c/mirror_zfs
1
0
Fork 0
mirror of https://git.proxmox.com/git/mirror_zfs.git synced 2026-05-22 18:40:43 +03:00
Code Issues Packages Projects Releases Wiki Activity

Implement allocation size ranges and use for gang leaves (#17111)

Browse Source 
When forced to resort to ganging, ZFS currently allocates three child
blocks, each one third of the size of the original. This is true
regardless of whether larger allocations could be made, which would
allow us to have fewer gang leaves. This improves performance when
fragmentation is high enough to require ganging, but not so high that
all the free ranges are only just big enough to hold a third of the
recordsize. This is also useful for improving the behavior of a future
change to allow larger gang headers.

We add the ability for the allocation codepath to allocate a range of
sizes instead of a single fixed size. We then use this to pre-allocate
the DVAs for the gang children. If those allocations fail, we fall back
to the normal write path, which will likely re-gang.

Signed-off-by: Paul Dagnelie <paul.dagnelie@klarasystems.com>
Co-authored-by: Paul Dagnelie <paul.dagnelie@klarasystems.com>
Reviewed-by: Alexander Motin <mav@FreeBSD.org>
Reviewed-by: Tony Hutter <hutter2@llnl.gov>
This commit is contained in:
Paul Dagnelie
2025-05-02 15:32:18 -07:00
committed by GitHub
parent a7de203c86
commit 246e5883bb
24 changed files with 392 additions and 107 deletions
Download Patch File Download Diff File Expand all files Collapse all files
module/zfs/vdev_draid.c
+12 -9
View File
@@ -578,7 +578,7 @@ vdev_draid_permute_id(vdev_draid_config_t *vdc,
* i.e. vdev_draid_psize_to_asize().
*/
static uint64_t
vdev_draid_asize(vdev_t *vd, uint64_t psize, uint64_t txg)
vdev_draid_psize_to_asize(vdev_t *vd, uint64_t psize, uint64_t txg)
{
(void) txg;
vdev_draid_config_t *vdc = vd->vdev_tsd;
@@ -599,8 +599,9 @@ vdev_draid_asize(vdev_t *vd, uint64_t psize, uint64_t txg)
* Deflate the asize to the psize, this includes stripping parity.
*/
uint64_t
vdev_draid_asize_to_psize(vdev_t *vd, uint64_t asize)
vdev_draid_asize_to_psize(vdev_t *vd, uint64_t asize, uint64_t txg)
{
(void) txg;
vdev_draid_config_t *vdc = vd->vdev_tsd;
ASSERT0(asize % vdc->vdc_groupwidth);
@@ -962,7 +963,7 @@ vdev_draid_map_alloc_row(zio_t *zio, raidz_row_t **rrp, uint64_t io_offset,
vdev_draid_config_t *vdc = vd->vdev_tsd;
uint64_t ashift = vd->vdev_top->vdev_ashift;
uint64_t io_size = abd_size;
uint64_t io_asize = vdev_draid_asize(vd, io_size, 0);
uint64_t io_asize = vdev_draid_psize_to_asize(vd, io_size, 0);
uint64_t group = vdev_draid_offset_to_group(vd, io_offset);
uint64_t start_offset = vdev_draid_group_to_offset(vd, group + 1);
@@ -972,7 +973,7 @@ vdev_draid_map_alloc_row(zio_t *zio, raidz_row_t **rrp, uint64_t io_offset,
*/
if (io_offset + io_asize > start_offset) {
io_size = vdev_draid_asize_to_psize(vd,
start_offset - io_offset);
start_offset - io_offset, 0);
}
/*
@@ -1117,7 +1118,7 @@ vdev_draid_map_alloc(zio_t *zio)
if (size < abd_size) {
vdev_t *vd = zio->io_vd;
io_offset += vdev_draid_asize(vd, size, 0);
io_offset += vdev_draid_psize_to_asize(vd, size, 0);
abd_offset += size;
abd_size -= size;
nrows++;
@@ -1770,7 +1771,7 @@ vdev_draid_need_resilver(vdev_t *vd, const dva_t *dva, size_t psize,
uint64_t phys_birth)
{
uint64_t offset = DVA_GET_OFFSET(dva);
uint64_t asize = vdev_draid_asize(vd, psize, 0);
uint64_t asize = vdev_draid_psize_to_asize(vd, psize, 0);
if (phys_birth == TXG_UNKNOWN) {
/*
@@ -1827,7 +1828,7 @@ vdev_draid_io_verify(vdev_t *vd, raidz_row_t *rr, int col)
zfs_range_seg64_t logical_rs, physical_rs, remain_rs;
logical_rs.rs_start = rr->rr_offset;
logical_rs.rs_end = logical_rs.rs_start +
vdev_draid_asize(vd, rr->rr_size, 0);
vdev_draid_psize_to_asize(vd, rr->rr_size, 0);
raidz_col_t *rc = &rr->rr_col[col];
vdev_t *cvd = vd->vdev_child[rc->rc_devidx];
@@ -2311,7 +2312,8 @@ vdev_ops_t vdev_draid_ops = {
.vdev_op_fini = vdev_draid_fini,
.vdev_op_open = vdev_draid_open,
.vdev_op_close = vdev_draid_close,
.vdev_op_asize = vdev_draid_asize,
.vdev_op_psize_to_asize = vdev_draid_psize_to_asize,
.vdev_op_asize_to_psize = vdev_draid_asize_to_psize,
.vdev_op_min_asize = vdev_draid_min_asize,
.vdev_op_min_alloc = vdev_draid_min_alloc,
.vdev_op_io_start = vdev_draid_io_start,
@@ -2801,7 +2803,8 @@ vdev_ops_t vdev_draid_spare_ops = {
.vdev_op_fini = vdev_draid_spare_fini,
.vdev_op_open = vdev_draid_spare_open,
.vdev_op_close = vdev_draid_spare_close,
.vdev_op_asize = vdev_default_asize,
.vdev_op_psize_to_asize = vdev_default_asize,
.vdev_op_asize_to_psize = vdev_default_psize,
.vdev_op_min_asize = vdev_default_min_asize,
.vdev_op_min_alloc = NULL,
.vdev_op_io_start = vdev_draid_spare_io_start,