Commit 8764cf4b authored by Mikhail Karpenko's avatar Mikhail Karpenko

Add a workaround patch to fix kernel panic after ESATA disk connection

parent 8fc41ee4
Patches from this directory are applied by 'meta-elphel393/recipes-kernel/linux/linux-xlnx' bitbake recipe. To add a new patch, append SRC_URI variable in the recipe.
Short description of the paches:
libata-eh.c.patch
This patch is a workaround for the situation when ESATA disk connection leads to kernel panic. Consider the following execution path in driver: libata-core.c:ata_exec_internal_sg -> error during command execution leading to command timeout -> libata-core.c:ata_qc_complete (or ata_port_freeze and finally to the same function) -> libata-core.c:ata_qc_schedule_eh where qc->scsicmd is a NULL pointer by the moment. This patch may not be needed when all hardware/software issues are rectified.
diff --git a/drivers/ata/libata-eh.c b/drivers/ata/libata-eh.c
index d2029a4..3fdf705 100644
--- a/drivers/ata/libata-eh.c
+++ b/drivers/ata/libata-eh.c
@@ -968,7 +968,7 @@ static void ata_eh_set_pending(struct ata_port *ap, int fastdrain)
void ata_qc_schedule_eh(struct ata_queued_cmd *qc)
{
struct ata_port *ap = qc->ap;
- struct request_queue *q = qc->scsicmd->device->request_queue;
+ struct request_queue *q;
unsigned long flags;
WARN_ON(!ap->ops->error_handler);
@@ -981,9 +981,12 @@ void ata_qc_schedule_eh(struct ata_queued_cmd *qc)
* Note that ATA_QCFLAG_FAILED is unconditionally set after
* this function completes.
*/
- spin_lock_irqsave(q->queue_lock, flags);
- blk_abort_request(qc->scsicmd->request);
- spin_unlock_irqrestore(q->queue_lock, flags);
+ if (qc->scsicmd != NULL) {
+ q = qc->scsicmd->device->request_queue;
+ spin_lock_irqsave(q->queue_lock, flags);
+ blk_abort_request(qc->scsicmd->request);
+ spin_unlock_irqrestore(q->queue_lock, flags);
+ }
}
/**
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment