On 20 November 2017 at 15:21, Mathieu Poirier mathieu.poirier@linaro.org wrote:
On 20 November 2017 at 07:48, Robert Walker robert.walker@arm.com wrote:
Hi Mathieu,
Thanks for the feedback, I'll update the patch and send a new version out.
On 17/11/17 19:01, Mathieu Poirier wrote:
Since cs_etm__sample() is used for trace decoding should we also apply this patch to the master branch?
I think a lot of the changes in this patch are closely coupled to Sebastian's original patches, so it may be hard to separate them out and apply them to master.
I wanted to test your implemenation but (8bbf7ee5d6cb perf inject: correct recording of branch address and destination) breaks normal CS trace decoding. Would you mind having a look?
Do you mean the trace decode from "perf script"? Do you have any more info on how it's broken and what you would expect? (I'm not familiar with perf script, so don't know what output it should produce).
I noticed that just doing a "perf report --stdio" on the autoFDO branch hangs with the commit I pointed out.
This hangs without Rob's patches too. --stdio --dump works, --stdio only hangs. Not sure we haven't seen this before?
Mike
Be mindful that I rebased the autoFDO tree to v4.14, so SHA's will have changed.
I'll use that as the baseline for any future patches.
Yes please. We don't have a choice, until all patches are upstream we have to rebase when to the newest kernel revision.
Regards
Rob
CoreSight mailing list CoreSight@lists.linaro.org https://lists.linaro.org/mailman/listinfo/coresight