mvneta: cannot request irq 25 on openblocks-ax3

Hi,

Here I've hit a bug on the recent kernel. As far as I know, this bug exists on 3.9-rc1 too.

When I tried the latest mvebu for-next tree
(git://git.infradead.org/users/jcooper/linux.git mvebu/for-next), I got below warning at bootup time and mvneta didn't work (link was never up). I ensured that "ifconfig ethX up" always caused that.

Does anyone succeed to boot openblocks-ax3 recently or hit same trouble?

------------[ cut here ]------------
WARNING: at /ssd/ksrc/linux-3/kernel/irq/manage.c:1417
request_threaded_irq+0x64/0x124()
Modules linked in:
Backtrace:
[<c00119a4>] (dump_backtrace+0x0/0x110) from [<c04b2bac>] (dump_stack+0x18/0x1c)
 r6:c05ad11a r5:00000589 r4:00000000 r3:ed80e000
[<c04b2b94>] (dump_stack+0x0/0x1c) from [<c0021484>]
(warn_slowpath_common+0x54/0x70)
[<c0021430>] (warn_slowpath_common+0x0/0x70) from [<c00214c4>]
(warn_slowpath_null+0x24/0x2c)
 r8:00000000 r7:ef0a4600 r6:ee9c4d40 r5:00000000 r4:c037378c

r3:00000009
[<c00214a0>] (warn_slowpath_null+0x0/0x2c) from [<c0084bb8>] (request_threaded_irq+0x64/0x124)
[<c0084b54>] (request_threaded_irq+0x0/0x124) from [<c037418c>] (mvneta_open+0x80/0x15c)
[<c037410c>] (mvneta_open+0x0/0x15c) from [<c03c9e3c>] (__dev_open+0xa0/0xf4)
r6:00001002 r5:c050bb24 r4:ee9c4800
[<c03c9d9c>] (__dev_open+0x0/0xf4) from [<c03ca070>] (__dev_change_flags+0x94/0x118)
r7:00000001 r6:00001002 r5:00001003 r4:ee9c4800
[<c03c9fdc>] (__dev_change_flags+0x0/0x118) from [<c03ca170>] (dev_change_flags+0x18/0x4c)
r7:00000000 r6:edd0d380 r5:00001002 r4:ee9c4800
[<c03ca158>] (dev_change_flags+0x0/0x4c) from [<c03d6168>] (do_setlink+0x2a4/0x7a0)
r6:edd0d380 r5:ed80fb4c r4:ee9c4800 r3:00001002
[<c03d5ec4>] (do_setlink+0x0/0x7a0) from [<c03d72d8>] (rtnl_newlink+0x248/0x45c) [<c03d7090>] (rtnl_newlink+0x0/0x45c) from [<c03d7060>] (rtnetlink_rcv_msg+0x248/0x278)
[<c03d6e18>] (rtnetlink_rcv_msg+0x0/0x278) from [<c03ead78>] (netlink_rcv_skb+0x58/0xb4)
[<c03ead20>] (netlink_rcv_skb+0x0/0xb4) from [<c03d5d28>] (rtnetlink_rcv+0x20/0x2c)
r6:eea75c00 r5:eea87240 r4:eea87240 r3:c03d5d08
[<c03d5d08>] (rtnetlink_rcv+0x0/0x2c) from [<c03ea780>] (netlink_unicast+0x15c/0x214)
r4:ef0a3000 r3:c03d5d08
[<c03ea624>] (netlink_unicast+0x0/0x214) from [<c03eab94>] (netlink_sendmsg+0x2b8/0x348)
[<c03ea8dc>] (netlink_sendmsg+0x0/0x348) from [<c03b2dbc>] (sock_sendmsg+0x88/0xa4) [<c03b2d34>] (sock_sendmsg+0x0/0xa4) from [<c03b3330>] (__sys_sendmsg+0x1c8/0x25c)
r7:00000000 r6:00000020 r5:eefccc80 r4:ed80ff64
[<c03b3168>] (__sys_sendmsg+0x0/0x25c) from [<c03b4f30>] (sys_sendmsg+0x44/0x68) [<c03b4eec>] (sys_sendmsg+0x0/0x68) from [<c000daa0>] (ret_fast_syscall+0x0/0x30)
r6:be9f367c r5:00000000 r4:5146f675

---[ end trace a8ceb54233e388b5 ]---

mvneta d0070000.ethernet eth0: cannot request irq 25

Thank you,

Masami HIRAMATSU
IT Management Research Dept. Linux Technology Center
Hitachi, Ltd., Yokohama Research Laboratory
E-mail: masami.hiramatsu.pt at hitachi.com

FYI: that branch isn't stable, it's used as a merge-test of arm-soc/for-next (also not stable) and any branches I am trying to push upstream that day.

Gregory has a patch in the works for this. Hopefully he'll submit it by the end of the week.

thx,

Jason.

