On 21. 12. 23, 4:09, Ahelenia Ziemiańska wrote:
We request non-blocking I/O in the generic copy_splice_read, but
"the tty layer doesn't actually honor the IOCB_NOWAIT flag for
various historical reasons.". This means that a tty->pipe splice
will happily sleep with the pipe locked forever, and any process
trying to take it (due to an open/read/write/&c.) will enter
uninterruptible sleep.
This also masks inconsistent wake-ups (usually every second line)
when splicing from ttys in icanon mode.
Link: https://lore.kernel.org/linux-fsdevel/CAHk-=wimmqG_wvSRtMiKPeGGDL816n65u=Mq2+H3-=uM2U6FmA@xxxxxxxxxxxxxx/
Signed-off-by: Ahelenia Ziemiańska <nabijaczleweli@xxxxxxxxxxxxxxxxxx>
---
drivers/tty/tty_io.c | 2 --
1 file changed, 2 deletions(-)
diff --git a/drivers/tty/tty_io.c b/drivers/tty/tty_io.c
index 06414e43e0b5..50c2957a9c7f 100644
--- a/drivers/tty/tty_io.c
+++ b/drivers/tty/tty_io.c
@@ -465,7 +465,6 @@ static const struct file_operations tty_fops = {
.llseek = no_llseek,
.read_iter = tty_read,
.write_iter = tty_write,
- .splice_read = copy_splice_read,
.splice_write = iter_file_splice_write,
This and the other patch effectively reverts dd78b0c483e33 and
9bb48c82aced0. I.e. it breaks "things". Especially:
commit 9bb48c82aced07698a2d08ee0f1475a6c4f6b266
Author: Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx>
Date: Tue Jan 19 11:41:16 2021 -0800
tty: implement write_iter
This makes the tty layer use the .write_iter() function instead of the
traditional .write() functionality.
That allows writev(), but more importantly also makes it possible to
enable .splice_write() for ttys, reinstating the "splice to tty"
functionality that was lost in commit 36e2c7421f02 ("fs: don't allow
splice read/write without explicit ops").
Fixes: 36e2c7421f02 ("fs: don't allow splice read/write without
explicit ops")
What are those "things" doing that "splice to tty", I don't recall and
the commit message above ^^^ does not spell that out. Linus?
thanks,
--
js
