drbd/0015-drbd-atomically-update-queue-limits-in-drbd_reconsid.patch
heming zhao d377bc2ee5 - Update DRBD version from 9.1.16 to 9.1.22
* Changelog from Linbit:
    9.1.22 (api:genl2/proto:86-121/transport:18)
    --------
     * Upgrade from partial resync to a full resync if necessary when the
       user manually resolves a split-brain situation
     * Fix a potential NULL deref when a disk fails while doing a
       forget-peer operation.
     * Fix a rcu_read_lock()/rcu_read_unlock() imbalance
     * Restart the open() syscall when a process auto promoting a drbd device gets
       interrupted by a signal
     * Remove a deadlock that caused DRBD to connect sometimes
       exceptionally slow
     * Make detach operations interruptible
     * Added dev_is_open to events2 status information
     * Improve log readability for 2PC state changes and drbd-threads
     * Updated compability code for Linux 6.9
    
    9.1.21 (api:genl2/proto:86-121/transport:18)
    --------
     * fix a deadlock that can trigger when deleting a connection and
       another connection going down in parallel. This is a regression of
       9.1.20
     * Fix an out-of-bounds access when scanning the bitmap. It leads to a
       crash when the bitmap ends on a page boundary, and this is also a
       regression in 9.1.20.
    
    9.1.20 (api:genl2/proto:86-121/transport:18)
    --------
     * Fix a kernel crash that is sometimes triggered when downing drbd
       resources in a specific, unusual order (was triggered by the
       Kubernetes CSI driver)
     * Fix a rarely triggering kernel crash upon adding paths to a
       connection by rehauling the path lists' locking
     * Fix the continuation of an interrupted initial resync
     * Fix the state engine so that an incapable primary does not outdate
       indirectly reachable secondary nodes
     * Fix a logic bug that caused drbd to pretend that a peer's disk is
       outdated when doing a manual disconnect on a down connection; with
       that cured impact on fencing and quorum.
     * Fix forceful demotion of suspended devices
     * Rehaul of the build system to apply compatibility patches out of
       place that allows one to build for different target kernels from a
       single drbd source tree
     * Updated compability code for Linux 6.8
    
    9.1.19 (api:genl2/proto:86-121/transport:18)
    --------
     * Fix a resync decision case where drbd wrongly decided to do a full
       resync, where a partial resync was sufficient; that happened in a
       specific connect order when all nodes were on the same data
       generation (UUID)
     * Fix the online resize code to obey cached size information about
       temporal unreachable nodes
     * Fix a rare corner case in which DRBD on a diskless primary node
       failed to re-issue a read request to another node with a backing
       disk upon connection loss on the connection where it shipped the
       read request initially
     * Make timeout during promotion attempts interruptible
     * No longer write activity-log updates on the secondary node in a
       cluster with precisely two nodes with backing disk; this is a
       performance optimization
     * Reduce CPU usage of acknowledgment processing
    
    9.1.18 (api:genl2/proto:86-121/transport:18)
    --------
     * Fixed connecting nodes with differently sized backing disks,
       specifically when the smaller node is primary, before establishing
       the connections
     * Fixed thawing a device that has I/O frozen after loss of quorum
       when a configuration change eases its quorum requirements
     * Properly fail TLS if requested (only available in drbd-9.2)
     * Fixed a race condition that can cause auto-demote to trigger right
       after an explicit promote
     * Fixed a rare race condition that could mess up the handshake result
       before it is committed to the replication state.
     * Preserve "tiebreaker quorum" over a reboot of the last node (3-node
       clusters only)
     * Update compatibility code for Linux 6.6
    
