[M] Change in flashrom[main]: libflashrom: Update the API for Logger Callback

[Dmitry Zhadinets (Code Review)]

Attention is currently required from: Anastasia Klimchuk, Peter Marheine.

Dmitry Zhadinets has posted comments on this change by Dmitry Zhadinets. ( https://review.coreboot.org/c/flashrom/+/86875?usp=email )

Change subject: libflashrom: Update the API for Logger Callback
......................................................................

Patch Set 7:

(1 comment)

File libflashrom.c:

https://review.coreboot.org/c/flashrom/+/86875/comment/3db70568_f84c6115?usp... :
PS7, Line 75: 	if (((unsigned)size+1) > sizeof(message))
            : 		return -ERANGE;
> A question here: I remember initial comment thread was about fallback to dymanic allocation if messa […]

Honestly, I would add assert here (if the message length is more then something), but there are no asserts in the project and I assume there is a reason for this. The length of the message depends on the abstract developer of flashrom project only. And I assume this developer really wants that flashrom will work faster, eat less memory, API will be useful and so on. This limitation (I mean the small size of the single message from the flashrom) is like an inner contract, a deal between developers. "Do not try to write "War & Peace" in the messages. Brevity is the sister of talent." BTW, really long messages are already generated by flashrom, without '\n' at the end, using a sequence of messages. So, at the moment we do not have such issue then it will be for future addons to the codebase. I tested the length with unittests and by hands. May be there is some server where we can perform full coverage test.
To be faster: I really want to add set_log_level function. Because right now even for any trace messages we will be calling user's callback. We cannot guarantee that everything is good there (some heavy construction can be used even to check necessity to do something). Just imaging crazy case. You are reading the next page of a flash memory and sending DEBUG message to print but he goes to the cloud to get settings about his current log level. You understand how long overall read will be. Even in my gflashrom it is several jump by pointers and a creation of C++ string. 
It will be my next commit to the chain but we a bit stuck here.