pve-qemu/debian/patches/pve/0007-PVE-Up-glusterfs-allow...

78 lines
2.9 KiB
Diff
Raw Normal View History

From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
2017-04-05 11:49:19 +03:00
From: Wolfgang Bumiller <w.bumiller@proxmox.com>
Date: Mon, 6 Apr 2020 12:16:38 +0200
Subject: [PATCH] PVE: [Up] glusterfs: allow partial reads
2017-04-05 11:49:19 +03:00
This should deal with qemu bug #1644754 until upstream
decides which way to go. The general direction seems to be
away from sector based block APIs and with that in mind, and
when comparing to other network block backends (eg. nfs)
treating partial reads as errors doesn't seem to make much
sense.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
2017-04-05 11:49:19 +03:00
---
block/gluster.c | 10 +++++++++-
1 file changed, 9 insertions(+), 1 deletion(-)
diff --git a/block/gluster.c b/block/gluster.c
index d0011085c4..2df3d6e35d 100644
2017-04-05 11:49:19 +03:00
--- a/block/gluster.c
+++ b/block/gluster.c
update submodule and patches to QEMU 8.0.0 Many changes were necessary this time around: * QAPI was changed to avoid redundant has_* variables, see commit 44ea9d9be3 ("qapi: Start to elide redundant has_FOO in generated C") for details. This affected many QMP commands added by Proxmox too. * Pending querying for migration got split into two functions, one to estimate, one for exact value, see commit c8df4a7aef ("migration: Split save_live_pending() into state_pending_*") for details. Relevant for savevm-async and PBS dirty bitmap. * Some block (driver) functions got converted to coroutines, so the Proxmox block drivers needed to be adapted. * Alloc track auto-detaching during PBS live restore got broken by AioContext-related changes resulting in a deadlock. The current, hacky method was replaced by a simpler one. Stefan apparently ran into a problem with that when he wrote the driver, but there were improvements in the stream job code since then and I didn't manage to reproduce the issue. It's a separate patch "alloc-track: fix deadlock during drop" for now, you can find the details there. * Async snapshot-related changes: - The pending querying got adapted to the above-mentioned split and a patch is added to optimize it/make it more similar to what upstream code does. - Added initialization of the compression counters (for future-proofing). - It's necessary the hold the BQL (big QEMU lock = iothread mutex) during the setup phase, because block layer functions are used there and not doing so leads to racy, hard-to-debug crashes or hangs. It's necessary to change some upstream code too for this, a version of the patch "migration: for snapshots, hold the BQL during setup callbacks" is intended to be upstreamed. - Need to take the bdrv graph read lock before flushing. * hmp_info_balloon was moved to a different file. * Needed to include a new headers from time to time to still get the correct functions. Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
2023-05-15 16:39:53 +03:00
@@ -58,6 +58,7 @@ typedef struct GlusterAIOCB {
2017-04-05 12:38:26 +03:00
int ret;
2017-04-05 11:49:19 +03:00
Coroutine *coroutine;
AioContext *aio_context;
+ bool is_write;
} GlusterAIOCB;
typedef struct BDRVGlusterState {
update submodule and patches to QEMU 8.0.0 Many changes were necessary this time around: * QAPI was changed to avoid redundant has_* variables, see commit 44ea9d9be3 ("qapi: Start to elide redundant has_FOO in generated C") for details. This affected many QMP commands added by Proxmox too. * Pending querying for migration got split into two functions, one to estimate, one for exact value, see commit c8df4a7aef ("migration: Split save_live_pending() into state_pending_*") for details. Relevant for savevm-async and PBS dirty bitmap. * Some block (driver) functions got converted to coroutines, so the Proxmox block drivers needed to be adapted. * Alloc track auto-detaching during PBS live restore got broken by AioContext-related changes resulting in a deadlock. The current, hacky method was replaced by a simpler one. Stefan apparently ran into a problem with that when he wrote the driver, but there were improvements in the stream job code since then and I didn't manage to reproduce the issue. It's a separate patch "alloc-track: fix deadlock during drop" for now, you can find the details there. * Async snapshot-related changes: - The pending querying got adapted to the above-mentioned split and a patch is added to optimize it/make it more similar to what upstream code does. - Added initialization of the compression counters (for future-proofing). - It's necessary the hold the BQL (big QEMU lock = iothread mutex) during the setup phase, because block layer functions are used there and not doing so leads to racy, hard-to-debug crashes or hangs. It's necessary to change some upstream code too for this, a version of the patch "migration: for snapshots, hold the BQL during setup callbacks" is intended to be upstreamed. - Need to take the bdrv graph read lock before flushing. * hmp_info_balloon was moved to a different file. * Needed to include a new headers from time to time to still get the correct functions. Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
