The private data variable 'dac' is used to hold the PCI base
address for the board. Use the 'iobase' variable provided by
the comedi_device instead.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The PCI device ids supported by this driver are used multiple
places in the code. To improve maintainability, create #define's
for them.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The functions cb_pcidda_ai_cmd(), cb_pcidda_ai_cmdtest(), and
cb_pcidda_ns_to_timer() are all partially coded stubs for command
support with the analog output subdevice. Just remove the code
since it is #if 0'ed out. It can be added back later if command
support is completed.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Pull spi fixes from Mark Brown:
"A bunch of fixes here, mostly minor except for the pl022 which has
just been a bit of a shambles all round, the recent runtime PM changes
have as far as I can tell never worked so they're just getting thrown
out."
* tag 'spi-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/misc:
spi/pl022: Revert recent runtime PM changes
spi: tsc2005: delete soon-obsolete e-mail address
spi: spi-rspi: fix build error for the latest shdma driver
Pull IOMMU fixes from Joerg Roedel:
"Two fixes this time:
1. Another fix for a broken BIOS to detect when AMD IOMMU interrupt
remapping can not work reliably
2. Typo fix for NVidia IOMMU driver"
* tag 'iommu-fixes-v3.7-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu:
iommu/tegra: smmu: Fix deadly typo
iommu/amd: Work around wrong IOAPIC device-id in IVRS table
Pull pinctrl fixes from Linus Walleij:
"This fixes a few pinctrl problems seen since v3.7-rc1:
- Section tagging for init code
- Use proper pointers to lookup struct device * in the bcm2835
(a.k.a. Raspberry Pi)
- Remove duplicate #includes
- Fix bad return values in errorpath
- Remove extraneous pull function from the sirf driver causing build
errors
- Provide compilation stubs for the Nomadik pinctrl driver when used
with legacy systems without PRCMU units
- Various irqdomain fixes in the Nomadik driver as predicted
- Various smallish bugs in the Tegra driver, most also targeted for
stable
- Removed a deadlocking mutex in the groups debugfs show function"
* tag 'pinctrl-v3.7-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-pinctrl:
pinctrl/nomadik: pass DT node to the irqdomain
pinctrl/nomadik: use zero as default irq_start
pinctrl: fix missing unlock on error in pinctrl_groups_show()
pinctrl/nomadik: use irq_create_mapping()
pinctrl: remove mutex lock in groups show
pinctrl: tegra: correct bank for pingroup and drv pingroup
pinctrl: tegra: set low power mode bank width to 2
dt: Document: correct tegra20/30 pinctrl slew-rate name
Pull apparmor bugfix from James Morris.
Fix a possibly unbounded recursion by iterating over the entries instead.
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/linux-security:
apparmor: fix IRQ stack overflow during free_profile
Pull amd64_edac fix from Borislav Petkov:
"An array out-of-bounds fix from Andrew when setting the scrub rate of
the memory controller."
* tag 'edac_scrubrates_fix' of git://git.kernel.org/pub/scm/linux/kernel/git/bp/bp:
amd64_edac:__amd64_set_scrub_rate(): avoid overindexing scrubrates[]
Pull cgroup fixes from Tejun Heo:
"This pull request contains three fixes.
Two are reverts of task_lock() removal in cgroup fork path. The
optimizations incorrectly assumed that threadgroup_lock can protect
process forks (as opposed to thread creations) too. Further cleanup
of cgroup fork path is scheduled.
The third fixes cgroup emptiness notification loss."
* 'for-3.7-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup:
Revert "cgroup: Remove task_lock() from cgroup_post_fork()"
Revert "cgroup: Drop task_lock(parent) on cgroup_fork()"
cgroup: notify_on_release may not be triggered in some cases
Pull workqueue fix from Tejun Heo:
"This pull request contains one patch from Dan Magenheimer to fix
cancel_delayed_work() regression introduced by its reimplementation
using try_to_grab_pending(). The reimplementation made it incorrectly
return %true when the work item is idle.
There aren't too many consumers of the return value but it broke at
least ramster."
* 'for-3.7-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq:
workqueue: cancel_delayed_work() should return %false if work item is idle
This should get rid of warnings of the type:
warning: passing argument 1 of '' discards qualifiers from pointer target type
note: expected 'void *' but argument is of type 'const void *'
Signed-off-by: Omar Ramirez Luna <omar.ramirez@copitl.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Change the type of sync_addr to 'void __iomem *' and ioremap the
physical address in the shared memory so we can access it using
_raw_*. While at it, drop 'dw_' prefix.
