Repository navigation
Update some debugging information - #351
Conversation
lyakh
left a comment
lyakh
left a comment
There was a problem hiding this comment.
the missing -l was there before, but since you already touch that text, we can fix it as well
marc-hb
left a comment
marc-hb
left a comment
There was a problem hiding this comment.
The term "mailbox" is all over the code and probably in the documentation too so it's useful to mention it here.
| .. note:: | ||
| debugfs files used by ``sof-logger`` | ||
|
|
||
| - ``etrace``: direct access (memcpy) to TRACE window of the SOF firmware |
There was a problem hiding this comment.
| - ``etrace``: direct access (memcpy) to TRACE window of the SOF firmware | |
| - ``etrace``: direct access to TRACE window of the SOF firmware in the shared memory "mailbox" |
There was a problem hiding this comment.
@marc-hb, the trace window is not within the 'mailbox' memory. We have a bigger shared area and within that area we have the 'mailbox' (up and down region), trace, debug, stream, regs and exception windows.
Some of these windows are optional and only used for debugging, like the trace region.
Would it be better if we use the term 'region' instead of 'window' when referring to the TRACE window/region of the shared area?
There was a problem hiding this comment.
However the sof-logger's help documents the t option:
sof-logger: -t DMA 'trace' stream instead of 'etrace' error mailbox
Probably something like this:
etrace: direct access to TRACE error mailbox of the firmware
There was a problem hiding this comment.
We have a bigger shared area and within that area we have the 'mailbox' (up and down region), trace, debug, stream, regs and exception windows.
Thanks, can you please share some pointers to this information? It looks like our terminology is inconsistent.
However the sof-logger's help documents the t option:
I recently changed this help message so it does not count :-) However I changed it based on symbols like CONFIG_TRACEM and MAILBOX_TRACE_BASE and other comments git grep -i -e mailbox -e mbox -- src/trace/; most of these are not from me.
So maybe "mailbox" is not the correct term here. On the other hand, "shared memory" is IMHO still better than "(memcpy)" in documentation.
There was a problem hiding this comment.
I would go with a simple:
etrace: direct access to TRACE window of the SOF firmware
If ever we will support non mem mapped type of DSP with SOF then this will be still valid.
We can use region instead of window, that should be fine as well.
The regions/windows are defined under xsram_window and they are really just sections within a shared memory between the CPU and DSP in currently supported platforms.
Imho to say that something is a mailbox it would take a bit more than a shared memory, a signaling mechanism like doorbells should be there. TRACE does not have associated doorbell afaik.
There was a problem hiding this comment.
I would go with a simple: etrace: direct access to TRACE window of the SOF firmware
I feel like this is a confusing way to say "non-DMA trace" or "shared memory". What's wrong with either of these?
There was a problem hiding this comment.
@lyakh Have your change requests been completed? @ujfalusi Are you ready for the document to be merged? thanks
@deb-intel yes, they have! I've approved the PR
There was a problem hiding this comment.
@marc-hb, can you check the new wording for etrace:
etrace: direct access to the shared TRACE window of the SOF firmware
There was a problem hiding this comment.
Good considering the next one mentions "DMA". I'm still frustrated by the lack of good, "non-DMA/mailbox" name in general (not just in these docs) but that should certainly not delay this useful addition.
Add a note section with brief information about the etrace and trace debugfs files. Signed-off-by: Peter Ujfalusi <peter.ujfalusi@linux.intel.com>
To avoid confusion, add link to the kernel documentation about the dynamic debug feature used by the Linux kernel. Signed-off-by: Peter Ujfalusi <peter.ujfalusi@linux.intel.com>
Distribution kernels ship with trace not enabled by default. Add instruction on how to enable the file in order to provide traces. Signed-off-by: Peter Ujfalusi <peter.ujfalusi@linux.intel.com>
c24b29d to
b44a2e5
Compare
|
@deb-intel fyi - ready for grammar check :) |
|
@lrgirdwo Sounds good. I plan to publish the site today. |
6 participants
Add clarifications and details for debugging SOF issues.
The updates are created based on issue #347 in hope that the information in the documentation will be more user friendly.