[dpdk-dev] [PATCH v3] bus/pci: forbid VA as IOVA mode if IOMMU address width too small

[dpdk-dev] [PATCH v3] bus/pci: forbid VA as IOVA mode if IOMMU address width too small

[Maxime Coquelin]

Intel VT-d supports different address widths for the IOVAs, from
39 bits to 56 bits.

While recent processors support at least 48 bits, VT-d emulation
currently only supports 39 bits. It makes DMA mapping to fail in this
case when using VA as IOVA mode, as user-space virtual addresses uses
up to 47 bits (see kernel's Documentation/x86/x86_64/mm.txt).

This patch parses VT-d CAP register value available in sysfs, and
forbid VA as IOVA mode if the GAW is 39 bits or unknown.

Fixes: f37dfab21c98 ("drivers/net: enable IOVA mode for Intel PMDs")

Cc: stable@dpdk.org
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
---

Changes in v3:
==============
- Rely on MGAW bitfiled instead of SAGAW (Qi)

Changes in v2:
==============
- Rework pci_one_device_iommu_support_va #ifdefery (Stephen)
- Don't inline introduced functions (Stephen)

 drivers/bus/pci/linux/pci.c | 90 ++++++++++++++++++++++++++++++++++++++++-----
 1 file changed, 81 insertions(+), 9 deletions(-)

diff --git a/drivers/bus/pci/linux/pci.c b/drivers/bus/pci/linux/pci.c
index 25f907e04..a89e8353d 100644
--- a/drivers/bus/pci/linux/pci.c
+++ b/drivers/bus/pci/linux/pci.c
@@ -547,6 +547,82 @@ pci_one_device_has_iova_va(void)
 	return 0;
 }
 
+#if defined(RTE_ARCH_X86)
+static bool
+pci_one_device_iommu_support_va(struct rte_pci_device *dev)
+{
+#define VTD_CAP_MGAW_SHIFT	16
+#define VTD_CAP_MGAW_MASK	(0x3fULL << VTD_CAP_MGAW_SHIFT)
+#define X86_VA_WIDTH 47 /* From Documentation/x86/x86_64/mm.txt */
+	struct rte_pci_addr *addr = &dev->addr;
+	char filename[PATH_MAX];
+	FILE *fp;
+	uint64_t mgaw, vtd_cap_reg = 0;
+
+	snprintf(filename, sizeof(filename),
+		 "%s/" PCI_PRI_FMT "/iommu/intel-iommu/cap",
+		 rte_pci_get_sysfs_path(), addr->domain, addr->bus, addr->devid,
+		 addr->function);
+	if (access(filename, F_OK) == -1) {
+		/* We don't have an Intel IOMMU, assume VA supported*/
+		return true;
+	}
+
+	/* We have an intel IOMMU */
+	fp = fopen(filename, "r");
+	if (fp == NULL) {
+		RTE_LOG(ERR, EAL, "%s(): can't open %s\n", __func__, filename);
+		return false;
+	}
+
+	if (fscanf(fp, "%lx", &vtd_cap_reg) != 1) {
+		RTE_LOG(ERR, EAL, "%s(): can't read %s\n", __func__, filename);
+		fclose(fp);
+		return false;
+	}
+
+	fclose(fp);
+
+	mgaw = ((vtd_cap_reg & VTD_CAP_MGAW_MASK) >> VTD_CAP_MGAW_SHIFT) + 1;
+	if (mgaw < X86_VA_WIDTH)
+		return false;
+
+	return true;
+}
+#elif defined(RTE_ARCH_PPC_64)
+static bool
+pci_one_device_iommu_support_va(struct rte_pci_device *dev)
+{
+	return false;
+}
+#else
+static bool
+pci_one_device_iommu_support_va(struct rte_pci_device *dev)
+{
+	return true;
+}
+#endif
+
+/*
+ * All devices IOMMUs support VA as IOVA
+ */
+static bool
+pci_devices_iommu_support_va(void)
+{
+	struct rte_pci_device *dev = NULL;
+	struct rte_pci_driver *drv = NULL;
+
+	FOREACH_DRIVER_ON_PCIBUS(drv) {
+		FOREACH_DEVICE_ON_PCIBUS(dev) {
+			if (!rte_pci_match(drv, dev))
+				continue;
+			if (!pci_one_device_iommu_support_va(dev))
+				return false;
+		}
+	}
+	return true;
+}
+
 /*
  * Get iommu class of PCI devices on the bus.
  */
@@ -557,12 +633,7 @@ rte_pci_get_iommu_class(void)
 	bool is_vfio_noiommu_enabled = true;
 	bool has_iova_va;
 	bool is_bound_uio;
-	bool spapr_iommu =
-#if defined(RTE_ARCH_PPC_64)
-		true;
-#else
-		false;
-#endif
+	bool iommu_no_va;
 