Fix the warning associated with dsp's sync_addr:
warning: passing argument 2 of '__raw_writel' makes pointer from integer without a cast
../io.h:88: note: expected 'volatile void *' but argument is of type 'u32'
Signed-off-by: Omar Ramirez Luna <omar.ramirez@copitl.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Currently per_pm_base and core_pm_base are declared as u32, however
_raw_* changed the data type, since:
195bbca ARM: 7500/1: io: avoid writeback addressing modes for __raw_ accessors
This should fix warnings for per and core accesses:
warning: passing argument 2 of '__raw_writel' makes pointer from integer without a cast
../io.h:88: note: expected 'volatile void *' but argument is of type 'u32'
Signed-off-by: Omar Ramirez Luna <omar.ramirez@copitl.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Custom mmu functions receive a 'const void __iomem *', all the
callers pass a 'void __iomem *', so drop the const to fix the
warnings like:
warning: passing argument 2 of '__raw_writel' discards qualifiers from pointer target type
../io.h:88: note: expected 'volatile void *' but argument is of type 'const void *'
Signed-off-by: Omar Ramirez Luna <omar.ramirez@copitl.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Requested irq for mmu is currently conflicting with a DMA irq
due to recent changes to irq header files, now the offset for the
start of the interrupt controller numbering has changed.
This should be removed during a future migration to omap-iommu,
for now it is hardcoded.
Signed-off-by: Omar Ramirez Luna <omar.ramirez@copitl.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The warning check for duplicate sysfs entries can cause a buffer overflow
when printing the warning, as strcat() doesn't check buffer sizes.
Use strlcat() instead.
Since strlcat() doesn't return a pointer to the passed buffer, unlike
strcat(), I had to convert the nested concatenation in sysfs_add_one() to
an admittedly more obscure comma operator construct, to avoid emitting code
for the concatenation if CONFIG_BUG is disabled.
Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: stable@vger.kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Update the prompt line and description for the COMEDI_AMPLC_DIO200_PCI
config option to add the new PCIe boards supported by the "amplc_dio200"
driver.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Change the `insn_read` and `insn_write` handlers of the '8254' counter
subdevices to support reading and writing multiple data values.
(Writing multiple values isn't very useful, but if that's what the user
wants....)
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Add clock source constants for the timer subdevice of the PCIe cards
supported by the "amplc_dio200" driver.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Implement the timer subdevice for the new PCIe boards. The subdevice was
previously marked as unused, but was reserved for this purpose.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
These are currently unused but will be used to support extra features of
the PCIe boards.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Add constants to "comedi.h" for the extra clock and gate sources
allowed by the new PCIe cards supported by the "amplc_dio200" driver.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
One of the enhancements of the new PCIe boards is the addition of extra
clock and gate sources. Allow clock and gate sources up to 31 for these
boards (compared to 7 for the old boards) although most of those are
reserved for future use.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The new PCIe boards PCIe215, PCIe236 and PCIe296 have some enhanced
features (over the older boards), although none of these features are
supported by this driver yet.
Enable the enhanced features in hardware and use `dio200_layouts[]` to
indicate which boards support the enhanced features.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The new PCIe boards PCIe215, PCIe236 and PCIe296 need a special register
setting to allow generation of interrupts on the PCIe bus. Add
`dio200_pcie_board_setup()` to do this and call it from
`dio200_attach_pci()` for those boards.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Add preliminary support for the above PCIe boards. The interrupt
subdevice is not working yet as the boards need some additional
initialization to enable interrupts. Reserve an unused subdevice that
will eventually become a timer subdevice.
Some unused subdevices have been added between the used subdevices on
the PCIe215 and PCIe236. This is so the subdevice numbers will match a
pattern-matching feature to be added at a later date, where the
pattern-match specifications for a subdevice include pattern-match
results from other subdevices, specified by index.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Add `mainshift` member to `struct dio200_board` to hold the amount of
left-shift required for main register offsets. This is 0 for all the
boards currently supported so it doesn't need initializing explicitly in
any current element of `dio200_boards[]`. It will be non-zero for some
new boards to be supported by this driver.
Modify the register access functions `dio200_read8()` and
`dio200_write8()` to take the shift into account.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Store the length of the main register region in new member `mainsize` of
`struct dio200_board` and initialize it in each element of
`dio200_boards[]`. For all currently supported boards this is
initialized to `DIO200_IO_SIZE` (0x20), but will be different for newer
boards to be supported in the future.
