aboutsummaryrefslogtreecommitdiff
path: root/fs/exofs/ore.c
diff options
context:
space:
mode:
Diffstat (limited to 'fs/exofs/ore.c')
-rw-r--r--fs/exofs/ore.c174
1 files changed, 106 insertions, 68 deletions
diff --git a/fs/exofs/ore.c b/fs/exofs/ore.c
index 49cf230554a..cfc0205d62c 100644
--- a/fs/exofs/ore.c
+++ b/fs/exofs/ore.c
@@ -58,9 +58,12 @@ int ore_verify_layout(unsigned total_comps, struct ore_layout *layout)
layout->parity = 1;
break;
case PNFS_OSD_RAID_PQ:
+ layout->parity = 2;
+ break;
case PNFS_OSD_RAID_4:
default:
- ORE_ERR("Only RAID_0/5 for now\n");
+ ORE_ERR("Only RAID_0/5/6 for now received-enum=%d\n",
+ layout->raid_algorithm);
return -EINVAL;
}
if (0 != (layout->stripe_unit & ~PAGE_MASK)) {
@@ -103,7 +106,7 @@ int ore_verify_layout(unsigned total_comps, struct ore_layout *layout)
layout->max_io_length =
(BIO_MAX_PAGES_KMALLOC * PAGE_SIZE - layout->stripe_unit) *
- layout->group_width;
+ (layout->group_width - layout->parity);
if (layout->parity) {
unsigned stripe_length =
(layout->group_width - layout->parity) *
@@ -112,6 +115,8 @@ int ore_verify_layout(unsigned total_comps, struct ore_layout *layout)
layout->max_io_length /= stripe_length;
layout->max_io_length *= stripe_length;
}
+ ORE_DBGMSG("max_io_length=0x%lx\n", layout->max_io_length);
+
return 0;
}
EXPORT_SYMBOL(ore_verify_layout);
@@ -286,7 +291,8 @@ int ore_get_rw_state(struct ore_layout *layout, struct ore_components *oc,
if (length) {
ore_calc_stripe_info(layout, offset, length, &ios->si);
ios->length = ios->si.length;
- ios->nr_pages = (ios->length + PAGE_SIZE - 1) / PAGE_SIZE;
+ ios->nr_pages = ((ios->offset & (PAGE_SIZE - 1)) +
+ ios->length + PAGE_SIZE - 1) / PAGE_SIZE;
if (layout->parity)
_ore_post_alloc_raid_stuff(ios);
}
@@ -401,7 +407,7 @@ static void _clear_bio(struct bio *bio)
struct bio_vec *bv;
unsigned i;
- __bio_for_each_segment(bv, bio, i, 0) {
+ bio_for_each_segment_all(bv, bio, i) {
unsigned this_count = bv->bv_len;
if (likely(PAGE_SIZE == this_count))
@@ -430,8 +436,12 @@ int ore_check_io(struct ore_io_state *ios, ore_on_dev_error on_dev_error)
if (likely(!ret))
continue;
- if (OSD_ERR_PRI_CLEAR_PAGES == osi.osd_err_pri) {
- /* start read offset passed endof file */
+ if ((OSD_ERR_PRI_CLEAR_PAGES == osi.osd_err_pri) &&
+ per_dev->bio) {
+ /* start read offset passed endof file.
+ * Note: if we do not have bio it means read-attributes
+ * In this case we should return error to caller.