 	is_bound = pci_one_device_is_bound();
 	if (!is_bound)
@@ -570,13 +641,14 @@ rte_pci_get_iommu_class(void)
 
 	has_iova_va = pci_one_device_has_iova_va();
 	is_bound_uio = pci_one_device_bound_uio();
+	iommu_no_va = !pci_devices_iommu_support_va();
 #ifdef VFIO_PRESENT
 	is_vfio_noiommu_enabled = rte_vfio_noiommu_is_enabled() == true ?
 					true : false;
 #endif
 
 	if (has_iova_va && !is_bound_uio && !is_vfio_noiommu_enabled &&
-			!spapr_iommu)
+			!iommu_no_va)
 		return RTE_IOVA_VA;
 
 	if (has_iova_va) {
@@ -585,8 +657,8 @@ rte_pci_get_iommu_class(void)
 			RTE_LOG(WARNING, EAL, "vfio-noiommu mode configured\n");
 		if (is_bound_uio)
 			RTE_LOG(WARNING, EAL, "few device bound to UIO\n");
-		if (spapr_iommu)
-			RTE_LOG(WARNING, EAL, "sPAPR IOMMU does not support IOVA as VA\n");
+		if (iommu_no_va)
+			RTE_LOG(WARNING, EAL, "IOMMU does not support IOVA as VA\n");
 	}
 
 	return RTE_IOVA_PA;
-- 
2.14.3

Re: [dpdk-dev] [PATCH v3] bus/pci: forbid VA as IOVA mode if IOMMU address width too small

[Burakov, Anatoly]

<snip>

> +#if defined(RTE_ARCH_X86)
> +static bool
> +pci_one_device_iommu_support_va(struct rte_pci_device *dev)
> +{
> +#define VTD_CAP_MGAW_SHIFT	16
> +#define VTD_CAP_MGAW_MASK	(0x3fULL << VTD_CAP_MGAW_SHIFT)
> +#define X86_VA_WIDTH 47 /* From Documentation/x86/x86_64/mm.txt */
> +	struct rte_pci_addr *addr = &dev->addr;
> +	char filename[PATH_MAX];
> +	FILE *fp;
> +	uint64_t mgaw, vtd_cap_reg = 0;
> +
> +	snprintf(filename, sizeof(filename),
> +		 "%s/" PCI_PRI_FMT "/iommu/intel-iommu/cap",
> +		 rte_pci_get_sysfs_path(), addr->domain, addr->bus, addr->devid,
> +		 addr->function);
> +	if (access(filename, F_OK) == -1) {
> +		/* We don't have an Intel IOMMU, assume VA supported*/
> +		return true;
> +	}
> +
> +	/* We have an intel IOMMU */
> +	fp = fopen(filename, "r");
> +	if (fp == NULL) {
> +		RTE_LOG(ERR, EAL, "%s(): can't open %s\n", __func__, filename);
> +		return false;
> +	}
> +
> +	if (fscanf(fp, "%lx", &vtd_cap_reg) != 1) {
> +		RTE_LOG(ERR, EAL, "%s(): can't read %s\n", __func__, filename);
> +		fclose(fp);
> +		return false;
> +	}
> +
> +	fclose(fp);

Hi Maxime,

You probably want to use eal_parse_sysfs_value() for this.


