kernel/debug/iommu/amd/devid
> $ cat /sys/kernel/debug/iommu/amd/devid
>
> Output:
> 0000:01:00.0
>
> Dump the device table entry for the input given
> $ cat /sys/kernel/debug/iommu/amd/irtbl
>
> Output:
> DeviceId 0000:01:00.0
> IRT[0000] 00000000000000200000000000000241
> IRT[0001] 00000000000000200000000000000841
> IRT[0002] 00000000000000200000000000002041
> IRT[0003] 00000000000000200000000000008041
> IRT[0004] 00000000000000200000000000020041
> ..........................................
> ..........................................
> ..........................................
>
> b. Write device id to dump device table entry for that device
> $ echo 01:00.0 > /sys/kernel/debug/iommu/amd/devid
> $ cat /sys/kernel/debug/iommu/amd/devid
>
> Output:
> 0000:01:00.0
>
> Dump the device table entry for the input given
> $ cat /sys/kernel/debug/iommu/amd/irttbl
>
> Output:
> Device 0000:01:00.0
> IRT[0000] 00000000000000200000000000000241
> IRT[0001] 00000000000000200000000000000841
> IRT[0002] 00000000000000200000000000002041
> IRT[0003] 00000000000000200000000000008041
> IRT[0004] 00000000000000200000000000020041
> ..........................................
> ..........................................
> ..........................................
>
> Open question:[Suggestion from Bjorn Helgaas] Instead of using device-id input with
> "/sys/kernel/debug/iommu/amd/devid", can we have something like other drivers
> have "/sys/kernel/debug/iommu/amd/<bdf>/".
>
> Each "/sys/kernel/debug/iommu/amd/<bdf>/" will have "devtbl" and "irqtbl" files
> for device table and interrupt remapping table dump file.
>
>
> Changes since v2:
> -> Patch 2/8 and 3/8: Use single file mmio and capability for input and output registers
> offset and registers dump respctively.
> -> Patch 8/8: Correct path for files "irqtbl" and "devtbl" in Documentation
>
> Changes since v1:
> -> Patch 2/8 and 3/8: Use kstrtou32_from_user() instead of memdup_user_nul() --> kstrtou32()
> -> Patch 4/8: Dump command buffer head and tail offset instead of head and tail pointer registers.
> -> Patch 8/8: Fix bot reported warning on v1 upstream.
>
> Dheeraj Kumar Srivastava (8):
> iommu/amd: Refactor AMD IOMMU debugfs initial setup
> iommu/amd: Add debugfs support to dump IOMMU MMIO registers
> iommu/amd: Add debugfs support to dump IOMMU Capability registers
> iommu/amd: Add debugfs support to dump IOMMU command buffer
> iommu/amd: Add support for device id user input
> iommu/amd: Add debugfs support to dump device table
> iommu/amd: Add debugfs support to dump IRT Table
> iommu/amd: Add documentation for AMD IOMMU debugfs support
>
> Documentation/ABI/testing/debugfs-amd-iommu | 114 +++++++
> drivers/iommu/amd/amd_iommu.h | 4 +-
> drivers/iommu/amd/amd_iommu_types.h | 7 +
> drivers/iommu/amd/debugfs.c | 350 +++++++++++++++++++-
> drivers/iommu/amd/init.c | 5 +-
> drivers/iommu/amd/iommu.c | 7 -
> 6 files changed, 466 insertions(+), 21 deletions(-)
> create mode 100644 Documentation/ABI/testing/debugfs-amd-iommu
>
^ permalink raw reply [flat|nested] 22+ messages in thread