On Wednesday 29 April 2009, Alan Jenkins wrote:
> Alan Jenkins wrote:
> > Johannes Berg wrote:
> >
> >> That doesn't seem relevant, this just does some initialisation. However,
> >> you definitely missed adding a call to wep_free().
> >>
> >>
> >
> > Hah, I should have realized something was wrong when I noticed I was
> > removing more lines that I added.
> >
> > The crypto init does cause the module load:
> >
> > wait_for_completion
> > call_usermodehelper_exec
> > __request_module
> > crypto_larval_lookup
> > ? extract_entropy
> > crypto_alg_mod_lookup
> > crypto_alloc_base
> > ieee80211_wep_init
> > ieee80211_register_hw
> >
>
> Here's a corrected patch complete with changelog. If there are no other
> problems with it, can you please apply this for 2.6.30 to keep my EeePC
> regression-free?
>
> Thanks
> Alan
>
> ------>
> From c5e9dc036247e70956d1a28e8850c3810385dda0 Mon Sep 17 00:00:00 2001
> From: Alan Jenkins <alan-jenkins@xxxxxxxxxxxxxx>
> Date: Wed, 29 Apr 2009 11:41:24 +0100
> Subject: [PATCH] mac80211: fix modprobe deadlock by not calling wep_init under rtnl_lock
>
> - ieee80211_wep_init(), which is called with rtnl_lock held, blocks in
> request_module() [waiting for modprobe to load a crypto module].
>
> - modprobe blocks in a call to flush_workqueue(), when it closes a TTY
> [presumably when it exits].
>
> - The workqueue item linkwatch_event() blocks on rtnl_lock.
>
> There's no reason for wep_init() to be called with rtnl_lock held, so
> just move it outside the critical section.
Has it been merged already or is it queued up for merging?
Rafael
> Signed-off-by: Alan Jenkins <alan-jenkins@xxxxxxxxxxxxxx>
> ---
> net/mac80211/main.c | 19 +++++++++----------
> 1 files changed, 9 insertions(+), 10 deletions(-)
>
> diff --git a/net/mac80211/main.c b/net/mac80211/main.c
> index fbcbed6..00968c2 100644
> --- a/net/mac80211/main.c
> +++ b/net/mac80211/main.c
> @@ -909,6 +909,13 @@ int ieee80211_register_hw(struct ieee80211_hw *hw)
> if (result < 0)
> goto fail_sta_info;
>
> + result = ieee80211_wep_init(local);
> + if (result < 0) {
> + printk(KERN_DEBUG "%s: Failed to initialize wep: %d\n",
> + wiphy_name(local->hw.wiphy), result);
> + goto fail_wep;
> + }
> +
> rtnl_lock();
> result = dev_alloc_name(local->mdev, local->mdev->name);
> if (result < 0)
> @@ -930,14 +937,6 @@ int ieee80211_register_hw(struct ieee80211_hw *hw)
> goto fail_rate;
> }
>
> - result = ieee80211_wep_init(local);
> -
> - if (result < 0) {
> - printk(KERN_DEBUG "%s: Failed to initialize wep: %d\n",
> - wiphy_name(local->hw.wiphy), result);
> - goto fail_wep;
> - }
> -
> /* add one default STA interface if supported */
> if (local->hw.wiphy->interface_modes & BIT(NL80211_IFTYPE_STATION)) {
> result = ieee80211_if_add(local, "wlan%d", NULL,
> @@ -967,13 +966,13 @@ int ieee80211_register_hw(struct ieee80211_hw *hw)
>
> return 0;
>
> -fail_wep:
> - rate_control_deinitialize(local);
> fail_rate:
> unregister_netdevice(local->mdev);
> local->mdev = NULL;
> fail_dev:
> rtnl_unlock();
> + ieee80211_wep_free(local);
> +fail_wep:
> sta_info_stop(local);
> fail_sta_info:
> debugfs_hw_del(local);
--
Everyone knows that debugging is twice as hard as writing a program
in the first place. So if you're as clever as you can be when you write it,
how will you ever debug it? --- Brian Kernighan
--
To unsubscribe from this list: send the line "unsubscribe kernel-testers" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
