It is possible for a mode set to fail if there aren't shared DPLLS that
match the new configuration requirement or other errors in clock
computation. If that step is executed after disabling crtcs, in the
failure case the hardware configuration is changed and needs to be
restored. Doing those things early will allow the mode set to fail
before actually touching the hardware.
Follow up patches will convert different platforms to use the new
infrastructure.
v2: Keep pll->new_config valid only during mode set (Ville)
Use kmemdup() in i915_shared_dpll_start_config() (Ville)
Restore old pll config if something fails before commit (Ville)
Don't set compute_clock hooks since dev_priv is kzalloc()'d (Ville)
Signed-off-by: Ander Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
The new struct will be used in a follow up patch to allow a current and
a staged config to exist for the same shared DPLL.
v2: Rebase on by mask_to_refcount()->hweight32() change. (Damien)
Signed-off-by: Ander Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
Reviewed-by: Damien Lespiau <damien.lespiau@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
This will be used in a follow up patch to properly release shared DPLLs
without relying on the shared_dpll field in pipe_config.
v2: Fix white space error (Ville)
Use hweight32() (Ville)
Signed-off-by: Ander Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
More concise. Noticed while reviewing Ander's patch which touched a
lot of the pipe_has_type checks.
v2: Use new_config in one place Ander spotted.
Cc: Ander Conselvan de Oliveira <conselvan2@gmail.com>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
The async command support should now work. Enable the hook up of the command
support functions in apci3120_auto_attach().
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Currently the DMA buffers are switched before writing the current samples to
the async buffer. This works but when the EOA event happens we end up with an
outstanding DMA operation in progress that gets terminated by the (*cancel).
Avoid the outstanding DMA operation by switching the DMA buffers after writing
the samples. The driver will detect the EOA event and not retart the DMA.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Pass the comedi_device and comedi_subdevice pointers to this function instead
of the irq number and void pointer.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Move this function from the included source file to the main driver
source file.
Remove the now empty included source file.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Use the helper function to calculate the number of samples in the DMA
buffer. For aesthetics, tidy up the local variables.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In apci3120_interrupt_dma(), use the async->events to report the hardware
error and allow the core to (*cancel) the command instead of calling the
(*cancel) function directly.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Move apci3120_ai_cmdtest() from the included source file to the main
driver source file. Tidy up Step 4 a bit during the move.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Regardless of the cmd->scan_begin_src, the minimum cmd->convert_arg
value is 10000 (10us). Fix the validation in apci3120_ai_cmdtest().
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The timer subdevice in this driver does not follow the comedi API.
Fix the (*insn_config) to correctly arm, disarm, set the mode, and
get the status of the timer.
Remove the unnecessary, and broken, (*insn_write).
The new (*insn_config) does not enable the interrupt for timer 2.
Remove the interrupt support code for the timer.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Timer 2 can be used to count the number of scans when the cmd->stop_src
is TRIG_COUNT. Unfortunately that timer can also be used by userspace and
this breaks the async command. Timer 2 also only works for counting the
number of scans when DMA is not used.
The number of scans is counted by the core using the private data membmer
'scans_done'. Use that to detect the End-Of-Scan for both DMA and none
DMA modes.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This function is only called during the attach of the driver and it's
called after the private data has been kzalloc'd. There is no reason
to clear any of the private data members.
Simplify the function by just writing '0' to the necessary registers
to disable all the interrupt sources and stop the timers.
For aesthetics, change the return type to void.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Move the common code used to initialize DMA to apci3120_init_dma().
This follows the programming procedure described in the APCI-3120
documentation.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Use the defines for the AMCC 5933 PCI controller registers and bits instead
of creating private defines in this driver.
Move the generic AGCSTS_TC_ENABLE define from this driver to the header.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Introduce a helper function to handle the common code that writes the DMA
start address and number of acquisitions to the AMCC Add-on registers.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This member of the private data is not really needed. The devpriv->ctrl
value can be checked to determine if the external trigger is enabled.
Remove the unnecessary member.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This member of the private data is not really needed. The interrupt mode
can be determined by checking the devpriv->mode value to see if the
interrupt is enabled. Remove the unnecessary member.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
There is no reason for the separate updates of the mode register in this
function. Refactor the code so that the mode register is only updated at
the end of the function after all the necessary bits have been set.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The 'SCAN_ENA' bit in the mode register needs to be set if the chanlist
has more than 1 channel. Set the bit in apci3120_set_chanlist() if needed.
The callers write the mode register after setting any additional bits,
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The APCI-3120 documentation says that the FIFO should be reset after the
chanlist sequence is programmed.
Reset the FIFO after programming the chanlist and remove the extra FIFO
resets in the driver.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Tidy up the programming of timers 0 and 1 so that only one local variable is
required.
Also, remove the unnecessary clear of devpriv->timer_mode. This value will get
set correctly by each apci3120_timer_set_mode() call. It's not necessary to
clear it first.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This functon is called by apci3120_ai_cmd() with one additional parameter,
'mode', The 'mode' is based on the cmd->scan_begin_src.
For aesthetics, absorb the function into apci3120_ai_cmd() and use the
cmd->scan_begin_src directly to determine the 'mode'.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
