This is merely a historical archive of years 2008-2021, before the migration to mailman3.
A maintained and still updated list archive can be found at https://lists.osmocom.org/hyperkitty/list/gerrit-log@lists.osmocom.org/.
Harald Welte gerrit-no-reply at lists.osmocom.orgHarald Welte has posted comments on this change. ( https://gerrit.osmocom.org/10313 )
Change subject: add synchronous UART transmission and use it in exceptions
......................................................................
Patch Set 1: Code-Review+1
(1 comment)
I think in general I would have preferred to simply use a separate Uart transmit function from the hardirq handlers. Either by explicitly changing those calls away from printf, or by catching/redirecting printf using a #define that's #included only in the exceptions.c file in order to bypass normal printf().
Now we have one conditional in every printf (not really relevant) and some global state that we'd have to switch back should we ever want to have "non-final" exception handling.
Another idea that just comes to my mind: Couldn't we simply check some CPU status flags to know that we're in an exception or in an interrupt? This way the decision to bypass the ring buffer could be entirely automatic without any explcit calls to disable the buffer, or any of the hacks I described above? I would suggest to look into this. If it turns out it's not that simple, please report back and we can merge your current patch as-is.
https://gerrit.osmocom.org/#/c/10313/1/firmware/libboard/common/source/uart_console.c
File firmware/libboard/common/source/uart_console.c:
https://gerrit.osmocom.org/#/c/10313/1/firmware/libboard/common/source/uart_console.c@169
PS1, Line 169: hift register */
: while (!(pUart->UART_SR & UART_SR_TXEMPTY)); /* Wait for transfer shift register to be empty (i.e. transfer is complete) */
why do we wait for this? IMHO this effectively disables the "1 byte fifo" of the transfer register? Shouldn't it be safe to transmit the next byte as soon as the hold register has been transferred to the shift register?
--
To view, visit https://gerrit.osmocom.org/10313
To unsubscribe, or for help writing mail filters, visit https://gerrit.osmocom.org/settings
Gerrit-Project: simtrace2
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I1b4ace5185cf2dc32684934ed12bf6a8682e9bad
Gerrit-Change-Number: 10313
Gerrit-PatchSet: 1
Gerrit-Owner: Kévin Redon <kredon at sysmocom.de>
Gerrit-Reviewer: Harald Welte <laforge at gnumonks.org>
Gerrit-Reviewer: Jenkins Builder
Gerrit-Comment-Date: Thu, 02 Aug 2018 21:13:24 +0000
Gerrit-HasComments: Yes
Gerrit-HasLabels: Yes
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osmocom.org/pipermail/gerrit-log/attachments/20180802/c3642c46/attachment.htm>