Skip to contentSkip to conversation

BUG_ON() Conversion in drivers/ide/

PATCH Bartlomiej Zolnierkiewicz 1 participant 1 message Feb 1, 2006 Kernel janitors

lore
Share & tools
lore
#0 lore raw

Hi,

Quoted 4 lines · Eric Sesterhenn / snakebyte

hi,

this changes if() BUG(); constructs to BUG_ON() which is cleaner, contains unlikely() and can better optimized away.

Patch looks fine but maybe we can try to fix few
underlying problems while at it?

drivers/ide/ide.c 2006-01-31 21:10:49.000000000 +0100 int generic_ide_ioctl(ide_drive_t *drive · line 1370

Quoted 11 lines
drivers/ide/ide.c 2006-01-31 21:10:49.000000000 +0100 renamed from drivers/ide/ide.c.orig 2006-01-31 21:09:17.000000000 +0100 +1 -2
@@ -1367,8 +1367,7 @@ int generic_ide_ioctl(ide_drive_t *drive
 
                        ide_abort(drive, "drive reset");
 
-                       if(HWGROUP(drive)->handler)
-                               BUG();
+                       BUG_ON(HWGROUP(drive)->handler);

minor: please use drive->hwif->hwgroup instead of HWGROUP() macro

drivers/ide/ide-cd.c 2006-01-31 21:13:39.000000000 +0100 static ide_startstop_t cdrom_newpc_intr( · line 1726

Quoted 13 lines
drivers/ide/ide.c 2006-01-31 21:10:49.000000000 +0100 renamed from drivers/ide/ide.c.orig 2006-01-31 21:09:17.000000000 +0100 +0 -0
                        /* Ensure nothing gets queued after we
                           drop the lock. Reset will clear the busy */
drivers/ide/ide-cd.c 2006-01-31 21:13:39.000000000 +0100 renamed from drivers/ide/ide-cd.c.orig 2006-01-31 21:13:27.000000000 +0100 +1 -2
@@ -1721,8 +1721,7 @@ static ide_startstop_t cdrom_newpc_intr(
                }
        }
 
-       if (HWGROUP(drive)->handler != NULL)
-               BUG();
+       BUG_ON(HWGROUP(drive)->handler != NULL);
 
        ide_set_handler(drive, cdrom_newpc_intr, rq->timeout, NULL);

__ide_set_handler() which is called internally by ide_set_handler() warns about ->handler != NULL.

We should add BUG_ON() to __ide_set_handler() and remove all BUG() checks which are done before calling ide_set_handler() (so this one and the ones from ide-floppy.c, ide-iops.c, ide-taskfile.c and pci/trm290.c).

Care to make a separate patch for this change?

Quoted 14 lines
drivers/ide/ide-cd.c 2006-01-31 21:13:39.000000000 +0100 renamed from drivers/ide/ide-cd.c.orig 2006-01-31 21:13:27.000000000 +0100 +0 -0
        return ide_started;
drivers/ide/ide-dma.c 2006-01-31 21:11:35.000000000 +0100 renamed from drivers/ide/ide-dma.c.orig 2006-01-31 21:11:13.000000000 +0100 +1 -2
@@ -206,8 +206,7 @@ int ide_build_sglist(ide_drive_t *drive,
        ide_hwif_t *hwif = HWIF(drive);
        struct scatterlist *sg = hwif->sg_table;
 
-       if ((rq->flags & REQ_DRIVE_TASKFILE) && rq->nr_sectors > 256)
-               BUG();
+       BUG_ON((rq->flags & REQ_DRIVE_TASKFILE) && rq->nr_sectors > 256);
 
        ide_map_sg(drive, rq);
 

drivers/ide/ide-dma.c 2006-01-31 21:11:35.000000000 +0100 void ide_setup_dma (ide_hwif_t *hwif, un · line 949

Quoted 6 lines
drivers/ide/ide-dma.c 2006-01-31 21:11:35.000000000 +0100 renamed from drivers/ide/ide-dma.c.orig 2006-01-31 21:11:13.000000000 +0100 +1 -2
@@ -947,8 +946,7 @@ void ide_setup_dma (ide_hwif_t *hwif, un
        }
        printk("\n");
 
-       if (!(hwif->dma_master))
-               BUG();
+       BUG_ON(!(hwif->dma_master));

We should just remove this BUG().

This check doesn't make much sense nowadays and is done too late in the initialisation sequence.

Care to make a separate patch?

Thanks,
Bartlomiej

_______________________________________________
Kernel-janitors mailing list
[email protected]
https://lists.osdl.org/mailman/listinfo/kernel-janitors