drop 22 patches based on RC review feedback
Dropped patches:
- "thunderbolt: Prevent XDomain delayed work use-after-free on disconnect"
Upstream:
2c5d2d3c3f70cde2565d7b279b544893a2035842
Queues: 7.1, 6.18, 6.12, 6.6
Reason: the queued backport contains none of the upstream fix -- it is a
semantically inert relocation of the unmodified update_xdomain() with no
upstream provenance. Upstream is 5 in-place guard hunks in xdomain.c plus
2 in include/linux/thunderbolt.h (xd->removing set under xd->lock in
tb_xdomain_remove(), checked at each queue_delayed_work() site); none of
that is present, so the use-after-free is not fixed. Requeue candidate:
the real upstream commit applies cleanly to all four branches.
Report: https://lore.kernel.org/stable/
2eae07c3-848e-4bc8-a444-
f8274692f653@oracle.com/
- "thunderbolt: Keep XDomain reference during the lifetime of a service" [dependency]
Upstream:
8b4060998637f06975fceee9b73845d8672d411e
Queues: 7.1, 6.18, 6.12, 6.6
Reason: depends on dropped patch above (Stable-dep-of:
2c5d2d3c3f70)
- "thunderbolt: Remove service debugfs entries during unregister" [dependency]
Upstream:
4d5fc3f4068568dfcb8cbe2852b4adc56394aa26
Queues: 7.1, 6.18, 6.12, 6.6
Reason: depends on dropped patch above (Stable-dep-of:
2c5d2d3c3f70)
- "thunderbolt: Remove XDomain from the bus without holding tb->lock" [dependency]
Upstream:
a8937f35cf39c39c64325aa84d0463d866850857
Queues: 7.1, 6.18, 6.12, 6.6
Reason: depends on dropped patch above (Stable-dep-of:
2c5d2d3c3f70)
- "thunderbolt: Handle lane bonding of Gen 4 XDomain links properly" [dependency]
Upstream:
36b6ad6ad0350554e611a8cb754ccd40857416a8
Queues: 6.6
Reason: depends on dropped patch above (Stable-dep-of:
2c5d2d3c3f70)
- "thunderbolt: Remove usage of the deprecated ida_simple_xx() API" [dependency]
Upstream:
dec6a613574cd3dea799170b7aaa8fd76e22f176
Queues: 6.6
Reason: depends on dropped patch above (Stable-dep-of:
2c5d2d3c3f70)
- "thunderbolt: Update property.c function documentation" [dependency]
Upstream:
d015642ad36d78e6eba12d8ab96cea6fd4602b49
Queues: 6.6
Reason: depends on dropped patch above (Stable-dep-of:
2c5d2d3c3f70)
- "bpf: Refactor {acquire,release}_reference_state"
Upstream:
769b0f1c821455ab29baf42491e1ea1d726451fa
Queues: 6.12
Reason: spurious dependency -- the queued patch carries none of
769b0f1c8214's semantics (no acquire_reference(), no
release_reference_nomark()) and has zero references to the fix it is
supposed to enable. Its diff is actually the first half of
201b62ccc831, which was adapted for 6.12 and split across two output
slots while the SHA metadata stayed aligned to the original commit list.
Report: https://lore.kernel.org/stable/
129ad991-e9fa-4f6b-bc1a-
a1d78eff6145@oracle.com/
- "bpf: Refactor check_ctx_access()" [dependency]
Upstream:
201b62ccc83153d2925d310a2afe762905e0c455
Queues: 6.12
Reason: carries the second half of the mis-split adaptation above and will
not apply without it
- "bpf: Reset register bounds before narrowing retval range in check_mem_access()" [dependency]
Upstream:
5e0b273e0a62cc04ec338c7b502797c66c2ed42a
Queues: 6.12
Reason: the target fix -- has a hard textual dependency on the mis-split
refactor above. Will be requeued next cycle with a corrected dependency
set.
Signed-off-by: Sasha Levin <sashal@kernel.org>