NAS-139155 / None / nvme-pci: use NSSR to avoid CAP.TO and optimize timeout handling #238
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
bugclerk
changed the title
ixhamza
force-pushed
the
NAS-139155-nssr-reset
branch
2 times, most recently
from
995ffa8 to
cbcfe20
Compare
amotin
approved these changes
Feb 4, 2026
Reduce hung device recovery time using NSSR and optimized timeouts: - Skip abort commands and go directly to reset (following FreeBSD) - Trigger NSSR on hotplug-capable slots for fast device removal - Skip explicit queue deletion during reset (saves admin_timeout) - Detect hung devices after first CAP.TO by checking controller state - Skip re-enable in reset_work for hung devices (saves second CAP.TO) - Remove namespaces immediately for hung devices after DEAD state NSSR path (NSSR supported + hotplug-capable): After io_timeout (30s), trigger NSSR which causes PCIe link down. pciehp detects link state change and calls nvme_remove() for immediate device removal. Total recovery time: ~30s. Non-NSSR path (NSSR not supported or non-hotplug slot): After io_timeout (30s), proceed with controller reset. Hung device detection after first CAP.TO prevents second CAP.TO wait in reset_work. With default CAP.TO (~45s), total recovery time: ~75s. Previously, hung device recovery took io_timeout×2 + admin_timeout + CAP.TO×2. With defaults (io_timeout=30s, admin_timeout=60s, CAP.TO=45s), this was ~210s. Namespace removal for hung devices cleans up stale block devices that previously remained visible despite dead controller, enabling proper drive replacement workflows. Signed-off-by: Ameer Hamza <ahamza@ixsystems.com>
ixhamza
force-pushed
the
NAS-139155-nssr-reset
branch
from
cbcfe20 to
72b5f00
Compare
|
This PR has been merged and conversations have been locked. |
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
3 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Reduce hung device recovery time using
NSSRand optimized timeouts:NSSRon hotplug-capable slots for fast device removaladmin_timeout)CAP.TOby checking controller stateCAP.TO)DEADstateNSSR path (NSSR supported + hotplug-capable):
After io_timeout (30s), trigger NSSR which causes PCIe link down. pciehp detects link state change and calls nvme_remove() for immediate device removal. Total recovery time: ~30s.
Non-NSSR path (NSSR not supported or non-hotplug slot):
After
io_timeout(30s), proceed with controller reset. Hung device detection after firstCAP.TOprevents secondCAP.TOwait inreset_work(). With defaultCAP.TO(~45s), total recovery time: ~75s.Previously, hung device recovery took
io_timeout×2+admin_timeout+CAP.TO×2. With defaults (io_timeout=30s, admin_timeout=60s, CAP.TO=45s), this was ~210s.Namespace removal for hung devices cleans up stale block devices that previously remained visible despite dead controller, enabling proper drive replacement workflows.
Testing