Donate to e Foundation | Murena handsets with /e/OS | Own a part of Murena! Learn more

Commit 8747d30a authored by Lars Ellenberg's avatar Lars Ellenberg Committed by Philipp Reisner
Browse files

drbd: a few more GFP_KERNEL -> GFP_NOIO



This has not yet been observed, but conceivably, when using GFP_KERNEL
allocations from drbd_md_sync(), drbd_flush_after_epoch() or
receive_SyncParam(), we could trigger additional IO to our own device,
or an other device in a criss-cross setup, and end up in a local
deadlock, or potentially a distributed deadlock in a criss-cross setup
involving the peer blocked in a similar way waiting for us to make
progress.

Signed-off-by: default avatarPhilipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: default avatarLars Ellenberg <lars.ellenberg@linbit.com>
parent bc891c9a
Loading
Loading
Loading
Loading
+1 −1
Original line number Diff line number Diff line
@@ -2294,7 +2294,7 @@ static inline void drbd_md_flush(struct drbd_conf *mdev)
	if (test_bit(MD_NO_FUA, &mdev->flags))
		return;

	r = blkdev_issue_flush(mdev->ldev->md_bdev, GFP_KERNEL, NULL);
	r = blkdev_issue_flush(mdev->ldev->md_bdev, GFP_NOIO, NULL);
	if (r) {
		set_bit(MD_NO_FUA, &mdev->flags);
		dev_err(DEV, "meta data flush failed with status %d, disabling md-flushes\n", r);
+1 −1
Original line number Diff line number Diff line
@@ -473,7 +473,7 @@ struct fifo_buffer *fifo_alloc(int fifo_size)
{
	struct fifo_buffer *fb;

	fb = kzalloc(sizeof(struct fifo_buffer) + sizeof(int) * fifo_size, GFP_KERNEL);
	fb = kzalloc(sizeof(struct fifo_buffer) + sizeof(int) * fifo_size, GFP_NOIO);
	if (!fb)
		return NULL;