diff options
| author | Thibaut VARÈNE <hacks@slashdirt.org> | 2020-05-08 13:39:10 +0200 |
|---|---|---|
| committer | Koen Vandeputte <koen.vandeputte@ncentric.com> | 2020-05-12 12:41:08 +0200 |
| commit | 1c6990fe6d198718850b6a37998416cac69f6f53 (patch) | |
| tree | 1a432e00a4be88036cbd81a0f5c96c15df0d2dd9 /target/linux/bcm63xx/dts/bcm6358-pirelli-a226g.dts | |
| parent | 0fed36dd484272eb546dd69f4dc84221ca5f01a5 (diff) | |
| download | upstream-1c6990fe6d198718850b6a37998416cac69f6f53.tar.gz upstream-1c6990fe6d198718850b6a37998416cac69f6f53.tar.bz2 upstream-1c6990fe6d198718850b6a37998416cac69f6f53.zip | |
ar71xx: mikrotik: bypass id check in __rb_get_wlan_data()
The id parameter in __rb_get_wlan_data() was incorrectly used on the
assumption that id "0" would always be tied to ath9k with RLE encoding
and positive id (in fact, only id "1" was valid) would always be tied to
("external") ath10k with LZO encoding.
Newer hardware revisions of supported devices prove this assumption to
be invalid, with ath9k caldata being now wrapped in MAGIC_ERD and LZO
compressed, so disable this check to allow newer hardware to correctly
decode caldata for ath9k. Since ath10k caldata is no longer pulled from
this implementation, this commit also disables the publication in sysfs
to avoid wasting memory.
Note: this patch assumes that ath9k caldata is never stored with the new
"LZOR" encoding scheme found on some ath10k devices.
Signed-off-by: Thibaut VARÈNE <hacks@slashdirt.org>
Diffstat (limited to 'target/linux/bcm63xx/dts/bcm6358-pirelli-a226g.dts')
0 files changed, 0 insertions, 0 deletions