From: Niels de Vos <ndevos@redhat.com>
To: Markus Armbruster <armbru@redhat.com>
Cc: kwolf@redhat.com, mitake.hitoshi@lab.ntt.co.jp,
qemu-devel@nongnu.org, qemu-block@nongnu.org,
namei.unix@gmail.com
Subject: Re: [Qemu-devel] [Qemu-block] [PATCH 12/15] gluster: Plug memory leaks in qemu_gluster_parse_json()
Date: Fri, 3 Mar 2017 00:17:43 -0800 [thread overview]
Message-ID: <20170303081743.GU16140@ndevos-x240.usersys.redhat.com> (raw)
In-Reply-To: <87pohy4xoa.fsf@dusky.pond.sub.org>
On Fri, Mar 03, 2017 at 08:38:45AM +0100, Markus Armbruster wrote:
> Niels de Vos <ndevos@redhat.com> writes:
>
> > On Thu, Mar 02, 2017 at 10:44:03PM +0100, Markus Armbruster wrote:
> >> To reproduce, run
> >>
> >> $ valgrind qemu-system-x86_64 --nodefaults -S --drive driver=gluster,volume=testvol,path=/a/b/c,server.0.type=xxx
> >>
> >> Signed-off-by: Markus Armbruster <armbru@redhat.com>
> >> ---
> >> block/gluster.c | 28 ++++++++++++++--------------
> >> 1 file changed, 14 insertions(+), 14 deletions(-)
> >>
> >> diff --git a/block/gluster.c b/block/gluster.c
> >> index 6fbcf9e..35a7abb 100644
> >> --- a/block/gluster.c
> >> +++ b/block/gluster.c
> >> @@ -480,7 +480,7 @@ static int qemu_gluster_parse_json(BlockdevOptionsGluster *gconf,
> >> QDict *options, Error **errp)
> >> {
> >> QemuOpts *opts;
> >> - GlusterServer *gsconf;
> >> + GlusterServer *gsconf = NULL;
> >> GlusterServerList *curr = NULL;
> >> QDict *backing_options = NULL;
> >> Error *local_err = NULL;
> >> @@ -529,17 +529,16 @@ static int qemu_gluster_parse_json(BlockdevOptionsGluster *gconf,
> >> }
> >>
> >> ptr = qemu_opt_get(opts, GLUSTER_OPT_TYPE);
> >> + if (!ptr) {
> >> + error_setg(&local_err, QERR_MISSING_PARAMETER, GLUSTER_OPT_TYPE);
> >> + error_append_hint(&local_err, GERR_INDEX_HINT, i);
> >> + goto out;
> >> +
> >> + }
> >> gsconf = g_new0(GlusterServer, 1);
> >> gsconf->type = qapi_enum_parse(GlusterTransport_lookup, ptr,
> >> - GLUSTER_TRANSPORT__MAX,
> >> - GLUSTER_TRANSPORT__MAX,
> >> + GLUSTER_TRANSPORT__MAX, 0,
> >
> > What is the reason to set the default to 0 and not the more readable
> > GLUSTER_TRANSPORT_UNIX?
>
> I chose 0 because the actual value must not matter: we don't want a
> default here.
>
> qapi_enum_parse() returns this argument when ptr isn't found. If ptr is
> non-null, it additionally sets an error. Since ptr can't be null here,
> the argument is only returned together with an error. But then we won't
> use *gsconf.
Ah, right, I that see now.
> Do you think -1 instead of 0 would be clearer?
Yes, it would be to me. -1 is clearly not part of the GLUSTER_TRANSPORT_*
enum, so it suggests it is a different case.
Thanks!
>
> >> &local_err);
> >> - if (!ptr) {
> >> - error_setg(&local_err, QERR_MISSING_PARAMETER, GLUSTER_OPT_TYPE);
> >> - error_append_hint(&local_err, GERR_INDEX_HINT, i);
> >> - goto out;
> >> -
> >> - }
> >> if (local_err) {
> >> error_append_hint(&local_err,
> >> "Parameter '%s' may be 'tcp' or 'unix'\n",
> >> @@ -626,8 +625,10 @@ static int qemu_gluster_parse_json(BlockdevOptionsGluster *gconf,
> >> curr->next->value = gsconf;
> >> curr = curr->next;
> >> }
> >> + gsconf = NULL;
> >>
> >> - qdict_del(backing_options, str);
> >> + QDECREF(backing_options);
> >> + backing_options = NULL;
> >> g_free(str);
> >> str = NULL;
> >> }
> >> @@ -636,11 +637,10 @@ static int qemu_gluster_parse_json(BlockdevOptionsGluster *gconf,
> >>
> >> out:
> >> error_propagate(errp, local_err);
> >> + qapi_free_GlusterServer(gsconf);
> >> qemu_opts_del(opts);
> >> - if (str) {
> >> - qdict_del(backing_options, str);
> >> - g_free(str);
> >> - }
> >> + g_free(str);
> >> + QDECREF(backing_options);
> >> errno = EINVAL;
> >> return -errno;
> >> }
> >> --
> >> 2.7.4
> >>
> >>
next prev parent reply other threads:[~2017-03-03 8:17 UTC|newest]
Thread overview: 61+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-03-02 21:43 [Qemu-devel] [PATCH 00/15] block: A bunch of fixes for Sheepdog and Gluster Markus Armbruster
