Now that pcpu_alloc_area() can allocate only from populated areas,
it's easy to add atomic allocation support to [__]alloc_percpu().
Update pcpu_alloc() so that it accepts @gfp and skips all the blocking
operations and allocates only from the populated areas if @gfp doesn't
contain GFP_KERNEL. New interface functions [__]alloc_percpu_gfp()
are added.
While this means that atomic allocations are possible, this isn't
complete yet as there's no mechanism to ensure that certain amount of
populated areas is kept available and atomic allocations may keep
failing under certain conditions.
Signed-off-by: Tejun Heo <tj@kernel.org>
The next patch will conditionalize the population block in
pcpu_alloc() which will end up making a rather large indentation
change obfuscating the actual logic change. This patch puts the block
under "if (true)" so that the next patch can avoid indentation
changes. The defintions of the local variables which are used only in
the block are moved into the block.
This patch is purely cosmetic.
Signed-off-by: Tejun Heo <tj@kernel.org>
Update pcpu_alloc_area() so that it can skip unpopulated areas if the
new parameter @pop_only is true. This is implemented by a new
function, pcpu_fit_in_area(), which determines the amount of head
padding considering the alignment and populated state.
@pop_only is currently always false but this will be used to implement
atomic allocation.
Signed-off-by: Tejun Heo <tj@kernel.org>
At first, the percpu allocator required a sleepable context for both
alloc and free paths and used pcpu_alloc_mutex to protect everything.
Later, pcpu_lock was introduced to protect the index data structure so
that the free path can be invoked from atomic contexts. The
conversion only updated what's necessary and left most of the
allocation path under pcpu_alloc_mutex.
The percpu allocator is planned to add support for atomic allocation
and this patch restructures locking so that the coverage of
pcpu_alloc_mutex is further reduced.
* pcpu_alloc() now grab pcpu_alloc_mutex only while creating a new
chunk and populating the allocated area. Everything else is now
protected soley by pcpu_lock.
After this change, multiple instances of pcpu_extend_area_map() may
race but the function already implements sufficient synchronization
using pcpu_lock.
This also allows multiple allocators to arrive at new chunk
creation. To avoid creating multiple empty chunks back-to-back, a
new chunk is created iff there is no other empty chunk after
grabbing pcpu_alloc_mutex.
* pcpu_lock is now held while modifying chunk->populated bitmap.
After this, all data structures are protected by pcpu_lock.
Signed-off-by: Tejun Heo <tj@kernel.org>
percpu-km instantiates the whole chunk on creation and doesn't make
use of chunk->populated bitmap and leaves it as zero. While this
currently doesn't cause any problem, the inconsistency makes it
difficult to build further logic on top of chunk->populated. This
patch makes percpu-km fill chunk->populated on creation so that the
bitmap is always consistent.
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Christoph Lameter <cl@linux.com>
Previously, pcpu_[de]populate_chunk() were called with the range which
may contain multiple target regions in it and
pcpu_[de]populate_chunk() iterated over the regions. This has the
benefit of batching up cache flushes for all the regions; however,
we're planning to add more bookkeeping logic around [de]population to
support atomic allocations and this delegation of iterations gets in
the way.
This patch moves the region iterations out of
pcpu_[de]populate_chunk() into its callers - pcpu_alloc() and
pcpu_reclaim() - so that we can later add logic to track more states
around them. This change may make cache and tlb flushes more frequent
but multi-region [de]populations are rare anyway and if this actually
becomes a problem, it's not difficult to factor out cache flushes as
separate callbacks which are directly invoked from percpu.c.
Signed-off-by: Tejun Heo <tj@kernel.org>
percpu-vm and percpu-km implement separate versions of
pcpu_[de]populate_chunk() and some part which is or should be common
are currently in the specific implementations. Make the following
changes.
* Allocate area clearing is moved from the pcpu_populate_chunk()
implementations to pcpu_alloc(). This makes percpu-km's version
noop.
