Skip to contentSkip to conversation

net/mlx5: Fix modify_cq_in alignment

PATCH In review Leon Romanovsky 3 participants 9 messages Jul 23, 2019 – Jul 25, 2019

Subsystems (5)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 net/mlx5: Fix modify_cq_in alignment

From: Edward Srouji <[email protected]>

Fix modify_cq_in alignment to match the device specification. After this fix the 'cq_umem_valid' field will be in the right offset.

Cc <[email protected]> # 4.19
Fixes bd37197554eb ("net/mlx5: Update mlx5_ifc with DEVX UID bits")
---
 include/linux/mlx5/mlx5_ifc.h | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
include/linux/mlx5/mlx5_ifc.h +4 -2
index b3d5752657d9..ec571fd7fcf8 100644
@@ -5975,10 +5975,12 @@ struct mlx5_ifc_modify_cq_in_bits {
 
 	struct mlx5_ifc_cqc_bits cq_context;
 
-	u8         reserved_at_280[0x40];
+	u8         reserved_at_280[0x60];
 
 	u8         cq_umem_valid[0x1];
-	u8         reserved_at_2c1[0x5bf];
+	u8         reserved_at_2e1[0x1f];
+
+	u8         reserved_at_300[0x580];
 
 	u8         pas[0][0x40];
 };

--
2.20.1

maintainer
#1 lore raw

From: Leon Romanovsky <[email protected]>
Date: Tue, 23 Jul 2019 10:12:55 +0300

Quoted 10 lines

From: Edward Srouji <[email protected]>

Fix modify_cq_in alignment to match the device specification. After this fix the 'cq_umem_valid' field will be in the right offset.

Cc <[email protected]> # 4.19
Fixes bd37197554eb ("net/mlx5: Update mlx5_ifc with DEVX UID bits")

Very confusing submission on many levels.

Coming from a Mellanox developer using a kernel.org email address.

Targetting the mlx5-next tree, yet CC:'ing stable.

A networking change, for which stable submissions are handled by me by hand and not via CC:'ing stable.

#2 lore raw
Quoted 14 lines · David Miller

From: Leon Romanovsky <[email protected]>
Date: Tue, 23 Jul 2019 10:12:55 +0300

Quoted 10 lines

From: Edward Srouji <[email protected]>

Fix modify_cq_in alignment to match the device specification. After this fix the 'cq_umem_valid' field will be in the right offset.

Cc <[email protected]> # 4.19
Fixes bd37197554eb ("net/mlx5: Update mlx5_ifc with DEVX UID bits")
Quoted 3 lines

Very confusing submission on many levels.

Coming from a Mellanox developer using a kernel.org email address.

It works for us and was proven internally as the best way to have setup which always works.

Quoted 2 lines

Targetting the mlx5-next tree, yet CC:'ing stable.

This patch was found by RDMA team, needed by RDMA but changes are located in code accessible by mlx5_core part. This is why mlx5-next.

Quoted 3 lines

A networking change, for which stable submissions are handled by me by hand and not via CC:'ing stable.

The intention was to have this patch in shared mlx5 branch, which is picked by RDMA too. This "Cc: stable@..." together with merge through RDMA will ensure that such patch will be part of stable automatically.

I can remove "Cc: ..." line if you think that it is inappropriate to have such line in patch in mlx5-next.

Thanks

maintainer
#3 lore raw

From: Leon Romanovsky <[email protected]>
Date: Tue, 23 Jul 2019 22:04:14 +0300

Quoted 3 lines

The intention was to have this patch in shared mlx5 branch, which is picked by RDMA too. This "Cc: stable@..." together with merge through RDMA will ensure that such patch will be part of stable automatically.

Why wouldn't it come via Saeed's usual mlx5 bug fix pull requests to me?

#4 lore raw
Quoted 10 lines · David Miller

From: Leon Romanovsky <[email protected]>
Date: Tue, 23 Jul 2019 22:04:14 +0300

Quoted 6 lines

The intention was to have this patch in shared mlx5 branch, which is
picked by RDMA too. This "Cc: stable@..." together with merge through
RDMA will ensure that such patch will be part of stable automatically.

Quoted 2 lines

Why wouldn't it come via Saeed's usual mlx5 bug fix pull requests to me?

That should have been the plan in first place, i will handle this, thanks Dave and sorry for any inconvenience.

I will apply this patch to my (mlx5) net queue, will submit to net shortly.

Leon can merge the next -rc when this patch lands there. meanwhile, Leon can also merge my (mlx5) net queue which is always based on latest -rc.
https://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux.git/log/?h=net-mlx5

#5 lore raw
Quoted 11 lines · Saeed Mahameed
Quoted 10 lines · David Miller

From: Leon Romanovsky <[email protected]>
Date: Tue, 23 Jul 2019 22:04:14 +0300

Quoted 6 lines

The intention was to have this patch in shared mlx5 branch, which is
picked by RDMA too. This "Cc: stable@..." together with merge through
RDMA will ensure that such patch will be part of stable automatically.

Quoted 8 lines
Quoted 2 lines

Why wouldn't it come via Saeed's usual mlx5 bug fix pull requests to me?

That should have been the plan in first place, i will handle this, thanks Dave and sorry for any inconvenience.

I will apply this patch to my (mlx5) net queue, will submit to net shortly.

OK, whatever works for you best.

Thanks

#6 lore raw
In reply to Leon Romanovsky
Quoted 21 lines · Leon Romanovsky
Quoted 16 lines · David Miller

From: Leon Romanovsky <[email protected]>
Date: Tue, 23 Jul 2019 10:12:55 +0300

Quoted 6 lines

From: Edward Srouji <[email protected]>

Fix modify_cq_in alignment to match the device specification. After this fix the 'cq_umem_valid' field will be in the right offset.

Quoted 6 lines
Cc <[email protected]> # 4.19
Fixes bd37197554eb ("net/mlx5: Update mlx5_ifc with DEVX UID

bits")

Quoted 3 lines

Very confusing submission on many levels.

Coming from a Mellanox developer using a kernel.org email address.

Quoted 9 lines

It works for us and was proven internally as the best way to have setup which always works.

Quoted 1 line

Targetting the mlx5-next tree, yet CC:'ing stable.

This patch was found by RDMA team, needed by RDMA but changes are located
in code accessible by mlx5_core part. This is why mlx5-next.

Leon,
mlx5-next "hence the -next" is NOT meant for fixes, it is indeed confusing what you are trying to do here, Dave's system works perfectly for us.

Quoted 11 lines
Quoted 3 lines

A networking change, for which stable submissions are handled by me by
hand and not via CC:'ing stable.

The intention was to have this patch in shared mlx5 branch, which is picked by RDMA too. This "Cc: stable@..." together with merge through RDMA will ensure that such patch will be part of stable automatically.

I can remove "Cc: ..." line if you think that it is inappropriate to have such line in patch in mlx5-next.

No, if this was meant to land in -stable then it should go to -rc via net branch not to mlx5-next, let's save everybody's time and energy here. no point in arguing..

I will take this to my net queue and submit to Dave's net branch, as we always do for mlx5 fixes.

for next time, mlx5 fixes should always go to net branch first, then it is just a matter of days to see the fix in -rc and queued up to -stable.

Quoted 2 lines

Thanks

#7 lore raw
In reply to Leon Romanovsky
Quoted 6 lines · Leon Romanovsky
Quoted 5 lines · David Miller

From: Leon Romanovsky <[email protected]>
Date: Tue, 23 Jul 2019 10:12:55 +0300

Quoted 2 lines

From: Edward Srouji <[email protected]>

Quoted 7 lines
Quoted 7 lines
Quoted 7 lines

Fix modify_cq_in alignment to match the device specification. After this fix the 'cq_umem_valid' field will be in the right offset.

Cc <[email protected]> # 4.19
Fixes bd37197554eb ("net/mlx5: Update mlx5_ifc with DEVX UID

bits")

Leon, I applied this patch to my tree, it got marked for -stable 4.20 and not 4.19, i checked manually and indeed the offending patch came to light only on 4.20

#8 lore raw
Quoted 11 lines · Saeed Mahameed
Quoted 10 lines · Leon Romanovsky
Quoted 9 lines · David Miller

From: Leon Romanovsky <[email protected]>
Date: Tue, 23 Jul 2019 10:12:55 +0300

Quoted 6 lines

From: Edward Srouji <[email protected]>

Fix modify_cq_in alignment to match the device specification. After this fix the 'cq_umem_valid' field will be in the right offset.

Quoted 7 lines
Quoted 3 lines
Quoted 3 lines
Quoted 3 lines
Cc <[email protected]> # 4.19
Fixes bd37197554eb ("net/mlx5: Update mlx5_ifc with DEVX UID

bits")

Leon, I applied this patch to my tree, it got marked for -stable 4.20 and not 4.19, i checked manually and indeed the offending patch came to light only on 4.20

Thanks

Quoted 1 line