On Tue, Dec 12, 2017 at 03:58:52PM -0800, Bjorn Andersson wrote: > A series of fixes for a few issues reported or seen with SMD. > > The first two fixes changes the trigger for creating devices for channels > found, which should cause the rpm-smd driver to probe on 8909 and 8994 > devices. >
.... and 8992. > The third patch fixes an issue when multiple callers end up waiting on events. > > The forth fix moves the check for a closed channel out of the tx-full loop, to > fail non-full send requests on a closed channel as well. > > The last fix improves the handling of concurrent send-requests, allowing > concurrent sends, in particular try_send, to progress even though we're > sleeping waiting for room in the tx fifo. > > Bjorn Andersson (5): > rpmsg: smd: Perform handshake during open > rpmsg: smd: Create device for all channels > rpmsg: smd: Wake up all waiters > rpmsg: smd: Fail send on a closed channel > rpmsg: smd: Don't hold the tx lock during wait > > drivers/rpmsg/qcom_smd.c | 60 > +++++++++++++++++++++++++++++++++++++----------- > 1 file changed, 47 insertions(+), 13 deletions(-) > Please add my tested by on this series as I have verified them on both my Nexus 5X and 6P. -jeremy
