On Fri, Feb 23, 2024 at 04:30:14PM +0530, Nilay Shroff wrote:
> I think if fatal signal is intercepted while running __blkdev_issue_write_zeroes() then we
> shouldn't need to re-enter the __blkdev_issue_zero_pages(). We may want to add following code:
>
> @@ -280,7 +306,7 @@ int blkdev_issue_zeroout(struct block_device *bdev, sector_t sector,
> bio_put(bio);
> }
> blk_finish_plug(&plug);
> - if (ret && try_write_zeroes) {
> + if (ret && ret != -EINTR && try_write_zeroes) {
> if (!(flags & BLKDEV_ZERO_NOFALLBACK)) {
> try_write_zeroes = false;
> goto retry;
Good point, I'll fold this in.
