Skip to contentSkip to conversation

hw/nvme: bug fixes and doc update

PATCH In review Ankit Kumar 3 participants 9 messages Aug 8, 2023 1 message not in the archive

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): CGME20230807160836epcas5p34398954fecd7388469012404b09b78f9@epcas5p3.samsung.com

#1 lore raw

cover hw/nvme: bug fixes and doc update

This series fixes two bugs
1. CRC64 generation when metadata buffer is used.
2. Protection information disable check for Type 3 protection.

This series also updates the documentaion for pi (protection information), and adds missing pif (protection information format) entry.

Ankit Kumar (3):
hw/nvme: fix CRC64 for guard tag
hw/nvme: fix disable pi checks for Type 3 protection
docs: update hw/nvme documentation for protection information

 docs/system/devices/nvme.rst | 10 +++++++---
 hw/nvme/dif.c                |  9 +++++----
 2 files changed, 12 insertions(+), 7 deletions(-)
Signature · 1 line
2.25.1
#2 lore raw

PATCH 1/3 hw/nvme: fix CRC64 for guard tag

The nvme CRC64 generator expects the caller to pass inverted seed value. Pass inverted crc value for metadata buffer.

---
 hw/nvme/dif.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
hw/nvme/dif.c +2 -2
index 63c44c86ab..01b19c3373 100644
@@ -115,7 +115,7 @@ static void nvme_dif_pract_generate_dif_crc64(NvmeNamespace *ns, uint8_t *buf,
         uint64_t crc = crc64_nvme(~0ULL, buf, ns->lbasz);
 
         if (pil) {
-            crc = crc64_nvme(crc, mbuf, pil);
+            crc = crc64_nvme(~crc, mbuf, pil);
         }
 
         dif->g64.guard = cpu_to_be64(crc);
@@ -246,7 +246,7 @@ static uint16_t nvme_dif_prchk_crc64(NvmeNamespace *ns, NvmeDifTuple *dif,
         uint64_t crc = crc64_nvme(~0ULL, buf, ns->lbasz);
 
         if (pil) {
-            crc = crc64_nvme(crc, mbuf, pil);
+            crc = crc64_nvme(~crc, mbuf, pil);
         }
 
         trace_pci_nvme_dif_prchk_guard_crc64(be64_to_cpu(dif->g64.guard), crc);
Signature · 1 line
2.25.1
Reviewed-by on v1 1/3
#3 lore raw

Re: [PATCH 1/3] hw/nvme: fix CRC64 for guard tag

Quoted 33 lines · Ankit Kumar

The nvme CRC64 generator expects the caller to pass inverted seed value. Pass inverted crc value for metadata buffer.

---
 hw/nvme/dif.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
hw/nvme/dif.c +2 -2
index 63c44c86ab..01b19c3373 100644
@@ -115,7 +115,7 @@ static void nvme_dif_pract_generate_dif_crc64(NvmeNamespace *ns, uint8_t *buf,
         uint64_t crc = crc64_nvme(~0ULL, buf, ns->lbasz);
 
         if (pil) {
-            crc = crc64_nvme(crc, mbuf, pil);
+            crc = crc64_nvme(~crc, mbuf, pil);
         }
 
         dif->g64.guard = cpu_to_be64(crc);
@@ -246,7 +246,7 @@ static uint16_t nvme_dif_prchk_crc64(NvmeNamespace *ns, NvmeDifTuple *dif,
         uint64_t crc = crc64_nvme(~0ULL, buf, ns->lbasz);
 
         if (pil) {
-            crc = crc64_nvme(crc, mbuf, pil);
+            crc = crc64_nvme(~crc, mbuf, pil);
         }
 
         trace_pci_nvme_dif_prchk_guard_crc64(be64_to_cpu(dif->g64.guard), crc);
Signature · 1 line
2.25.1

Good catch, thanks!

#4 lore raw

PATCH 2/3 hw/nvme: fix disable pi checks for Type 3 protection

In reply to Ankit Kumar

As per the NVM command set specification, the protection information checks for Type 3 protection are disabled, only when both application and reference tag have all bits set to 1.