Hi Jason,

Thanks! could you tell me which branch is stable? (however, I'd like to try new fixes/features on my device too :))

Thank you,

Masami HIRAMATSU
IT Management Research Dept. Linux Technology Center
Hitachi, Ltd., Yokohama Research Laboratory
E-mail: masami.hiramatsu.pt at hitachi.com

Generally, we advise using one of Linus' tags (eg v3.9-rc3). Only if that doesn't work, seek out a branch containing the fix/feature. When in doubt, I'll do:

$ ./scripts/get_maintainer.pl -f drivers/net/ethernet/marvell/mvneta.c

and ask those listed which branch might contain a fix or feature.

In this case, it'll probably go through David Miller and the netdev mailinglist, but it isn't there yet. ;-)

hth,

Jason.

Hi Masami,

This is a known bug. Gregory Clement already has a fix and he will submit it soon. In case you need this fixed ASAP, I'm attaching you a patch with a fix.

Please note the attached patch is not ready for mainline inclusion, as I said Gregory will submit a cleaner version soon.

Ezequiel Garc?a, Free Electrons
Embedded Linux, Kernel and Android Engineering
http://free-electrons.com
-------------- next part --------------

From: Gregory CLEMENT <[email protected]>
Date: Sat, 9 Feb 2013 22:07:54 +0100
Subject: [PATCH] net: mvneta: convert to percpu interrupt

---
 drivers/net/ethernet/marvell/mvneta.c |   12 +++++++++---
 1 files changed, 9 insertions(+), 3 deletions(-)

index b6025c3..7f63dd4 100644
@@ -1800,7 +1800,7 @@ static void mvneta_set_rx_mode(struct net_device *dev)
 /* Interrupt handling - the callback for request_irq() */
 static irqreturn_t mvneta_isr(int irq, void *dev_id)
 {
-	struct mvneta_port *pp = (struct mvneta_port *)dev_id;
+	struct mvneta_port *pp = *(struct mvneta_port **)dev_id;
 
 	/* Mask all interrupts */
 	mvreg_write(pp, MVNETA_INTR_NEW_MASK, 0);

@@ -2368,6 +2368,7 @@ static void mvneta_mdio_remove(struct mvneta_port *pp)
 	phy_disconnect(pp->phy_dev);
 	pp->phy_dev = NULL;
 }
+static struct mvneta_port __percpu **percpu_pp;
 
 static int mvneta_open(struct net_device *dev)
 {

@@ -2386,9 +2387,14 @@ static int mvneta_open(struct net_device *dev)
 	if (ret)
 		goto err_cleanup_rxqs;
 
+	percpu_pp = alloc_percpu(struct mvneta_port *);
+	*__this_cpu_ptr(percpu_pp) = pp;
+
 	/* Connect to port interrupt line */
-	ret = request_irq(pp->dev->irq, mvneta_isr, 0,
-			  MVNETA_DRIVER_NAME, pp);
+	ret = request_percpu_irq(pp->dev->irq, mvneta_isr,
+			  MVNETA_DRIVER_NAME, percpu_pp);
+	enable_percpu_irq(pp->dev->irq, 0);
+
 	if (ret) {
 		netdev_err(pp->dev, "cannot request irq %d\n", pp->dev->irq);
 		goto err_cleanup_txqs;

1.7.8.6

Hi Ezequiel,

Thanks! I'll try that.

Yeah, I look forward to it :)

Thank you,

Masami HIRAMATSU
IT Management Research Dept. Linux Technology Center
Hitachi, Ltd., Yokohama Research Laboratory
E-mail: masami.hiramatsu.pt at hitachi.com

Hi Masami,

You can try this patch if you want.
I don't have the hardware today so I didn't test it.
If you (and also Florian and Ezequiel) can test it and if it fixed the bug, then I will be able send a proper email for it,

Thanks,

Gregory Clement, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com

-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0001-net-mvneta-convert-to-local-interrupt.patch
Type: text/x-diff
Size: 3140 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20130319/46ad8983/attachment.bin>

I just tested your patch on RD-A370-A1 and DB-MV784MP-GP and it works fine, thanks! Feel free to add my Tested-by: Florian Fainelli <[email protected]> to your next submission.

Thanks!

Florian

Thanks for having tested it, but with this patch there is still an issue when we try to release the network interface. In this case the the interrupt line is not properly freed.
Unfortunately the state of the percpu_irq API prevents the driver to use it.

Indeed the interrupts have to be freed when the .stop() function is called. As the free_percpu_irq() function don't disable the interrupt line, we have to do it on each CPU before calling this. The function disable_percpu_irq() only disable the percpu on the current CPU and there is no function which allows to disable a percpu irq on a given CPU. Waiting for the extension of the percpu_irq API, this fix allows to use again the mvneta driver.

I have just wrote and tested a new patch to fix this issue by just go back on the old way we managed the irq for the mvenat as global irq and not as local ones.

I am going to send the fix on the LAKML in a few minutes.

Thanks,

Gregory

Hi Gregory,

Thank you for giving me the latest patch.
I've tested it on OpenBlocks-AX4 and that fixed the problem!

Thanks,

Masami HIRAMATSU
IT Management Research Dept. Linux Technology Center
Hitachi, Ltd., Yokohama Research Laboratory
E-mail: masami.hiramatsu.pt at hitachi.com