| Message ID | 163177339181.682366.8713781325929549256.stgit@devnote2 (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | bootconfig: Fixes to bootconfig memory management etc. | expand |
On Thu, 16 Sep 2021 15:23:12 +0900 Masami Hiramatsu <mhiramat@kernel.org> wrote: > Free unused memblock in a error case to fix memblock leak > in xbc_make_cmdline(). > > Fixes: 51887d03aca1 ("bootconfig: init: Allow admin to use bootconfig for kernel command line") > Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org> > --- > init/main.c | 1 + > 1 file changed, 1 insertion(+) > > diff --git a/init/main.c b/init/main.c > index 3f7216934441..0b054fff8e92 100644 > --- a/init/main.c > +++ b/init/main.c > @@ -382,6 +382,7 @@ static char * __init xbc_make_cmdline(const char *key) > ret = xbc_snprint_cmdline(new_cmdline, len + 1, root); > if (ret < 0 || ret > len) { > pr_err("Failed to print extra kernel cmdline.\n"); > + memblock_free_ptr(new_cmdline, len + 1); > return NULL; > } > Hmm, looking at my patch queue, I noticed that this did not get applied. I'm thinking I may have been confused with the other memory freeing that was put into the xbc_destroy(), thinking this was part of that. But now that I look at this patch in the context of the code, it looks like this patch is required, as "new_cmdline" never gets exposed on this error. Masami, I just want to confirm, that this patch is still relevant, right? Thanks! -- Steve
On Wed, 6 Oct 2021 21:02:16 -0400 Steven Rostedt <rostedt@goodmis.org> wrote: > On Thu, 16 Sep 2021 15:23:12 +0900 > Masami Hiramatsu <mhiramat@kernel.org> wrote: > > > Free unused memblock in a error case to fix memblock leak > > in xbc_make_cmdline(). > > > > Fixes: 51887d03aca1 ("bootconfig: init: Allow admin to use bootconfig for kernel command line") > > Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org> > > --- > > init/main.c | 1 + > > 1 file changed, 1 insertion(+) > > > > diff --git a/init/main.c b/init/main.c > > index 3f7216934441..0b054fff8e92 100644 > > --- a/init/main.c > > +++ b/init/main.c > > @@ -382,6 +382,7 @@ static char * __init xbc_make_cmdline(const char *key) > > ret = xbc_snprint_cmdline(new_cmdline, len + 1, root); > > if (ret < 0 || ret > len) { > > pr_err("Failed to print extra kernel cmdline.\n"); > > + memblock_free_ptr(new_cmdline, len + 1); > > return NULL; > > } > > > > Hmm, looking at my patch queue, I noticed that this did not get > applied. I'm thinking I may have been confused with the other memory > freeing that was put into the xbc_destroy(), thinking this was part of > that. But now that I look at this patch in the context of the code, it > looks like this patch is required, as "new_cmdline" never gets exposed > on this error. > > Masami, I just want to confirm, that this patch is still relevant, right? Yes, with other 2 patches in this series ([1/4]-[3/4]), I thought you already queued it in your tree as you said in [1]; > I'm going to leave this patch out, and just review and accept the first three patches > in the series. [1] https://lore.kernel.org/all/20210916164805.32592423@gandalf.local.home/T/#u So, my next cleanup series [2] (including xbc_destroy_all() -> xbc_exit()) was based on the [1]'s first 3 patches. [2] https://lore.kernel.org/all/163187294400.2366983.7393164788107844569.stgit@devnote2/T/#u If it helps, I can make these series to one series and rebase on top of your for-next (or ftrace/core) branch. Thank you, > > Thanks! > > -- Steve
On Thu, 7 Oct 2021 10:43:57 +0900 Masami Hiramatsu <mhiramat@kernel.org> wrote: > On Wed, 6 Oct 2021 21:02:16 -0400 > Steven Rostedt <rostedt@goodmis.org> wrote: > > > On Thu, 16 Sep 2021 15:23:12 +0900 > > Masami Hiramatsu <mhiramat@kernel.org> wrote: > > > > > Free unused memblock in a error case to fix memblock leak > > > in xbc_make_cmdline(). > > > > > > Fixes: 51887d03aca1 ("bootconfig: init: Allow admin to use bootconfig for kernel command line") > > > Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org> > > > --- > > > init/main.c | 1 + > > > 1 file changed, 1 insertion(+) > > > > > > diff --git a/init/main.c b/init/main.c > > > index 3f7216934441..0b054fff8e92 100644 > > > --- a/init/main.c > > > +++ b/init/main.c > > > @@ -382,6 +382,7 @@ static char * __init xbc_make_cmdline(const char *key) > > > ret = xbc_snprint_cmdline(new_cmdline, len + 1, root); > > > if (ret < 0 || ret > len) { > > > pr_err("Failed to print extra kernel cmdline.\n"); > > > + memblock_free_ptr(new_cmdline, len + 1); > > > return NULL; > > > } > > > > > > > Hmm, looking at my patch queue, I noticed that this did not get > > applied. I'm thinking I may have been confused with the other memory > > freeing that was put into the xbc_destroy(), thinking this was part of > > that. But now that I look at this patch in the context of the code, it > > looks like this patch is required, as "new_cmdline" never gets exposed > > on this error. > > > > Masami, I just want to confirm, that this patch is still relevant, right? > > Yes, with other 2 patches in this series ([1/4]-[3/4]), I thought you already > queued it in your tree as you said in [1]; > > > I'm going to leave this patch out, and just review and accept the first three patches > > in the series. > > [1] https://lore.kernel.org/all/20210916164805.32592423@gandalf.local.home/T/#u Bah, doing all my rebases with Linus's "nack" probably caused me to accidentally drop this one. > > So, my next cleanup series [2] (including xbc_destroy_all() -> xbc_exit()) was > based on the [1]'s first 3 patches. > > [2] https://lore.kernel.org/all/163187294400.2366983.7393164788107844569.stgit@devnote2/T/#u > I just went through my queue from as far back as August, to pick up everything that I left behind to worry about Linux Plumbers and Open Source Summit, and found this set in that queue. I'll be processing all these patches in the next few days. > If it helps, I can make these series to one series and rebase on top of your > for-next (or ftrace/core) branch. No, this patch should be added to my urgent tree and pushed onto Linus (and stable). I'm working on those patches first, then will move on to my for-next tree. Thanks, -- Steve
diff --git a/init/main.c b/init/main.c index 3f7216934441..0b054fff8e92 100644 --- a/init/main.c +++ b/init/main.c @@ -382,6 +382,7 @@ static char * __init xbc_make_cmdline(const char *key) ret = xbc_snprint_cmdline(new_cmdline, len + 1, root); if (ret < 0 || ret > len) { pr_err("Failed to print extra kernel cmdline.\n"); + memblock_free_ptr(new_cmdline, len + 1); return NULL; }
Free unused memblock in a error case to fix memblock leak in xbc_make_cmdline(). Fixes: 51887d03aca1 ("bootconfig: init: Allow admin to use bootconfig for kernel command line") Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org> --- init/main.c | 1 + 1 file changed, 1 insertion(+)