Skip to contentSkip to conversation

scsi: lpfc: use unsigned type for num_sge

PATCH Applied Daniel Wagner 3 participants 6 messages Dec 20, 2023 – Feb 6, 2024

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 scsi: lpfc: use unsigned type for num_sge

From: Hannes Reinecke <[email protected]>

LUNs going into “failed ready running” state observed on >1T and on even numbers of size (2T, 4T, 6T, 8T and 10T). The issue occurs when DIF is enabled at the host.

The kernel logs:

Cannot setup S/G List for HBAIO segs 1/1 SGL 512 SCSI 256: 3 0

The host lpfc driver is failing to setup scatter/gather list (protection data) for the IO's.

The return type lpfc_bg_setup_sgl()/lpfc_bg_setup_sgl_prot() causes the compiler to remove the most significant bit. Use an unsigned type instead.

[dwagner: added commit message]
---
 drivers/scsi/lpfc/lpfc_scsi.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)
drivers/scsi/lpfc/lpfc_scsi.c +6 -6
index d26941b131fd..bf879d81846b 100644
@@ -1918,7 +1918,7 @@ lpfc_bg_setup_bpl_prot(struct lpfc_hba *phba, struct scsi_cmnd *sc,
  *
  * Returns the number of SGEs added to the SGL.
  **/
-static int
+static uint32_t
 lpfc_bg_setup_sgl(struct lpfc_hba *phba, struct scsi_cmnd *sc,
 		struct sli4_sge *sgl, int datasegcnt,
 		struct lpfc_io_buf *lpfc_cmd)
@@ -1926,8 +1926,8 @@ lpfc_bg_setup_sgl(struct lpfc_hba *phba, struct scsi_cmnd *sc,
 	struct scatterlist *sgde = NULL; /* s/g data entry */
 	struct sli4_sge_diseed *diseed = NULL;
 	dma_addr_t physaddr;
-	int i = 0, num_sge = 0, status;
-	uint32_t reftag;
+	int i = 0, status;
+	uint32_t reftag, num_sge = 0;
 	uint8_t txop, rxop;
 #ifdef CONFIG_SCSI_LPFC_DEBUG_FS
 	uint32_t rc;
@@ -2099,7 +2099,7 @@ lpfc_bg_setup_sgl(struct lpfc_hba *phba, struct scsi_cmnd *sc,
  *
  * Returns the number of SGEs added to the SGL.
  **/
-static int
+static uint32_t
 lpfc_bg_setup_sgl_prot(struct lpfc_hba *phba, struct scsi_cmnd *sc,
 		struct sli4_sge *sgl, int datacnt, int protcnt,
 		struct lpfc_io_buf *lpfc_cmd)
@@ -2123,8 +2123,8 @@ lpfc_bg_setup_sgl_prot(struct lpfc_hba *phba, struct scsi_cmnd *sc,
 	uint32_t rc;
 #endif
 	uint32_t checking = 1;
-	uint32_t dma_offset = 0;
-	int num_sge = 0, j = 2;
+	uint32_t dma_offset = 0, num_sge = 0;
+	int j = 2;
 	struct sli4_hybrid_sgl *sgl_xtra = NULL;
 
 	sgpe = scsi_prot_sglist(sc);
Signature · 1 line
2.43.0
#1 lore raw

The change is good, however, I don't think this was really the problem.

We would like to know if this patch really solved an issue they observed.

A good data point to know is what adapter they're using. If that adapter supports hybrid sgl (i.e. phba->cfg_xpsgl), then we would have set the max sg_tablesize = LPFC_MAX_SG_TABLESIZE = 0xffff.

But even then, this patch implies that dma_map_sg() returned a crazy huge amount with the MSB set.

Quoted 76 lines · Daniel Wagner

From: Hannes Reinecke <[email protected]>

LUNs going into “failed ready running” state observed on >1T and on even numbers of size (2T, 4T, 6T, 8T and 10T). The issue occurs when DIF is enabled at the host.

The kernel logs:

Cannot setup S/G List for HBAIO segs 1/1 SGL 512 SCSI 256: 3 0

The host lpfc driver is failing to setup scatter/gather list (protection data) for the IO's.

The return type lpfc_bg_setup_sgl()/lpfc_bg_setup_sgl_prot() causes the compiler to remove the most significant bit. Use an unsigned type instead.

[dwagner: added commit message]
---
 drivers/scsi/lpfc/lpfc_scsi.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)
drivers/scsi/lpfc/lpfc_scsi.c +1 -1
index d26941b131fd..bf879d81846b 100644
@@ -1918,7 +1918,7 @@ lpfc_bg_setup_bpl_prot(struct lpfc_hba *phba, struct
 scsi_cmnd *sc,
  *
  * Returns the number of SGEs added to the SGL.
  **/
-static int
+static uint32_t
 lpfc_bg_setup_sgl(struct lpfc_hba *phba, struct scsi_cmnd *sc,
                struct sli4_sge *sgl, int datasegcnt,

struct lpfc_io_buf *lpfc_cmd)

drivers/scsi/lpfc/lpfc_scsi.c +2 -2
@@ -1926,8 +1926,8 @@ lpfc_bg_setup_sgl(struct lpfc_hba *phba, struct
 scsi_cmnd *sc,
        struct scatterlist *sgde = NULL; /* s/g data entry */
        struct sli4_sge_diseed *diseed = NULL;
        dma_addr_t physaddr;
-       int i = 0, num_sge = 0, status;
-       uint32_t reftag;
+       int i = 0, status;
+       uint32_t reftag, num_sge = 0;
        uint8_t txop, rxop;
 #ifdef CONFIG_SCSI_LPFC_DEBUG_FS
        uint32_t rc;
drivers/scsi/lpfc/lpfc_scsi.c +1 -1
@@ -2099,7 +2099,7 @@ lpfc_bg_setup_sgl(struct lpfc_hba *phba, struct
 scsi_cmnd *sc,
  *
  * Returns the number of SGEs added to the SGL.
  **/
-static int
+static uint32_t
 lpfc_bg_setup_sgl_prot(struct lpfc_hba *phba, struct scsi_cmnd *sc,
                struct sli4_sge *sgl, int datacnt, int protcnt,

struct lpfc_io_buf *lpfc_cmd)

drivers/scsi/lpfc/lpfc_scsi.c +2 -2
@@ -2123,8 +2123,8 @@ lpfc_bg_setup_sgl_prot(struct lpfc_hba *phba, struct
 scsi_cmnd *sc,
        uint32_t rc;
 #endif
        uint32_t checking = 1;
-       uint32_t dma_offset = 0;
-       int num_sge = 0, j = 2;
+       uint32_t dma_offset = 0, num_sge = 0;
+       int j = 2;
        struct sli4_hybrid_sgl *sgl_xtra = NULL;
 
        sgpe = scsi_prot_sglist(sc);

--
2.43.0

Signature · 11 lines
This electronic communication and the information and any files transmitted 
with it, or attached to it, are confidential and are intended solely for 
the use of the individual or entity to whom it is addressed and may contain 
information that is confidential, legally privileged, protected by privacy 
laws, or otherwise restricted from disclosure to anyone else. If you are 
not the intended recipient or the person responsible for delivering the 
e-mail to the intended recipient, you are hereby notified that any use, 
copying, distributing, dissemination, forwarding, printing, or copying of 
this e-mail is strictly prohibited. If you received this e-mail in error, 
please return the e-mail to the sender, delete it from your computer, and 
destroy any printed copy of it.
#2 lore raw

Hi Dick,

Quoted 2 lines · Dick Kennedy

The change is good, however, I don't think this was really the problem.

I tried to write the commit message based on the bug report we got. So yes, it's possible the it is not correct as I was not really involved and might missinterpret it.

Quoted 2 lines

We would like to know if this patch really solved an issue they observed.

Yes, it fixes the reported problem.

Quoted 1 line

A good data point to know is what adapter they're using.

A bunch of different HPE cards which show this log entry: SN1700E, SN1610E and SN1200E.

Quoted 5 lines

If that adapter supports hybrid sgl (i.e. phba->cfg_xpsgl), then we would have set the max sg_tablesize = LPFC_MAX_SG_TABLESIZE = 0xffff.

But even then, this patch implies that dma_map_sg() returned a crazy huge amount with the MSB set.

Sure, though this seems to be the case.

One noteworthy information is that DIF needs to be enabled to trigger it:

# cat /sys/module/lpfc/parameters/lpfc_enable_bg
1
# cat /sys/module/lpfc/parameters/lpfc_prot_guard
2

Thanks,
Daniel

#3 lore raw
Quoted 9 lines · Daniel Wagner

Hi Dick,

Quoted 2 lines · Dick Kennedy

The change is good, however, I don't think this was really the problem.

I tried to write the commit message based on the bug report we got. So yes, it's possible the it is not correct as I was not really involved and might missinterpret it.

Any chance to get this moving forward?

#4 lore raw

Sorry, I got distracted.
The change is fine.

Quoted 13 lines · Daniel Wagner
Quoted 9 lines · Daniel Wagner

Hi Dick,

Quoted 2 lines · Dick Kennedy

The change is good, however, I don't think this was really the problem.

I tried to write the commit message based on the bug report we got. So yes, it's possible the it is not correct as I was not really involved and might missinterpret it.

Any chance to get this moving forward?

Signature · 11 lines
This electronic communication and the information and any files transmitted 
with it, or attached to it, are confidential and are intended solely for 
the use of the individual or entity to whom it is addressed and may contain 
information that is confidential, legally privileged, protected by privacy 
laws, or otherwise restricted from disclosure to anyone else. If you are 
not the intended recipient or the person responsible for delivering the 
e-mail to the intended recipient, you are hereby notified that any use, 
copying, distributing, dissemination, forwarding, printing, or copying of 
this e-mail is strictly prohibited. If you received this e-mail in error, 
please return the e-mail to the sender, delete it from your computer, and 
destroy any printed copy of it.
#5 lore raw
In reply to Daniel Wagner
Quoted 9 lines · Daniel Wagner

LUNs going into “failed ready running” state observed on >1T and on even numbers of size (2T, 4T, 6T, 8T and 10T). The issue occurs when DIF is enabled at the host.

The kernel logs:

Cannot setup S/G List for HBAIO segs 1/1 SGL 512 SCSI 256: 3 0

[...]

Applied to 6.8/scsi-fixes, thanks!

[1/1] scsi: lpfc: use unsigned type for num_sge

      https://git.kernel.org/mkp/scsi/c/d6c1b19153f9
Signature · 1 line
Martin K. Petersen	Oracle Linux Engineering