The Linux Kernel Mailing List
 help / color / mirror / Atom feed
* [PATCH v2 0/8] char: xillybus: Fix issues reported by Sashiko
@ 2026-11-17 11:58 Eli Billauer
  2026-11-17 11:58 ` [PATCH v2 1/8] char: xillybus: Improve sanity check Eli Billauer
                   ` (7 more replies)
  0 siblings, 8 replies; 9+ messages in thread
From: Eli Billauer @ 2026-11-17 11:58 UTC (permalink / raw)
  To: gregkh; +Cc: arnd, linux-kernel, Eli Billauer

This series is a follow-up to the previously submitted patchset [1].
It addresses Sashiko review comments on surrounding code. This v2 series
is the result of Sashiko's comments on v1 [2].

Two Sashiko comments from [2] are not addressed in this patch set:

(1) "Can unbinding the device while a file descriptor is open cause a
    use-after-free in xillybus_write()?" This remark is correct, but
    will be fixed in a future set of patches, which makes the PCIe
    driver hot-pluggable.

(2) "Should autoflush_skipped only be consumed when
    channel->dont_autoflush reaches zero?" Sashiko correctly recognizes
    that if two concurrent calls to xillybus_write() are made, both of
    them might block indefinitely in a situation where one of them could
    have returned. Concurrent calls to xillybus_write() are not an
    expected usage scenario, and blocking indefinitely when the buffers
    are full isn't wrong, only suboptimal. The user will never know one
    of the calls is blocking in vain. On the other hand, postponing the
    compensation of the missed autoflush until the last xillybus_write()
    returns can theoretically postpone it forever if the user application
    always has a few such calls in flight, effectively disabling the
    autoflush mechanism. Either way, the usage scenarios are bizarre, and
    suboptimal performance is preferred over breaking the usage model.

Regards,
  Eli

[1] https://lore.kernel.org/all/20261009074557.19996-1-eli.billauer@gmail.com/
[2] https://sashiko.dev/#/patchset/20261109112800.48663-1-eli.billauer@gmail.com

Eli Billauer (8):
  char: xillybus: Improve sanity check
  char: xillybus: Safer access of unaligned pointers
  char: xillybus: Return error in fops read method handler
  char: xillybus: Rely on FMODE flags in poll method
  char: xillybus: Refactor flush function
  char: xillybus: Correct handling of wait queue
  char: xillybus: Avoid indefinite sleep of work item
  char: xillybus: Cancel delayed work items before teardown

 drivers/char/xillybus/xillybus.h      |   3 +
 drivers/char/xillybus/xillybus_core.c | 109 ++++++++++++-----
 drivers/char/xillybus/xillyusb.c      | 166 +++++++++++---------------
 3 files changed, 153 insertions(+), 125 deletions(-)

-- 
2.17.1


^ permalink raw reply	[flat|nested] 9+ messages in thread

end of thread, other threads:[~2026-11-17 11:59 UTC | newest]

Thread overview: 9+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-11-17 11:58 [PATCH v2 0/8] char: xillybus: Fix issues reported by Sashiko Eli Billauer
2026-11-17 11:58 ` [PATCH v2 1/8] char: xillybus: Improve sanity check Eli Billauer
2026-11-17 11:58 ` [PATCH v2 2/8] char: xillybus: Safer access of unaligned pointers Eli Billauer
2026-11-17 11:58 ` [PATCH v2 3/8] char: xillybus: Return error in fops read method handler Eli Billauer
2026-11-17 11:58 ` [PATCH v2 4/8] char: xillybus: Rely on FMODE flags in poll method Eli Billauer
2026-11-17 11:58 ` [PATCH v2 5/8] char: xillybus: Refactor flush function Eli Billauer
2026-11-17 11:59 ` [PATCH v2 6/8] char: xillybus: Correct handling of wait queue Eli Billauer
2026-11-17 11:59 ` [PATCH v2 7/8] char: xillybus: Avoid indefinite sleep of work item Eli Billauer
2026-11-17 11:59 ` [PATCH v2 8/8] char: xillybus: Cancel delayed work items before teardown Eli Billauer

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox