| Message ID | 20220720142456.1414262-3-hch@lst.de (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [1/5] block: move ->bio_split to the gendisk | expand |
On 7/20/22 23:24, Christoph Hellwig wrote: > As we start out with a default of 0, this needs a min_not_zero to > actually work. > > Signed-off-by: Christoph Hellwig <hch@lst.de> > --- > block/blk-settings.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/block/blk-settings.c b/block/blk-settings.c > index 8bb9eef5310eb..9f6e271ca67f4 100644 > --- a/block/blk-settings.c > +++ b/block/blk-settings.c > @@ -554,7 +554,7 @@ int blk_stack_limits(struct queue_limits *t, struct queue_limits *b, > t->max_dev_sectors = min_not_zero(t->max_dev_sectors, b->max_dev_sectors); > t->max_write_zeroes_sectors = min(t->max_write_zeroes_sectors, > b->max_write_zeroes_sectors); > - t->max_zone_append_sectors = min(t->max_zone_append_sectors, > + t->max_zone_append_sectors = min_not_zero(t->max_zone_append_sectors, > b->max_zone_append_sectors); Hmmm... Given that max_zone_append_sectors should never be zero for any zoned block device, that is OK. However, DM targets combining zoned and non-zoned devices to create a non zoned logical drive, e.g. dm-zoned with a regular ssd for metadata, should not have a non-zero max_zone_append_sectors. So I am not confident this change leads to correct limits in all cases. > t->bounce = max(t->bounce, b->bounce); >
On Fri, Jul 22, 2022 at 02:59:30PM +0900, Damien Le Moal wrote: > Hmmm... Given that max_zone_append_sectors should never be zero for any > zoned block device, that is OK. However, DM targets combining zoned and > non-zoned devices to create a non zoned logical drive, e.g. dm-zoned with > a regular ssd for metadata, should not have a non-zero > max_zone_append_sectors. So I am not confident this change leads to > correct limits in all cases. Good point. I think we can drop this patch, as I just need to call blk_set_stacking_limits instead of blk_set_default_limits in the btrfs code, which should sort all this out.
diff --git a/block/blk-settings.c b/block/blk-settings.c index 8bb9eef5310eb..9f6e271ca67f4 100644 --- a/block/blk-settings.c +++ b/block/blk-settings.c @@ -554,7 +554,7 @@ int blk_stack_limits(struct queue_limits *t, struct queue_limits *b, t->max_dev_sectors = min_not_zero(t->max_dev_sectors, b->max_dev_sectors); t->max_write_zeroes_sectors = min(t->max_write_zeroes_sectors, b->max_write_zeroes_sectors); - t->max_zone_append_sectors = min(t->max_zone_append_sectors, + t->max_zone_append_sectors = min_not_zero(t->max_zone_append_sectors, b->max_zone_append_sectors); t->bounce = max(t->bounce, b->bounce);
As we start out with a default of 0, this needs a min_not_zero to actually work. Signed-off-by: Christoph Hellwig <hch@lst.de> --- block/blk-settings.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)