af_unix: Random improvements for GC.
Status from this message: This series was applied to netdev/net-next.git (main)
Get series with b4
b4 am [email protected]Versions newest first · linked by author and title
- v5 Latest5 patches 0 repliesApplied 5/5
- v44 patches 2 repliesSuperseded
- v3 posting 2/20 of 4 patches 0 repliesSuperseded
Earlier versions · 3
| Version | Posted | Patches | Replies | Status |
|---|---|---|---|---|
| v3 posting 1/2 | 4 patches | 0 replies | Superseded | |
| v2 | 4 patches | 7 replies | Superseded | |
| v1 | 4 patches | 0 replies | Superseded |
Changes in v5 from the cover letter
* Rebase on the latest
* Add patch 1
Patches in v5 All 5 applied
| # | Subject and files | Size | Replies | Applied |
|---|---|---|---|---|
| 1/5 | af_unix: Annotate data-race of gc_in_progress in wait_for_unix_gc().net/ | +1 | 0 replies | Applied |
| 2/5 | af_unix: Do not use atomic ops for unix_sk(sk)->inflight.include/ net/ net/ +1 more | +16 | 0 replies | Applied |
| 3/5 | af_unix: Return struct unix_sock from unix_get_socket().include/ net/ net/ | +15 | 0 replies | Applied |
| 4/5 | af_unix: Run GC on only one CPU.net/ | +27 | 0 replies | Applied |
| 5/5 | af_unix: Try to run GC async.include/ include/ net/ +2 more | +29 | 0 replies | Applied |
Reviews trailers per patch
2 of 5 patches have a Reviewed-by or Acked-by.
| Patch | Simon Horman | Pavel Begunkov |
|---|---|---|
| 1/5 | ||
| 2/5 | R-b | |
| 3/5 | R-b | A-b |
| 4/5 | ||
| 5/5 |
Trailers given in replies are filled; those carried in the patches from earlier versions are outlined with dashes.
Cover letter [PATCH v5 net-next 0/5] af_unix: Random improvements for GC.
If more than 16000 inflight AF_UNIX sockets exist on a host, each sendmsg() will be forced to wait for unix_gc() even if a process is not sending any FD.
This series tries not to impose such a penalty on sane users who do not send AF_UNIX FDs or do not have inflight sockets more than SCM_MAX_FD * 8.
The first patch can be backported to -stable.
Cleanup patches for commit 69db702c8387 ("io_uring/af_unix: disable sending io_uring over sockets") and large refactoring of GC will be followed later.
Changes:
v5:
* Rebase on the latest
* Add patch 1
v4: https://lore.kernel.org/netdev/[email protected]/
* Rebase on the latest
v3: https://lore.kernel.org/netdev/[email protected]/
* Patch 3
* Reuse gc_in_progress flag.
* Call flush_work() only when gc is queued or in progress.
* Patch 4
* Bump UNIX_INFLIGHT_SANE_USER to (SCM_MAX_FD * 8).
v2: https://lore.kernel.org/netdev/[email protected]/
* Patch 4
* Fix build error when CONFIG_UNIX=n
v1: https://lore.kernel.org/netdev/[email protected]/
Kuniyuki Iwashima (5):
af_unix: Annotate data-race of gc_in_progress in wait_for_unix_gc().
af_unix: Do not use atomic ops for unix_sk(sk)->inflight.
af_unix: Return struct unix_sock from unix_get_socket().
af_unix: Run GC on only one CPU.
af_unix: Try to run GC async.
include/net/af_unix.h | 14 +++++-- include/net/scm.h | 1 + net/core/scm.c | 5 +++ net/unix/af_unix.c | 10 +++-- net/unix/garbage.c | 98 ++++++++++++++++++++++--------------------- net/unix/scm.c | 27 ++++++------ 6 files changed, 85 insertions(+), 70 deletions(-)
Signature · 1 line
2.30.2