[coreboot] usbrom debugging

Jason Wang wangqingpei at gmail.com
Mon Aug 24 03:32:56 CEST 2009


On Sun, Aug 23, 2009 at 5:45 AM, Kevin O'Connor <kevin at koconnor.net> wrote:

> On Sun, Aug 23, 2009 at 04:03:40AM +0800, Jason Wang wrote:
> > Hi all,
> >     I am trying to debug why the grub crashed (under usbrom) after
> loading
> > certain blocks into the memory.
> > My new found is that while usbrom is trying to load (block
> lba=0x0019a9e0).
> > it crashed. the flow is
> > usb_new_read(0,0x0019a9e0,0x20,0xfffa2000), 0x20 is the count which
> should
> > be read, and the 0xfffa2000 is the
> > virtual address of (0x70000).
> > usb_new_read call readwrite_blocks then execute_command. after executed
> > wrap_cbw in execute_command.
> > it should call dev->controller->bulk, this address is a wrong address. my
> > debug information are below:
>
> Where in memory is "dev->controller->bulk"?  You way want to check if
> its contents were overwritten by something (eg, grub, linux, seabios,
> or a data block loaded by the usb rom).
>
dev->controller-bulk is memory is 5d2d, but the point which direct to
dev->controller is 6Ef30,which physical memory is
6EF30+CE000>0x100000,what's why grub kernel can load some blocks to 0x100000
but crashed later. grub overwrite the
space which usbrom use.
So the problem is be confirmed with memory.
Kevin, how can i use PMM by seabios from usbrom? it should be much about 1M
memory that should be used.
Does PMM can suppport so large or just 40K of low memory/64K high memory

>
> -Kevin
>



-- 
Jason Wang
Peking University
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.coreboot.org/pipermail/coreboot/attachments/20090824/07e479ba/attachment.html>


More information about the coreboot mailing list