linux-kernel - [PATCH v2 13/11] tty: splice

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <fvxufyqixohx65lcusrkkfoxs5cnlsuv7kajv6bnngcoewsodx@tarta.nabijaczleweli.xyz>
Date: Sun, 24 Dec 2023 06:01:49 +0100
From: 
	Ahelenia Ziemiańska <nabijaczleweli@...ijaczleweli.xyz>
Cc: Jens Axboe <axboe@...nel.dk>, Christian Brauner <brauner@...nel.org>, 
	Alexander Viro <viro@...iv.linux.org.uk>, linux-fsdevel@...r.kernel.org, linux-kernel@...r.kernel.org, 
	Greg Kroah-Hartman <gregkh@...uxfoundation.org>, Jiri Slaby <jirislaby@...nel.org>, linux-serial@...r.kernel.org
Subject: [PATCH v2 13/11] tty: splice_write: disable

Given:
	cat > ttyW.c <<^D
	#define _GNU_SOURCE
	#include <fcntl.h>
	#include <stdlib.h>
	int main()
	{
		int pt = posix_openpt(O_RDWR);
		grantpt(pt);
		unlockpt(pt);
		int cl = open(ptsname(pt), O_WRONLY);
		for (;;)
			splice(0, 0, cl, 0, 128 * 1024 * 1024, 0);
	}
	^D
	cc ttyW.c -o ttyW
	mkfifo fifo
	truncate 32M 32M
	./ttyW < fifo &
	cp 32M fifo &
	sleep 0.1
	read -r _ < fifo
ttyW used to sleep in splice and the shell used to enter an
uninterruptible sleep in open("fifo");
now the splice returns -EINVAL and the whole program completes.

This is also symmetric with the splice_read removal.

Signed-off-by: Ahelenia Ziemiańska <nabijaczleweli@...ijaczleweli.xyz>
---
It hit me that I should actually probably exhaustively re-evaluate
splice_write as well since re-evaluating splice_read went so well.

fs/fuse/dev.c:  .splice_write   = fuse_dev_splice_write,
drivers/char/virtio_console.c:  .splice_write = port_fops_splice_write,
locks, takes some pages, unlocks, writes, so OK

drivers/char/mem.c:	.splice_write	= splice_write_null,
drivers/char/mem.c:	.splice_write	= splice_write_zero,
no-op

drivers/char/random.c:	.splice_write = iter_file_splice_write,
drivers/char/random.c:	.splice_write = iter_file_splice_write,
AFAICT write_pool_user is okay to invoke like this?

net/socket.c:   .splice_write = splice_to_socket,
already dealt with in 11/11

drivers/tty/tty_io.c:   .splice_write   = iter_file_splice_write,
drivers/tty/tty_io.c:   .splice_write   = iter_file_splice_write,
they do lock the pipe and try the write with the lock held;
we already killed splice_read so just kill splice_write for symmetry
(13/11)

fs/fuse/file.c: .splice_write   = iter_file_splice_write,
same logic as splice_read applies (14/11)

 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 50c2957a9c7f..d931c34ddcbf 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_write	= iter_file_splice_write,
 	.poll		= tty_poll,
 	.unlocked_ioctl	= tty_ioctl,
 	.compat_ioctl	= tty_compat_ioctl,
@@ -479,7 +478,6 @@ static const struct file_operations console_fops = {
 	.llseek		= no_llseek,
 	.read_iter	= tty_read,
 	.write_iter	= redirected_tty_write,
-	.splice_write	= iter_file_splice_write,
 	.poll		= tty_poll,
 	.unlocked_ioctl	= tty_ioctl,
 	.compat_ioctl	= tty_compat_ioctl,
-- 
2.39.2

Download attachment "signature.asc" of type "application/pgp-signature" (834 bytes)