+ */
_clear_bio(per_dev->bio);
ORE_DBGMSG("start read offset passed end of file "
"offset=0x%llx, length=0x%llx\n",
@@ -536,24 +546,28 @@ void ore_calc_stripe_info(struct ore_layout *layout, u64 file_offset,
u64 H = LmodS - G * T;
u32 N = div_u64(H, U);
+ u32 Nlast;
/* "H - (N * U)" is just "H % U" so it's bound to u32 */
u32 C = (u32)(H - (N * U)) / stripe_unit + G * group_width;
+ u32 first_dev = C - C % group_width;
div_u64_rem(file_offset, stripe_unit, &si->unit_off);
si->obj_offset = si->unit_off + (N * stripe_unit) +
(M * group_depth * stripe_unit);
+ si->cur_comp = C - first_dev;
+ si->cur_pg = si->unit_off / PAGE_SIZE;
if (parity) {
u32 LCMdP = lcm(group_width, parity) / parity;
/* R = N % LCMdP; */
u32 RxP = (N % LCMdP) * parity;
- u32 first_dev = C - C % group_width;
si->par_dev = (group_width + group_width - parity - RxP) %
group_width + first_dev;
- si->dev = (group_width + C - RxP) % group_width + first_dev;
+ si->dev = (group_width + group_width + C - RxP) %
+ group_width + first_dev;
si->bytes_in_stripe = U;
si->first_stripe_start = M * S + G * T + N * U;
} else {
@@ -568,6 +582,10 @@ void ore_calc_stripe_info(struct ore_layout *layout, u64 file_offset,
si->length = T - H;
if (si->length > length)
si->length = length;
+
+ Nlast = div_u64(H + si->length + U - 1, U);
+ si->maxdevUnits = Nlast - N;
+
si->M = M;
}
EXPORT_SYMBOL(ore_calc_stripe_info);
@@ -583,13 +601,16 @@ int _ore_add_stripe_unit(struct ore_io_state *ios, unsigned *cur_pg,
int ret;
if (per_dev->bio == NULL) {
- unsigned pages_in_stripe = ios->layout->group_width *
- (ios->layout->stripe_unit / PAGE_SIZE);
- unsigned nr_pages = ios->nr_pages * ios->layout->group_width /
- (ios->layout->group_width -
- ios->layout->parity);
- unsigned bio_size = (nr_pages + pages_in_stripe) /
- ios->layout->group_width;
+ unsigned bio_size;
+
+ if (!ios->reading) {
+ bio_size = ios->si.maxdevUnits;
+ } else {
+ bio_size = (ios->si.maxdevUnits + 1) *
+ (ios->layout->group_width - ios->layout->parity) /
+ ios->layout->group_width;
+ }
+ bio_size *= (ios->layout->stripe_unit / PAGE_SIZE);
per_dev->bio = bio_kmalloc(GFP_KERNEL, bio_size);
if (unlikely(!per_dev->bio)) {
@@ -609,8 +630,12 @@ int _ore_add_stripe_unit(struct ore_io_state *ios, unsigned *cur_pg,
added_len = bio_add_pc_page(q, per_dev->bio, pages[pg],
pglen, pgbase);
if (unlikely(pglen != added_len)) {
- ORE_DBGMSG("Failed bio_add_pc_page bi_vcnt=%u\n",
- per_dev->bio->bi_vcnt);
+ /* If bi_vcnt == bi_max then this is a SW BUG */
+ ORE_DBGMSG("Failed bio_add_pc_page bi_vcnt=0x%x "
+ "bi_max=0x%x BIO_MAX=0x%x cur_len=0x%x\n",
+ per_dev->bio->bi_vcnt,
+ per_dev->bio->bi_max_vecs,
+ BIO_MAX_PAGES_KMALLOC, cur_len);
ret = -ENOMEM;
goto out;
}
@@ -632,6 +657,43 @@ out: /* we fail the complete unit on an error eg don't advance
return ret;
}
+static int _add_parity_units(struct ore_io_state *ios,
+ struct ore_striping_info *si,
+ unsigned dev, unsigned first_dev,
+ unsigned mirrors_p1, unsigned devs_in_group,
+ unsigned cur_len)
+{
+ unsigned do_parity;
+ int ret = 0;
+
+ for (do_parity = ios->layout->parity; do_parity; --do_parity) {
+ struct ore_per_dev_state *per_dev;
+
+ per_dev = &ios->per_dev[dev - first_dev];
+ if (!per_dev->length && !per_dev->offset) {
+ /* Only/always the parity unit of the first
+ * stripe will be empty. So this is a chance to
+ * initialize the per_dev info.
