| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <20061210192506.GF29871@p15091797.pureserver.info>
Date: Sun, 10 Dec 2006 20:25:06 +0100
From: Ulrich Kunitz <kune@...ne-taler.de>
To: Maxime Austruy <maxime@...lhalla.org>
Cc: zd1211-devs@...ts.sourceforge.net,
matthieu castet <castet.matthieu@...e.fr>,
Linux Kernel list <linux-kernel@...r.kernel.org>,
Daniel Drake <dsd@...too.org>
Subject: Re: 2.6.19 lot's of oops in mmx_copy_page/ mmx_clear_page functions
On 06-12-10 11:02 Maxime Austruy wrote:
>
> Matthieu,
>
> On Sun, Dec 10, 2006 at 12:32:08PM +0100, matthieu castet wrote:
> > Hi,
> >
> > with 2.6.19 I got some random crash and I got kernel opps.
> > There all happens in mmx_copy_page/ mmx_clear_page functions with
> > differents process : Xorg, cat, mozilla, ...
> >
> >
> > What could be the cause of these crashes ?
> > What can I provide in order to debug this ?
>
> FWIW, I've seen the same oops on my box. My instance was caused by a
> combination of:
> . zd1211rw calling ieee80211_rx in irq context while it's supposed to
> be called in softirq,
> . crypto code only expecting to be called in softirq/user context but
> not enforcing it. Consequently, it ends up doing
> v = kmap_atomic(..., KM_USER0);
> even when invoked by the zd1211rw driver in irq context, causing some
> corruption.
>
> Given that you have zd1211rw insmod'd, that could be it. I have some
> hacks that work around this bug, but Ulrich has a patch ready that
> should fix this. Thanks,
>
> Max
Here is the patch against vanilla 2.6.19.
For Linus' latest tree I have sent also patches on netdev.
Users of my out-of-kernel git tree, should look at the branch rx.
Uli
[PATCH] zd1211rw: Call ieee80211_rx in tasklet
The driver called ieee80211_rx in hardware irq context. This
caused a problem withe crypto libraries and was against the
documentation/intention of the ieee80211_rx function. This patch
fixes that.
Signed-off-by: Ulrich Kunitz <kune@...ne-taler.de>
---
drivers/net/wireless/zd1211rw/zd_mac.c | 88 ++++++++++++++++++++++++--------
drivers/net/wireless/zd1211rw/zd_mac.h | 6 ++
drivers/net/wireless/zd1211rw/zd_usb.c | 4 +
3 files changed, 72 insertions(+), 26 deletions(-)
diff --git a/drivers/net/wireless/zd1211rw/zd_mac.c b/drivers/net/wireless/zd1211rw/zd_mac.c
index a7d29bd..f52dd3a 100644
--- a/drivers/net/wireless/zd1211rw/zd_mac.c
+++ b/drivers/net/wireless/zd1211rw/zd_mac.c
@@ -37,6 +37,8 @@ static void housekeeping_init(struct zd_
static void housekeeping_enable(struct zd_mac *mac);
static void housekeeping_disable(struct zd_mac *mac);
+static void do_rx(unsigned long mac_ptr);
+
int zd_mac_init(struct zd_mac *mac,
struct net_device *netdev,
struct usb_interface *intf)
@@ -47,6 +49,10 @@ int zd_mac_init(struct zd_mac *mac,
spin_lock_init(&mac->lock);
mac->netdev = netdev;
+ skb_queue_head_init(&mac->rx_queue);
+ tasklet_init(&mac->rx_tasklet, do_rx, (unsigned long)mac);
+ tasklet_disable(&mac->rx_tasklet);
+
ieee_init(ieee);
softmac_init(ieee80211_priv(netdev));
zd_chip_init(&mac->chip, netdev, intf);
@@ -132,6 +138,8 @@ out:
void zd_mac_clear(struct zd_mac *mac)
{
+ skb_queue_purge(&mac->rx_queue);
+ tasklet_kill(&mac->rx_tasklet);
zd_chip_clear(&mac->chip);
ZD_ASSERT(!spin_is_locked(&mac->lock));
ZD_MEMCLEAR(mac, sizeof(struct zd_mac));
@@ -160,6 +168,8 @@ int zd_mac_open(struct net_device *netde
struct zd_chip *chip = &mac->chip;
int r;
+ tasklet_enable(&mac->rx_tasklet);
+
r = zd_chip_enable_int(chip);
if (r < 0)
goto out;
@@ -210,6 +220,8 @@ int zd_mac_stop(struct net_device *netde
*/
zd_chip_disable_rx(chip);
+ skb_queue_purge(&mac->rx_queue);
+ tasklet_disable(&mac->rx_tasklet);
housekeeping_disable(mac);
ieee80211softmac_stop(netdev);
@@ -873,45 +885,75 @@ static int fill_rx_stats(struct ieee8021
return 0;
}
-int zd_mac_rx(struct zd_mac *mac, const u8 *buffer, unsigned int length)
+static void zd_mac_rx(struct zd_mac *mac, struct sk_buff *skb)
{
int r;
struct ieee80211_device *ieee = zd_mac_to_ieee80211(mac);
struct ieee80211_rx_stats stats;
const struct rx_status *status;
- struct sk_buff *skb;
- if (length < ZD_PLCP_HEADER_SIZE + IEEE80211_1ADDR_LEN +
- IEEE80211_FCS_LEN + sizeof(struct rx_status))
- return -EINVAL;
+ if (skb->len < ZD_PLCP_HEADER_SIZE + IEEE80211_1ADDR_LEN +
+ IEEE80211_FCS_LEN + sizeof(struct rx_status))
+ {
+ dev_warn(zd_mac_dev(mac), "Packet with length %u to small.\n",
+ skb->len);
+ goto free_skb;
+ }
- r = fill_rx_stats(&stats, &status, mac, buffer, length);
- if (r)
- return r;
+ r = fill_rx_stats(&stats, &status, mac, skb->data, skb->len);
+ if (r) {
+ /* Only packets with rx errors are included here. */
+ goto free_skb;
+ }
- length -= ZD_PLCP_HEADER_SIZE+IEEE80211_FCS_LEN+
- sizeof(struct rx_status);
- buffer += ZD_PLCP_HEADER_SIZE;
+ __skb_pull(skb, ZD_PLCP_HEADER_SIZE);
+ __skb_trim(skb, skb->len -
+ (IEEE80211_FCS_LEN + sizeof(struct rx_status)));
- update_qual_rssi(mac, buffer, length, stats.signal, stats.rssi);
+ update_qual_rssi(mac, skb->data, skb->len, stats.signal,
+ status->signal_strength);
- r = filter_rx(ieee, buffer, length, &stats);
- if (r <= 0)
- return r;
+ r = filter_rx(ieee, skb->data, skb->len, &stats);
+ if (r <= 0) {
+ if (r < 0)
+ dev_dbg_f(zd_mac_dev(mac), "Error in packet.\n");
+ goto free_skb;
+ }
- skb = dev_alloc_skb(sizeof(struct zd_rt_hdr) + length);
- if (!skb)
- return -ENOMEM;
if (ieee->iw_mode == IW_MODE_MONITOR)
- fill_rt_header(skb_put(skb, sizeof(struct zd_rt_hdr)), mac,
+ fill_rt_header(skb_push(skb, sizeof(struct zd_rt_hdr)), mac,
&stats, status);
- memcpy(skb_put(skb, length), buffer, length);
r = ieee80211_rx(ieee, skb, &stats);
- if (!r) {
- ZD_ASSERT(in_irq());
- dev_kfree_skb_irq(skb);
+ if (r)
+ return;
+free_skb:
+ /* We are always in a soft irq. */
+ dev_kfree_skb(skb);
+}
+
+static void do_rx(unsigned long mac_ptr)
+{
+ struct zd_mac *mac = (struct zd_mac *)mac_ptr;
+ struct sk_buff *skb;
+
+ while ((skb = skb_dequeue(&mac->rx_queue)) != NULL)
+ zd_mac_rx(mac, skb);
+}
+
+int zd_mac_rx_irq(struct zd_mac *mac, const u8 *buffer, unsigned int length)
+{
+ struct sk_buff *skb;
+
+ skb = dev_alloc_skb(sizeof(struct zd_rt_hdr) + length);
+ if (!skb) {
+ dev_warn(zd_mac_dev(mac), "Could not allocate skb.\n");
+ return -ENOMEM;
}
+ skb_reserve(skb, sizeof(struct zd_rt_hdr));
+ memcpy(__skb_put(skb, length), buffer, length);
+ skb_queue_tail(&mac->rx_queue, skb);
+ tasklet_schedule(&mac->rx_tasklet);
return 0;
}
diff --git a/drivers/net/wireless/zd1211rw/zd_mac.h b/drivers/net/wireless/zd1211rw/zd_mac.h
index b8ea3de..c1bea87 100644
--- a/drivers/net/wireless/zd1211rw/zd_mac.h
+++ b/drivers/net/wireless/zd1211rw/zd_mac.h
@@ -133,6 +133,10 @@ struct zd_mac {
/* Unlocked reading possible */
struct iw_statistics iw_stats;
struct housekeeping housekeeping;
+
+ struct tasklet_struct rx_tasklet;
+ struct sk_buff_head rx_queue;
+
unsigned int stats_count;
u8 qual_buffer[ZD_MAC_STATS_BUFFER_SIZE];
u8 rssi_buffer[ZD_MAC_STATS_BUFFER_SIZE];
@@ -174,7 +178,7 @@ int zd_mac_open(struct net_device *netde
int zd_mac_stop(struct net_device *netdev);
int zd_mac_set_mac_address(struct net_device *dev, void *p);
-int zd_mac_rx(struct zd_mac *mac, const u8 *buffer, unsigned int length);
+int zd_mac_rx_irq(struct zd_mac *mac, const u8 *buffer, unsigned int length);
int zd_mac_set_regdomain(struct zd_mac *zd_mac, u8 regdomain);
u8 zd_mac_get_regdomain(struct zd_mac *zd_mac);
diff --git a/drivers/net/wireless/zd1211rw/zd_usb.c b/drivers/net/wireless/zd1211rw/zd_usb.c
index 3faaeb2..4a5f5d5 100644
--- a/drivers/net/wireless/zd1211rw/zd_usb.c
+++ b/drivers/net/wireless/zd1211rw/zd_usb.c
@@ -599,13 +599,13 @@ static void handle_rx_packet(struct zd_u
n = l+k;
if (n > length)
return;
- zd_mac_rx(mac, buffer+l, k);
+ zd_mac_rx_irq(mac, buffer+l, k);
if (i >= 2)
return;
l = (n+3) & ~3;
}
} else {
- zd_mac_rx(mac, buffer, length);
+ zd_mac_rx_irq(mac, buffer, length);
}
}
--
1.4.1
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists