Marcel reported a warning, which quite obviously comes
from an oversight in the code handling deauth frames,
and which resulted in multiple follow-up warnings due
to this missing handling. This patch adds the missing
deauth handling (telling cfg80211 about it) and also
removes the follow-up warnings since they could happen
due to races even if nothing is wrong. I've explained
the races in the comments.
Signed-off-by: Johannes Berg <[email protected]>
Reported-by: Marcel Holtmann <[email protected]>
---
net/mac80211/mlme.c | 21 +++++++++++++++++----
1 file changed, 17 insertions(+), 4 deletions(-)
--- wireless-testing.orig/net/mac80211/mlme.c 2009-07-25 11:03:12.000000000 +0200
+++ wireless-testing/net/mac80211/mlme.c 2009-07-25 11:55:35.000000000 +0200
@@ -2003,6 +2003,9 @@ static void ieee80211_sta_rx_queued_mgmt
case RX_MGMT_CFG80211_ASSOC:
cfg80211_send_rx_assoc(sdata->dev, (u8 *) mgmt, skb->len);
break;
+ case RX_MGMT_CFG80211_DEAUTH:
+ cfg80211_send_deauth(sdata->dev, (u8 *)mgmt, skb->len, NULL);
+ break;
default:
WARN(1, "unexpected: %d", rma);
}
@@ -2498,8 +2501,13 @@ int ieee80211_mgd_deauth(struct ieee8021
}
}
- /* cfg80211 should catch this... */
- if (WARN_ON(!bssid)) {
+ /*
+ * cfg80211 should catch this ... but it's racy since
+ * we can receive a deauth frame, process it, hand it
+ * to cfg80211 while that's in a locked section already
+ * trying to tell us that the user wants to disconnect.
+ */
+ if (!bssid) {
mutex_unlock(&ifmgd->mtx);
return -ENOLINK;
}
@@ -2524,8 +2532,13 @@ int ieee80211_mgd_disassoc(struct ieee80
mutex_lock(&ifmgd->mtx);
- /* cfg80211 should catch that */
- if (WARN_ON(&ifmgd->associated->cbss != req->bss)) {
+ /*
+ * cfg80211 should catch this ... but it's racy since
+ * we can receive a disassoc frame, process it, hand it
+ * to cfg80211 while that's in a locked section already
+ * trying to tell us that the user wants to disconnect.
+ */
+ if (&ifmgd->associated->cbss != req->bss) {
mutex_unlock(&ifmgd->mtx);
return -ENOLINK;
}
Hi Johannes,
> > Marcel reported a warning, which quite obviously comes
> > from an oversight in the code handling deauth frames,
> > and which resulted in multiple follow-up warnings due
> > to this missing handling. This patch adds the missing
> > deauth handling (telling cfg80211 about it) and also
> > removes the follow-up warnings since they could happen
> > due to races even if nothing is wrong. I've explained
> > the races in the comments.
> >
> > Signed-off-by: Johannes Berg <[email protected]>
> > Reported-by: Marcel Holtmann <[email protected]>
>
> so me current wireless-testing tree now contains the tree from today
> with the oops fix from Luiz, your lockdep fix, Reinette's iwlwifi
> updates and this one. No oopses so far and my connection to the AP stays
> stable. No immediate disconnects anymore. Everything looks good again.
while this fixes the oops and helps with my connect/disconnect issues I
have seen, it doesn't fully fix them. As soon as the AP is something
like more than 5m or 10m away, it connects, disconnects, re-connects
etc. all over again :(
Regards
Marcel
On Sat, 2009-07-25 at 11:58 +0200, Johannes Berg wrote:
> Marcel reported a warning, which quite obviously comes
> from an oversight in the code handling deauth frames,
> and which resulted in multiple follow-up warnings due
> to this missing handling. This patch adds the missing
> deauth handling (telling cfg80211 about it) and also
> removes the follow-up warnings since they could happen
> due to races even if nothing is wrong. I've explained
> the races in the comments.
>
> Signed-off-by: Johannes Berg <[email protected]>
> Reported-by: Marcel Holtmann <[email protected]>
> ---
> net/mac80211/mlme.c | 21 +++++++++++++++++----
> 1 file changed, 17 insertions(+), 4 deletions(-)
>
> --- wireless-testing.orig/net/mac80211/mlme.c 2009-07-25 11:03:12.000000000 +0200
> +++ wireless-testing/net/mac80211/mlme.c 2009-07-25 11:55:35.000000000 +0200
> @@ -2003,6 +2003,9 @@ static void ieee80211_sta_rx_queued_mgmt
> case RX_MGMT_CFG80211_ASSOC:
> cfg80211_send_rx_assoc(sdata->dev, (u8 *) mgmt, skb->len);
> break;
> + case RX_MGMT_CFG80211_DEAUTH:
> + cfg80211_send_deauth(sdata->dev, (u8 *)mgmt, skb->len, NULL);
> + break;
> default:
> WARN(1, "unexpected: %d", rma);
> }
> @@ -2498,8 +2501,13 @@ int ieee80211_mgd_deauth(struct ieee8021
> }
> }
>
> - /* cfg80211 should catch this... */
> - if (WARN_ON(!bssid)) {
> + /*
> + * cfg80211 should catch this ... but it's racy since
> + * we can receive a deauth frame, process it, hand it
> + * to cfg80211 while that's in a locked section already
> + * trying to tell us that the user wants to disconnect.
> + */
> + if (!bssid) {
> mutex_unlock(&ifmgd->mtx);
> return -ENOLINK;
> }
> @@ -2524,8 +2532,13 @@ int ieee80211_mgd_disassoc(struct ieee80
>
> mutex_lock(&ifmgd->mtx);
>
> - /* cfg80211 should catch that */
> - if (WARN_ON(&ifmgd->associated->cbss != req->bss)) {
> + /*
> + * cfg80211 should catch this ... but it's racy since
> + * we can receive a disassoc frame, process it, hand it
> + * to cfg80211 while that's in a locked section already
> + * trying to tell us that the user wants to disconnect.
> + */
> + if (&ifmgd->associated->cbss != req->bss) {
> mutex_unlock(&ifmgd->mtx);
> return -ENOLINK;
> }
Don't know if related, but I see this on latest wireless testing with 2
patches applied:
[PATCH]_nl80211:_add_missing_parameter_clearing
[PATCH]_mac80211:_fix_an_oops_in_ieee80211_scan_state_set_channel
<1>[ 315.219232] BUG: unable to handle kernel NULL pointer dereference a 0000000000000048
<1>[ 315.219242] IP: [<ffffffffa0a467c8>] cfg80211_mlme_disassoc+0x98/0x110 [cfg80211]
<4>[ 315.219271] PGD 7061b067 PUD 7e9c0067 PMD 0
<0>[ 315.219280] Oops: 0000 [#1] PREEMPT SMP
<0>[ 315.219287] last sysfs file: /sys/devices/system/cpu/cpu0/cpufreq/scaling_cur_freq
<4>[ 315.219293] CPU 0
<4>[ 315.219297] Modules linked in: af_packet sco bridge stp llc bnep l2cap bluetooth nfsd exportfs nfs lockd nfs_acl auth_rpcgss sunrpc usb_storage usb_libusual cpufreq_powersave cpufreq_conservative cpufreq_userspace acpi_cpufreq coretemp joydev sbp2 snd_hda_codec_realtek iwl3945 iwlcore snd_hda_intel snd_hda_codec uvcvideo mac80211 videodev snd_hwdep psmouse v4l1_compat v4l2_compat_ioctl32 snd_pcm acer_wmi backlight cfg80211 serio_raw snd_timer tg3 iTCO_wdt uhci_hcd rfkill nvidia(P) ohci1394 snd_page_alloc ehci_hcd sdhci_pci sdhci libphy iTCO_vendor_support usbcore wmi evdev fuse
<6>[ 315.219383] Pid: 4078, comm: wpa_supplicant Tainted: P 2.6.31-rc4-wl #32 Aspire 5720
<6>[ 315.219388] RIP: 0010:[<ffffffffa0a467c8>] [<ffffffffa0a467c8>] cfg80211_mlme_disassoc+0x98/0x110 [cfg80211]
<6>[ 315.219411] RSP: 0018:ffff880070611988 EFLAGS: 00010296
<6>[ 315.219415] RAX: 0000000000000000 RBX: 00000000ffffff95 RCX: 0000000000000006
<6>[ 315.219420] RDX: ffff880067e61828 RSI: 0000000000000048 RDI: ffff880067e61828
<6>[ 315.219425] RBP: ffff8800706119f8 R08: ffff8800706119a8 R09: 0000000000000000
<6>[ 315.219430] R10: ffff88007f8df000 R11: 0000000000000000 R12: ffff88007f8df590
<6>[ 315.219435] R13: ffff88007f8df000 R14: 0000000000000000 R15: 0000000000000000
