Re: u-boot debug, was: Re: U-boot on RPI3, sees disk but won't boot it

From: Mark Millard <marklmi_at_yahoo.com>
Date: Thu, 29 Sep 2022 19:55:00 UTC
On 2022-Sep-29, at 10:09, bob prohaska <fbsd@www.zefox.net> wrote:

> On Thu, Sep 29, 2022 at 09:38:04AM -0700, Mark Millard wrote:
>> On 2022-Sep-29, at 08:19, bob prohaska <fbsd@www.zefox.net> wrote:
>> 
>>> 
>>> I'll try to gather a few more examples.
>>> 
>> 
>> FYI:
>> 
>> Building based on the latest patch-common_usb__storage.c
>> that I sent (by itself) would avoid the huge number of
>> sequences like:
>> 
>> usb_read: udev 0
>> 
> [snip]
>> 
>> and make the log files much smaller. Any other messages
>> from usb_storage.c would still still be present.
>> 
> 
> 
> From my point of view the size isn't a serious problem, the
> script command makes capturing output easy. It seemed more
> valuable to capture maximum detail until the problem is 
> better understood. If it's understood well enough now just
> say the word and I'll recompile.
> 
> Size might become an issue, but there are larger capacity 
> webservers available. I'm using www.zefox.net out of habit.

In part I was going by your indication to avoid "bloat",
which I had guessed was a reference to the logging output
instead of the size of u-boot.bin :

QUOTE
The maximum logging level was 4 out of 7, fearing bloat.
END QUOTE

The first "usb_read:" is after the "1 Storage Device(s)
found". Prior to that there is one example of a:

COMMAND phase
DATA phase
STATUS phase

sequence and one of a:

COMMAND phase
STATUS phase
COMMAND phase
DATA phase
STATUS phase

sequence. I'd be surprised if these 8 lines proved useful
for the "0 Storage Device(s) found" issue.

For issues after "Storage Device(s) found" the status is
less clear and the output would be more likely to
contribute.

As long as you are happy with the size of the output, it
should be fine either way for "Storage Device(s) found"
observations.


On a different issue . . .

I'm also hoping that you will happen to also capture
examples of automatic reset/start-over sequences. It
may be that those requiring getting to the U-Boot prompt
and typing commands. (Has it ever happened otherwise?)


===
Mark Millard
marklmi at yahoo.com