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
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date:   Tue, 28 Feb 2017 08:07:52 -0800
From:   Michael Zoran <mzoran@...wfest.net>
To:     Baruch Siach <baruch@...s.co.il>, linux-usb@...r.kernel.org
Cc:     John Youn <johnyoun@...opsys.com>, netdev@...r.kernel.org,
        Olivier Sobrie <olivier@...rie.be>,
        linux-rpi-kernel@...ts.infradead.org,
        linux-arm-kernel@...ts.infradead.org
Subject: Re: usb/net/hso: WARNING: ungligned urb->setup_dma

On Tue, 2017-02-28 at 14:01 +0200, Baruch Siach wrote:
> Hi linux-usb list,
> 
> (Dropped Jan's bouncing address, added Rpi3 platform lists)
> 
> On Tue, Feb 28, 2017 at 10:28:10AM +0200, Baruch Siach wrote:
> > Hi linux-usb list,
> > 
> > I'm hitting this warning consistently on my Raspberry Pi 3 running
> > kernel 
> > v4.10.1 with some unrelated device tree changes, and a debug print
> > (below). 
> > The device identifies as "GlobeTrotter HSDPA Modem", VID: 0af0,
> > PID: 6971.
> > The warning triggers consistently on first write access to
> > /dev/ttyHS0 that 
> > ModemManager attempts. The first line in the log is my debug print.
> 
> I tested the same hardware successfully on an i.MX6 CuBox-i (ARM32)
> using the 
> same kernel version (4.10.1), and on an x86_64 PC (4.9). So this
> seems to be 
> platform specific. I don't have any other ARM64 machine at the
> moment, though.
> 
> Has anyone noticed USB host issues on Raspberry Pi 3 with 64bit
> kernel?
> 
> baruch
> 
> > [   65.004892] dwc2_map_urb_for_dma: setup_dma: f5635066
> > [   65.010718] ------------[ cut here ]------------
> > [   65.016079] WARNING: CPU: 1 PID: 794 at
> > drivers/usb/dwc2/hcd.c:2631 dwc2_map_urb_for_dma+0x94/0x130
> > [   65.025930] 
> > [   65.028143] CPU: 1 PID: 794 Comm: ModemManager Not tainted
> > 4.10.1-00008-g76809ec3947f-dirty #719
> > [   65.037747] Hardware name: Raspberry Pi 3 Model B (DT)
> > [   65.043656] task: ffff8000362e3e80 task.stack: ffff80003556c000
> > [   65.050359] PC is at dwc2_map_urb_for_dma+0x94/0x130
> > [   65.056090] LR is at dwc2_map_urb_for_dma+0x34/0x130
> > [   65.061819] pc : [<ffff0000083ff0ac>] lr : [<ffff0000083ff04c>]
> > pstate: 000001c5
> > [   65.070018] sp : ffff80003556fa70
> > [   65.074080] x29: ffff80003556fa70 x28: 0000000000000005 
> > [   65.080167] x27: ffff80003563d000 x26: ffff8000350d4600 
> > [   65.086252] x25: ffff800035686780 x24: 0000000000000002 
> > [   65.092247] x23: 0000000000000001 x22: 0000000001080020 
> > [   65.097635] x21: ffff800035425800 x20: 0000000001080020 
> > [   65.103024] x19: ffff800035b72400 x18: 0000000000000010 
> > [   65.108413] x17: 0000ffff908f5a40 x16: ffff0000081abad8 
> > [   65.113801] x15: 0000000000000006 x14: ffff00008882c047 
> > [   65.119191] x13: ffff00000882c055 x12: ffff00000882e458 
> > [   65.124579] x11: ffff80003556f7d0 x10: ffff0000087c9a38 
> > [   65.129968] x9 : ffff0000083613e0 x8 : 363566203a616d64 
> > [   65.135356] x7 : 5f7075746573203a x6 : 00000000000002fc 
> > [   65.140744] x5 : 0000000000000000 x4 : 0000000000000000 
