northbound wrote:silvah wrote:so... .12 is the same driver as .14 or are there some differences?
I have not had the time to do a git diff they are supposed to be the same.
They're not. .13 added another change as well, which is still there in .14.
bjorn@nemi:/usr/local/src/git/mwlwifi$ git log --oneline origin|head
0e3478e3405c Commit mwlwifi driver 10.3.0.14.
716010d6cfbe Commit mwlwifi driver 10.3.0.13.
30e6b06de659 Commit mwlwifi driver 10.3.0.12.
7a6aff718d2d Commit mwlwifi driver 10.3.0.10.
59e0d3d167a5 Commit mwlwifi driver 10.3.0.8
aaf953e1baf9 Remove unnecessary debug messages.
36e430f93174 Update F/W to version 7.2.8.6.
2b93ae5bf32e Commit mwlwifi driver 10.3.0.3
f6f2b26593a3 Change Chip to modules.
010dd6fba522 Modify code to pass checkpatch.pl.
bjorn@nemi:/usr/local/src/git/mwlwifi$ git diff --stat 30e6b06de659..0e3478e3405c
debugfs.c | 8 --------
dev.h | 3 ++-
fwcmd.c | 19 ++++++++++---------
hostcmd.h | 2 +-
mac80211.c | 98 ++++++++++++++++++++++++++++++++++++++++----------------------------------------------------------
main.c | 2 +-
rx.c | 82 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
sysadpt.h | 2 +-
tx.c | 25 ++++++++++++++++++++-----
9 files changed, 157 insertions(+), 84 deletions(-)
This is a problem with how they use git. It would have been much, much, much better if they commited each change separately instead of collecting a number of them and then commiting a whole new release as a major mess of changes. This makes it very hard to pinpoint regression sources, and makes powerful tools like "git bisect" mostly useless. The recent flow of small releases improves the situation, but there are still problems like the one shown above: The 10.3.0.13 commit was really two separate changes merged together.
But I beleive the driver should be ready for mainline soon, and then this problem is going away. There is no way that kind of commits are going to be accepted after the driver is in mainline.
