* [PATCH] libmultipath: fix crash in print_foreign_topology
@ 2025-08-20 18:39 Benjamin Marzinski
2025-08-25 21:56 ` Martin Wilck
0 siblings, 1 reply; 2+ messages in thread
From: Benjamin Marzinski @ 2025-08-20 18:39 UTC (permalink / raw)
To: Christophe Varoqui; +Cc: device-mapper development, Martin Wilck
print_foreign_topology() called get_paths() to get a vector of
(struct gen_path *) items and then called get_multipath_layout__(),
which expects a vector of (struct gen_multipath *) items, with the path
vector. This can easily end badly. Fix it to correctly call
get_path_layout__(), and rename width to p_width in the functions that
end up calling snprint_multipath_topology__(), which is expecting to get
passed the path field widths.
Fixes: d54f271ef ("libmultipath: introduce width argument for pretty-printing functions")
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
---
libmultipath/foreign.c | 16 ++++++++--------
1 file changed, 8 insertions(+), 8 deletions(-)
diff --git a/libmultipath/foreign.c b/libmultipath/foreign.c
index 138b2eb3..72fef8b0 100644
--- a/libmultipath/foreign.c
+++ b/libmultipath/foreign.c
@@ -488,7 +488,7 @@ void foreign_multipath_layout(fieldwidth_t *width)
}
static int snprint_foreign_topology__(struct strbuf *buf, int verbosity,
- const fieldwidth_t *width)
+ const fieldwidth_t *p_width)
{
struct foreign *fgn;
int i;
@@ -506,7 +506,7 @@ static int snprint_foreign_topology__(struct strbuf *buf, int verbosity,
if (vec != NULL) {
vector_foreach_slot(vec, gm, j) {
if (snprint_multipath_topology__(
- gm, buf, verbosity, width) < 0)
+ gm, buf, verbosity, p_width) < 0)
break;
}
}
@@ -518,7 +518,7 @@ static int snprint_foreign_topology__(struct strbuf *buf, int verbosity,
}
int snprint_foreign_topology(struct strbuf *buf, int verbosity,
- const fieldwidth_t *width)
+ const fieldwidth_t *p_width)
{
int rc;
@@ -528,7 +528,7 @@ int snprint_foreign_topology(struct strbuf *buf, int verbosity,
return 0;
}
pthread_cleanup_push(unlock_foreigns, NULL);
- rc = snprint_foreign_topology__(buf, verbosity, width);
+ rc = snprint_foreign_topology__(buf, verbosity, p_width);
pthread_cleanup_pop(1);
return rc;
}
@@ -538,9 +538,9 @@ void print_foreign_topology(int verbosity)
STRBUF_ON_STACK(buf);
struct foreign *fgn;
int i;
- fieldwidth_t *width __attribute__((cleanup(cleanup_ucharp))) = NULL;
+ fieldwidth_t *p_width __attribute__((cleanup(cleanup_ucharp))) = NULL;
- if ((width = alloc_path_layout()) == NULL)
+ if ((p_width = alloc_path_layout()) == NULL)
return;
rdlock_foreigns();
if (foreigns == NULL) {
@@ -554,11 +554,11 @@ void print_foreign_topology(int verbosity)
fgn->lock(fgn->context);
pthread_cleanup_push(fgn->unlock, fgn->context);
vec = fgn->get_paths(fgn->context);
- get_multipath_layout__(vec, LAYOUT_RESET_NOT, width);
+ get_path_layout__(vec, LAYOUT_RESET_NOT, p_width);
fgn->release_paths(fgn->context, vec);
pthread_cleanup_pop(1);
}
- snprint_foreign_topology__(&buf, verbosity, width);
+ snprint_foreign_topology__(&buf, verbosity, p_width);
pthread_cleanup_pop(1);
printf("%s", get_strbuf_str(&buf));
}
--
2.50.1
^ permalink raw reply related [flat|nested] 2+ messages in thread
* Re: [PATCH] libmultipath: fix crash in print_foreign_topology
2025-08-20 18:39 [PATCH] libmultipath: fix crash in print_foreign_topology Benjamin Marzinski
@ 2025-08-25 21:56 ` Martin Wilck
0 siblings, 0 replies; 2+ messages in thread
From: Martin Wilck @ 2025-08-25 21:56 UTC (permalink / raw)
To: Benjamin Marzinski, Christophe Varoqui; +Cc: device-mapper development
On Wed, 2025-08-20 at 14:39 -0400, Benjamin Marzinski wrote:
> print_foreign_topology() called get_paths() to get a vector of
> (struct gen_path *) items and then called get_multipath_layout__(),
> which expects a vector of (struct gen_multipath *) items, with the
> path
> vector. This can easily end badly. Fix it to correctly call
> get_path_layout__(), and rename width to p_width in the functions
> that
> end up calling snprint_multipath_topology__(), which is expecting to
> get
> passed the path field widths.
Thanks for catching this!
I don't like the p_width rename so much, as we generally don't have
this kind of variable naming convention, but that's not a problem.
Reviewed-by: Martin Wilck <mwilck@suse.com>
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2025-08-25 21:56 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2025-08-20 18:39 [PATCH] libmultipath: fix crash in print_foreign_topology Benjamin Marzinski
2025-08-25 21:56 ` Martin Wilck
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).