Hi
Here I send a patch that removes some impossible code from dm-zoned. The
target also contains large number of GFP_KERNEL allocations, they should
be changed to GFP_NOIO, if I/O processing could be delayed because of the
allocation (i.e. anywhere except initialization you should use GFP_NOIO).
Regarding this "if (!nr_sectors && bio_op(bio) != REQ_OP_WRITE)" - could
this be removed too? Does it really process some zero-sized requests that
are not flushes? It doesn't seem so.
Mikulas
From: Mikulas Patocka <mpatocka@xxxxxxxxxx>
The value REQ_OP_FLUSH is used by the block code in the request structure.
It is not used in the bio structure.
This patch removes the tests for REQ_OP_FLUSH from dm-zoned-target.
Signed-off-by: Mikulas Patocka <mpatocka@xxxxxxxxxx>
---
drivers/md/dm-zoned-target.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
Index: linux-2.6/drivers/md/dm-zoned-target.c
===================================================================
--- linux-2.6.orig/drivers/md/dm-zoned-target.c
+++ linux-2.6/drivers/md/dm-zoned-target.c
@@ -588,7 +588,7 @@ static int dmz_map(struct dm_target *ti,
bio->bi_bdev = dev->bdev;
- if (!nr_sectors && (bio_op(bio) != REQ_OP_FLUSH) && (bio_op(bio) != REQ_OP_WRITE))
+ if (!nr_sectors && bio_op(bio) != REQ_OP_WRITE)
return DM_MAPIO_REMAPPED;
/* The BIO should be block aligned */
@@ -603,7 +603,7 @@ static int dmz_map(struct dm_target *ti,
bioctx->status = BLK_STS_OK;
/* Set the BIO pending in the flush list */
- if (bio_op(bio) == REQ_OP_FLUSH || (!nr_sectors && bio_op(bio) == REQ_OP_WRITE)) {
+ if (!nr_sectors && bio_op(bio) == REQ_OP_WRITE) {
spin_lock(&dmz->flush_lock);
bio_list_add(&dmz->flush_list, bio);
spin_unlock(&dmz->flush_lock);
--
dm-devel mailing list
dm-devel@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/dm-devel
