arch/xtensa/platforms/iss/network.c | 1 + 1 file changed, 1 insertion(+)
We fixed a bug where the variable dev
was not freed before return.
Signed-off-by: Minjie Du <duminjie@vivo.com>
---
arch/xtensa/platforms/iss/network.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/arch/xtensa/platforms/iss/network.c b/arch/xtensa/platforms/iss/network.c
index 9ac46ab3a..e33f3b37c 100644
--- a/arch/xtensa/platforms/iss/network.c
+++ b/arch/xtensa/platforms/iss/network.c
@@ -540,6 +540,7 @@ static void iss_net_configure(int index, char *init)
rtnl_unlock();
pr_err("%s: error registering net device!\n", dev->name);
platform_device_unregister(&lp->pdev);
+ free_netdev(dev);
return;
}
rtnl_unlock();
--
2.39.0
Hi Minjie, On Mon, Jul 3, 2023 at 5:33 AM Minjie Du <duminjie@vivo.com> wrote: > > We fixed a bug where the variable dev > was not freed before return. > > Signed-off-by: Minjie Du <duminjie@vivo.com> > --- > arch/xtensa/platforms/iss/network.c | 1 + > 1 file changed, 1 insertion(+) Applied to my xtensa tree with the following commit message: xtensa: ISS: fix etherdev leak in error return path iss_net_configure frees etherdev in all error return paths except one where register_netdevice fails. Add missing free_netdev to that path. -- Thanks. -- Max
On Mon, Jul 3, 2023 at 7:19 AM Max Filippov <jcmvbkbc@gmail.com> wrote: > On Mon, Jul 3, 2023 at 5:33 AM Minjie Du <duminjie@vivo.com> wrote: > > > > We fixed a bug where the variable dev > > was not freed before return. > > > > Signed-off-by: Minjie Du <duminjie@vivo.com> > > --- > > arch/xtensa/platforms/iss/network.c | 1 + > > 1 file changed, 1 insertion(+) > > Applied to my xtensa tree with the following commit message: > > xtensa: ISS: fix etherdev leak in error return path > > iss_net_configure frees etherdev in all error return paths except one > where register_netdevice fails. Add missing free_netdev to that path. > After testing it I realized that in that error return path the etherdev is freed by the iss_net_pdev_release callback and that this patch introduces freeing an active timer on that path. I've dropped this patch and committed another one that adds a comment about what's going on at that place. -- Thanks. -- Max
© 2016 - 2026 Red Hat, Inc.