Message ID | 20180310010606.19539-1-jsmart2021@gmail.com (mailing list archive) |
---|---|
State | Superseded |
Headers | show |
Really, the commit message has to be descriptive. This makes me think this is about some annoying policy of making everything static rather than a critical bug. What's wrong with lpfc: add missing unlock on defer WQFULL path ? You can then expand on the static checker discovery in the main message. Since this looks to be a pretty common occurrence, perhaps this should be folded with a rebase? James
On 3/9/2018 5:23 PM, James Bottomley wrote: > Really, the commit message has to be descriptive. This makes me think > this is about some annoying policy of making everything static rather > than a critical bug. > > What's wrong with > > lpfc: add missing unlock on defer WQFULL path > > ? You can then expand on the static checker discovery in the main > message. > > Since this looks to be a pretty common occurrence, perhaps this should > be folded with a rebase? > > James > yeah - the title was pretty poor. Too long of a day on a friday. reposted a v2. -- james
diff --git a/drivers/scsi/lpfc/lpfc_nvmet.c b/drivers/scsi/lpfc/lpfc_nvmet.c index 66218afa1453..07f89524c320 100644 --- a/drivers/scsi/lpfc/lpfc_nvmet.c +++ b/drivers/scsi/lpfc/lpfc_nvmet.c @@ -1047,6 +1047,7 @@ lpfc_nvmet_xmt_fcp_abort(struct nvmet_fc_target_port *tgtport, lpfc_nvmet_unsol_fcp_issue_abort(phba, ctxp, ctxp->sid, ctxp->oxid); wq = phba->sli4_hba.nvme_wq[ctxp->wqeq->hba_wqidx]; + spin_unlock_irqrestore(&ctxp->ctxlock, flags); lpfc_nvmet_wqfull_flush(phba, wq, ctxp); return; }