2023-05-15 16:39:53 +03:00
@@ -753,8 +754,10 @@ static void gluster_finish_aiocb(struct glfs_fd *fd, ssize_t ret,
2017-04-05 11:49:19 +03:00
acb->ret = 0; /* Success */
} else if (ret < 0) {
acb->ret = -errno; /* Read/Write failed */
+ } else if (acb->is_write) {
+ acb->ret = -EIO; /* Partial write - fail it */
} else {
- acb->ret = -EIO; /* Partial read/write - fail it */
+ acb->ret = 0; /* Success */
}
2017-04-05 12:38:26 +03:00
aio_co_schedule(acb->aio_context, acb->coroutine);
update submodule and patches to QEMU 8.0.0 Many changes were necessary this time around: * QAPI was changed to avoid redundant has_* variables, see commit 44ea9d9be3 ("qapi: Start to elide redundant has_FOO in generated C") for details. This affected many QMP commands added by Proxmox too. * Pending querying for migration got split into two functions, one to estimate, one for exact value, see commit c8df4a7aef ("migration: Split save_live_pending() into state_pending_*") for details. Relevant for savevm-async and PBS dirty bitmap. * Some block (driver) functions got converted to coroutines, so the Proxmox block drivers needed to be adapted. * Alloc track auto-detaching during PBS live restore got broken by AioContext-related changes resulting in a deadlock. The current, hacky method was replaced by a simpler one. Stefan apparently ran into a problem with that when he wrote the driver, but there were improvements in the stream job code since then and I didn't manage to reproduce the issue. It's a separate patch "alloc-track: fix deadlock during drop" for now, you can find the details there. * Async snapshot-related changes: - The pending querying got adapted to the above-mentioned split and a patch is added to optimize it/make it more similar to what upstream code does. - Added initialization of the compression counters (for future-proofing). - It's necessary the hold the BQL (big QEMU lock = iothread mutex) during the setup phase, because block layer functions are used there and not doing so leads to racy, hard-to-debug crashes or hangs. It's necessary to change some upstream code too for this, a version of the patch "migration: for snapshots, hold the BQL during setup callbacks" is intended to be upstreamed. - Need to take the bdrv graph read lock before flushing. * hmp_info_balloon was moved to a different file. * Needed to include a new headers from time to time to still get the correct functions. Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
2023-05-15 16:39:53 +03:00
@@ -1021,6 +1024,7 @@ static coroutine_fn int qemu_gluster_co_pwrite_zeroes(BlockDriverState *bs,
2017-04-05 11:49:19 +03:00
acb.ret = 0;
acb.coroutine = qemu_coroutine_self();
acb.aio_context = bdrv_get_aio_context(bs);
+ acb.is_write = true;
ret = glfs_zerofill_async(s->fd, offset, bytes, gluster_finish_aiocb, &acb);
2017-04-05 11:49:19 +03:00
if (ret < 0) {
update submodule and patches to QEMU 8.0.0 Many changes were necessary this time around: * QAPI was changed to avoid redundant has_* variables, see commit 44ea9d9be3 ("qapi: Start to elide redundant has_FOO in generated C") for details. This affected many QMP commands added by Proxmox too. * Pending querying for migration got split into two functions, one to estimate, one for exact value, see commit c8df4a7aef ("migration: Split save_live_pending() into state_pending_*") for details. Relevant for savevm-async and PBS dirty bitmap. * Some block (driver) functions got converted to coroutines, so the Proxmox block drivers needed to be adapted. * Alloc track auto-detaching during PBS live restore got broken by AioContext-related changes resulting in a deadlock. The current, hacky method was replaced by a simpler one. Stefan apparently ran into a problem with that when he wrote the driver, but there were improvements in the stream job code since then and I didn't manage to reproduce the issue. It's a separate patch "alloc-track: fix deadlock during drop" for now, you can find the details there. * Async snapshot-related changes: - The pending querying got adapted to the above-mentioned split and a patch is added to optimize it/make it more similar to what upstream code does. - Added initialization of the compression counters (for future-proofing). - It's necessary the hold the BQL (big QEMU lock = iothread mutex) during the setup phase, because block layer functions are used there and not doing so leads to racy, hard-to-debug crashes or hangs. It's necessary to change some upstream code too for this, a version of the patch "migration: for snapshots, hold the BQL during setup callbacks" is intended to be upstreamed. - Need to take the bdrv graph read lock before flushing. * hmp_info_balloon was moved to a different file. * Needed to include a new headers from time to time to still get the correct functions. Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
2023-05-15 16:39:53 +03:00
@@ -1201,9 +1205,11 @@ static coroutine_fn int qemu_gluster_co_rw(BlockDriverState *bs,
