| Message ID | 20210630183617.3093690-1-sukadev@linux.ibm.com (mailing list archive) |
|---|---|
| State | Accepted |
| Commit | 4f408e1fa6e10b6da72691233369172bac7d9e9b |
| Delegated to: | Netdev Maintainers |
| Headers | show |
| Series | [1/1] ibmvnic: retry reset if there are no other resets | expand |
| Context | Check | Description |
|---|---|---|
| netdev/cover_letter | success | Link |
| netdev/fixes_present | success | Link |
| netdev/patch_count | success | Link |
| netdev/tree_selection | success | Guessed tree name to be net-next |
| netdev/subject_prefix | warning | Target tree name not specified in the subject |
| netdev/cc_maintainers | warning | 7 maintainers not CCed: tlfalcon@linux.ibm.com paulus@samba.org benh@kernel.crashing.org linuxppc-dev@lists.ozlabs.org mpe@ellerman.id.au davem@davemloft.net kuba@kernel.org |
| netdev/source_inline | success | Was 0 now: 0 |
| netdev/verify_signedoff | success | Link |
| netdev/module_param | success | Was 0 now: 0 |
| netdev/build_32bit | success | Errors and warnings before: 39 this patch: 39 |
| netdev/kdoc | success | Errors and warnings before: 0 this patch: 0 |
| netdev/verify_fixes | success | Link |
| netdev/checkpatch | warning | WARNING: networking block comments don't use an empty /* line, use /* Comment... |
| netdev/build_allmodconfig_warn | success | Errors and warnings before: 32 this patch: 32 |
| netdev/header_inline | success | Link |
On 2021-06-30 11:36, Sukadev Bhattiprolu wrote: > Normally, if a reset fails due to failover or other communication error > there is another reset (eg: FAILOVER) in the queue and we would process > that reset. But if we are unable to communicate with PHYP or VIOS after > H_FREE_CRQ, there would be no other resets in the queue and the adapter > would be in an undefined state even though it was in the OPEN state > earlier. While starting the reset we set the carrier to off state so > we won't even get the timeout resets. > > If the last queued reset fails, retry it as a hard reset (after the > usual 60 second settling time). > > Signed-off-by: Sukadev Bhattiprolu <sukadev@linux.ibm.com> Reviewed-by: Dany Madden <drt@linux.ibm.com> > ---
Hello: This patch was applied to netdev/net.git (refs/heads/master): On Wed, 30 Jun 2021 14:36:17 -0400 you wrote: > Normally, if a reset fails due to failover or other communication error > there is another reset (eg: FAILOVER) in the queue and we would process > that reset. But if we are unable to communicate with PHYP or VIOS after > H_FREE_CRQ, there would be no other resets in the queue and the adapter > would be in an undefined state even though it was in the OPEN state > earlier. While starting the reset we set the carrier to off state so > we won't even get the timeout resets. > > [...] Here is the summary with links: - [1/1] ibmvnic: retry reset if there are no other resets https://git.kernel.org/netdev/net/c/4f408e1fa6e1 You are awesome, thank you! -- Deet-doot-dot, I am a bot. https://korg.docs.kernel.org/patchwork/pwbot.html
diff --git a/drivers/net/ethernet/ibm/ibmvnic.c b/drivers/net/ethernet/ibm/ibmvnic.c index 697b9714fc76..ff49cda142b0 100644 --- a/drivers/net/ethernet/ibm/ibmvnic.c +++ b/drivers/net/ethernet/ibm/ibmvnic.c @@ -2420,9 +2420,10 @@ static int do_passive_init(struct ibmvnic_adapter *adapter) static void __ibmvnic_reset(struct work_struct *work) { - struct ibmvnic_rwi *rwi; struct ibmvnic_adapter *adapter; bool saved_state = false; + struct ibmvnic_rwi *tmprwi; + struct ibmvnic_rwi *rwi; unsigned long flags; u32 reset_state; int rc = 0; @@ -2489,7 +2490,7 @@ static void __ibmvnic_reset(struct work_struct *work) } else { rc = do_reset(adapter, rwi, reset_state); } - kfree(rwi); + tmprwi = rwi; adapter->last_reset_time = jiffies; if (rc) @@ -2497,8 +2498,23 @@ static void __ibmvnic_reset(struct work_struct *work) rwi = get_next_rwi(adapter); + /* + * If there is another reset queued, free the previous rwi + * and process the new reset even if previous reset failed + * (the previous reset could have failed because of a fail + * over for instance, so process the fail over). + * + * If there are no resets queued and the previous reset failed, + * the adapter would be in an undefined state. So retry the + * previous reset as a hard reset. + */ + if (rwi) + kfree(tmprwi); + else if (rc) + rwi = tmprwi; + if (rwi && (rwi->reset_reason == VNIC_RESET_FAILOVER || - rwi->reset_reason == VNIC_RESET_MOBILITY)) + rwi->reset_reason == VNIC_RESET_MOBILITY || rc)) adapter->force_reset_recovery = true; }
Normally, if a reset fails due to failover or other communication error there is another reset (eg: FAILOVER) in the queue and we would process that reset. But if we are unable to communicate with PHYP or VIOS after H_FREE_CRQ, there would be no other resets in the queue and the adapter would be in an undefined state even though it was in the OPEN state earlier. While starting the reset we set the carrier to off state so we won't even get the timeout resets. If the last queued reset fails, retry it as a hard reset (after the usual 60 second settling time). Signed-off-by: Sukadev Bhattiprolu <sukadev@linux.ibm.com> --- drivers/net/ethernet/ibm/ibmvnic.c | 22 +++++++++++++++++++--- 1 file changed, 19 insertions(+), 3 deletions(-)