mirror of https://gitee.com/openkylin/linux.git
a3476ac678
This set contains a change to the ABI for the hid-sensors drivers to bring them in line with the long published documentation. Unfortunately, rather than reporting true scale and offset values via sysfs they were reporting some magic numbers that could only be converted to anything useful using the HID sensors specification. I missed this entirely through the introduction of a number of drivers, only picking up on it recently. Srinivas has had user feedback about this as well. The patch set is too large to go as a fix at this stage in the cycle and is not a regression fix as this was never right and so will have to wait for the next merge window. Srinivas assures me that there are relatively few pieces of hardware out there and he has had a number of people contact him to point out that the drivers did not obey the ABI. Hence hopefully the fallout of this, if any will be minor. If we don't fix it now, it will only get worse going forward. There is no sensible way of maintaining the incorrect ABI as it is simply returning the wrong values through the standard interfaces. Non IIO elements * Introduce devm_kmemdup. Does what it says on the tin. New drivers: * hid-sensors rotation devices (output as quaternion) * Freescale MPL115A2 presure and temperature sensor. * Melexis mlx90614 contactless infrared sensor. * Freescale MMA8452Q 3-axis accelerometer. New functionality: * Addition of multiple element callback to allow for sysfs interfaces to access elements such as quaternions which have no useful meaning if all 4 elements are not presented together. Other future usecases for this include rotation matrices. * Support for multiple element buffer entries for exactly the same uses as the sysfs related elements described above. * Quaternion support via the quaternion IIO modifier. * TEMP_AMBIENT and TEMP_OBJECT modifiers to distinguish cases with thermopile devices. * hid-sensors gain sysfs access to the sensor readings. Previously these drivers used the buffered interface only. This change involves some additional hid-sensors core support to read poll values back from the devices to allow the drivers to know roughly how long to wait for a result when polling the sensor. There is also an associated hid-sensors abi to allow the devices to be turned off between reads and powered up on demand. Cleanups and fixes * Hid sensors fix as described above. Result is to make the _scale and _offset attributes applicable in the same way as for all other IIO drivers. * Some additional documentation - mostly covering stuff that graduated from staging without managing to take it's ABI docs with it. * A series of little tidy ups to the exynos_adc driver that make the code nicer to read and improve handling of some corner cases. * A tidy up to mag3110 (logical fix rather than a real one ;). Also enable user offset calibration for this device. * Drop some left over IS_ERR() checks from ad799x that snuck through during the cleanup in the last IIO patch set. * Fix a naming issue from clashing patches in ak8975 - note the clash only occured in the last IIO patch set, hence the fix needs to go through this tree. * A format string missmatch fix in ad7280.c. Unlikely to have ever had an impact so not worth rushing through. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (GNU/Linux) iQIcBAABAgAGBQJTZ4D6AAoJEFSFNJnE9BaI48QP/AwIAmbImfIHH+992SbL0yU4 O/O5DnectH11fH0aJftISelBmXdzySigMX/7cEGzGKpFyRm3u8zBXRh2f4bS2JZW V5//sBe7bNTTeNm6RDdgABos6boeLZFW0f+UNY1jCsIq3WvdRlVtMGLrOCPRoKEL jyjo5kpt14Rhz5Sitk91aEfHk9IFv2HGeUkKFQ+y8tdQmpza7O1o9PLw6HWeVelv yLfqO/1hIqBCnJKal39n+gqGSVmyXNlog7D6hZCx+Qb/u0uJU6Iw57BVo3vbYuiV IcV0dgweEaQFxS32ueRfj5FuxAw3DS2pMF/y1j1fhW3UNJl5JBxfAeRSUDzsKjIT H602WTl6C/20sii/EZmbX5AfffYW1HRSvXEnjQIOfcYcbasn+zx4hlzXKS1vJyxF T4OyKVQIZ1gnwDkbcYnIgcuY+p6vJNj3kE0884WqJA4xYATzQ5BqFvTLlxZ2wp73 RyhqbYlmvckl4376vJacqkLkSQqpSfqWRV8035hXZaY/1QMnuUBqFpjXdcvr1VJ8 +4KzDGzzvLrOlJMsI3JZS2MXh72t+Rwl7/m4kocRApG0waEj6ykFRUJjJTdTfxmC jSfivW7rdctlPFnFQdTchOyg95RzRD5+Ynj/FFHxz6l4SPJzY6RrzuOSHJ4d4yl9 /QHkdJV4UpR/YN2ECpe/ =ghZG -----END PGP SIGNATURE----- Merge tag 'iio-for-3.16b' of git://git.kernel.org/pub/scm/linux/kernel/git/jic23/iio into staging-next Jonathan writes: Second set of IIO new drivers, cleanups and functionality for the 3.16 cycle. This set contains a change to the ABI for the hid-sensors drivers to bring them in line with the long published documentation. Unfortunately, rather than reporting true scale and offset values via sysfs they were reporting some magic numbers that could only be converted to anything useful using the HID sensors specification. I missed this entirely through the introduction of a number of drivers, only picking up on it recently. Srinivas has had user feedback about this as well. The patch set is too large to go as a fix at this stage in the cycle and is not a regression fix as this was never right and so will have to wait for the next merge window. Srinivas assures me that there are relatively few pieces of hardware out there and he has had a number of people contact him to point out that the drivers did not obey the ABI. Hence hopefully the fallout of this, if any will be minor. If we don't fix it now, it will only get worse going forward. There is no sensible way of maintaining the incorrect ABI as it is simply returning the wrong values through the standard interfaces. Non IIO elements * Introduce devm_kmemdup. Does what it says on the tin. New drivers: * hid-sensors rotation devices (output as quaternion) * Freescale MPL115A2 presure and temperature sensor. * Melexis mlx90614 contactless infrared sensor. * Freescale MMA8452Q 3-axis accelerometer. New functionality: * Addition of multiple element callback to allow for sysfs interfaces to access elements such as quaternions which have no useful meaning if all 4 elements are not presented together. Other future usecases for this include rotation matrices. * Support for multiple element buffer entries for exactly the same uses as the sysfs related elements described above. * Quaternion support via the quaternion IIO modifier. * TEMP_AMBIENT and TEMP_OBJECT modifiers to distinguish cases with thermopile devices. * hid-sensors gain sysfs access to the sensor readings. Previously these drivers used the buffered interface only. This change involves some additional hid-sensors core support to read poll values back from the devices to allow the drivers to know roughly how long to wait for a result when polling the sensor. There is also an associated hid-sensors abi to allow the devices to be turned off between reads and powered up on demand. Cleanups and fixes * Hid sensors fix as described above. Result is to make the _scale and _offset attributes applicable in the same way as for all other IIO drivers. * Some additional documentation - mostly covering stuff that graduated from staging without managing to take it's ABI docs with it. * A series of little tidy ups to the exynos_adc driver that make the code nicer to read and improve handling of some corner cases. * A tidy up to mag3110 (logical fix rather than a real one ;). Also enable user offset calibration for this device. * Drop some left over IS_ERR() checks from ad799x that snuck through during the cleanup in the last IIO patch set. * Fix a naming issue from clashing patches in ak8975 - note the clash only occured in the last IIO patch set, hence the fix needs to go through this tree. * A format string missmatch fix in ad7280.c. Unlikely to have ever had an impact so not worth rushing through. |
||
---|---|---|
.. | ||
acpi | ||
asm-generic | ||
clocksource | ||
crypto | ||
drm | ||
dt-bindings | ||
keys | ||
kvm | ||
linux | ||
math-emu | ||
media | ||
memory | ||
misc | ||
net | ||
pcmcia | ||
ras | ||
rdma | ||
rxrpc | ||
scsi | ||
sound | ||
target | ||
trace | ||
uapi | ||
video | ||
xen | ||
Kbuild |