mthca: work around -Wenum-conversion warning

PATCH In review Arnd Bergmann 2 participants 4 messages Oct 26, 2020 – Nov 12, 2020 InfiniBand

lore

Share & tools

lore

Arnd Bergmann

author

2020-10-26 21:16 UTC #0 lore raw

PATCH 1/1 mthca: work around -Wenum-conversion warning

From: Arnd Bergmann <[email protected]>

gcc points out a suspicious mixing of enum types in a function that converts from MTHCA_OPCODE_* values to IB_WC_* values:

drivers/infiniband/hw/mthca/mthca_cq.c: In function 'mthca_poll_one':
drivers/infiniband/hw/mthca/mthca_cq.c:607:21: warning: implicit conversion from 'enum <anonymous>' to 'enum ib_wc_opcode' [-Wenum-conversion]
607 | entry->opcode = MTHCA_OPCODE_INVALID;

Nothing seems to ever check for MTHCA_OPCODE_INVALID again, no idea if this is meaningful, but it seems harmless as it deals with an invalid input.

Add a cast to suppress the warning.

Fixes 2a4443a69934 ("[PATCH] IB/mthca: fill in opcode field for send completions")

Signed-off-by Arnd Bergmann <[email protected]>

---
 drivers/infiniband/hw/mthca/mthca_cq.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

drivers/infiniband/hw/mthca/mthca_cq.c +1 -1

index c3cfea243af8..319b8aa63f36 100644
@@ -604,7 +604,7 @@ static inline int mthca_poll_one(struct mthca_dev *dev,
 			entry->byte_len  = MTHCA_ATOMIC_BYTE_LEN;
 			break;
 		default:
-			entry->opcode    = MTHCA_OPCODE_INVALID;
+			entry->opcode    = (u8)MTHCA_OPCODE_INVALID;
 			break;
 		}
 	} else {

Signature · 1 line

2.27.0

Jason Gunthorpe

maintainer

2020-11-12 15:57 UTC #1 lore raw

Quoted 24 lines · Arnd Bergmann

From: Arnd Bergmann <[email protected]>
gcc points out a suspicious mixing of enum types in a function that converts from MTHCA_OPCODE_* values to IB_WC_* values:
drivers/infiniband/hw/mthca/mthca_cq.c: In function 'mthca_poll_one':
drivers/infiniband/hw/mthca/mthca_cq.c:607:21: warning: implicit conversion from 'enum <anonymous>' to 'enum ib_wc_opcode' [-Wenum-conversion]
607 | entry->opcode = MTHCA_OPCODE_INVALID;
Nothing seems to ever check for MTHCA_OPCODE_INVALID again, no idea if this is meaningful, but it seems harmless as it deals with an invalid input.
Add a cast to suppress the warning.
Fixes 2a4443a69934 ("[PATCH] IB/mthca: fill in opcode field for send completions")
Signed-off-by Arnd Bergmann <[email protected]>
 drivers/infiniband/hw/mthca/mthca_cq.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/infiniband/hw/mthca/mthca_cq.c +1 -0
index c3cfea243af8..319b8aa63f36 100644
+++ b/drivers/infiniband/hw/mthca/mthca_cq.c

drivers/infiniband/hw/mthca/mthca_cq.c static inline int mthca_poll_one(struct mthca_dev *dev, · line 608

Quoted 6 lines

drivers/infiniband/hw/mthca/mthca_cq.c +1 -1

@@ -604,7 +604,7 @@ static inline int mthca_poll_one(struct mthca_dev *dev,
 			entry->byte_len  = MTHCA_ATOMIC_BYTE_LEN;
 			break;
 		default:
-			entry->opcode    = MTHCA_OPCODE_INVALID;
+			entry->opcode    = (u8)MTHCA_OPCODE_INVALID;
 			break;

This code is completely bogus, sigh

Is this OK as far as the warning goes?

drivers/infiniband/hw/mthca/mthca_cq.c +1 -1

index 319b8aa63f3645..36416f937b696c 100644
@@ -604,7 +604,7 @@ static inline int mthca_poll_one(struct mthca_dev *dev,
 			entry->byte_len  = MTHCA_ATOMIC_BYTE_LEN;
 			break;
 		default:
-			entry->opcode    = (u8)MTHCA_OPCODE_INVALID;
+			entry->opcode = 0xFF;
 			break;
 		}
 	} else {

drivers/infiniband/hw/mthca/mthca_dev.h +0 -1

index 9dbbf4d16796a4..a445160de3e16c 100644
@@ -105,7 +105,6 @@ enum {
 	MTHCA_OPCODE_ATOMIC_CS      = 0x11,
 	MTHCA_OPCODE_ATOMIC_FA      = 0x12,
 	MTHCA_OPCODE_BIND_MW        = 0x18,
-	MTHCA_OPCODE_INVALID        = 0xff
 };
 
