brd: use radix_tree_maybe_preload instead of radix_tree_preload

Unconditionally calling radix_tree_preload_end() results in a OOPS
message as the preload is only conditionally called for
gfpflags_allow_blocking().

[   20.267323] BUG: using smp_processor_id() in preemptible [00000000] code: fio/416
[   20.267837] caller is brd_insert_page.part.0+0xbe/0x190 [brd]
[   20.269436] Call Trace:
[   20.269598]  <TASK>
[   20.269742]  dump_stack_lvl+0x32/0x50
[   20.269982]  check_preemption_disabled+0xd1/0xe0
[   20.270289]  brd_insert_page.part.0+0xbe/0x190 [brd]
[   20.270664]  brd_submit_bio+0x33f/0xf40 [brd]

Use radix_tree_maybe_preload() which does preload only if
gfpflags_allow_blocking() is true but also takes the lock. Therefore,
unconditionally calling radix_tree_preload_end() should not create any
issues and the message disappears.

Fixes: 6ded703c56c2 ("brd: check for REQ_NOWAIT and set correct page allocation mask")
Signed-off-by: Pankaj Raghav <p.raghav@samsung.com>
Link: https://lore.kernel.org/r/20230217121442.33914-1-p.raghav@samsung.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>

diff --git a/drivers/block/brd.c b/drivers/block/brd.c
index 740631d..a8a77a1 100644 (file)
--- a/drivers/block/brd.c
+++ b/drivers/block/brd.c
@@ -94,7 +94,7 @@ static int brd_insert_page(struct brd_device *brd, sector_t sector, gfp_t gfp)
        if (!page)
                return -ENOMEM;
 
-       if (gfpflags_allow_blocking(gfp) && radix_tree_preload(gfp)) {
+       if (radix_tree_maybe_preload(gfp)) {
                __free_page(page);
                return -ENOMEM;
        }