    9.1.17 (api:genl2/proto:86-121/transport:18)
    --------
     * fix a potential crash when configuring drbd to bind to a
       non-existent local IP address (this is a regression of drbd-9.1.8)
     * Cure a very seldom triggering race condition bug during
       establishing connections; when you triggered it, you got an OOPS
       hinting to list corruption
     * fix a race condition regarding operations on the bitmap while
       forgetting a bitmap slot and a pointless warning
     * Fix handling of unexpected (on a resource in secondary role) write
       requests
     * Fix a corner case that can cause a process to hang when closing the
       DRBD device, while a connection gets re-established
     * Correctly block signal delivery during auto-demote
     * Improve the reliability of establishing connections
     * Do not clear the transport with `net-options --set-defaults`. This
       fix avoids unexpected disconnect/connect cycles upon an `adjust`
       when using the 'lb-tcp' or 'rdma' transports in drbd-9.2.
     * New netlink packet to report path status to drbdsetup
     * Improvements to the content and rate-limiting of many log messages
     * Update compatibility code and follow Linux upstream development
       until Linux 6.5
  * remove patches which already included in the new version:
     0001-drbd-allow-transports-to-take-additional-krefs-on-a-.patch
     0002-drbd-improve-decision-about-marking-a-failed-disk-Ou.patch
     0003-drbd-fix-error-path-in-drbd_get_listener.patch
     0004-drbd-build-fix-spurious-re-build-attempt-of-compat.p.patch
     0005-drbd-log-error-code-when-thread-fails-to-start.patch
     0006-drbd-log-numeric-value-of-drbd_state_rv-as-well-as-s.patch
     0007-drbd-stop-defining-__KERNEL_SYSCALLS__.patch
     0008-compat-block-introduce-holder-ops.patch
     0009-drbd-reduce-net_ee-not-empty-info-to-a-dynamic-debug.patch
     0010-drbd-do-not-send-P_CURRENT_UUID-to-DRBD-8-peer-when-.patch
     0011-compat-block-pass-a-gendisk-to-open.patch
     0012-drbd-Restore-DATA_CORKED-and-CONTROL_CORKED-bits.patch
     0013-drbd-remove-unused-extern-for-conn_try_outdate_peer.patch
     0014-drbd-include-source-of-state-change-in-log.patch
     0015-compat-block-use-the-holder-as-indication-for-exclus.patch
     0016-drbd-Fix-net-options-set-defaults-to-not-clear-the-t.patch
     0017-drbd-propagate-exposed-UUIDs-only-into-established-c.patch
     0018-drbd-rework-autopromote.patch
     0019-compat-block-remove-the-unused-mode-argument-to-rele.patch
     0020-drbd-do-not-allow-auto-demote-to-be-interrupted-by-s.patch
     0021-compat-sock-Remove-sendpage-in-favour-of-sendmsg-MSG.patch
     0022-compat-block-replace-fmode_t-with-a-block-specific-t.patch
     0023-compat-genetlink-remove-userhdr-from-struct-genl_inf.patch
     0024-compat-fixup-FMODE_READ-FMODE_WRITE-usage.patch
     0025-compat-drdb-Convert-to-use-bdev_open_by_path.patch
     0026-compat-gate-blkdev_-patches-behind-bdev_open_by_path.patch
     boo1230635_01-compat-fix-nla_nest_start_noflag-test.patch
     boo1230635_02-drbd-port-block-device-access-to-file.patch
   
  * removed patches which are not needed anymore:
     boo1229062-re-enable-blk_queue_max_hw_sectors.patch
     bsc1226510-fix-build-err-against-6.9.3.patch
   
  * update:
     drbd_git_revision
     suse-coccinelle.patch
     drbd.spec
   
  * add upstream patches to align commit 13ada1be201e:
     0001-drbd-properly-rate-limit-resync-progress-reports.patch
     0002-drbd-inherit-history-UUIDs-from-sync-source-when-res.patch
     0003-build-compat-fix-line-offset-in-annotation-pragmas-p.patch
     0004-drbd-fix-exposed_uuid-going-backward.patch
     0005-drbd-Proper-locking-around-new_current_uuid-on-a-dis.patch
     0006-build-CycloneDX-fix-bom-ref-add-purl.patch
     0007-build-Another-update-to-the-spdx-files.patch
     0008-build-generate-spdx.json-not-tag-value-format.patch
     0009-compat-fix-gen_patch_names-for-bdev_file_open_by_pat.patch
     0010-compat-fix-nla_nest_start_noflag-test.patch
     0011-compat-fix-blk_alloc_disk-rule.patch
     0012-drbd-remove-const-from-function-return-type.patch
     0013-drbd-don-t-set-max_write_zeroes_sectors-in-decide_on.patch
     0014-drbd-split-out-a-drbd_discard_supported-helper.patch
     0015-drbd-atomically-update-queue-limits-in-drbd_reconsid.patch
     0016-compat-test-and-patch-for-queue_limits_start_update.patch
     0017-compat-specify-which-essential-change-was-not-made.patch
     0018-gen_patch_names-reorder-blk_mode_t.patch
     0019-compat-fix-blk_queue_update_readahead-patch.patch
     0020-compat-test-and-patch-for-que_limits-max_hw_discard_.patch
     0021-compat-fixup-write_zeroes__no_capable.patch
     0022-compat-fixup-queue_flag_discard__yes_present.patch
     0023-drbd-move-flags-to-queue_limits.patch
     0024-compat-test-and-patch-for-queue_limits.features.patch
     0025-drbd-Annotate-struct-fifo_buffer-with-__counted_by.patch
     0026-compat-test-and-patch-for-__counted_by.patch
     0027-drbd-fix-function-cast-warnings-in-state-machine.patch
     0028-Add-missing-documentation-of-peer_device-parameter-t.patch
     0030-drbd-kref_put-path-when-kernel_accept-fails.patch
     0031-build-fix-typo-in-Makefile.spatch.patch
     0032-drbd-open-do-not-delay-open-if-already-Primary.patch
   