* Quick exit tests in pcpu_[de]populate_chunk() of percpu-vm are moved
to their respective callers so that they are applied to percpu-km
too. This doesn't make any meaningful difference as both functions
are noop for percpu-km; however, this is more consistent and will
help implementing atomic allocation support.
Signed-off-by: Tejun Heo <tj@kernel.org>
pcpu_get_pages() creates the temp pages array if not already allocated
and returns the pointer to it. As the function is called from both
[de]population paths and depopulation can only happen after at least
one successful population, the param doesn't make any difference - the
allocation will always happen on the population path anyway.
Remove @may_alloc from pcpu_get_pages(). Also, add an lockdep
assertion pcpu_alloc_mutex instead of vaguely stating that the
exclusion is the caller's responsibility.
Signed-off-by: Tejun Heo <tj@kernel.org>
percpu-vm uses pcpu_get_pages_and_bitmap() to acquire temp pages array
and populated bitmap and uses the two during [de]population. The temp
bitmap is used only to build the new bitmap that is copied to
chunk->populated after the operation succeeds; however, the new bitmap
can be trivially set after success without using the temp bitmap.
This patch removes the temp populated bitmap usage from percpu-vm.c.
* pcpu_get_pages_and_bitmap() is renamed to pcpu_get_pages() and no
longer hands out the temp bitmap.
* @populated arugment is dropped from all the related functions.
@populated updates in pcpu_[un]map_pages() are dropped.
* Two loops in pcpu_map_pages() are merged.
* pcpu_[de]populated_chunk() modify chunk->populated bitmap directly
from @page_start and @page_end after success.
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Christoph Lameter <cl@linux.com>
`struct pci230_private` has two members to manage the enabled interrupt
sources. `int_en` is the interrupt sources we want to be enabled and
`ier` is a shadow of the write-only interrupt enable register. They
have the same value most of the time. They differ in the interrupt
handler (`pci230_interrupt()`) itself when it temporarily clears bits in
the interrupt enable register and the `ier` member in order to unlatch
them in hardware, but leaves the `int_en` member alone. They also
differ in `pci230_ai_stop()` and `pci230_ao_stop()` which clear bits in
the `int_en` member and wait for the interrupt handler to finish before
copying the value to the `ier` member and the interrupt enable register.
Simplify the handling a bit, by making the `ier` member take on the role
of the `int_en` member, and allowing the value to differ from the
interrupt enable register while the interrupt handler is running.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Change the return type of `pci230_handle_ao_fifo()` from `int` to
`bool`. A return value of `true` indicates the AO command is still
running.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Some counter channels may be required for AI commands and AO commands.
Depending on how the commands are set up, it may not be possible to run
both at the same time, so we keep some state and code to find out if the
required resources are busy or not.
The existing code is a bit unwieldy - the code for claiming resources
involves two `for` loops for example. Rewrite it to make it simpler.
The new code just has a bit-mask value for each shared resource (counter
channels), and an array indexed by resource "owners" (AI and AO
commands), so the code for claiming resources now just has a single loop
that checks that none of the other owners have claimed the wanted
resources.
Rename the functions involved, because the old names involving 'put' and
'get' suggested some sort of usage counting.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The `state` member of `struct pci230_private` is used with the atomic
bit-op functions and has a couple of bits defined, `AI_CMD_STARTED` and
`AO_CMD_STARTED`. Spin-locks are used to protect the clearing of these
bits and other stuff. No special protection is used for setting these
bits. Replace the `state` member with a couple of new, single-bit
bitfield members, `ai_cmd_started` and `ao_cmd_started` to save some
space.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Change the `intr_running` member of `struct pci230_private` into a
single-bit bitfield of type `bool` to save a bit of space.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Use the inline `comedi_range_is_bipolar()` function from "comedidev.h"
to decide whether a range is bipolar or unipolar instead of using the
local static arrays `pci230_ai_bipolar[]` and `pci230_ao_bipolar[]`
which can then be removed.
