Skip to contentSkip to conversation

bcm43xx: Drain TX status before starting IRQs

PATCH In review Michael Buesch 6 participants 6 messages Oct 19, 2006 – Nov 4, 2006 3 messages not in the archive

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 bcm43xx: Drain TX status before starting IRQs

Drain the Microcode TX-status-FIFO before we enable IRQs. This is required, because the FIFO may still have entries left from a previous run. Those would immediately fire after enabling IRQs and would lead to an oops in the DMA TXstatus handling code.

--

Please consider also pushing this into the -stable tree. The bug is not likely to trigger, but at least ben triggered it in the past. Anyway, it can't hurt much to drain the FIFO before running the device.

Note that this is diffed against 2.6.18.1 and not 2.6.18 as the diff prolog suggests. I just forgot to rename the directory. ;)

drivers/net/wireless/bcm43xx/bcm43xx_main.c +16 -0
Index: linux-2.6.18/drivers/net/wireless/bcm43xx/bcm43xx_main.c
===================================================================
@@ -1463,6 +1463,21 @@ static void handle_irq_transmit_status(s
 	}
 }
 
+static void drain_txstatus_queue(struct bcm43xx_private *bcm)
+{
+	u32 dummy;
+
+	/* Read all entries from the microcode TXstatus FIFO
+	 * and throw them away.
+	 */
+	while (1) {
+		dummy = bcm43xx_read32(bcm, BCM43xx_MMIO_XMITSTAT_0);
+		if (!dummy)
+			break;
+		dummy = bcm43xx_read32(bcm, BCM43xx_MMIO_XMITSTAT_1);
+	}
+}
+
 static void bcm43xx_generate_noise_sample(struct bcm43xx_private *bcm)
 {
 	bcm43xx_shm_write16(bcm, BCM43xx_SHM_SHARED, 0x408, 0x7F7F);
@@ -3509,6 +3524,7 @@ int bcm43xx_select_wireless_core(struct 
 	bcm43xx_macfilter_clear(bcm, BCM43xx_MACFILTER_ASSOC);
 	bcm43xx_macfilter_set(bcm, BCM43xx_MACFILTER_SELF, (u8 *)(bcm->net_dev->dev_addr));
 	bcm43xx_security_init(bcm);
+	drain_txstatus_queue(bcm);
 	ieee80211softmac_start(bcm->net_dev);
 
 	/* Let's go! Be careful after enabling the IRQs.
Signature · 1 line
Greetings Michael.
#2 lore raw
In reply to a message not in the archive
Quoted 6 lines · Michael Buesch

Drain the Microcode TX-status-FIFO before we enable IRQs. This is required, because the FIFO may still have entries left from a previous run. Those would immediately fire after enabling IRQs and would lead to an oops in the DMA TXstatus handling code.

Great, thanks. Note that I haven't yet hit the bug since I updated the firmware, so it could be a mix of firmware and that problem, though it's not been long enough to be a clear results.

In any case, that patch should go in.

Cheers,
Ben.

#3 lore raw
In reply to Michael Buesch
Quoted 9 lines · Michael Buesch

Drain the Microcode TX-status-FIFO before we enable IRQs. This is required, because the FIFO may still have entries left from a previous run. Those would immediately fire after enabling IRQs and would lead to an oops in the DMA TXstatus handling code.

--

Quoted 8 lines

Please consider also pushing this into the -stable tree. The bug is not likely to trigger, but at least ben triggered it in the past. Anyway, it can't hurt much to drain the FIFO before running the device.

Note that this is diffed against 2.6.18.1 and not 2.6.18 as the diff prolog suggests. I just forgot to rename the directory. ;)

Has anyone used this patch, particularly with WPA encryption? When I try it, wpa_supplicant immediately uses 90+% of the cpu and never actually authenticates with my AP. I wonder if it is something with my system.

Thanks,

Larry

#5 lore raw
In reply to a message not in the archive
Quoted 3 lines · Larry Finger

Has anyone used this patch, particularly with WPA encryption? When I try it, wpa_supplicant immediately uses 90+% of the cpu and never actually authenticates with my AP. I wonder if it is something with my system.

Unless wpa supplicant is stupid [1] this cannot cause it any problems.

johannes

[1] by that, I mean sending out frames, turning off and on the interface, and then hoping to get a transmit status for frames sent before off/on cycle.

#6 lore raw
Quoted 4 lines · Johannes Berg
Quoted 3 lines · Larry Finger

Has anyone used this patch, particularly with WPA encryption? When I try it, wpa_supplicant immediately uses 90+% of the cpu and never actually authenticates with my AP. I wonder if it is something with my system.

What does wpa_supplicant do here? Have you looked at the debug log? Is it in some kind of busy loop doing something?

Quoted 7 lines

Unless wpa supplicant is stupid [1] this cannot cause it any problems.

johannes

[1] by that, I mean sending out frames, turning off and on the interface, and then hoping to get a transmit status for frames sent before off/on cycle.

wpa_supplicant does not receive or query TX status (at least yet; in the user space MLME case this is likely to change at some point).

Signature · 1 line
Jouni Malinen                                            PGP id EFC895FA
#8 lore raw
In reply to a message not in the archive
Quoted 8 lines · Jouni Malinen
Quoted 4 lines · Johannes Berg
Quoted 3 lines · Larry Finger

Has anyone used this patch, particularly with WPA encryption? When I try it, wpa_supplicant immediately uses 90+% of the cpu and never actually authenticates with my AP. I wonder if it is something with my system.

What does wpa_supplicant do here? Have you looked at the debug log? Is it in some kind of busy loop doing something?

I started wpa_supplicant in the foreground with the -w and -ddd options. When the BCM4306 card was plugged in, the following was output (this is hand transcribed and may not be exact):

Added interface wlan0
Ignore event for foreign ifindex 8
l2_packet_receive - recvfrom: Network is down
Ignore event for foreign ifindex 8
State: DISCONNECTED -> SCANNING
Starting AP scan (specific SSID)
Scan SSID - hexdump_ascii (len = 6)

         6c 77 66 64 6a 66               lwfdjf

ioctl[SIOGSIWSCAN]: No such device
Failed to initiate AP scan.
Setting scan request: 10 sec 0 usec
Ignore event for foreign ifindex 8

At this point, the system was in the state described earlier. I could not kill wpa_supplicant nor do anything else, and had to power off to recover.

Larry