  * add patch to fix kernel imcompatibility issue (boo#1231290):
     boo1231290_fix_drbd_build_error_against_kernel_v6.11.0.patch

OBS-URL: https://build.opensuse.org/package/show/network:ha-clustering:Factory/drbd?expand=0&rev=153
2024-10-11 04:45:18 +00:00

195 lines
7.3 KiB
Diff

From 5931286cb71114c0f34e51fc7f357dceec139ca1 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Christoph=20B=C3=B6hmwalder?=
<christoph.boehmwalder@linbit.com>
Date: Tue, 13 Aug 2024 12:26:02 +0200
Subject: [PATCH 15/32] drbd: atomically update queue limits in
drbd_reconsider_queue_parameters
Switch drbd_reconsider_queue_parameters to set up the queue parameters
in an on-stack queue_limits structure and apply the limits atomically.
Remove various helpers that have become so trivial that they can be
folded into drbd_reconsider_queue_parameters.
Same idea as upstream kernel commit e6dfe748f09e ("drbd: atomically
update queue limits in drbd_reconsider_queue_parameters"), but with some
adaptations to DRBD 9.
---
drbd/drbd_nl.c | 136 ++++++++++++++++++++++---------------------------
1 file changed, 60 insertions(+), 76 deletions(-)
diff --git a/drbd/drbd_nl.c b/drbd/drbd_nl.c
index e3f508cd0421..7193ba0c77e3 100644
--- a/drbd/drbd_nl.c
+++ b/drbd/drbd_nl.c
@@ -2018,11 +2018,6 @@ static u32 common_connection_features(struct drbd_resource *resource)
return features;
}
-static void blk_queue_discard_granularity(struct request_queue *q, unsigned int granularity)
-{
- q->limits.discard_granularity = granularity;
-}
-
static unsigned int drbd_max_discard_sectors(struct drbd_resource *resource)
{
struct drbd_connection *connection;
@@ -2059,73 +2054,20 @@ static bool drbd_discard_supported(struct drbd_device *device,
return true;
}
-static void decide_on_discard_support(struct drbd_device *device,
- struct drbd_backing_dev *bdev)
-{
- struct request_queue *q = device->rq_queue;
- unsigned int max_discard_sectors;
-
- if (!drbd_discard_supported(device, bdev)) {
- blk_queue_discard_granularity(q, 0);
- blk_queue_max_discard_sectors(q, 0);
- return;
- }
-
- /*
- * We don't care for the granularity, really.
- *
- * Stacking limits below should fix it for the local device. Whether or
- * not it is a suitable granularity on the remote device is not our
- * problem, really. If you care, you need to use devices with similar
- * topology on all peers.
- */
- blk_queue_discard_granularity(q, 512);
- max_discard_sectors = drbd_max_discard_sectors(device->resource);
- blk_queue_max_discard_sectors(q, max_discard_sectors);
-}
-
-static void fixup_write_zeroes(struct drbd_device *device, struct request_queue *q)
-{
- /* Fixup max_write_zeroes_sectors after blk_stack_limits():
- * if we can handle "zeroes" efficiently on the protocol,
- * we want to do that, even if our backend does not announce
- * max_write_zeroes_sectors itself. */
-
- /* If all peers announce WZEROES support, use it. Otherwise, rather
- * send explicit zeroes than rely on some discard-zeroes-data magic. */
- if (common_connection_features(device->resource) & DRBD_FF_WZEROES)
- q->limits.max_write_zeroes_sectors = DRBD_MAX_BBIO_SECTORS;
- else
- q->limits.max_write_zeroes_sectors = 0;
-}
-
-static void fixup_discard_support(struct drbd_device *device, struct request_queue *q)
-{
- unsigned int max_discard = device->rq_queue->limits.max_discard_sectors;
- unsigned int discard_granularity = device->rq_queue->limits.discard_granularity >> SECTOR_SHIFT;
-
- if (discard_granularity > max_discard) {
- blk_queue_discard_granularity(q, 0);
- blk_queue_max_discard_sectors(q, 0);
- }
-}
-
-void drbd_reconsider_queue_parameters(struct drbd_device *device, struct drbd_backing_dev *bdev)
+static void get_common_queue_limits(struct queue_limits *common_limits,
+ struct drbd_device *device)
{
- struct request_queue * const q = device->rq_queue;