---
 hw/nvme/dif.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
hw/nvme/dif.c +3 -2
index 01b19c3373..f9bd29a2a6 100644
@@ -157,7 +157,8 @@ static uint16_t nvme_dif_prchk_crc16(NvmeNamespace *ns, NvmeDifTuple *dif,
 {
     switch (NVME_ID_NS_DPS_TYPE(ns->id_ns.dps)) {
     case NVME_ID_NS_DPS_TYPE_3:
-        if (be32_to_cpu(dif->g16.reftag) != 0xffffffff) {
+        if ((be32_to_cpu(dif->g16.reftag) != 0xffffffff) ||
+            (be16_to_cpu(dif->g16.apptag) != 0xffff)) {
             break;
         }
 
@@ -225,7 +226,7 @@ static uint16_t nvme_dif_prchk_crc64(NvmeNamespace *ns, NvmeDifTuple *dif,
 
     switch (NVME_ID_NS_DPS_TYPE(ns->id_ns.dps)) {
     case NVME_ID_NS_DPS_TYPE_3:
-        if (r != 0xffffffffffff) {
+        if (r != 0xffffffffffff || (be16_to_cpu(dif->g64.apptag) != 0xffff)) {
             break;
         }
 
Signature · 1 line
2.25.1
#5 lore raw

Re: [PATCH 2/3] hw/nvme: fix disable pi checks for Type 3 protection

Quoted 14 lines · Ankit Kumar

As per the NVM command set specification, the protection information checks for Type 3 protection are disabled, only when both application and reference tag have all bits set to 1.

---
 hw/nvme/dif.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
hw/nvme/dif.c +0 -0
index 01b19c3373..f9bd29a2a6 100644

hw/nvme/dif.c static uint16_t nvme_dif_prchk_crc16(NvmeNamespace *ns, NvmeDifTuple *dif, · line 163

Quoted 8 lines
hw/nvme/dif.c +2 -1
@@ -157,7 +157,8 @@ static uint16_t nvme_dif_prchk_crc16(NvmeNamespace *ns, NvmeDifTuple *dif,
 {
     switch (NVME_ID_NS_DPS_TYPE(ns->id_ns.dps)) {
     case NVME_ID_NS_DPS_TYPE_3:
-        if (be32_to_cpu(dif->g16.reftag) != 0xffffffff) {
+        if ((be32_to_cpu(dif->g16.reftag) != 0xffffffff) ||
+            (be16_to_cpu(dif->g16.apptag) != 0xffff)) {
             break;
         }

For type 3, if reftag is 0xffffffff the NVME_ID_NS_DPS_TYPE_3 case will fallthrough to the next cases (_TYPE_1 and _TYPE_2), checking if apptag is 0xffff, and disable checking if so.

hw/nvme/dif.c static uint16_t nvme_dif_prchk_crc64(NvmeNamespace *ns, NvmeDifTuple *dif, · line 231

Quoted 9 lines
hw/nvme/dif.c +1 -1
@@ -225,7 +226,7 @@ static uint16_t nvme_dif_prchk_crc64(NvmeNamespace *ns, NvmeDifTuple *dif,
 
     switch (NVME_ID_NS_DPS_TYPE(ns->id_ns.dps)) {
     case NVME_ID_NS_DPS_TYPE_3:
-        if (r != 0xffffffffffff) {
+        if (r != 0xffffffffffff || (be16_to_cpu(dif->g64.apptag) != 0xffff)) {
             break;
         }

Same here.

#6 lore raw

PATCH 3/3 docs: update hw/nvme documentation for protection information

In reply to Ankit Kumar

Add missing entry for pif ("protection information format"). Protection information size can be 8 or 16 bytes, Update the pil entry as per the NVM command set specification.

---
 docs/system/devices/nvme.rst | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)
docs/system/devices/nvme.rst +7 -3
index 2a3af268f7..30d46d9338 100644
@@ -271,9 +271,13 @@ The virtual namespace device supports DIF- and DIX-based protection information
 
 ``pil=UINT8`` (default: ``0``)
   Controls the location of the protection information within the metadata. Set
-  to ``1`` to transfer protection information as the first eight bytes of
-  metadata. Otherwise, the protection information is transferred as the last
-  eight bytes.
+  to ``1`` to transfer protection information as the first bytes of metadata.
+  Otherwise, the protection information is transferred as the last bytes of
+  metadata.
+
+``pif=UINT8`` (default: ``0``)
+  By default, the namespace device uses 16 bit guard protection information
+  format. Set to ``2`` to enable 64 bit guard protection information format.
 
 Virtualization Enhancements and SR-IOV (Experimental Support)
 -------------------------------------------------------------
Signature · 1 line
2.25.1
Reviewed-by on v1 3/3
#7 lore raw

Re: [PATCH 3/3] docs: update hw/nvme documentation for protection information

Quoted 18 lines · Ankit Kumar

Add missing entry for pif ("protection information format"). Protection information size can be 8 or 16 bytes, Update the pil entry as per the NVM command set specification.

---
 docs/system/devices/nvme.rst | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)
docs/system/devices/nvme.rst +0 -1
index 2a3af268f7..30d46d9338 100644
@@ -271,9 +271,13 @@ The virtual namespace device supports DIF- and DIX-based protection information
 
 ``pil=UINT8`` (default: ``0``)
   Controls the location of the protection information within the metadata. Set
-  to ``1`` to transfer protection information as the first eight bytes of

docs/system/devices/nvme.rst The virtual namespace device supports DIF- and DIX-based protection information · line 281

Quoted 10 lines
docs/system/devices/nvme.rst +7 -2
-  metadata. Otherwise, the protection information is transferred as the last
-  eight bytes.
+  to ``1`` to transfer protection information as the first bytes of metadata.
+  Otherwise, the protection information is transferred as the last bytes of
+  metadata.
+
+``pif=UINT8`` (default: ``0``)
+  By default, the namespace device uses 16 bit guard protection information
+  format. Set to ``2`` to enable 64 bit guard protection information format.
 

I'll add a small note that pif=1 (32b guard) is not supported.

Thanks,

#8 lore raw
In reply to Ankit Kumar
Quoted 11 lines · Ankit Kumar

This series fixes two bugs
1. CRC64 generation when metadata buffer is used.
2. Protection information disable check for Type 3 protection.

This series also updates the documentaion for pi (protection information), and adds missing pif (protection information format) entry.

Ankit Kumar (3):
hw/nvme: fix CRC64 for guard tag
hw/nvme: fix disable pi checks for Type 3 protection
docs: update hw/nvme documentation for protection information

At least the CRC64 change smells like a -stable material, - the bug is present in, for example, qemu-7.2 too. But I don't know how important it is to keep nvme updated in 8.0 or before, and what the outcome of this bug is, to begin with. Somehow I think nvme was in preliminary shape before 8.0.

Are the other changes also relevant for -stable?

Please keep Cc: [email protected] for anything you think is worth to have in previous/stable releases.

Thanks,

/mjt

#9 lore raw
Quoted 19 lines · Michael Tokarev
Quoted 11 lines · Ankit Kumar

This series fixes two bugs
1. CRC64 generation when metadata buffer is used.
2. Protection information disable check for Type 3 protection.

This series also updates the documentaion for pi (protection information), and adds missing pif (protection information format) entry.

Ankit Kumar (3):
hw/nvme: fix CRC64 for guard tag
hw/nvme: fix disable pi checks for Type 3 protection
docs: update hw/nvme documentation for protection information

At least the CRC64 change smells like a -stable material, - the bug is present in, for example, qemu-7.2 too. But I don't know how important it is to keep nvme updated in 8.0 or before, and what the outcome of this bug is, to begin with. Somehow I think nvme was in preliminary shape before 8.0.

Quoted 5 lines

Are the other changes also relevant for -stable?

Please keep Cc: [email protected] for anything you think is worth to have in previous/stable releases.

Hi Michael,

Yes, this is stable worthy.

I'll add Cc: stable and fixes tags on relevant patches on the pull.

Thanks for the heads up!