Change the types of the `ai_bipolar` and `ao_bipolar` members of `struct
pci230_private` to `bool` to match the return value of
`comedi_range_is_bipolar()` and change them into single-bit bitfields to
save a bit of space. Also change the type and name of some local
variables in `pci230_ai_check_chanlist()` that hold the result of
`comedi_range_is_bipolar()`.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Most functions in "amplc_pci230" are named with the prefix `pci230_`,
apart from one or two that have the prefix `amplc_pci230_` and a few
odd-balls with no particular prefix. Rename the ones without a prefix
for consistency.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Rename the AI subdevice "insn_read" handler function `pci230_ai_rinsn()`
to `pci230_ai_insn_read()` for consistency.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Some functions in "amplc_pci230.c" are declared `inline`. Remove the
`inline` specifiers and let the compiler do what it wants with them.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
`pci230_ai_read()` reads a sample from the ADC data register and
converts it to a comedi sample value. The AI sample may have 12 or 16
bits of resolution, depending on the board type, but 12-bit sample
values are in bits 15 to 4 of the register. The hardware value is
signed, 2's complement if set to a bipolar mode, or unsigned, straight
binary if set to a unipolar mode. To convert to a Comedi sample value
it may need shifting right by 4 bits, and the top bit of the sample
value may need to be toggled.
Simplify the existing code by doing the 2's complement to straight
binary conversion before the shift. That way, it is always bit 15 that
is inverted regardless of the resolution.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
`pci230_ao_mangle_datum()` converts comedi sample values for the AO
subdevice to hardware register values. The comedi sample value will be
an unsigned value in the range 0 to 4095 (assuming 12-bit resolution).
The hardware wants the value shifted so the m.s. bit of the sample in in
bit 15. If set to a bipolar range, it also expects a 2's complement
value, so the top bit of the sample value needs to be inverted in that
case.
Simplify the existing code by doing the 2's complement conversion after
the shift. That way, it is always bit 15 that is inverted regardless of
the resolution.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The `ai_bits`, `ao_bits`, and `min_hwver` members of `struct
pci230_board` are only set to small, non-negative values, so make them
`unsigned char`. The `have_dio` member is used as a boolean so change
it to a bitfield of type `bool`.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The PCI230(+) has an AO subdevice with 2 channels, but the PCI260(+) has
none.
The `ao_chans` member of `struct pci230_board` indicates whether the
board has an AO subdevice and the number of AO channels. The
`ao_bits` member indicates the AO sample width in bits and will only be
non-zero for boards with an AO subdevice.
Use `ao_bits` to indicate whether the board has an AO subdevice. If it
has, assume the the number of AO channels is 2. Then the `ao_chans`
member becomes redundant and can be removed.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
All boards supported by the "amplc_pci230" driver have 16 AI channels,
so the `ai_chans` member of `struct pci230_board` is superfluous and can
be removed.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
`pci230_alloc_private()` is now only called from `pci230_auto_attach()`
to allocate private device storage and initialize various spin-lock
members therein. Absorb the body of `pci230_alloc_private()` into
`pci230_auto_attach()` itself.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The comedi core module calls `comedi_set_hw_dev()` to associate the
hardware `struct device` with the `struct comedi_device` before it calls
the comedi driver's "auto_attach" hook `pci230_auto_attach()`. There is
no need for `pci230_auto_attach()` to call `comedi_set_hw_dev()` itself,
so remove the call.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
`pci230_attach_common()` is now only called from `pci230_auto_attach()`,
so absorb it into that function.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Since the comedi driver's "detach`" handler `pci230_detach()` now merely
calls `comedi_pci_detach()` with the same parameter, use
`comedi_pci_detach()` itself as the "detach" handler.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This driver no longer supports a "legacy" attach mechanism that searches
for a suitable PCI device and increments it's reference count, but since
the common "detach" handler `pci230_detach()` still has a left-over
`pci_dev_put()`, a matching `pci_dev_get()` is needed in the
"auto_attach" handler `pci230_auto_attach()`. There is no longer any
reason to "get" and "put" the PCI device, so those calls can be removed.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The "amplc_pci230" driver currently retains the legacy attach mechanism
to allow devices to be attached manually via the `COMEDI_DEVCONFIG`
ioctl. The only real use for this is to pretend that a PCI230+ or
PCI260+ is a PCI230 or PCI260 for backwards compatibility, as they have
different number of bits of resolution on the AI subdevice. Since the
card would be automatically configured as a PCI230+ or PCI260+ at PCI
probe time anyway, hopefully any users who want it to appear as a PCI230
or PCI260 would have got tired of removing the automatically configured
device and configuring it manually by now and will have updated their
software to cope with the PCI230+ or PCI260+.