2017-04-05 11:49:19 +03:00
acb.aio_context = bdrv_get_aio_context(bs);
if (write) {
+ acb.is_write = true;
ret = glfs_pwritev_async(s->fd, qiov->iov, qiov->niov, offset, 0,
gluster_finish_aiocb, &acb);
} else {
+ acb.is_write = false;
ret = glfs_preadv_async(s->fd, qiov->iov, qiov->niov, offset, 0,
gluster_finish_aiocb, &acb);
}
update submodule and patches to QEMU 8.0.0 Many changes were necessary this time around: * QAPI was changed to avoid redundant has_* variables, see commit 44ea9d9be3 ("qapi: Start to elide redundant has_FOO in generated C") for details. This affected many QMP commands added by Proxmox too. * Pending querying for migration got split into two functions, one to estimate, one for exact value, see commit c8df4a7aef ("migration: Split save_live_pending() into state_pending_*") for details. Relevant for savevm-async and PBS dirty bitmap. * Some block (driver) functions got converted to coroutines, so the Proxmox block drivers needed to be adapted. * Alloc track auto-detaching during PBS live restore got broken by AioContext-related changes resulting in a deadlock. The current, hacky method was replaced by a simpler one. Stefan apparently ran into a problem with that when he wrote the driver, but there were improvements in the stream job code since then and I didn't manage to reproduce the issue. It's a separate patch "alloc-track: fix deadlock during drop" for now, you can find the details there. * Async snapshot-related changes: - The pending querying got adapted to the above-mentioned split and a patch is added to optimize it/make it more similar to what upstream code does. - Added initialization of the compression counters (for future-proofing). - It's necessary the hold the BQL (big QEMU lock = iothread mutex) during the setup phase, because block layer functions are used there and not doing so leads to racy, hard-to-debug crashes or hangs. It's necessary to change some upstream code too for this, a version of the patch "migration: for snapshots, hold the BQL during setup callbacks" is intended to be upstreamed. - Need to take the bdrv graph read lock before flushing. * hmp_info_balloon was moved to a different file. * Needed to include a new headers from time to time to still get the correct functions. Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
2023-05-15 16:39:53 +03:00
@@ -1266,6 +1272,7 @@ static coroutine_fn int qemu_gluster_co_flush_to_disk(BlockDriverState *bs)
2017-04-05 11:49:19 +03:00
acb.ret = 0;
acb.coroutine = qemu_coroutine_self();
acb.aio_context = bdrv_get_aio_context(bs);
+ acb.is_write = true;
ret = glfs_fsync_async(s->fd, gluster_finish_aiocb, &acb);
if (ret < 0) {
update submodule and patches to QEMU 8.0.0 Many changes were necessary this time around: * QAPI was changed to avoid redundant has_* variables, see commit 44ea9d9be3 ("qapi: Start to elide redundant has_FOO in generated C") for details. This affected many QMP commands added by Proxmox too. * Pending querying for migration got split into two functions, one to estimate, one for exact value, see commit c8df4a7aef ("migration: Split save_live_pending() into state_pending_*") for details. Relevant for savevm-async and PBS dirty bitmap. * Some block (driver) functions got converted to coroutines, so the Proxmox block drivers needed to be adapted. * Alloc track auto-detaching during PBS live restore got broken by AioContext-related changes resulting in a deadlock. The current, hacky method was replaced by a simpler one. Stefan apparently ran into a problem with that when he wrote the driver, but there were improvements in the stream job code since then and I didn't manage to reproduce the issue. It's a separate patch "alloc-track: fix deadlock during drop" for now, you can find the details there. * Async snapshot-related changes: - The pending querying got adapted to the above-mentioned split and a patch is added to optimize it/make it more similar to what upstream code does. - Added initialization of the compression counters (for future-proofing). - It's necessary the hold the BQL (big QEMU lock = iothread mutex) during the setup phase, because block layer functions are used there and not doing so leads to racy, hard-to-debug crashes or hangs. It's necessary to change some upstream code too for this, a version of the patch "migration: for snapshots, hold the BQL during setup callbacks" is intended to be upstreamed. - Need to take the bdrv graph read lock before flushing. * hmp_info_balloon was moved to a different file. * Needed to include a new headers from time to time to still get the correct functions. Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
2023-05-15 16:39:53 +03:00
@@ -1314,6 +1321,7 @@ static coroutine_fn int qemu_gluster_co_pdiscard(BlockDriverState *bs,
2017-04-05 11:49:19 +03:00
acb.ret = 0;
acb.coroutine = qemu_coroutine_self();
acb.aio_context = bdrv_get_aio_context(bs);
+ acb.is_write = true;
ret = glfs_discard_async(s->fd, offset, bytes, gluster_finish_aiocb, &acb);
2017-04-05 11:49:19 +03:00
if (ret < 0) {