2017-03-02 21:43 ` [Qemu-devel] [PATCH 01/15] sheepdog: Defuse time bomb in sd_open() error handling Markus Armbruster
2017-03-02 22:46 ` Eric Blake
2017-03-03 5:18 ` Markus Armbruster
2017-03-02 21:43 ` [Qemu-devel] [PATCH 02/15] sheepdog: Fix error handling in sd_snapshot_delete() Markus Armbruster
2017-03-02 23:13 ` Eric Blake
2017-03-03 5:22 ` Markus Armbruster
2017-03-03 13:07 ` Kevin Wolf
2017-03-03 13:31 ` Markus Armbruster
2017-03-02 21:43 ` [Qemu-devel] [PATCH 03/15] sheepdog: Fix error handling sd_create() Markus Armbruster
2017-03-02 23:16 ` Eric Blake
2017-03-03 0:07 ` Philippe Mathieu-Daudé
2017-03-03 13:13 ` Kevin Wolf
2017-03-02 21:43 ` [Qemu-devel] [PATCH 04/15] sheepdog: Mark sd_snapshot_delete() lossage FIXME Markus Armbruster
2017-03-02 23:18 ` Eric Blake
2017-03-02 21:43 ` [Qemu-devel] [PATCH 05/15] sheepdog: Fix snapshot ID parsing in _open(), _create, _goto() Markus Armbruster
2017-03-02 23:30 ` Eric Blake
2017-03-03 13:25 ` Kevin Wolf
2017-03-03 13:41 ` Markus Armbruster
2017-03-02 21:43 ` [Qemu-devel] [PATCH 06/15] sheepdog: Don't truncate long VDI name in _open(), _create() Markus Armbruster
2017-03-02 23:32 ` Eric Blake
2017-03-03 0:25 ` Philippe Mathieu-Daudé
2017-03-03 5:21 ` Markus Armbruster
2017-03-03 5:21 ` Markus Armbruster
2017-03-03 0:10 ` Philippe Mathieu-Daudé
2017-03-02 21:43 ` [Qemu-devel] [PATCH 07/15] sheepdog: Report errors in pseudo-filename more usefully Markus Armbruster
2017-03-03 13:36 ` Kevin Wolf
2017-03-03 14:53 ` Markus Armbruster
2017-03-03 13:49 ` Kevin Wolf
2017-03-03 14:57 ` Markus Armbruster
2017-03-02 21:43 ` [Qemu-devel] [PATCH 08/15] sheepdog: Use SocketAddress and socket_connect() Markus Armbruster
2017-03-03 13:47 ` Kevin Wolf
2017-03-02 21:44 ` [Qemu-devel] [PATCH 09/15] sheepdog: Implement bdrv_parse_filename() Markus Armbruster
2017-03-03 20:17 ` Eric Blake
2017-03-02 21:44 ` [Qemu-devel] [PATCH 10/15] gluster: Drop assumptions on SocketTransport names Markus Armbruster
2017-03-03 6:40 ` [Qemu-devel] [Qemu-block] " Niels de Vos
2017-03-03 7:31 ` Markus Armbruster
2017-03-02 21:44 ` [Qemu-devel] [PATCH 11/15] gluster: Don't duplicate qapi-util.c's qapi_enum_parse() Markus Armbruster
2017-03-03 6:35 ` [Qemu-devel] [Qemu-block] " Niels de Vos
2017-03-02 21:44 ` [Qemu-devel] [PATCH 12/15] gluster: Plug memory leaks in qemu_gluster_parse_json() Markus Armbruster
2017-03-03 7:11 ` [Qemu-devel] [Qemu-block] " Niels de Vos
2017-03-03 7:38 ` Markus Armbruster
2017-03-03 8:17 ` Niels de Vos [this message]
2017-03-03 8:35 ` Markus Armbruster
2017-03-03 17:06 ` Niels de Vos
2017-03-02 21:44 ` [Qemu-devel] [PATCH 13/15] qapi-schema: Rename GlusterServer to SocketAddressFlat Markus Armbruster
2017-03-03 16:31 ` Eric Blake
2017-03-03 17:05 ` Markus Armbruster
2017-03-03 18:33 ` Eric Blake
2017-03-02 21:44 ` [Qemu-devel] [PATCH 14/15] qapi-schema: Rename SocketAddressFlat's variant tcp to inet Markus Armbruster
2017-03-03 18:35 ` Eric Blake
2017-03-03 20:03 ` Markus Armbruster
2017-03-06 15:00 ` Markus Armbruster
2017-03-02 21:44 ` [Qemu-devel] [PATCH 15/15] sheepdog: Support blockdev-add Markus Armbruster
2017-03-03 18:42 ` Eric Blake
2017-03-02 23:35 ` [Qemu-devel] [PATCH 00/15] block: A bunch of fixes for Sheepdog and Gluster Eric Blake
2017-03-03 5:39 ` Markus Armbruster
2017-03-03 16:27 ` Eric Blake
2017-03-03 17:14 ` Peter Maydell
2017-03-03 18:37 ` Markus Armbruster
2017-03-03 18:50 ` Peter Maydell
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20170303081743.GU16140@ndevos-x240.usersys.redhat.com \
--to=ndevos@redhat.com \
--cc=armbru@redhat.com \
--cc=kwolf@redhat.com \
--cc=mitake.hitoshi@lab.ntt.co.jp \
--cc=namei.unix@gmail.com \
--cc=qemu-block@nongnu.org \
--cc=qemu-devel@nongnu.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
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).