| Message ID | 1500973046-14268-1-git-send-email-jk@ozlabs.org |
|---|---|
| State | Awaiting Upstream |
| Headers | show |
diff --git a/include/linux/tty_flip.h b/include/linux/tty_flip.h index c28dd52..15d03a1 100644 --- a/include/linux/tty_flip.h +++ b/include/linux/tty_flip.h @@ -26,7 +26,7 @@ static inline int tty_insert_flip_char(struct tty_port *port, *char_buf_ptr(tb, tb->used++) = ch; return 1; } - return tty_insert_flip_string_flags(port, &ch, &flag, 1); + return tty_insert_flip_string_fixed_flag(port, &ch, flag, 1); } static inline int tty_insert_flip_string(struct tty_port *port,
In change acc0f67f30, we introduced flip buffers that skipped allocation of the flags buffer for characters received with TTY_NORMAL flags. However, the slow path of tty_insert_flip_char() calls tty_insert_flip_string_flags() (providing a flag buffer pointer), which forces the buffer code to allocate a !TTYB_NORMAL buffer. If we took the slow path due to running out of buffer space, rather than seeing !TTY_NORMAL flags, we've needlessly allocated a flags buffer. This change uses tty_insert_flip_string_fixed_flag instead, which will allocate TTYB_NORMAL buffers if flag == TTY_NORMAL. Since we're only inserting one character, it's fine for the flag to be "fixed". Signed-off-by: Jeremy Kerr <jk@ozlabs.org> CC: Peter Hurley <peter@hurleysoftware.com> --- RFC: I'm certainly no expert on the tty layer, and perhaps there's a good reason to always allocate a flags buffer. However, this seems to relieve buffer pressure in my tests. --- include/linux/tty_flip.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)