Get rid of the legacy attach mechanism by removing the Comedi driver
"attach" handler `pci230_attach()` and associated code. Also remove the
"wildcard" entry from the board table `pci230_boards[]` as it is no
longer needed. Don't bother initializing the `board_name`, `offset`,
and `num_names` members of `struct comedi_driver amplc_pci230_driver`
any longer as they are only needed when configuring the device manually.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Where the only thing in an `else { ... }` block is another `if`
statement, collapse it to an `else if {` block where it makes sense to
do so.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Remove some pairs of parentheses that don't really improve readability.
Also, reduce the amount of leading whitespace in a few places.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Read MAC address from the EEPROM.
This version two corrects a flaw in the result code returning that
did exist in the first version.
Signed-off-by: Olli Salonen <olli.salonen@iki.fi>
Reviewed-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <m.chehab@samsung.com>
TechnoTrend TT-connect CT2-4650 CI (0b48:3012) is an USB DVB-T2/C tuner with
the following components:
USB interface: Cypress CY7C68013A-56LTXC
Demodulator: Silicon Labs Si2168-A20
Tuner: Silicon Labs Si2158-A20
CI chip: CIMaX SP2HF
The firmware for the tuner is the same as for TechnoTrend TT-TVStick CT2-4400.
See https://www.mail-archive.com/linux-media@vger.kernel.org/msg76944.html
The demodulator needs a firmware that can be extracted from the Windows drivers.
File ttConnect4650_64.sys should be extracted from
http://www.tt-downloads.de/bda-treiber_4.1.0.4.zip (MD5 sum below).
3464bfc37a47b4032568718bacba23fb ttConnect4650_64.sys
Then the firmware can be extracted:
dd if=ttConnect4650_64.sys ibs=1 skip=273376 count=6424 of=dvb-demod-si2168-a20-01.fw
The SP2 CI module requires a definition of a function cxusb_tt_ct2_4650_ci_ctrl
that is passed on to the SP2 driver and called back for CAM operations.
[crope@iki.fi: meld USB ID define patch to this]
Signed-off-by: Olli Salonen <olli.salonen@iki.fi>
Reviewed-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <m.chehab@samsung.com>
Driver for the CIMaX SP2 common interface chip. It is very much based on
the existing cimax2 driver for cx23885, but should be more reusable. The
product has been sold with name Atmel T90FJR as well and the data sheets
for that chip seem to be publicly available.
It seems that the USB device that I have and the cx23885 based devices will
need to interact differently with the chip for the CAM operations. Thus
there is one callback function that is passed on to the sp2 driver
(see function sp2_ci_op_cam for that one).
IRQ functionality is not included currently (not needed by USB devices
and I don't have a PCIe device for development).
Signed-off-by: Olli Salonen <olli.salonen@iki.fi>
Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <m.chehab@samsung.com>
TS mode must be set in the existing TechnoTrend CT2-4400 driver.
Signed-off-by: Olli Salonen <olli.salonen@iki.fi>
Reviewed-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <m.chehab@samsung.com>
