firewire: core: no need to track irq flags in bm_work
authorStefan Richter <stefanr@s5r6.in-berlin.de>
Mon, 21 Jun 2010 21:23:52 +0000 (23:23 +0200)
committerStefan Richter <stefanr@s5r6.in-berlin.de>
Thu, 8 Jul 2010 14:45:54 +0000 (16:45 +0200)
This is a workqueue job and always entered with IRQs enabled.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
drivers/firewire/core-card.c

index 156b098b0ada773ca040bae6b05120e38db76865..11fc81500f8287da0b7dc2b42dfaa164fddaa931 100644 (file)
@@ -239,17 +239,16 @@ static void fw_card_bm_work(struct work_struct *work)
        struct fw_card *card = container_of(work, struct fw_card, work.work);
        struct fw_device *root_device;
        struct fw_node *root_node;
-       unsigned long flags;
        int root_id, new_root_id, irm_id, local_id;
        int gap_count, generation, grace, rcode;
        bool do_reset = false;
        bool root_device_is_running;
        bool root_device_is_cmc;
 
-       spin_lock_irqsave(&card->lock, flags);
+       spin_lock_irq(&card->lock);
 
        if (card->local_node == NULL) {
-               spin_unlock_irqrestore(&card->lock, flags);
+               spin_unlock_irq(&card->lock);
                goto out_put_card;
        }
 
@@ -291,7 +290,7 @@ static void fw_card_bm_work(struct work_struct *work)
                card->bm_transaction_data[0] = cpu_to_be32(0x3f);
                card->bm_transaction_data[1] = cpu_to_be32(local_id);
 
-               spin_unlock_irqrestore(&card->lock, flags);
+               spin_unlock_irq(&card->lock);
 
                rcode = fw_run_transaction(card, TCODE_LOCK_COMPARE_SWAP,
                                irm_id, generation, SCODE_100,
@@ -322,7 +321,7 @@ static void fw_card_bm_work(struct work_struct *work)
                        goto out;
                }
 
-               spin_lock_irqsave(&card->lock, flags);
+               spin_lock_irq(&card->lock);
 
                if (rcode != RCODE_COMPLETE) {
                        /*
@@ -341,7 +340,7 @@ static void fw_card_bm_work(struct work_struct *work)
                 * We weren't BM in the last generation, and the last
                 * bus reset is less than 125ms ago.  Reschedule this job.
                 */
-               spin_unlock_irqrestore(&card->lock, flags);
+               spin_unlock_irq(&card->lock);
                fw_schedule_bm_work(card, DIV_ROUND_UP(HZ, 8));
                goto out;
        }
@@ -364,7 +363,7 @@ static void fw_card_bm_work(struct work_struct *work)
                 * If we haven't probed this device yet, bail out now
                 * and let's try again once that's done.
                 */
-               spin_unlock_irqrestore(&card->lock, flags);
+               spin_unlock_irq(&card->lock);
                goto out;
        } else if (root_device_is_cmc) {
                /*
@@ -402,7 +401,7 @@ static void fw_card_bm_work(struct work_struct *work)
            (card->gap_count != gap_count || new_root_id != root_id))
                do_reset = true;
 
-       spin_unlock_irqrestore(&card->lock, flags);
+       spin_unlock_irq(&card->lock);
 
        if (do_reset) {
                fw_notify("phy config: card %d, new root=%x, gap_count=%d\n",