- 07 Jun, 2012 18 commits
-
-
Lars-Peter Clausen authored
Use scnprint instead of snprintf, because snprintf returns the number of bytes that would have been written to the buffer if there was enough space, and as a result writing to buf[len-1] might cause a access beyond the buffers limits. Reported-by:
Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by:
Lars-Peter Clausen <lars@metafoo.de> Acked-by:
Jonathan Cameron <jic23@kernel.org> Signed-off-by:
Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-
Michael Hennerich authored
drivers/iio/frequency/adf4350.c:316:32: warning: ‘val’ may be used uninitialized in this function [-Wuninitialized] Signed-off-by:
Michael Hennerich <michael.hennerich@analog.com> Reported-by:
-
Samuel Iglesias Gonsalvez authored
Signed-off-by:
Samuel Iglesias Gonsalvez <siglesias@igalia.com> Signed-off-by:
-
Samuel Iglesias Gonsalvez authored
Signed-off-by:
-
Samuel Iglesias Gonsalvez authored
When uninstalling a device, the call to the ipack_bus_ops remove() frees resources in the ipack device driver but without unregistering the device. It generates a kernel oops when somebody wants to unregister the device. Signed-off-by:
-
Samuel Iglesias Gonsalvez authored
tpci200->info is used later when uninstalling the module. As there is another kfree in the proper place, this patch removes the wrong one. Signed-off-by:
-
Samuel Iglesias Gonsalvez authored
When uninstalling a device, there is a loop of calls that produces, at the end, two calls to __ipoctal_remove() function with the same ipack_device argument. The first time works fine, but the second will fail in tty_unregister_driver() To avoid this situation, the call to __ipoctal_remove() it is done only from the ipack bus driver and not from the ipack device driver. Signed-off-by:
-
Samuel Iglesias Gonsalvez authored
When a bus driver calls ipack_driver_register(), it should manages the returning NULL value to undo all the operations it did before this call, and print the corresponding trace. It is not a task for the ipack driver to call the remove() function here. Signed-off-by:
-
Samuel Iglesias Gonsalvez authored
Use idr interface to give the bus number. That way, we remove the limitation of 64 buses. The mutex is removed because the ida interface uses spinlocks inside, so it is not needed an extra lock. Signed-off-by:
-
Samuel Iglesias Gonsalvez authored
When ipack_device_register() is called, the variable tpci200->slots[dev->slot].dev has not assigned a value and it gives an error when the mezzanine driver is reading a register from the board for the match() function, as all the I/O functions call check_slot(). Signed-off-by:
-
Samuel Iglesias Gonsalvez authored
This field is not needed at all, as the IRQ is registered for the carrier not for the mezzanine. Signed-off-by:
-
Samuel Iglesias Gonsalvez authored
Signed-off-by:
-
Samuel Iglesias Gonsalvez authored
To avoid a dereference of a NULL pointer, the availability of the function is checked before its use. Signed-off-by:
-
Samuel Iglesias Gonsalvez authored
tpci200_slot_map_space() should return 0 if the operation was properly done. If not, the caller will think that something wrong happened. This patch establish the returned value to 0. It is overwritten in case of error. Signed-off-by:
-
Samuel Iglesias Gonsalvez authored
tpci200_slot_unmap_space() should return 0 if the operation was properly done. If not, the caller will think that something wrong happened. Signed-off-by:
-
Samuel Iglesias Gonsalvez authored
It should return the same value given by the match function of the ipack_driver that has been called. Returning 0 here, means that the match has failed and it could be succeed. Signed-off-by:
-
Samuel Iglesias Gonsalvez authored
To perform the installation of a mezzanine it was needed to write on these files, now it is not needed at all as the device model is properly implemented. Signed-off-by:
-
Sachin Kamat authored
version.h header file is no longer required. Signed-off-by:
Sachin Kamat <sachin.kamat@linaro.org> Signed-off-by:
-
- 05 Jun, 2012 22 commits
-
-
Shubhrajyoti D authored
The memory regions of the probe and remove are move to __devinit and __devexit respectively. Acked-by:
Shubhrajyoti D <shubhrajyoti@ti.com> Signed-off-by:
-
Shubhrajyoti D authored
The memory regions of the probe and remove are move to __devinit and __devexit respectively. Acked-by:
-
Lars-Peter Clausen authored
The IIO DAC drivers are in a reasonably good shape. They all make use of channel spec and non of them provides non-documented sysfs attributes. Code style should be OK as well, both checkpatch and coccicheck only report trivial issues. So lets move the whole folder out of staging. Signed-off-by:
-
Lars-Peter Clausen authored
The header for this driver contains some private structs and defines, which do not have to be shared between multiple source files, as well as the platform data struct for this driver, which has to be shared with other source files. Since there is no need to expose those private structs and defines move them to the source file. Signed-off-by:
-
Lars-Peter Clausen authored
The header for this driver contains some private structs and defines, which do not have to be shared between multiple source files, as well as the platform data struct for this driver, which has to be shared with other source files. Since there is no need to expose those private structs and defines move them to the source file. Signed-off-by:
-
Lars-Peter Clausen authored
IRQ support is optional, so we have to check whether it was actually requested before we attempt to free it. Signed-off-by:
-
Lars-Peter Clausen authored
Use strtobool for parsing the powerdown value instead of strict_strtol, since the powerdown attribute is a boolean value. Signed-off-by:
-
Lars-Peter Clausen authored
Use strtobool for parsing the powerdown value instead of strict_strtol, since the powerdown attribute is a boolean value. Signed-off-by:
-
Lars-Peter Clausen authored
Use strtobool for parsing the powerdown value instead of strict_strtol, since the powerdown attribute is a boolean value. Signed-off-by:
-
Lars-Peter Clausen authored
Now that all drivers have been converted to channel spec, we can finally remove the dac.h file since it is no longer used. Signed-off-by:
-
Lars-Peter Clausen authored
Convert the max517 driver to channel spec. As part of the conversion the "out_voltage_1&2_raw" property, which updates both channel 1 and 2 simultaneously with the same value, is lost, since this is not really covered by the IIO spec and has only a limited use case in practice. Also the channel index for the sysfs files is now zero based instead of one based, which means all channel numbers will be lower by one. E.g. "out_voltage_1_scale" instead of "out_voltage_2_scale" Signed-off-by:
Roland Stigge <stigge@antcom.de> Signed-off-by:
-
Lars-Peter Clausen authored
None of these drivers use anything from dac.h, so remove the include. Signed-off-by:
-
Lars-Peter Clausen authored
Use extended channel attributes instead of raw sysfs files for the additional channel attributes. This allows us to remove some boilerplate code. Signed-off-by:
-
Lars-Peter Clausen authored
Use extended channel attributes instead of raw sysfs files for the additional channel attributes. This allows us to remove some boilerplate code. Signed-off-by:
-
Lars-Peter Clausen authored
Use extended channel attributes instead of raw sysfs files for the additional channel attributes. This allows us to remove some boilerplate code. Signed-off-by:
-
Lars-Peter Clausen authored
Use extended channel attributes instead of raw sysfs files for the additional channel attributes. This allows us to remove some boilerplate code. Signed-off-by:
-
Lars-Peter Clausen authored
Use extended channel attributes instead of raw sysfs files for the additional channel attributes. This allows us to remove some boilerplate code. Signed-off-by:
-
Lars-Peter Clausen authored
This allows us to remove some boilerplate code for comparing and formatting the enum strings. Signed-off-by:
-
Lars-Peter Clausen authored
This allows us to remove some boilerplate code for comparing and formatting the enum strings. Signed-off-by:
-
Lars-Peter Clausen authored
We often have the case were we do have a enum style channel attribute. These attributes have in common that they are a list of string values which usually map in a 1-to-1 fashion to integer values. This patch implements some common helper code for implementing enum style channel attributes using extended channel attributes. The helper functions take care of converting between the string and integer values, as well providing a function for "_available" attributes which list all available enum items. Signed-off-by:
-
Lars-Peter Clausen authored
Use iio_device_get and iio_device_put instead of open-coding it. Signed-off-by:
-
Lars-Peter Clausen authored
This patch add the iio_device_get() function, which increases the reference count of a iio device. The matching function to decrease the reference count - iio_device_put() - already exists. Signed-off-by:
-
