Remove a leftover hunk to switch from sequential zones to random
zones when selecting a reclaim zone; the logic has moved into the
caller and this patch is now pointless.
Fixes: 34f5affd04c4 ("dm zoned: separate random and cache zones")
Signed-off-by: Hannes Reinecke unmap_rnd_list) {
-
On 2020/05/22 16:32, Hannes Reinecke wrote:
> Remove a leftover hunk to switch from sequential zones to random
...from random zones to sequential zones...
> zones when selecting a reclaim zone; the logic has moved into the
> caller and this patch is now pointless.
s/this patch/this hunk/ ?
>
On 2020/05/14 15:58, Damien Le Moal wrote:
> When devices with different physical sector sizes are stacked, the
> largest value is used as the stacked device physical sector size. For
> the optimal IO size, the lowest common multiple (lcm) of the underlying
> devices is used for the stacked
>>> + if (t->io_opt & (t->physical_block_size - 1))
>>> + t->io_opt = lcm(t->io_opt, t->physical_block_size);
>
>> Any comment on this patch ? Note: the patch the patch "nvme: Fix
>> io_opt limit setting" is already queued for 5.8.
>
> Setting io_opt to the physical block size is
Instead of having one reclaim workqueue for the entire set we should
be allocating a reclaim workqueue per device; that will reduce
contention and should boost performance for a multi-device setup.
Signed-off-by: Hannes Reinecke
---
drivers/md/dm-zoned-reclaim.c | 70
Random and sequential zones should be part of the respective
device structure to make arbitration between devices possible.
Signed-off-by: Hannes Reinecke
---
drivers/md/dm-zoned-metadata.c | 143 +
drivers/md/dm-zoned.h | 10 +++
2 files
Remove the hard-coded limit of two devices and support an unlimited
number of additional zoned devices.
With that we need to increase the device-mapper version number to
3.0.0 as we've modified the interface.
Signed-off-by: Hannes Reinecke
---
drivers/md/dm-zoned-metadata.c | 68
The secondary superblock must reside on the same device than the
primary superblock, so there's no need to re-calculate the device.
Signed-off-by: Hannes Reinecke
---
drivers/md/dm-zoned-metadata.c | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git
Checking the teriary superblock just consists of validating UUIDs,
crcs, and the generation number; it doesn't have contents which
would be required during the actual operation.
So we should use an on-stack superblock and avoid having to store
it together with the 'real' superblocks.
When reclaiming zones we should arbitrate between the zoned
devices to get a better throughput. So implement a simple
round-robin load balancer between the zoned devices.
Signed-off-by: Hannes Reinecke
---
drivers/md/dm-zoned-metadata.c | 8 +++-
1 file changed, 7 insertions(+), 1
The zones array is getting really large, and large arrays
tend to wreak havoc with the caches.
So convert it to xarray to become more cache friendly.
Signed-off-by: Hannes Reinecke
---
drivers/md/dm-zoned-metadata.c | 98 +++---
1 file changed, 73
Add a pointer to the containing device to struct dm_zone and
kill dmz_zone_to_dev().
Signed-off-by: Hannes Reinecke
---
drivers/md/dm-zoned-metadata.c | 47 --
drivers/md/dm-zoned-reclaim.c | 18 +++-
drivers/md/dm-zoned-target.c | 7
Instead of counting the number of reserved zones in dmz_free_zone()
we should mark the zone as 'reserved' during allocation and simplify
dmz_free_zone().
Signed-off-by: Hannes Reinecke
---
drivers/md/dm-zoned-metadata.c | 4 ++--
drivers/md/dm-zoned.h | 2 ++
2 files changed, 4
Instead of just reporting the errno this patch adds some more
verbose debugging message in the reclaim path.
Signed-off-by: Hannes Reinecke
---
drivers/md/dm-zoned-reclaim.c | 13 ++---
1 file changed, 10 insertions(+), 3 deletions(-)
diff --git a/drivers/md/dm-zoned-reclaim.c
Signed-off-by: Hannes Reinecke
---
drivers/md/dm-zoned-metadata.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/drivers/md/dm-zoned-metadata.c b/drivers/md/dm-zoned-metadata.c
index 4a2e351365c5..b0d3ed4ac56a 100644
--- a/drivers/md/dm-zoned-metadata.c
+++
Hi all,
on the risk of boring you to death, here's yet another RFC to update
dm-zoned. As it has seen only light testing and has some areas which
need to be improved I'd consider it RFC material.
I'm just putting it out now to get some feedback and get it ready for
the next merge window.
So,
Add a metadata pointer to struct dmz_dev and use it as argument
for blkdev_report_zones() instead of the metadata itself.
Signed-off-by: Hannes Reinecke
---
drivers/md/dm-zoned-metadata.c | 14 +-
drivers/md/dm-zoned.h | 7 ---
2 files changed, 13 insertions(+), 8
Damien,
>> +if (t->io_opt & (t->physical_block_size - 1))
>> +t->io_opt = lcm(t->io_opt, t->physical_block_size);
> Any comment on this patch ? Note: the patch the patch "nvme: Fix
> io_opt limit setting" is already queued for 5.8.
Setting io_opt to the physical block size is
18 matches
Mail list logo