- struct queue_limits common_limits = { 0 }; /* sizeof(struct queue_limits) ~ 110 bytes */
- struct queue_limits peer_limits = { 0 };
struct drbd_peer_device *peer_device;
- struct request_queue *b = NULL;
+ struct queue_limits peer_limits = { 0 };
- blk_set_stacking_limits(&common_limits);
+ blk_set_stacking_limits(common_limits);
/* This is the workaround for "bio would need to, but cannot, be split" */
- common_limits.seg_boundary_mask = PAGE_SIZE - 1;
- common_limits.max_hw_sectors = device->device_conf.max_bio_size >> SECTOR_SHIFT;
- common_limits.max_sectors = device->device_conf.max_bio_size >> SECTOR_SHIFT;
- common_limits.physical_block_size = device->device_conf.block_size;
- common_limits.logical_block_size = device->device_conf.block_size;
- common_limits.io_min = device->device_conf.block_size;
+ common_limits->seg_boundary_mask = PAGE_SIZE - 1;
+ common_limits->max_hw_sectors = device->device_conf.max_bio_size >> SECTOR_SHIFT;
+ common_limits->max_sectors = device->device_conf.max_bio_size >> SECTOR_SHIFT;
+ common_limits->physical_block_size = device->device_conf.block_size;
+ common_limits->logical_block_size = device->device_conf.block_size;
+ common_limits->io_min = device->device_conf.block_size;
rcu_read_lock();
for_each_peer_device_rcu(peer_device, device) {
@@ -2140,21 +2082,63 @@ void drbd_reconsider_queue_parameters(struct drbd_device *device, struct drbd_ba
peer_limits.io_opt = peer_device->q_limits.io_opt;
peer_limits.max_hw_sectors = peer_device->q_limits.max_bio_size >> SECTOR_SHIFT;
peer_limits.max_sectors = peer_device->q_limits.max_bio_size >> SECTOR_SHIFT;
- blk_stack_limits(&common_limits, &peer_limits, 0);
+ blk_stack_limits(common_limits, &peer_limits, 0);
}
rcu_read_unlock();
+}
+
+void drbd_reconsider_queue_parameters(struct drbd_device *device, struct drbd_backing_dev *bdev)
+{
+ struct request_queue * const q = device->rq_queue;
+ struct queue_limits lim;
+ struct request_queue *b = NULL;
+
+ lim = queue_limits_start_update(q);
+ get_common_queue_limits(&lim, device);
+
+ /*
+ * We don't care for the granularity, really.
+ * Stacking limits below should fix it for the local device. Whether or
+ * not it is a suitable granularity on the remote device is not our
+ * problem, really. If you care, you need to use devices with similar
+ * topology on all peers.
+ */
+ if (drbd_discard_supported(device, bdev)) {
+ lim.discard_granularity = 512;
+ lim.max_hw_discard_sectors = drbd_max_discard_sectors(device->resource);
+ } else {
+ lim.discard_granularity = 0;
+ lim.max_hw_discard_sectors = 0;
+ }
if (bdev) {
b = bdev->backing_bdev->bd_disk->queue;
- blk_stack_limits(&common_limits, &b->limits, 0);
- disk_update_readahead(device->vdisk);
+ blk_stack_limits(&lim, &b->limits, 0);
+ }
+
+ /*
+ * If we can handle "zeroes" efficiently on the protocol,
+ * we want to do that, even if our backend does not announce
+ * max_write_zeroes_sectors itself.
+ */
+ if (common_connection_features(device->resource) & DRBD_FF_WZEROES)
+ lim.max_write_zeroes_sectors = DRBD_MAX_BBIO_SECTORS;
+ else
+ lim.max_write_zeroes_sectors = 0;
+
+ if ((lim.discard_granularity >> SECTOR_SHIFT) >
+ lim.max_hw_discard_sectors) {
+ /*
+ * discard_granularity is the smallest supported unit of a
+ * discard. If that is larger than the maximum supported discard
+ * size, we need to disable discards altogether.
+ */
+ lim.discard_granularity = 0;
+ lim.max_hw_discard_sectors = 0;
}
- q->limits = common_limits;
- blk_queue_max_hw_sectors(q, common_limits.max_hw_sectors);
- decide_on_discard_support(device, bdev);
- fixup_write_zeroes(device, q);
- fixup_discard_support(device, q);
+ if (queue_limits_commit_update(q, &lim))
+ drbd_err(device, "setting new queue limits failed\n");
}
/* Make sure IO is suspended before calling this function(). */
--
2.35.3