drop 7 patches based on RC review feedback

drop 7 patches based on RC review feedback

Dropped patches:

- "rtla: Use str_has_prefix() for prefix checks"
  Queues: 7.0
  Reason: build break -- str_has_prefix() helper (upstream 0f4bc9d67a64)
    not backported to 7.0; reported by RTLA co-maintainer Tomas Glozar.
  Report: https://lore.kernel.org/stable/CAP4=nvTUdVQN8j_qJnfytGOJMa4LaX2h3Z9-cmDopBhKE87_nA@mail.gmail.com/

- "rtla/trace: Fix write loop in trace_event_save_hist()"  [dependency]
  Queues: 7.0
  Reason: queued atop the dropped str_has_prefix conversion; Tomas
    offered to send a rebased version.

- "rtla: Simplify code by caching string lengths"  [dependency]
  Queues: 7.0
  Reason: sibling Stable-dep-of 4bf4ef5292b9; dropped together with the
    other two rtla patches per Sasha's expanded DROP_WITH_DEPS.

- "libbpf: Stringify errno in log messages in libbpf.c"
  Queues: 6.12
  Reason: build break -- errstr() helper not declared in 6.12 (would
    require c68b6fdc3600 plus a chain of refactors that don't apply
    cleanly). Reported by Salvatore Bonaccorso and Florian Fainelli.
  Reports:
    https://lore.kernel.org/stable/ag4vSWzIUCsRlpKv@eldamar.lan/
    https://lore.kernel.org/stable/54f12da1-32e2-48dc-bf84-3bdaf8ef0f6a@gmail.com/

- "libbpf: Prevent double close and leak of btf objects"  [dependency]
  Queues: 6.12
  Reason: adds additional errstr() call sites; cannot stay without
    the stringify patch.

- "libbpf: Change log level of btf loading error message"  [dependency]
  Queues: 6.12
  Reason: only queued as Stable-dep-of 380044c40b16; loses purpose once
    that target is dropped.

- "erofs: verify metadata accesses for file-backed mounts"
  Queues: 7.0, 6.18, 6.12
  Reason: EROFS maintainer / patch author Gao Xiang requested drop
    across all three trees -- introduces an Android SELinux/permission
    regression and the upstream follow-up fix is still under discussion.
  Reports:
    https://lore.kernel.org/stable/75e6d1c8-e989-4eb7-aca3-37a40318e888@linux.alibaba.com/
    https://lore.kernel.org/stable/5a4afec4-fe39-419e-8b2b-4e9901eb93be@linux.alibaba.com/
    https://lore.kernel.org/stable/ag3qlMOcTYM2FBUQ@debian/
    https://lore.kernel.org/stable/ag3txj1lJNTGZQjp@debian/
    https://lore.kernel.org/stable/ag37lbajmtyv9xBd@debian/

- "sched: Make class_schedulers avoid pushing current, and get rid of
  proxy_tag_curr()"
  Queues: 6.18
  Reason: patch author John Stultz asked to drop from 6.18-stable -- it
    only matters for the EXPERT-gated, still-in-development proxy-exec
    feature, and the downside of not taking it is non-ideal balancing.
    Keeping it queued in 7.0 per Stultz's literal request scope.
  Report: https://lore.kernel.org/stable/CANDhNCpZWMk6GWubK8+E0rxKUqtuhOtrjqxunS=Kmho-UbR0UA@mail.gmail.com/

- "riscv: misaligned: Make enabling delegation depend on NONPORTABLE"
  Queues: 7.0, 6.18
  Reason: patch author Vivian Wang retracts her Cc: stable request --
    boot-time crashes seen on 7.1-rc / linux-next on some hardware /
    firmware combinations; asks for time to investigate before
    re-submitting.
  Report: https://lore.kernel.org/stable/99c8c715-b37f-4f2a-8100-5ea4970ff34d@iscas.ac.cn/

- "perf tools: Fix module symbol resolution for non-zero .text sh_addr"
  Queues: 6.6, 6.1, 5.15, 5.10
  Reason: build break -- patch calls dso__rel() inline which doesn't
    exist on these versions (the underlying dso refcount refactor was
    never backported). 7.0/6.18/6.12 have the inline and remain queued.
    Reported by Florian Fainelli.
  Report: https://lore.kernel.org/stable/fbe52f15-5b34-4e03-88e0-005ae6200a60@gmail.com/

Signed-off-by: Sasha Levin <sashal@kernel.org>