JTAG debugging of the kernel

JTAG debugging of the kernel

[Peter Barada]

I've been trying to use JTAG (and an Abatron BDI-2000) to debug the ALSA
driver to determine why madplay can't keep up when pulling .mp3 from an
SD card, and have not had any luck trying to halt via JTAG and set/stop
at breakpoints within the kernelthe kernel.  I've found that JTAG works
for a while on kernel startup, but stops working at a non-reproducible
point in the startup code.

Have people had success using a BDI-2000 to debug the kernel after it
has completed startup?

Any suggestions on what can cause the JTAG to stop working after the
kernel starts?

Thanks in advance!

RE: JTAG debugging of the kernel

[Syed Mohammed, Khasim]

> -----Original Message-----
> From: linux-omap-owner@vger.kernel.org [mailto:linux-omap-owner@vger.kernel.org] On Behalf Of Peter
> Barada
> Sent: Friday, August 15, 2008 8:34 PM
> To: linux-omap@vger.kernel.org
> Subject: JTAG debugging of the kernel
> 
> I've been trying to use JTAG (and an Abatron BDI-2000) to debug the ALSA
> driver to determine why madplay can't keep up when pulling .mp3 from an
> SD card, and have not had any luck trying to halt via JTAG and set/stop
> at breakpoints within the kernelthe kernel.  I've found that JTAG works
> for a while on kernel startup, but stops working at a non-reproducible
> point in the startup code.
> 
> Have people had success using a BDI-2000 to debug the kernel after it
> has completed startup?
> 
> Any suggestions on what can cause the JTAG to stop working after the
> kernel starts?
> 

I think you should disable Power Management. Which kernel are you using now?

Regards,
Khasim

RE: JTAG debugging of the kernel

[Peter Barada]

On Fri, 2008-08-15 at 23:04 +0530, Syed Mohammed, Khasim wrote:
> 
> > -----Original Message-----
> > From: linux-omap-owner@vger.kernel.org [mailto:linux-omap-owner@vger.kernel.org] On Behalf Of Peter
> > Barada
> > Sent: Friday, August 15, 2008 8:34 PM
> > To: linux-omap@vger.kernel.org
> > Subject: JTAG debugging of the kernel
> > 
> > I've been trying to use JTAG (and an Abatron BDI-2000) to debug the ALSA
> > driver to determine why madplay can't keep up when pulling .mp3 from an
> > SD card, and have not had any luck trying to halt via JTAG and set/stop
> > at breakpoints within the kernelthe kernel.  I've found that JTAG works
> > for a while on kernel startup, but stops working at a non-reproducible
> > point in the startup code.
> > 
> > Have people had success using a BDI-2000 to debug the kernel after it
> > has completed startup?
> > 
> > Any suggestions on what can cause the JTAG to stop working after the
> > kernel starts?
> > 
> 
> I think you should disable Power Management. Which kernel are you using now?

2.6.22.18.  I'm planning to move to 2.6.26 after the dust settles...

I'll try w/o power management, but any idea why power management might
affect JTAG?

> Regards,
> Khasim

Re: JTAG debugging of the kernel

[Josh Karabin]

Peter Barada wrote:
> On Fri, 2008-08-15 at 23:04 +0530, Syed Mohammed, Khasim wrote:
>>> -----Original Message-----
>>> From: linux-omap-owner@vger.kernel.org [mailto:linux-omap-owner@vger.kernel.org] On Behalf Of Peter
>>> Barada

>>> Any suggestions on what can cause the JTAG to stop working after the
>>> kernel starts?
>>>
>> I think you should disable Power Management. Which kernel are you using now?
> 
> 2.6.22.18.  I'm planning to move to 2.6.26 after the dust settles...
> 
> I'll try w/o power management, but any idea why power management might
> affect JTAG?

The guy I share an office with (who's out on vacation) worked through 
this with their tech support, who suggested replacing WFI instructions 
with NOPs.

Look at cpu_v7_do_idle in arch/arm/mm/proc-v7.S.  Changing the WFI to a 
NOP did the trick for him.

Presumably power management was off, but even WFI (wait for interrupt) 
put the CPU in a state that wasn't handled by the BDI firmware at the 
time.  I'm not sure if they resolved the issue in a later firmware 
update or not.  We didn't think the workaround was worthy of a Kconfig 
option - hopefully updated firmware for the BDI box will address the issue.


- Josh

Re: JTAG debugging of the kernel

[Peter Barada]

On Fri, 2008-08-15 at 14:57 -0400, Josh Karabin wrote:
> Peter Barada wrote:
> > On Fri, 2008-08-15 at 23:04 +0530, Syed Mohammed, Khasim wrote:
> >>> -----Original Message-----
> >>> From: linux-omap-owner@vger.kernel.org [mailto:linux-omap-owner@vger.kernel.org] On Behalf Of Peter
> >>> Barada
> 
> >>> Any suggestions on what can cause the JTAG to stop working after the
> >>> kernel starts?
> >>>
> >> I think you should disable Power Management. Which kernel are you using now?
> > 
> > 2.6.22.18.  I'm planning to move to 2.6.26 after the dust settles...
> > 
> > I'll try w/o power management, but any idea why power management might
> > affect JTAG?
> 
> The guy I share an office with (who's out on vacation) worked through 
> this with their tech support, who suggested replacing WFI instructions 
> with NOPs.
> 
> Look at cpu_v7_do_idle in arch/arm/mm/proc-v7.S.  Changing the WFI to a 
> NOP did the trick for him.
> 
> Presumably power management was off, but even WFI (wait for interrupt) 
> put the CPU in a state that wasn't handled by the BDI firmware at the 
> time.  I'm not sure if they resolved the issue in a later firmware 
> update or not.  We didn't think the workaround was worthy of a Kconfig 
> option - hopefully updated firmware for the BDI box will address the issue.

Hmm, trying both disabling power management and changing the WFI to NOP
in cpu_v7_do_idle does not work, the BDI times out when I try to halt
the processor.  I'm running with version 1.07 of the BDI firmware.  Is
this what other people are successfully using to debug the kernel with a
BDI-2000?

> 
> - Josh