AM35x USB Host EHCI stalling

Krunal Patel_11

Other Parts Discussed in Thread: AM3505, OMAP3530

Hello TI,

We are using AM35xx Sitara processor based custom board. In our use case we need to display capture data on LCD through DSS as well as stream the data using USB to Ethernet converter.

We observe one issue that when there is heavy traffic through USB and simultaneously DSS displays data on LCD, then after approx 15 minutes we get USB Stall.

We referred AM3505 Errata - Advisory 1.1.36 and we found that this can happen due to DSS consuming high bandwidth resulting in USB Host EHCI stall. So we disabled DSS with following sysfs command:

echo 0 > /sys/devices/omapdss/display0/enabled

But still after disabling DSS, we observed that USB stall now happens after approx an hour and we get below kernel messages.

[12363.164337] WARNING: at net/sched/sch_generic.c:258 dev_watchdog+0x164/0x254()

[12363.171966] NETDEV WATCHDOG: eth1 (smsc95xx): transmit queue 0 timed out

[12363.179046] Modules linked in: ifb act_mirred sch_ingress cls_u32 sch_prio ipt_MASQUERADE iptable_nat nf_nat iptable_mangle xt_mark xt_tcpudp wl12xx_sdio wl12xx mac80211 cfg80211 c]

[12363.221618] [<c008b440>] (unwind_backtrace+0x0/0xec) from [<c00ae2b8>] (warn_slowpath_common+0x4c/0x64)

[12363.231536] [<c00ae2b8>] (warn_slowpath_common+0x4c/0x64) from [<c00ae350>] (warn_slowpath_fmt+0x2c/0x3c)

[12363.241638] [<c00ae350>] (warn_slowpath_fmt+0x2c/0x3c) from [<c03636a4>] (dev_watchdog+0x164/0x254)

[12363.251190] [<c03636a4>] (dev_watchdog+0x164/0x254) from [<c00b7c64>] (run_timer_softirq+0x1d8/0x28c)

[12363.260925] [<c00b7c64>] (run_timer_softirq+0x1d8/0x28c) from [<c00b35c0>] (__do_softirq+0x74/0xfc)

[12363.270446] [<c00b35c0>] (__do_softirq+0x74/0xfc) from [<c00b3688>] (irq_exit+0x40/0x8c)

[12363.278991] [<c00b3688>] (irq_exit+0x40/0x8c) from [<c007b078>] (asm_do_IRQ+0x78/0x98)

[12363.287384] [<c007b078>] (asm_do_IRQ+0x78/0x98) from [<c0411874>] (__irq_svc+0x34/0x80)

[12363.295806] Exception stack(0xc056bf80 to 0xc056bfc8)

[12363.301147] bf80: 00000000 40000013 40000013 00003505 c056a000 c05b95c0 c002e0e4 c056e844

[12363.309753] bfa0: 8002c6ec 411fc087 0000001f 00000000 c05816f8 c056bfc8 c0094d78 c0095388

[12363.318359] bfc0: 60000013 ffffffff

[12363.322052] [<c0411874>] (__irq_svc+0x34/0x80) from [<c0095388>] (omap3_pm_idle+0x48/0x4c)

[12363.330780] [<c0095388>] (omap3_pm_idle+0x48/0x4c) from [<c0086c38>] (cpu_idle+0x48/0x88)

[12363.339416] [<c0086c38>] (cpu_idle+0x48/0x88) from [<c0008bdc>] (start_kernel+0x254/0x2a8)

[12363.348144] [<c0008bdc>] (start_kernel+0x254/0x2a8) from [<80008034>] (0x80008034)

[12363.356109] ---[ end trace abe4b08a822dc09e ]---

[ 3] 4360.0-4365.0 sec 7.84 MB[12363.833435] smsc95xx 1-1.2.3:1.0: eth1: Failed to read register index 0x00000114

ytes 13.2 Mbits[12363.842498] smsc95xx 1-1.2.3:1.0: eth1: MII is busy in smsc95xx_mdio_read

/sec

[12363.851287] ehci-omap ehci-omap.0: remove, state 1

[12363.856658] usb usb1: USB disconnect, address 1

[12363.861419] usb 1-1: USB disconnect, address 2

[12363.866149] usb 1-1.2: USB disconnect, address 3

[12363.879180] usb 1-1.2.1: USB disconnect, address 6

[12363.886352] usb 1-1.2.2: USB disconnect, address 8

Can anyone suggest how can we avoid USB stalling?

Thanks,

Krunal

over 11 years ago

0 Biser Gatchev-XID over 11 years ago

TI__Guru**** 393215 points

Hi Krunal,

I will forward this to the USB experts.

0 Krunal Patel_11 over 11 years ago in reply to Biser Gatchev-XID

Expert 2250 points

Hello TI,

Any updates on this issue?

Thanks,

Krunal

0 Biser Gatchev-XID over 11 years ago in reply to Krunal Patel_11

TI__Guru**** 393215 points

No feedback so far.

0 Devang panchal over 11 years ago in reply to Biser Gatchev-XID

Expert 1050 points

Hi Biser,

Any update on this?
It will be great if we get any input for this issue as it is very critical for us.

Thanks,
Devang.

0 Biser Gatchev-XID over 11 years ago in reply to Devang panchal

TI__Guru**** 393215 points

I have asked again.

0 ranchu over 9 years ago in reply to Biser Gatchev-XID

Guru 20755 points

Hello,

I am having similiar relation between usb ehci & dss.
What's the connection in HW between these 2 modules ?

Thank you,
Ran

0 Biser Gatchev-XID over 9 years ago in reply to ranchu

TI__Guru**** 393215 points

The thing in common is the L3 interconnect.

0 ranchu over 9 years ago in reply to Biser Gatchev-XID

Guru 20755 points

Helllo Biser,

Thank you very much for the reply.
We have some issue in omap3530 (I know it's not the same is sitara, but I have no other forum to try to ask this question :) )
We see that graphic video (DSS, same DSS as in this thread, Right ? ) are not displayed if USB EHCI is supported in kernel.
I understand from the following link that there is an errata & patch on this issue, but I see that our code already contain the patch for this issue:
e2e.ti.com/.../29680

Yet, the graphic does not function with USB EHCI.
Is there some way to overcome this failure ?

Thank you,
Ran

0 Biser Gatchev-XID over 9 years ago in reply to ranchu

TI__Guru**** 393215 points

I am sorry, we do not support OMAP devices on this forum. There is no one on our Sitara software team that could help on this. I suggest you contact your local TI representative for support on this.

Processors

Processors forum

AM35x USB Host EHCI stalling