Skip to content

Comments

A pull request to cherry-pick changes#1

Open
CRImier wants to merge 14 commits intoZeroPhone:masterfrom
jwrdegoede:master
Open

A pull request to cherry-pick changes#1
CRImier wants to merge 14 commits intoZeroPhone:masterfrom
jwrdegoede:master

Conversation

@CRImier
Copy link
Member

@CRImier CRImier commented Nov 26, 2017

No description provided.

Icenowy and others added 14 commits May 18, 2016 02:11
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
The definition of IEEE80211_NUM_BANDS and several other band-related
quantities are removed in kernel 4.7.

Inspired by a similar patch for the rtl8723bs driver by
Larry Finger <Larry.Finger@lwfinger.net>.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Replaced a check for interface offchannel that referenced a header
outside of include/ with the supported one through mac80211.h.  The right way
to check was added two kernel releases before the version ifdef'd block
that uses the illicit structure.  WTF Espressif?
First loading a dummy driver to power-up the device is not necessary
with upstream kernels + mmc-pwrseq, stop calling the empty
sif_platform_target_poweron() stub and remove the entire dummy-driver
dance.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Since sif_platform_rescan_card() is just an empty stub we need (hack)
broken-cd in the devicetree with upstream kernels to redetect the card
after the initial firmware load. When leaving the redection to the
mmc-core like this, the probe function will get called again
automatically when the mmc-core does the redetection.

This makes the entire unregister / register and semaphore dance
unnecessary, therefor this commit removes it.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
I've send a patch to the upstream kernel which adds a
mmc_force_detect_change() function to the upstream kernel which
allows things to work with non-removable in the devicetree instead
of using broken-cd as a hack.

This commit adds support for it, as well as adding a stub implementation
to use until this gets merged upstream.

Note that the stub does *not* fix the issue of the mmc core powering
down the device between the first and second probe, so until this is
merged upstream the regulators will still need regulator-always-on.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
sif_platform_check_r1_ready() is empty, waiting for r1 has been fixed
for all sdio devices in the upstream kernel by:
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/drivers/mmc/core?id=5d3f6ef04d65dda92a3573841a7215fb05c347de
"mmc: Wait for card_busy before starting sdio requests"

This commit removes the calls to the empty sif_platform_check_r1_ready
stub, all the other stub functions are already unused, so sdio_stub.c
can be completely removed.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
These are android only -> remove.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Change a bunch of messages which are clearly debug messages from
error / show messages to debug messages.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants