[PATCH 2/2] bcachefs: only console_trylock in bch2_print_string_as_lines

[PATCH 2/2] bcachefs: only console_trylock in bch2_print_string_as_lines

[Daniel Vetter]

console_lock is the outermost subsystem lock for a lot of subsystems,
which means get/put_user must nest within. Which means it cannot be
acquired somewhere deeply nested in other locks, and most definitely
not while holding fs locks potentially needed to resolve faults.

console_trylock is the best we can do here.

Including printk folks since even trylock feels realyl iffy here to
me.

Reported-by: syzbot+6cebc1af246fe020a2f0@syzkaller.appspotmail.com
References: https://lore.kernel.org/dri-devel/00000000000026c1ff061cd0de12@google.com/
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Fixes: a8f354284304 ("bcachefs: bch2_print_string_as_lines()")
Cc: <stable@vger.kernel.org> # v6.7+
Cc: Kent Overstreet <kent.overstreet@linux.dev>
Cc: Brian Foster <bfoster@redhat.com>
Cc: linux-bcachefs@vger.kernel.org
Cc: Petr Mladek <pmladek@suse.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: John Ogness <john.ogness@linutronix.de>
Cc: Sergey Senozhatsky <senozhatsky@chromium.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
---
 fs/bcachefs/util.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/fs/bcachefs/util.c b/fs/bcachefs/util.c
index de331dec2a99..02381c653603 100644
--- a/fs/bcachefs/util.c
+++ b/fs/bcachefs/util.c
@@ -255,13 +255,14 @@ void bch2_prt_u64_base2(struct printbuf *out, u64 v)
 void bch2_print_string_as_lines(const char *prefix, const char *lines)
 {
 	const char *p;
+	int locked;
 
 	if (!lines) {
 		printk("%s (null)\n", prefix);
 		return;
 	}
 
-	console_lock();
+	locked = console_trylock();
 	while (1) {
 		p = strchrnul(lines, '\n');
 		printk("%s%.*s\n", prefix, (int) (p - lines), lines);
@@ -269,7 +270,8 @@ void bch2_print_string_as_lines(const char *prefix, const char *lines)
 			break;
 		lines = p + 1;
 	}
-	console_unlock();
+	if (locked)
+		console_unlock();
 }
 
 int bch2_save_backtrace(bch_stacktrace *stack, struct task_struct *task, unsigned skipnr,
-- 
2.45.2

Re: [PATCH 2/2] bcachefs: only console_trylock in bch2_print_string_as_lines

[John Ogness]

On 2024-07-10, Daniel Vetter <daniel.vetter@ffwll.ch> wrote:
> console_lock is the outermost subsystem lock for a lot of subsystems,
> which means get/put_user must nest within. Which means it cannot be
> acquired somewhere deeply nested in other locks, and most definitely
> not while holding fs locks potentially needed to resolve faults.
>
> console_trylock is the best we can do here.
>
> Including printk folks since even trylock feels realyl iffy here to
> me.

Using the console lock here at all is wrong. The console lock does not
prevent other CPUs from calling printk() and inserting lines in between.

There is no way to guarantee a contiguous ringbuffer block using
multiple printk() calls.

The console_lock usage should be removed.

John Ogness

[PATCH] bcachefs: no console_lock in bch2_print_string_as_lines

[Daniel Vetter]

console_lock is the outermost subsystem lock for a lot of subsystems,
which means get/put_user must nest within. Which means it cannot be
acquired somewhere deeply nested in other locks, and most definitely
not while holding fs locks potentially needed to resolve faults.

console_trylock is the best we can do here. But John pointed out on a
previous version that this is futile:

"Using the console lock here at all is wrong. The console lock does not
prevent other CPUs from calling printk() and inserting lines in between.

"There is no way to guarantee a contiguous ringbuffer block using
multiple printk() calls.

"The console_lock usage should be removed."

https://lore.kernel.org/lkml/87frsh33xp.fsf@jogness.linutronix.de/

Do that.

Reported-by: syzbot+6cebc1af246fe020a2f0@syzkaller.appspotmail.com
References: https://lore.kernel.org/dri-devel/00000000000026c1ff061cd0de12@google.com/
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Fixes: a8f354284304 ("bcachefs: bch2_print_string_as_lines()")
Cc: <stable@vger.kernel.org> # v6.7+
Cc: Kent Overstreet <kent.overstreet@linux.dev>
Cc: Brian Foster <bfoster@redhat.com>
Cc: linux-bcachefs@vger.kernel.org
Cc: Petr Mladek <pmladek@suse.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: John Ogness <john.ogness@linutronix.de>
Cc: Sergey Senozhatsky <senozhatsky@chromium.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
--
v2: Dont trylock, drop console_lock entirely
---
 fs/bcachefs/util.c | 3 ---
 1 file changed, 3 deletions(-)

diff --git a/fs/bcachefs/util.c b/fs/bcachefs/util.c
index de331dec2a99..dc891563d502 100644
--- a/fs/bcachefs/util.c
+++ b/fs/bcachefs/util.c
@@ -8,7 +8,6 @@
 
 #include <linux/bio.h>
 #include <linux/blkdev.h>
-#include <linux/console.h>
 #include <linux/ctype.h>
 #include <linux/debugfs.h>
 #include <linux/freezer.h>
@@ -261,7 +260,6 @@ void bch2_print_string_as_lines(const char *prefix, const char *lines)
 		return;
 	}
 
-	console_lock();
 	while (1) {
 		p = strchrnul(lines, '\n');
 		printk("%s%.*s\n", prefix, (int) (p - lines), lines);
@@ -269,7 +267,6 @@ void bch2_print_string_as_lines(const char *prefix, const char *lines)
 			break;
 		lines = p + 1;
 	}
-	console_unlock();
 }
 
 int bch2_save_backtrace(bch_stacktrace *stack, struct task_struct *task, unsigned skipnr,
-- 
2.45.2

Re: [PATCH] bcachefs: no console_lock in bch2_print_string_as_lines

[John Ogness]

On 2024-07-10, Daniel Vetter <daniel.vetter@ffwll.ch> wrote:
> console_lock is the outermost subsystem lock for a lot of subsystems,
> which means get/put_user must nest within. Which means it cannot be
> acquired somewhere deeply nested in other locks, and most definitely
> not while holding fs locks potentially needed to resolve faults.
>
> console_trylock is the best we can do here. But John pointed out on a
> previous version that this is futile:
>
> "Using the console lock here at all is wrong. The console lock does not
> prevent other CPUs from calling printk() and inserting lines in between.
>
> "There is no way to guarantee a contiguous ringbuffer block using
> multiple printk() calls.
>
> "The console_lock usage should be removed."
>
> https://lore.kernel.org/lkml/87frsh33xp.fsf@jogness.linutronix.de/
>
> Do that.
>
> Reported-by: syzbot+6cebc1af246fe020a2f0@syzkaller.appspotmail.com
> References: https://lore.kernel.org/dri-devel/00000000000026c1ff061cd0de12@google.com/
> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
> Fixes: a8f354284304 ("bcachefs: bch2_print_string_as_lines()")

Reviewed-by: John Ogness <john.ogness@linutronix.de>

Re: [PATCH] bcachefs: no console_lock in bch2_print_string_as_lines

[John Ogness]

On 2024-07-10, Daniel Vetter <daniel.vetter@ffwll.ch> wrote:
> console_lock is the outermost subsystem lock for a lot of subsystems,
> which means get/put_user must nest within. Which means it cannot be
> acquired somewhere deeply nested in other locks, and most definitely
> not while holding fs locks potentially needed to resolve faults.
>
> console_trylock is the best we can do here. But John pointed out on a
> previous version that this is futile:
>
> "Using the console lock here at all is wrong. The console lock does not
> prevent other CPUs from calling printk() and inserting lines in between.
>
> "There is no way to guarantee a contiguous ringbuffer block using
> multiple printk() calls.
>
> "The console_lock usage should be removed."
>
> https://lore.kernel.org/lkml/87frsh33xp.fsf@jogness.linutronix.de/
>
> Do that.

Note that there is more of this incorrect usage of console lock in:

fs/bcachefs/debug.c:bch2_btree_verify_replica()

fs/bcachefs/bset.c:bch2_dump_btree_node()

from commit 1c6fdbd8f246("bcachefs: Initial commit")

... and its parent bcache:

drivers/md/bcache/debug.c:bch_btree_verify()

drivers/md/bcache/bset.c:bch_dump_bucket()

from commit cafe56359144("bcache: A block layer cache")

These should also be removed. Although Kent should verify that the
console lock is not providing some sort of necessary side-effect
synchronization.

John Ogness

Re: [PATCH] bcachefs: no console_lock in bch2_print_string_as_lines

[Daniel Vetter]

On Wed, Jul 10, 2024 at 04:19:53PM +0206, John Ogness wrote:
> On 2024-07-10, Daniel Vetter <daniel.vetter@ffwll.ch> wrote:
> > console_lock is the outermost subsystem lock for a lot of subsystems,
> > which means get/put_user must nest within. Which means it cannot be
> > acquired somewhere deeply nested in other locks, and most definitely
> > not while holding fs locks potentially needed to resolve faults.
> >
> > console_trylock is the best we can do here. But John pointed out on a
> > previous version that this is futile:
> >
> > "Using the console lock here at all is wrong. The console lock does not
> > prevent other CPUs from calling printk() and inserting lines in between.
> >
> > "There is no way to guarantee a contiguous ringbuffer block using
> > multiple printk() calls.
> >
> > "The console_lock usage should be removed."
> >
> > https://lore.kernel.org/lkml/87frsh33xp.fsf@jogness.linutronix.de/
> >
> > Do that.
> 
> Note that there is more of this incorrect usage of console lock in:
> 
> fs/bcachefs/debug.c:bch2_btree_verify_replica()
> 
> fs/bcachefs/bset.c:bch2_dump_btree_node()
> 
> from commit 1c6fdbd8f246("bcachefs: Initial commit")
> 
> ... and its parent bcache:
> 
> drivers/md/bcache/debug.c:bch_btree_verify()
> 
> drivers/md/bcache/bset.c:bch_dump_bucket()
> 
> from commit cafe56359144("bcache: A block layer cache")
> 
> These should also be removed. Although Kent should verify that the
> console lock is not providing some sort of necessary side-effect
> synchronization.

I'll take a look, at least some of them seem doable to audit without deep
bcachefs understanding. Thanks for pointing them out, I should have looked
a bit more at git grep ...
-Sima
-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch