[PATCH net 0/2] sfc: TC probe fixes

[PATCH net 0/2] sfc: TC probe fixes

[edward.cree]

From: Edward Cree <ecree.xilinx@gmail.com>

Fix a couple of minor infelicities in the error paths of EF100 TC
 offload setup at probe time.  Both found by code inspection.
Patch #1 will produce a conflict when merging net into net-next
 (with 3bf969e88ada ("sfc: add MAE table machinery for conntrack table"));
 the resolution is appended.

Edward Cree (2):
  sfc: don't unregister flow_indr if it was never registered
  sfc: don't fail probe if MAE/TC setup fails

 drivers/net/ethernet/sfc/ef100_nic.c | 2 +-
 drivers/net/ethernet/sfc/tc.c        | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

---

diff --cc drivers/net/ethernet/sfc/tc.c
index fe268b6c1cac,246657222958..000000000000
--- a/drivers/net/ethernet/sfc/tc.c
+++ b/drivers/net/ethernet/sfc/tc.c
@@@ -1657,11 -2087,17 +2087,17 @@@ int efx_init_tc(struct efx_nic *efx
  	rc = efx_tc_configure_fallback_acts_reps(efx);
  	if (rc)
  		return rc;
- 	rc = flow_indr_dev_register(efx_tc_indr_setup_cb, efx);
+ 	rc = efx_mae_get_tables(efx);
  	if (rc)
  		return rc;
 -	efx->tc->up = true;
+ 	rc = flow_indr_dev_register(efx_tc_indr_setup_cb, efx);
+ 	if (rc)
+ 		goto out_free;
 +	efx->tc->up = true;
  	return 0;
+ out_free:
+ 	efx_mae_free_tables(efx);
+ 	return rc;
  }
  
  void efx_fini_tc(struct efx_nic *efx)

[PATCH net 1/2] sfc: don't unregister flow_indr if it was never registered

[edward.cree]

From: Edward Cree <ecree.xilinx@gmail.com>

In efx_init_tc(), move the setting of efx->tc->up after the
 flow_indr_dev_register() call, so that if it fails, efx_fini_tc()
 won't call flow_indr_dev_unregister().

Fixes: 5b2e12d51bd8 ("sfc: bind indirect blocks for TC offload on EF100")
Suggested-by: Pieter Jansen van Vuuren <pieter.jansen-van-vuuren@amd.com>
Reviewed-by: Martin Habets <habetsm.xilinx@gmail.com>
Signed-off-by: Edward Cree <ecree.xilinx@gmail.com>
---
 drivers/net/ethernet/sfc/tc.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/sfc/tc.c b/drivers/net/ethernet/sfc/tc.c
index 15ebd3973922..fe268b6c1cac 100644
--- a/drivers/net/ethernet/sfc/tc.c
+++ b/drivers/net/ethernet/sfc/tc.c
@@ -1657,10 +1657,10 @@ int efx_init_tc(struct efx_nic *efx)
 	rc = efx_tc_configure_fallback_acts_reps(efx);
 	if (rc)
 		return rc;
-	efx->tc->up = true;
 	rc = flow_indr_dev_register(efx_tc_indr_setup_cb, efx);
 	if (rc)
 		return rc;
+	efx->tc->up = true;
 	return 0;
 }
 

Re: [PATCH net 1/2] sfc: don't unregister flow_indr if it was never registered

[patchwork-bot+netdevbpf]

Hello:

This series was applied to netdev/net.git (main)
by Jakub Kicinski <kuba@kernel.org>:

On Tue, 15 Aug 2023 16:57:27 +0100 you wrote:
> From: Edward Cree <ecree.xilinx@gmail.com>
> 
> In efx_init_tc(), move the setting of efx->tc->up after the
>  flow_indr_dev_register() call, so that if it fails, efx_fini_tc()
>  won't call flow_indr_dev_unregister().
> 
> Fixes: 5b2e12d51bd8 ("sfc: bind indirect blocks for TC offload on EF100")
> Suggested-by: Pieter Jansen van Vuuren <pieter.jansen-van-vuuren@amd.com>
> Reviewed-by: Martin Habets <habetsm.xilinx@gmail.com>
> Signed-off-by: Edward Cree <ecree.xilinx@gmail.com>
> 
> [...]

Here is the summary with links:
  - [net,1/2] sfc: don't unregister flow_indr if it was never registered
    https://git.kernel.org/netdev/net/c/fa165e194997
  - [net,2/2] sfc: don't fail probe if MAE/TC setup fails
    https://git.kernel.org/netdev/net/c/54c9016eb8ed

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html

[PATCH net 2/2] sfc: don't fail probe if MAE/TC setup fails

[edward.cree]

From: Edward Cree <ecree.xilinx@gmail.com>

Existing comment in the source explains why we don't want efx_init_tc()
 failure to be fatal.  Cited commit erroneously consolidated failure
 paths causing the probe to be failed in this case.

Fixes: 7e056e2360d9 ("sfc: obtain device mac address based on firmware handle for ef100")
Reviewed-by: Martin Habets <habetsm.xilinx@gmail.com>
Signed-off-by: Edward Cree <ecree.xilinx@gmail.com>
---
 drivers/net/ethernet/sfc/ef100_nic.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/sfc/ef100_nic.c b/drivers/net/ethernet/sfc/ef100_nic.c
index 7adde9639c8a..35d8e9811998 100644
--- a/drivers/net/ethernet/sfc/ef100_nic.c
+++ b/drivers/net/ethernet/sfc/ef100_nic.c
@@ -1194,7 +1194,7 @@ int ef100_probe_netdev_pf(struct efx_nic *efx)
 		net_dev->features |= NETIF_F_HW_TC;
 		efx->fixed_features |= NETIF_F_HW_TC;
 	}
-	return rc;
+	return 0;
 }
 
 int ef100_probe_vf(struct efx_nic *efx)

Re: [PATCH net 0/2] sfc: TC probe fixes

[Simon Horman]

On Tue, Aug 15, 2023 at 04:57:26PM +0100, edward.cree@amd.com wrote:
> From: Edward Cree <ecree.xilinx@gmail.com>
> 
> Fix a couple of minor infelicities in the error paths of EF100 TC
>  offload setup at probe time.  Both found by code inspection.
> Patch #1 will produce a conflict when merging net into net-next
>  (with 3bf969e88ada ("sfc: add MAE table machinery for conntrack table"));
>  the resolution is appended.
> 
> Edward Cree (2):
>   sfc: don't unregister flow_indr if it was never registered
>   sfc: don't fail probe if MAE/TC setup fails
> 
>  drivers/net/ethernet/sfc/ef100_nic.c | 2 +-
>  drivers/net/ethernet/sfc/tc.c        | 2 +-
>  2 files changed, 2 insertions(+), 2 deletions(-)

For series:

Reviewed-by: Simon Horman <horms@kernel.org>