tag name | iio-fixes-for-5.9a (5baf84e211edd8e8fcfe982e8aff1778da9f3621) |
tag date | 2020-08-31 10:49:33 +0100 |
tagged by | Jonathan Cameron <Jonathan.Cameron@huawei.com> |
tagged object | commit 3f1093d83d... |
download | iio-fixes-for-5.9a.tar.gz |
---|
First set of IIO fixes for the 5.9 cycle
Most of the fixes this time are for a long term issue that Lars-Peter
Clausen identified recently.
IIO assumes that any data pushed to the buffer interface (either kfifo
or another in kernel consumer) are naturally aligned. Unfortunately
this isn't true in a number of drivers, mostly by failing to ensure
the buffer used is aligned suitably for an s64 timestamp.
For the ones covered this time we use a structure to enforce this
alignment, with the added need for __aligned(8) to ensure 8 byte
alignment of the timestamp on x86_32 and similar platforms where
it would be 4 byte aligned giving different padded from some other
architectures.
Patches to make this requirement clearer and potentially cause an
error print will follow once we've cleaned up existing cases.
Note that it is a very hard to hit problem and as a result of this
as we only have one bug report despite the problem being present
for many years.
Other fixes:
* cros_ec:
- set gyroscope default frequency. For some older boards not having
this set can lead to a choice that doesn't work.
* counter,microchip-tcb-capture:
- Fix a wrong value check in error check.
* mcp3422
- Fix locking to protect against race condition.
* meson-adc
- Use right device when looking up fuse values for calibration.
* rockchip-adc
- Fix missing Kconfig dependency.
* ti-ads1015:
- Fix reading when CONFIG_PM not set.
-----BEGIN PGP SIGNATURE-----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=WLun
-----END PGP SIGNATURE-----