Skip to contentSkip to conversation

dm: fix uninitialized variable reference

PATCH In review Arnd Bergmann 2 participants 3 messages Dec 11, 2017 Device mapper

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 dm: fix uninitialized variable reference

The last bugfix apparently introduced another problem, as shown by this gcc warning:

drivers/md/dm.c: In function '__send_changing_extent_only':
drivers/md/dm.c:1365:28: error: 'ti' is used uninitialized in this function [-Werror=uninitialized]

This restores the intialization of the 'ti' variable.

Fixes aecefd4919de ("dm: fix __send_changing_extent_only() to send first bio and chain remainder")

---
I did not test this or give the patch much thought, this just seemed to be the most likely fix I could come up with in a short time, so please review carefully, and ignore if the solution is something else.

---
 drivers/md/dm.c | 4 ++++
 1 file changed, 4 insertions(+)
drivers/md/dm.c +4 -0
index 05aa9c094352..f19e9787076e 100644
@@ -1356,6 +1356,10 @@ static int __send_changing_extent_only(struct clone_info *ci,
 	unsigned len;
 	unsigned num_bios;
 
+	ti = dm_table_find_target(ci->map, ci->sector);
+	if (!dm_target_is_valid(ti))
+		return -EIO;
+
 	/*
 	 * Even though the device advertised support for this type of
 	 * request, that does not mean every target supports it, and
Signature · 1 line
2.9.0
#1 lore raw
Quoted 37 lines · Arnd Bergmann

The last bugfix apparently introduced another problem, as shown by this gcc warning:

drivers/md/dm.c: In function '__send_changing_extent_only':
drivers/md/dm.c:1365:28: error: 'ti' is used uninitialized in this function [-Werror=uninitialized]

This restores the intialization of the 'ti' variable.

Fixes aecefd4919de ("dm: fix __send_changing_extent_only() to send first bio and chain remainder")

---
I did not test this or give the patch much thought, this just seemed to be the most likely fix I could come up with in a short time, so please review carefully, and ignore if the solution is something else.

---
 drivers/md/dm.c | 4 ++++
 1 file changed, 4 insertions(+)
drivers/md/dm.c +4 -0
index 05aa9c094352..f19e9787076e 100644
@@ -1356,6 +1356,10 @@ static int __send_changing_extent_only(struct clone_info *ci,
 	unsigned len;
 	unsigned num_bios;
 
+	ti = dm_table_find_target(ci->map, ci->sector);
+	if (!dm_target_is_valid(ti))
+		return -EIO;
+
 	/*
 	 * Even though the device advertised support for this type of
 	 * request, that does not mean every target supports it, and
Signature · 1 line
2.9.0

Already resolved this thanks to Stephen Rothwell's earlier (substantially more discrete) mail.

I always enjoy a good public shaming but this cc list is particularly wide. Why?

Anyway, I resolved this differently:
https://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm.git/commit/?h=dm-4.16&id=496cc64f4d10ba112aa0679a3bf29ae6ea74ff3d

#2 lore raw
Quoted 9 lines · Mike Snitzer
Quoted 1 line · Arnd Bergmann

Already resolved this thanks to Stephen Rothwell's earlier (substantially more discrete) mail.

I always enjoy a good public shaming but this cc list is particularly wide. Why?

;-)

Sorry about that, I just used the defaults that scripts/get_maintainer.pl offered,
which is often easier than coming up with a smaller distribution list. You are right this was a lot more than necessary in this case.

Quoted 2 lines

Anyway, I resolved this differently:
https://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm.git/commit/?h=dm-4.16&id=496cc64f4d10ba112aa0679a3bf29ae6ea74ff3d

Ok, thanks!

Arnd