 enum {

Arnd Bergmann

author Acked-by

2020-11-12 16:45 UTC #2 lore raw

Quoted 35 lines · Jason Gunthorpe

Quoted 24 lines · Arnd Bergmann
From: Arnd Bergmann <[email protected]>
gcc points out a suspicious mixing of enum types in a function that converts from MTHCA_OPCODE_* values to IB_WC_* values:
drivers/infiniband/hw/mthca/mthca_cq.c: In function 'mthca_poll_one':
drivers/infiniband/hw/mthca/mthca_cq.c:607:21: warning: implicit conversion from 'enum <anonymous>' to 'enum ib_wc_opcode' [-Wenum-conversion]
607 | entry->opcode = MTHCA_OPCODE_INVALID;
Nothing seems to ever check for MTHCA_OPCODE_INVALID again, no idea if this is meaningful, but it seems harmless as it deals with an invalid input.
Add a cast to suppress the warning.
Fixes 2a4443a69934 ("[PATCH] IB/mthca: fill in opcode field for send completions")
Signed-off-by Arnd Bergmann <[email protected]>
 drivers/infiniband/hw/mthca/mthca_cq.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/infiniband/hw/mthca/mthca_cq.c +1 -0
index c3cfea243af8..319b8aa63f36 100644
+++ b/drivers/infiniband/hw/mthca/mthca_cq.c
drivers/infiniband/hw/mthca/mthca_cq.c static inline int mthca_poll_one(struct mthca_dev *dev, · line 608
Quoted 6 lines
drivers/infiniband/hw/mthca/mthca_cq.c +1 -1
@@ -604,7 +604,7 @@ static inline int mthca_poll_one(struct mthca_dev *dev,
                      entry->byte_len  = MTHCA_ATOMIC_BYTE_LEN;
                      break;
              default:
-                     entry->opcode    = MTHCA_OPCODE_INVALID;
+                     entry->opcode    = (u8)MTHCA_OPCODE_INVALID;
                      break;
This code is completely bogus, sigh

Quoted 1 line

Is this OK as far as the warning goes?

Yes, I'm sure your patch fixes it and it makes more sense than my version.

Acked-by Arnd Bergmann <[email protected]>

Jason Gunthorpe

maintainer

2020-11-12 16:46 UTC #3 lore raw

Quoted 36 lines · Arnd Bergmann

Quoted 35 lines · Jason Gunthorpe
Quoted 24 lines · Arnd Bergmann
From: Arnd Bergmann <[email protected]>
gcc points out a suspicious mixing of enum types in a function that converts from MTHCA_OPCODE_* values to IB_WC_* values:
drivers/infiniband/hw/mthca/mthca_cq.c: In function 'mthca_poll_one':
drivers/infiniband/hw/mthca/mthca_cq.c:607:21: warning: implicit conversion from 'enum <anonymous>' to 'enum ib_wc_opcode' [-Wenum-conversion]
607 | entry->opcode = MTHCA_OPCODE_INVALID;
Nothing seems to ever check for MTHCA_OPCODE_INVALID again, no idea if this is meaningful, but it seems harmless as it deals with an invalid input.
Add a cast to suppress the warning.
Fixes 2a4443a69934 ("[PATCH] IB/mthca: fill in opcode field for send completions")
Signed-off-by Arnd Bergmann <[email protected]>
 drivers/infiniband/hw/mthca/mthca_cq.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/infiniband/hw/mthca/mthca_cq.c +1 -0
index c3cfea243af8..319b8aa63f36 100644
+++ b/drivers/infiniband/hw/mthca/mthca_cq.c
drivers/infiniband/hw/mthca/mthca_cq.c static inline int mthca_poll_one(struct mthca_dev *dev, · line 608
Quoted 6 lines
drivers/infiniband/hw/mthca/mthca_cq.c +1 -1
@@ -604,7 +604,7 @@ static inline int mthca_poll_one(struct mthca_dev *dev,
                      entry->byte_len  = MTHCA_ATOMIC_BYTE_LEN;
                      break;
              default:
-                     entry->opcode    = MTHCA_OPCODE_INVALID;
+                     entry->opcode    = (u8)MTHCA_OPCODE_INVALID;
                      break;
This code is completely bogus, sigh

Quoted 5 lines

Quoted 1 line
Is this OK as far as the warning goes?
Yes, I'm sure your patch fixes it and it makes more sense than my version.
Acked-by Arnd Bergmann <[email protected]>

Okay, I'll revise it, thanks

Jason