Function read_dump() memory maps the input via grab_file(), but fails to call
the corresponding unmap function. Add the missing call to release_file().
Detected by Coverity: CID 1192419
Signed-off-by: Christian Engelmayer <[email protected]>
---
Compile tested - Applies against v3.14 as well as linux-next.
---
scripts/mod/modpost.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/scripts/mod/modpost.c b/scripts/mod/modpost.c
index 0663556..ea3e2bd 100644
--- a/scripts/mod/modpost.c
+++ b/scripts/mod/modpost.c
@@ -2113,8 +2113,10 @@ static void read_dump(const char *fname, unsigned int kernel)
s->preloaded = 1;
sym_update_crc(symname, mod, crc, export_no(export));
}
+ release_file(file, size);
return;
fail:
+ release_file(file, size);
fatal("parse error in symbol dump file\n");
}
--
1.8.3.2
On Sun, Apr 06, 2014 at 12:36:49AM +0200, Christian Engelmayer wrote:
> Function read_dump() memory maps the input via grab_file(), but fails to call
> the corresponding unmap function. Add the missing call to release_file().
> Detected by Coverity: CID 1192419
It doesn't matter at all. All memory is freed on exit() in this program.
-Andi
On 2014-04-06 02:05, Andi Kleen wrote:
> On Sun, Apr 06, 2014 at 12:36:49AM +0200, Christian Engelmayer wrote:
>> Function read_dump() memory maps the input via grab_file(), but fails to call
>> the corresponding unmap function. Add the missing call to release_file().
>> Detected by Coverity: CID 1192419
>
> It doesn't matter at all. All memory is freed on exit() in this program.
The leak can disturb in valgrind output.
Michal
Michal Marek <[email protected]> writes:
> On 2014-04-06 02:05, Andi Kleen wrote:
>> On Sun, Apr 06, 2014 at 12:36:49AM +0200, Christian Engelmayer wrote:
>>> Function read_dump() memory maps the input via grab_file(), but fails to call
>>> the corresponding unmap function. Add the missing call to release_file().
>>> Detected by Coverity: CID 1192419
>>
>> It doesn't matter at all. All memory is freed on exit() in this program.
>
> The leak can disturb in valgrind output.
Yeah. If release_file() didn't exist at all, I'd agree with Andi. As
it is, I've queued the patch (with typo in title fixed).
Thanks!
Rusty.