> > [   65.146132] x3 : 0000000000000000 x2 : ffff8000362e3e80 
> > [   65.151521] x1 : 0000000000000001 x0 : ffff00000880e000 
> > [   65.156907] 
> > [   65.158414] ---[ end trace a78d20eaecac455a ]---
> > [   65.163093] Call trace:
> > [   65.165573] Exception stack(0xffff80003556f8a0 to
> > 0xffff80003556f9d0)
> > [   65.172108] f8a0: ffff800035b72400 0001000000000000
> > ffff80003556fa70 ffff0000083ff0ac
> > [   65.180052] f8c0: ffff0000086f8f70 ffff000008795000
> > ffff00000882dce8 ffff0000087fb318
> > [   65.187996] f8e0: ffff00000882c048 000000010882bae8
> > ffff80003556f990 ffff0000080dc664
> > [   65.195940] f900: ffff800035b72400 0000000001080020
> > ffff800035425800 0000000001080020
> > [   65.203883] f920: 0000000000000001 0000000000000002
> > ffff800035686780 ffff8000350d4600
> > [   65.211827] f940: ffff00000880e000 0000000000000001
> > ffff8000362e3e80 0000000000000000
> > [   65.219771] f960: 0000000000000000 0000000000000000
> > 00000000000002fc 5f7075746573203a
> > [   65.227715] f980: 363566203a616d64 ffff0000083613e0
> > ffff0000087c9a38 ffff80003556f7d0
> > [   65.235659] f9a0: ffff00000882e458 ffff00000882c055
> > ffff00008882c047 0000000000000006
> > [   65.243601] f9c0: ffff0000081abad8 0000ffff908f5a40
> > [   65.248548] [<ffff0000083ff0ac>] dwc2_map_urb_for_dma+0x94/0x130
> > [   65.254643] [<ffff0000083e8984>] usb_hcd_submit_urb+0x8c/0x8c0
> > [   65.260560] [<ffff0000083eb0b8>] usb_submit_urb+0x248/0x480
> > [   65.266215] [<ffff0000083d2d2c>] mux_device_request+0xdc/0x1f8
> > [   65.272134] [<ffff0000083d2e70>]
> > hso_mux_serial_write_data+0x28/0x38
> > [   65.278581] [<ffff0000083d2fa8>] hso_kick_transmit+0x88/0xb8
> > [   65.284323] [<ffff0000083d4dc8>] hso_serial_write+0x60/0xc0
> > [   65.289979] [<ffff000008345e80>] n_tty_write+0x300/0x3e0
> > [   65.295369] [<ffff000008340a50>] tty_write+0x170/0x2b8
> > [   65.300585] [<ffff0000081ab6ac>] __vfs_write+0x1c/0x100
> > [   65.305884] [<ffff0000081ab92c>] vfs_write+0x9c/0x1a8
> > [   65.311010] [<ffff0000081abb1c>] SyS_write+0x44/0xa0
> > [   65.316049] [<ffff000008082730>] el0_svc_naked+0x24/0x28
> > 
> > The debug print otherwise just shows:
> > 
> > [   64.476785] dwc2_map_urb_for_dma: setup_dma: 00000000
> > 
> > So it looks like urb->setup_dma content is garbage.
> > 
> > Any thoughts?
> > 
> > Thanks,
> > baruch
> 

I've seen similar issues twice before:

1. Back when arm64 was initially being ported to the RPI 3 by people on
the net back in March/April of last year and only the serial console
worked, DMA was broken in the dwc2 driver.  The dwc2 driver would send
and receive garbage data.  The reason was that some of the memory
manager DMA functionality for arm64 wasn't implemented yet.

2. I've also seen in complex cases things can run out of coherent DMA
pool.  If this is what is happening, which I doubt, you can try
increasing the size with the "coherent_pool=" command line option.

I don't know what the device tree changes are, but the dma-ranges in
the device tree is what determines how dma is setup on an arm64
machine.

Powered by blists - more mailing lists