-- 
Thanks,
Anatoly

Re: [dpdk-dev] [PATCH v3] bus/pci: forbid VA as IOVA mode if IOMMU address width too small

[Maxime Coquelin]

Hi Anatoly,

On 01/12/2018 12:10 PM, Burakov, Anatoly wrote:
> <snip>
> 
>> +#if defined(RTE_ARCH_X86)
>> +static bool
>> +pci_one_device_iommu_support_va(struct rte_pci_device *dev)
>> +{
>> +#define VTD_CAP_MGAW_SHIFT    16
>> +#define VTD_CAP_MGAW_MASK    (0x3fULL << VTD_CAP_MGAW_SHIFT)
>> +#define X86_VA_WIDTH 47 /* From Documentation/x86/x86_64/mm.txt */
>> +    struct rte_pci_addr *addr = &dev->addr;
>> +    char filename[PATH_MAX];
>> +    FILE *fp;
>> +    uint64_t mgaw, vtd_cap_reg = 0;
>> +
>> +    snprintf(filename, sizeof(filename),
>> +         "%s/" PCI_PRI_FMT "/iommu/intel-iommu/cap",
>> +         rte_pci_get_sysfs_path(), addr->domain, addr->bus, addr->devid,
>> +         addr->function);
>> +    if (access(filename, F_OK) == -1) {
>> +        /* We don't have an Intel IOMMU, assume VA supported*/
>> +        return true;
>> +    }
>> +
>> +    /* We have an intel IOMMU */
>> +    fp = fopen(filename, "r");
>> +    if (fp == NULL) {
>> +        RTE_LOG(ERR, EAL, "%s(): can't open %s\n", __func__, filename);
>> +        return false;
>> +    }
>> +
>> +    if (fscanf(fp, "%lx", &vtd_cap_reg) != 1) {
>> +        RTE_LOG(ERR, EAL, "%s(): can't read %s\n", __func__, filename);
>> +        fclose(fp);
>> +        return false;
>> +    }
>> +
>> +    fclose(fp);
> 
> Hi Maxime,
> 
> You probably want to use eal_parse_sysfs_value() for this.

I initially planned to use it, but the sysfs value is hexadecimal, but
not prefixed with "0x". For example:

# cat /sys/devices/pci0000\:00/0000\:00\:02.0/iommu/intel-iommu/cap
1c0000c40660462

So strtoul() assumes the value is decimal in this case, as explained in
its man page:

"
DESCRIPTION
<snip\>
If base is zero or 16, the string may then include a "0x" prefix, and
the number will be read in base 16; otherwise, a zero base is taken as
10 (decimal)
"

Thanks,
Maxime

Re: [dpdk-dev] [PATCH v3] bus/pci: forbid VA as IOVA mode if IOMMU address width too small

[Chas Williams]

Tested-by: Chas Williams <chas@att.com>

On Fri, Jan 12, 2018 at 8:18 AM, Maxime Coquelin <maxime.coquelin@redhat.com
> wrote:

> Hi Anatoly,
>
>
> On 01/12/2018 12:10 PM, Burakov, Anatoly wrote:
>
>> <snip>
>>
>> +#if defined(RTE_ARCH_X86)
>>> +static bool
>>> +pci_one_device_iommu_support_va(struct rte_pci_device *dev)
>>> +{
>>> +#define VTD_CAP_MGAW_SHIFT    16
>>> +#define VTD_CAP_MGAW_MASK    (0x3fULL << VTD_CAP_MGAW_SHIFT)
>>> +#define X86_VA_WIDTH 47 /* From Documentation/x86/x86_64/mm.txt */
>>> +    struct rte_pci_addr *addr = &dev->addr;
>>> +    char filename[PATH_MAX];
>>> +    FILE *fp;
>>> +    uint64_t mgaw, vtd_cap_reg = 0;
>>> +
>>> +    snprintf(filename, sizeof(filename),
>>> +         "%s/" PCI_PRI_FMT "/iommu/intel-iommu/cap",
>>> +         rte_pci_get_sysfs_path(), addr->domain, addr->bus, addr->devid,
>>> +         addr->function);
>>> +    if (access(filename, F_OK) == -1) {
>>> +        /* We don't have an Intel IOMMU, assume VA supported*/
>>> +        return true;
>>> +    }
>>> +
>>> +    /* We have an intel IOMMU */
>>> +    fp = fopen(filename, "r");
>>> +    if (fp == NULL) {
>>> +        RTE_LOG(ERR, EAL, "%s(): can't open %s\n", __func__, filename);
>>> +        return false;
>>> +    }
>>> +
>>> +    if (fscanf(fp, "%lx", &vtd_cap_reg) != 1) {
>>> +        RTE_LOG(ERR, EAL, "%s(): can't read %s\n", __func__, filename);
>>> +        fclose(fp);
>>> +        return false;
>>> +    }
>>> +
>>> +    fclose(fp);
>>>
>>
>> Hi Maxime,
>>
>> You probably want to use eal_parse_sysfs_value() for this.
>>
>
> I initially planned to use it, but the sysfs value is hexadecimal, but
> not prefixed with "0x". For example:
>
> # cat /sys/devices/pci0000\:00/0000\:00\:02.0/iommu/intel-iommu/cap
> 1c0000c40660462
>
> So strtoul() assumes the value is decimal in this case, as explained in
> its man page:
>
> "
> DESCRIPTION
> <snip\>
> If base is zero or 16, the string may then include a "0x" prefix, and
> the number will be read in base 16; otherwise, a zero base is taken as
> 10 (decimal)
> "
>
> Thanks,
> Maxime
>

Re: [dpdk-dev] [dpdk-stable] [PATCH v3] bus/pci: forbid VA as IOVA mode if IOMMU address width too small

[Thomas Monjalon]

12/01/2018 11:22, Maxime Coquelin:
> Intel VT-d supports different address widths for the IOVAs, from
> 39 bits to 56 bits.
> 
> While recent processors support at least 48 bits, VT-d emulation
> currently only supports 39 bits. It makes DMA mapping to fail in this
> case when using VA as IOVA mode, as user-space virtual addresses uses
> up to 47 bits (see kernel's Documentation/x86/x86_64/mm.txt).
> 
> This patch parses VT-d CAP register value available in sysfs, and
> forbid VA as IOVA mode if the GAW is 39 bits or unknown.
> 
> Fixes: f37dfab21c98 ("drivers/net: enable IOVA mode for Intel PMDs")
> 
> Cc: stable@dpdk.org
> Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
[...]
> +	if (fscanf(fp, "%lx", &vtd_cap_reg) != 1) {

Compilation error on 32-bit. Fix:

-       if (fscanf(fp, "%lx", &vtd_cap_reg) != 1) {
+       if (fscanf(fp, "%" PRIx64, &vtd_cap_reg) != 1) {

[...]
> +#elif defined(RTE_ARCH_PPC_64)
> +static bool
> +pci_one_device_iommu_support_va(struct rte_pci_device *dev)
> +{
> +       return false;
> +}
> +#else
> +static bool
> +pci_one_device_iommu_support_va(struct rte_pci_device *dev)
> +{
> +	return true;
> +}
> +#endif

Compilation error on non-x86. Fix:

 #elif defined(RTE_ARCH_PPC_64)
 static bool
-pci_one_device_iommu_support_va(struct rte_pci_device *dev)
+pci_one_device_iommu_support_va(__rte_unused struct rte_pci_device *dev)
 {
        return false;
 }
 #else
 static bool
-pci_one_device_iommu_support_va(struct rte_pci_device *dev)
+pci_one_device_iommu_support_va(__rte_unused struct rte_pci_device *dev)
 {
        return true;
 }

Applied with above fixes, thanks.