+ */
+ per_dev->dev = dev;
+ per_dev->offset = si->obj_offset - si->unit_off;
+ }
+
+ ret = _ore_add_parity_unit(ios, si, per_dev, cur_len,
+ do_parity == 1);
+ if (unlikely(ret))
+ break;
+
+ if (do_parity != 1) {
+ dev = ((dev + mirrors_p1) % devs_in_group) + first_dev;
+ si->cur_comp = (si->cur_comp + 1) %
+ ios->layout->group_width;
+ }
+ }
+
+ return ret;
+}
+
static int _prepare_for_striping(struct ore_io_state *ios)
{
struct ore_striping_info *si = &ios->si;
@@ -641,7 +703,6 @@ static int _prepare_for_striping(struct ore_io_state *ios)
unsigned devs_in_group = group_width * mirrors_p1;
unsigned dev = si->dev;
unsigned first_dev = dev - (dev % devs_in_group);
- unsigned dev_order;
unsigned cur_pg = ios->pages_consumed;
u64 length = ios->length;
int ret = 0;
@@ -653,16 +714,13 @@ static int _prepare_for_striping(struct ore_io_state *ios)
BUG_ON(length > si->length);
- dev_order = _dev_order(devs_in_group, mirrors_p1, si->par_dev, dev);
- si->cur_comp = dev_order;
- si->cur_pg = si->unit_off / PAGE_SIZE;
-
while (length) {
- unsigned comp = dev - first_dev;
- struct ore_per_dev_state *per_dev = &ios->per_dev[comp];
+ struct ore_per_dev_state *per_dev =
+ &ios->per_dev[dev - first_dev];
unsigned cur_len, page_off = 0;
- if (!per_dev->length) {
+ if (!per_dev->length && !per_dev->offset) {
+ /* First time initialize the per_dev info. */
per_dev->dev = dev;
if (dev == si->dev) {
WARN_ON(dev == si->par_dev);
@@ -671,13 +729,7 @@ static int _prepare_for_striping(struct ore_io_state *ios)
page_off = si->unit_off & ~PAGE_MASK;
BUG_ON(page_off && (page_off != ios->pgbase));
} else {
- if (si->cur_comp > dev_order)
- per_dev->offset =
- si->obj_offset - si->unit_off;
- else /* si->cur_comp < dev_order */
- per_dev->offset =
- si->obj_offset + stripe_unit -
- si->unit_off;
+ per_dev->offset = si->obj_offset - si->unit_off;
cur_len = stripe_unit;
}
} else {
@@ -691,11 +743,9 @@ static int _prepare_for_striping(struct ore_io_state *ios)
if (unlikely(ret))
goto out;
- dev += mirrors_p1;
- dev = (dev % devs_in_group) + first_dev;
-
length -= cur_len;
+ dev = ((dev + mirrors_p1) % devs_in_group) + first_dev;
si->cur_comp = (si->cur_comp + 1) % group_width;
if (unlikely((dev == si->par_dev) || (!length && ios->sp2d))) {
if (!length && ios->sp2d) {
@@ -703,23 +753,16 @@ static int _prepare_for_striping(struct ore_io_state *ios)
* stripe. then operate on parity dev.
*/
dev = si->par_dev;
- }
- if (ios->sp2d)
- /* In writes cur_len just means if it's the
- * last one. See _ore_add_parity_unit.
- */
- cur_len = length;
- per_dev = &ios->per_dev[dev - first_dev];
- if (!per_dev->length) {
- /* Only/always the parity unit of the first
- * stripe will be empty. So this is a chance to
- * initialize the per_dev info.
- */
- per_dev->dev = dev;
- per_dev->offset = si->obj_offset - si->unit_off;
+ /* If last stripe operate on parity comp */
+ si->cur_comp = group_width - ios->layout->parity;
}
- ret = _ore_add_parity_unit(ios, si, per_dev, cur_len);
+ /* In writes cur_len just means if it's the
+ * last one. See _ore_add_parity_unit.
+ */
+ ret = _add_parity_units(ios, si, dev, first_dev,
+ mirrors_p1, devs_in_group,
+ ios->sp2d ? length : cur_len);
if (unlikely(ret))
goto out;
@@ -730,18 +773,14 @@ static int _prepare_for_striping(struct ore_io_state *ios)
/* Next stripe, start fresh */
si->cur_comp = 0;
si->cur_pg = 0;
+ si->obj_offset += cur_len;
+ si->unit_off = 0;
}
}
out:
ios->numdevs = devs_in_group;
ios->pages_consumed = cur_pg;
- if (unlikely(ret)) {
- if (length == ios->length)
- return ret;
- else
- ios->length -= length;
- }
- return 0;
+ return ret;
}
int ore_create(struct ore_io_state *ios)
@@ -820,8 +859,8 @@ static int _write_mirror(struct ore_io_state *ios, int cur_comp)
struct bio *bio;
if (per_dev != master_dev) {
- bio = bio_kmalloc(GFP_KERNEL,
- master_dev->bio->bi_max_vecs);
+ bio = bio_clone_kmalloc(master_dev->bio,
+ GFP_KERNEL);
if (unlikely(!bio)) {
ORE_DBGMSG(
"Failed to allocate BIO size=%u\n",
@@ -830,7 +869,6 @@ static int _write_mirror(struct ore_io_state *ios, int cur_comp)
goto out;
}
- __bio_clone(bio, master_dev->bio);
bio->bi_bdev = NULL;
bio->bi_next = NULL;
per_dev->offset = master_dev->offset;
@@ -843,11 +881,11 @@ static int _write_mirror(struct ore_io_state *ios, int cur_comp)
bio->bi_rw |= REQ_WRITE;
}
- osd_req_write(or, _ios_obj(ios, dev), per_dev->offset,
- bio, per_dev->length);
+ osd_req_write(or, _ios_obj(ios, cur_comp),
+ per_dev->offset, bio, per_dev->length);
ORE_DBGMSG("write(0x%llx) offset=0x%llx "
"length=0x%llx dev=%d\n",
- _LLU(_ios_obj(ios, dev)->id),
+ _LLU(_ios_obj(ios, cur_comp)->id),
_LLU(per_dev->offset),
_LLU(per_dev->length), dev);
} else if (ios->kern_buff) {
@@ -859,20 +897,20 @@ static int _write_mirror(struct ore_io_state *ios, int cur_comp)
(ios->si.unit_off + ios->length >
ios->layout->stripe_unit));
- ret = osd_req_write_kern(or, _ios_obj(ios, per_dev->dev),
+ ret = osd_req_write_kern(or, _ios_obj(ios, cur_comp),
per_dev->offset,
ios->kern_buff, ios->length);
if (unlikely(ret))
goto out;
ORE_DBGMSG2("write_kern(0x%llx) offset=0x%llx "
"length=0x%llx dev=%d\n",
- _LLU(_ios_obj(ios, dev)->id),
+ _LLU(_ios_obj(ios, cur_comp)->id),
_LLU(per_dev->offset),
_LLU(ios->length), per_dev->dev);
} else {
- osd_req_set_attributes(or, _ios_obj(ios, dev));
+ osd_req_set_attributes(or, _ios_obj(ios, cur_comp));
ORE_DBGMSG2("obj(0x%llx) set_attributes=%d dev=%d\n",
- _LLU(_ios_obj(ios, dev)->id),
+ _LLU(_ios_obj(ios, cur_comp)->id),
ios->out_attr_len, dev);
}
@@ -1105,7 +1143,7 @@ int ore_truncate(struct ore_layout *layout, struct ore_components *oc,
size_attr->attr = g_attr_logical_length;
size_attr->attr.val_ptr = &size_attr->newsize;
- ORE_DBGMSG("trunc(0x%llx) obj_offset=0x%llx dev=%d\n",
+ ORE_DBGMSG2("trunc(0x%llx) obj_offset=0x%llx dev=%d\n",
_LLU(oc->comps->obj.id), _LLU(obj_size), i);
ret = _truncate_mirrors(ios, i * ios->layout->mirrors_p1,
&size_attr->attr);
generated by cgit v1.2.3-70-g09d2 (git 2.46.0) at 2025-11-09 12:09:01 +0000