For ISA boards, this is the actual region length requested, but for PCI
boards it's the minimum expected region length. Verify that the PCI BAR
region length is sufficient in `dio200_attach_pci()`.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Add `mainbar` member to `struct dio200_board` to hold the PCI BAR number
for the main registers. This is `2` for the PCI boards currently
supported (PCI215 and PCI272) but will be different for some new boards
to be supported later.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The boards currently supported by this module all use port I/O. Support
memory-mapped I/O as well for future PCI/PCIe cards.
Define `struct dio200_region` to hold the type of register access and
either the port I/O base address or an ioremapped MMIO address. Add a
member `io` to the comedi device private data (`struct dio200_private`)
to hold this. Use this instead of `dev->iobase`. Memory-mapped
registers are mapped in `dio200_pci_attach()` and unmapped in
`dio200_detach()`.
`dio200_detach()` now uses the private data pointer `devpriv` set to
`dev->private` but can return early if it is `NULL` because no clean-up
needs to be done in that case.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Don't pass the I/O base address to `dio200_common_attach()`. The only
thing it does with it is set `dev->iobase` to the passed in value. Do
that before calling `dio200_common_attach()` in order to simplify
upcoming support for different register access methods.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Add custom functions to access hardware registers and call them instead
of calling `inb()` and `outb()` directly. When additional cards are
supported by this driver requiring different register access methods,
the new functions will localize the register access differences to just
these functions.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The "amplc_dio200" module no longer depends on the "8255" module, so the
'COMEDI_AMPLC_DIO200' Kconfig setting no longer needs to select
'COMEDI_8255'.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Implement the '8255' DIO subdevice internally to this module instead of
using the external "8255" module. I plan to add support for additional
cards to this driver that would require the I/O callback functionality
of the 8255 module, but the existing callback functions do not have much
context to handle this elegantly. The additional cards also have extra
DIO features which cannot be handled by the existing "8255" module and
that I'd like to support some time in the future.
The bottom line is I _could_ continue using the "8255" module for a
while with a callback function, but it would turn out to be a very ugly
callback function and I'd have to ditch the use of the "8255" module as
soon as I added an extra feature to the DIO subdevice.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Add our own functions to manipulate the '8254' counter chip instead of
the inline ones from "8253.h". This will make the code less messy when
we add code to support new boards later.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Replace the absolute, modified I/O base addresses in the private data
structures for the '8254' counter and 'intr' DIO subdevices with offsets
from the main I/O base address. `dio200_subdev_intr_init()` now needs
the offset instead of the absolute address. `dio200_subdev_8254_init()`
now only needs the offset instead of both the absolute address and the
offset. The '8255' DIO subdevices are unaffected as they are handled by
the external "8255" module.
This change is not useful by itself but is an intermediate step for
later changes.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The `has_clk_gat_sce` member of `struct dio200_subdev_8254` indicates
whether the board has clock and gate source selection registers. The
same information can be obtained from read-only board layout data so
doesn't need to be set per '8254' counter subdevice. Eliminate the
member and use the read-only data instead. The last parameter of
`dio200_subdev_8254_init()` is used to initialize this member, so remove
that parameter as well.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Add inline helper function `dio200_board_layout(board)` to get a pointer
to the board layout data for a board. Add inline helper function
`dio200_dev_layout(dev)` to get a pointer to the board layout data for a
comedi device (this function is currently unused but will be used by a
later change).
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Now that the tag name `dio200_layout` is available, rename `struct
dio200_layout_struct` to `struct dio200_layout` as the `_struct` suffix
is a bit redundant and I plan to use this type in other places in the
module.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Rename `enum dio200_layout` to `enum dio200_layout_idx`. It's only used
once and frees up the `dio200_layout` tag for something else later.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Rename the clock and gate configuration functions for the '8254' counter
subdevices and pass in the pointers to the comedi device and comedi
subdevice. This is just preparing the way for later changes.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Absorb `dio200_pci_common_attach()` into `dio200_attach_pci()` since
that's the only place it is called from.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The call to `comedi_set_hw_dev()` from `dio200_pci_common_attach()` is
now unnecessary since `dio200_pci_common_attach()` is now only called
from this driver's `attach_pci` hook `dio200_attach_pci()` and the
comedi core now calls `comedi_set_hw_dev()` before calling that. Remove
the unnecessary call.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Now that this driver no longer supports "manual" attachment of PCI
devices in its `attach` hook (`dio200_attach()`), it no longer has code
that searches for a suitable PCI device and increments its reference
count. Since the driver no longer has any reason for incrementing and
decrementing the PCI device's reference count, the calls to
`pci_dev_get()` and `pci_dev_put()` can be removed.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