<6>[ 315.219441] FS: 00007fd2a264f6f0(0000) GS:ffff8800016a3000(0000) knlGS:0000000000000000
<6>[ 315.219446] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
<6>[ 315.219450] CR2: 0000000000000048 CR3: 00000000704e1000 CR4: 00000000000006f0
<6>[ 315.219455] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
<6>[ 315.219460] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
<4>[ 315.219465] Process wpa_supplicant (pid: 4078, threadinfo ffff880070610000, task ffff88007056c230)
<0>[ 315.219469] Stack:
<4>[ 315.219472] ffff8800706119a8 ffff880067e61828 ffff88007f330000 ffff88007f8df5b8
<4>[ 315.219480] <0> 0000000000000000 0000000000000000 0000000000000000 0000000000000003
<4>[ 315.219489] <0> ffff8800706119f8 00000000ffffffea ffff880070611a58 ffff880067e61800
<0>[ 315.219499] Call Trace:
<4>[ 315.219522] [<ffffffffa0a3f287>] nl80211_disassociate+0x127/0x140 [cfg80211]
<4>[ 315.219533] [<ffffffff81340b06>] genl_rcv_msg+0x1b6/0x1f0
<4>[ 315.219541] [<ffffffff81340950>] ? genl_rcv_msg+0x0/0x1f0
<4>[ 315.219547] [<ffffffff813408e9>] netlink_rcv_skb+0x89/0xb0
<4>[ 315.219553] [<ffffffff81340937>] genl_rcv+0x27/0x40
<4>[ 315.219559] [<ffffffff81340449>] ? netlink_sendmsg+0x159/0x300
<4>[ 315.219566] [<ffffffff813402da>] netlink_unicast+0x2da/0x2f0
<4>[ 315.219574] [<ffffffff8131fefe>] ? __alloc_skb+0x6e/0x170
<4>[ 315.219581] [<ffffffff813404ee>] netlink_sendmsg+0x1fe/0x300
<4>[ 315.219591] [<ffffffff810ac158>] ? generic_file_buffered_write+0x128/0x340
<4>[ 315.219600] [<ffffffff81316ed7>] sock_sendmsg+0x127/0x140
<4>[ 315.219608] [<ffffffff81063ee0>] ? autoremove_wake_function+0x0/0x40
<4>[ 315.219616] [<ffffffff810ad382>] ? generic_file_aio_write+0x72/0xd0
<4>[ 315.219623] [<ffffffff81315b1b>] ? move_addr_to_kernel+0x2b/0x40
<4>[ 315.219629] [<ffffffff8132151c>] ? verify_iovec+0x3c/0xd0
<4>[ 315.219636] [<ffffffff81317079>] sys_sendmsg+0x189/0x320
<4>[ 315.219643] [<ffffffff81063ee0>] ? autoremove_wake_function+0x0/0x40
<4>[ 315.219651] [<ffffffff810c4d28>] ? handle_mm_fault+0x1d8/0x830
<4>[ 315.219660] [<ffffffff810e194f>] ? vfs_write+0x13f/0x1a0
<4>[ 315.219669] [<ffffffff8100beeb>] system_call_fastpath+0x16/0x1b
<0>[ 315.219673] Code: 5d c8 4c 89 75 b8 49 63 c7 b9 06 00 00 00 48 89 45 c0 48 8b 7d 98 4d 8b 8c 24 d8 00 00 00 4c 8d 45 b0 49 8d 71 48 bb 95 ff ff ff <f3> a6 0f 97 c2 0f 92 c0 38 c2 75 2a 48 8b 55 a0 49 8d 41 40 48
<1>[ 315.219743] RIP [<ffffffffa0a467c8>] cfg80211_mlme_disassoc+0x98/0x110 [cfg80211]
<4>[ 315.219764] RSP <ffff880070611988>
<0>[ 315.219767] CR2: 0000000000000048
<4>[ 315.219772] ---[ end trace f7733b1bb80b3185 ]---
NM associates, but then if I manually ask it to associate again, this happens. I use nl80211 in wpa_supplicant
Best regards,
Maxim Levitsky
Hi Johannes,
> Marcel reported a warning, which quite obviously comes
> from an oversight in the code handling deauth frames,
> and which resulted in multiple follow-up warnings due
> to this missing handling. This patch adds the missing
> deauth handling (telling cfg80211 about it) and also
> removes the follow-up warnings since they could happen
> due to races even if nothing is wrong. I've explained
> the races in the comments.
>
> Signed-off-by: Johannes Berg <[email protected]>
> Reported-by: Marcel Holtmann <[email protected]>
so me current wireless-testing tree now contains the tree from today
with the oops fix from Luiz, your lockdep fix, Reinette's iwlwifi
updates and this one. No oopses so far and my connection to the AP stays
stable. No immediate disconnects anymore. Everything looks good again.
Tested-by: Marcel Holtmann <[email protected]>
Regards
Marcel