On io completion, the driver is taking an adapter wide lock and nulling the scsi command back pointer. The nulling of the back pointer is to signify the io was completed and the scsi_done() routine was called. However, the routine makes no check to see if the abort routine had done the same thing and possibly nulled the pointer. Thus it may doubly-complete the io.
Make the following mods: - Check to make sure forward progress (call scsi_done()) only happens if the command pointer was non-null. - As the taking of the lock, which is adapter wide, is very costly on a system under load, null the pointer using an xchg operation rather than under lock. Signed-off-by: Dick Kennedy <[email protected]> Signed-off-by: James Smart <[email protected]> --- drivers/scsi/lpfc/lpfc_scsi.c | 14 +++++++++++--- 1 file changed, 11 insertions(+), 3 deletions(-) diff --git a/drivers/scsi/lpfc/lpfc_scsi.c b/drivers/scsi/lpfc/lpfc_scsi.c index 5c7858e735c9..200b5bca1f5f 100644 --- a/drivers/scsi/lpfc/lpfc_scsi.c +++ b/drivers/scsi/lpfc/lpfc_scsi.c @@ -4158,9 +4158,17 @@ lpfc_scsi_cmd_iocb_cmpl(struct lpfc_hba *phba, struct lpfc_iocbq *pIocbIn, } lpfc_scsi_unprep_dma_buf(phba, lpfc_cmd); - spin_lock_irqsave(&phba->hbalock, flags); - lpfc_cmd->pCmd = NULL; - spin_unlock_irqrestore(&phba->hbalock, flags); + /* If pCmd was set to NULL from abort path, do not call scsi_done */ + if (xchg(&lpfc_cmd->pCmd, NULL) == NULL) { + lpfc_printf_vlog(vport, KERN_INFO, LOG_FCP, + "0711 FCP cmd already NULL, sid: 0x%06x, " + "did: 0x%06x, oxid: 0x%04x\n", + vport->fc_myDID, + (pnode) ? pnode->nlp_DID : 0, + phba->sli_rev == LPFC_SLI_REV4 ? + lpfc_cmd->cur_iocbq.sli4_xritag : 0xffff); + return; + } /* The sdev is not guaranteed to be valid post scsi_done upcall. */ cmd->scsi_done(cmd); -- 2.13.1
