Hi all,
I noticed a xhci warning on Stratus fault-tolerant box running automated
surprise device removal tests over the weekend:
irq event 95: bogus return value ffffff94
CPU: 0 PID: 31710 Comm: kworker/u97:2 Tainted: PF O-------------- 3.10.0-229.el7.x86_64 #1
Hardware name: Stratus ftServer 6800/G7LYY, BIOS BIOS Version 8.0:38 03/30/2015
Workqueue: events_unbound async_run_entry_fn
ffff88037dd8ada0 000000009c52d75c ffff88085fa03cd8 ffffffff81603f36
ffff88085fa03d00 ffffffff8110d852 ffff88037dd8ada0 000000000000005f
00000000ffffff94 ffff88085fa03d40 ffffffff8110dc1c 000000009c52d75c
Call Trace:
<IRQ> [<ffffffff81603f36>] dump_stack+0x19/0x1b
[<ffffffff8110d852>] __report_bad_irq+0x32/0xd0
[<ffffffff8110dc1c>] note_interrupt+0xdc/0x1f0
[<ffffffff8110b381>] handle_irq_event_percpu+0xe1/0x1e0
[<ffffffff8110b4bd>] handle_irq_event+0x3d/0x60
[<ffffffff8110e157>] handle_edge_irq+0x77/0x130
[<ffffffff81015c9f>] handle_irq+0xbf/0x150
[<ffffffff81077d27>] ? irq_enter+0x17/0xa0
[<ffffffff8161626f>] do_IRQ+0x4f/0xf0
[<ffffffff8160b4ed>] common_interrupt+0x6d/0x6d
[<ffffffff810abe65>] ? sched_clock_cpu+0xb5/0x100
[<ffffffff81077adb>] ? __do_softirq+0x9b/0x280
[<ffffffff816156dc>] call_softirq+0x1c/0x30
[<ffffffff81015d95>] do_softirq+0x65/0xa0
[<ffffffff81077ec5>] irq_exit+0x115/0x120
[<ffffffff81616355>] smp_apic_timer_interrupt+0x45/0x60
[<ffffffff81614a1d>] apic_timer_interrupt+0x6d/0x80
<EOI> [<ffffffff8107089f>] ? vprintk_emit+0x1bf/0x530
[<ffffffff815fd9ef>] printk+0x77/0x8e
[<ffffffffa0390340>] ? _scsih_suspend+0xb0/0xb0 [mpt3sas]
[<ffffffffa038b144>] mpt3sas_port_enable+0x24/0x100 [mpt3sas]
[<ffffffffa039038f>] _scsih_scan_start+0x4f/0x70 [mpt3sas]
[<ffffffff813fc6c7>] do_scsi_scan_host+0x37/0xa0
[<ffffffff813fc8fc>] do_scan_async+0x1c/0x150
[<ffffffff8109e839>] async_run_entry_fn+0x39/0x120
[<ffffffff8108f0ab>] process_one_work+0x17b/0x470
[<ffffffff8108fe8b>] worker_thread+0x11b/0x400
[<ffffffff8108fd70>] ? rescuer_thread+0x400/0x400
[<ffffffff8109726f>] kthread+0xcf/0xe0
[<ffffffff810971a0>] ? kthread_create_on_node+0x140/0x140
[<ffffffff81613cfc>] ret_from_fork+0x7c/0xb0
[<ffffffff810971a0>] ? kthread_create_on_node+0x140/0x140
handlers:
[<ffffffff81458c70>] xhci_msi_irq
drivers/usb/host/xhci-ring.c :
irqreturn_t xhci_irq(struct usb_hcd *hcd)
{
struct xhci_hcd *xhci = hcd_to_xhci(hcd);
u32 status;
u64 temp_64;
union xhci_trb *event_ring_deq;
dma_addr_t deq;
spin_lock(&xhci->lock);
/* Check if the xHC generated the interrupt, or the irq is shared */
status = readl(&xhci->op_regs->status);
if (status == 0xffffffff)
goto hw_died;
if (!(status & STS_EINT)) {
spin_unlock(&xhci->lock);
return IRQ_NONE;
}
if (status & STS_FATAL) {
xhci_warn(xhci, "WARNING: Host System Error\n");
xhci_halt(xhci);
hw_died:
spin_unlock(&xhci->lock);
return -ESHUTDOWN;
}
...
So -ESHUTDOWN = -108 (0xffffff94) provoked bad_action_ret into reporting
a bogus return value and stack trace above.
This error message and stack trace has been reported in the past [1], and
I was wondering about the value of the stack trace (at least in the
hw_died case). A quick survey of the other host controller drivers that
bother checking for 0xffffffff shows they all do something like this:
if (status == ~(u32) 0) {
ehci_dbg (ehci, "device removed\n");
goto dead;
}
...
dead:
...
return IRQ_HANDLED;
but xhci-ring.c looks to be the only one to return !IRQ_XXX from its irq
handler.
Granted, most folks aren't hotplugging USB host controllers, but this is
supported on Stratus HW. Are there any options about the following patch?
Regards,
-- Joe
[1] https://bugzilla.redhat.com/show_bug.cgi?id=692425
--->8-- -->8-- -->8--
>From ff69f1bb5601ce5f0e70bb2e97c65456e13dc38e Mon Sep 17 00:00:00 2001
From: Joe Lawrence <joe.lawrence@xxxxxxxxxxx>
Date: Mon, 20 Apr 2015 11:14:47 -0400
Subject: [PATCH] xhci: gracefully handle xhci_irq dead device
If the xHCI host controller has died (ie, device removed), then xhci_irq
should quietly handle this condition with IRQ_HANDLED.
Signed-off-by: Joe Lawrence <joe.lawrence@xxxxxxxxxxx>
---
drivers/usb/host/xhci-ring.c | 9 +++++----
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/drivers/usb/host/xhci-ring.c b/drivers/usb/host/xhci-ring.c
index f5397a517c54..42bd3e2670bf 100644
--- a/drivers/usb/host/xhci-ring.c
+++ b/drivers/usb/host/xhci-ring.c
@@ -2626,11 +2626,13 @@ irqreturn_t xhci_irq(struct usb_hcd *hcd)
dma_addr_t deq;
spin_lock(&xhci->lock);
- /* Check if the xHC generated the interrupt, or the irq is shared */
status = readl(&xhci->op_regs->status);
- if (status == 0xffffffff)
- goto hw_died;
+ if (status == 0xffffffff) {
+ spin_unlock(&xhci->lock);
+ return IRQ_HANDLED;
+ }
+ /* Check if the xHC generated the interrupt, or the irq is shared */
if (!(status & STS_EINT)) {
spin_unlock(&xhci->lock);
return IRQ_NONE;
@@ -2638,7 +2640,6 @@ irqreturn_t xhci_irq(struct usb_hcd *hcd)
if (status & STS_FATAL) {
xhci_warn(xhci, "WARNING: Host System Error\n");
xhci_halt(xhci);
-hw_died:
spin_unlock(&xhci->lock);
return -ESHUTDOWN;
}
--
1.8.3.1
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
