Now it looks like we finally know enough about the output devices to give
them proper names. As VIA_96 is often referred to as DVP0 rename it to
VIA_DVP0. As VIA_6C and VIA_93 seem to exist only on CLE266 and "replace"
DVP0 and DVP1 there rename them to VIA_LDVP0 and VIA_LDVP1 (L as legacy).
The proc names were changed accordingly which should be harmless as they
were just introduced and not beyond RFC state.
This patch should make things a bit more comfortable and less scary.
Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de>
Acked-by: Jonathan Corbet <corbet@lwn.net>
Cc: Joseph Chan <JosephChan@via.com.tw>
Cc: Andrew Morton <akpm@linux-foundation.org>
This patch maps supported output devices to IGP versions. This list may
contain errors as most of it is derived of the driver source but it should
be correct enough to provide a good help. The devices are exported via a
proc entry in the same format as those showing the output devices per IGA.
Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de>
Acked-by: Jonathan Corbet <corbet@lwn.net>
Cc: Joseph Chan <JosephChan@via.com.tw>
Cc: Andrew Morton <akpm@linux-foundation.org>
This patch sets the sync polarity for all output devices, not only CRT.
This may give some people a working screen but only if lcd scaling and
centering are not used as it is currently too dificult to propagate a
different resolution (from what the application thinks) to the correct
output device. Hopefully this does not introduce regressions as the
polarity of non-CRT devices was completly ignored before.
Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de>
Cc: Joseph Chan <JosephChan@via.com.tw>
At the moment only the sync polarity for CRT is handled but there are
also bits for controlling the sync polarity for other output devices.
Add a function to change those similar to the other output device
functions.
There is no runtime change yet as the code still handles only CRT.
Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de>
Cc: Joseph Chan <JosephChan@via.com.tw>
This patch reduces the value for I2C timeout and udelay.
The udelay was reduced to 10 (old: 40) which is still very high as for
standard-mode I2C even 5 should work. This gives a speedup of factor 4
when talking to I2C devices.
The timeout was reduced to 2 (old: 20) which is taken from the radeon
driver so it should work as well. This gives a speedup of factor 10 when
detecting that there is no I2C device we want to talk to.
This causes a huge improvement of device initialization time.
Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de>
Acked-by: Jonathan Corbet <corbet@lwn.net>
Cc: Joseph Chan <JosephChan@via.com.tw>
This patch enables the I2C port to talk with the CRT. This allows adding
EDID reading and parsing capability for CRTs.
Just a small preparation, no notable changes in user experience yet.
Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de>
Acked-by: Jonathan Corbet <corbet@lwn.net>
Cc: Joseph Chan <JosephChan@via.com.tw>
i2c_transfer returns negative errno on error and number of messages
processed on success. Just returning this value would give a poor
interface as it is not obvious that you must compare with 2 after reading
1 or n bytes and with 1 after writing 1 byte to determine if it was
successful. To avoid this error prone interface convert the error code
of a successful read/write to zero and all other non-negative values to
an negative error code.
This fixes a regression introduced by
via: Rationalize vt1636 detection
which resulted in no longer detecting a VT1636 chip and therefore has
broken the output in configurations which contain this chip.
Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de>
Acked-by: Jonathan Corbet <corbet@lwn.net>
Cc: Joseph Chan <JosephChan@via.com.tw>
Cc: stable@kernel.org
This patch merges tbl1636 into vt1636 and cleans it up as the data was
only used there anyway. No runtime changes are expected.
Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de>
Cc: Joseph Chan <JosephChan@via.com.tw>
This patch moves common parts of dvi.c, lcd.c and vt1636.c to hw.c to
start a per output device power management. There should be no runtime
changes aside that this patch enables the proc interface to enable/disable
devices when needed which greatly increases the chances that changes to
the output device configuration will work. However the power management is
not yet complete so it might fail on some configurations. As this area is
quite complex and touches undocumented things there is a slight chance of
regressions.
Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de>
Cc: Joseph Chan <JosephChan@via.com.tw>
This patch cleans the LCD code up. It forbids overwritting global
decissions like what IGA should be used as the source and whether to
enable the second display channel. Additionally it removes a bit code
duplication.
Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de>
Cc: Joseph Chan <JosephChan@via.com.tw>
This patch extends the proc entry to contain a possibility to view and
change the output devices for each IGA. This is useful for debugging
output problems as it provides a reliable way to query which low level
devices are active after VIAs output device configuration nightmare
happended. It's as well suitable for daily use as one can change the
output configuration on the fly for example to connect a projector.
At the moment it's still unstable. The reason is that we have to handle
a bunch of undocumented output devices (those without a proper name) and
that this patch is the first step to collect and verify the needed
information. Basically the only configuration change that is expected to
work at the moment is switching output devices between IGA1 and IGA2.
Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de>
Acked-by: Jonathan Corbet <corbet@lwn.net>
Cc: Joseph Chan <JosephChan@via.com.tw>
Cc: Andrew Morton <akpm@linux-foundation.org>
This patch merges the remaining functionality of the output path
function in the associated enabling functions. This is very natural as
most of the remaining code does actually enable the device.
Just some more or less intelligent code merge. If no stupid mistakes
occured there should be no regressions.
Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de>
Cc: Joseph Chan <JosephChan@via.com.tw>
This patch uses the iga{1,2}_devices variables to select which IGA
should be the source. Doing this is convinient, more powerfull
than the older scheme and easy extendable to support further output
devices. It is not yet completed as the device on/off selection needs
to be converted to the same scheme to take full advantage.
No visible changes yet as we want to complete the transition before
anouncing any unstabke interface.
Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de>
Cc: Joseph Chan <JosephChan@via.com.tw>
This patch adds a new output device management that stores for each
IGA which output devices are routed to it and a compatiblity layer
that converts the old per-output device values in the new format.
Bounding the output devices to each IGA is a central idea of the
cleanup. Doing it this way should be easier and make much more sense
than the old format which happily mixed different output devices
together and did not even take into account that some devices are no
longer available on newer chipsets.
Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de>
Cc: Joseph Chan <JosephChan@via.com.tw>
The result of viafb_set_iga_path can change after init only in the
hotplug ioctl. So we can move it there rather than doing it always
when a new mode is set. The setup during init is done via a call
from the init chip funtion.
This change enables a stable mapping between the old device scheme
and a new more powerfull one.
Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de>
Cc: Joseph Chan <JosephChan@via.com.tw>
There are a lot of init functions which are not marked as such.
Fix this.
Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de>
Cc: Joseph Chan <JosephChan@via.com.tw>
This patch rips the device routing out of the 3 main functions to
separate functions to make them available for transition to a better
controlling scheme.
Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de>
Cc: Joseph Chan <JosephChan@via.com.tw>
Move the enabling of the second display channel to the generic
output routing function and do it exactly if something is using it.
This unifies a zoo of variants how to do it and does no longer
disable it if LCD is not on the second display which is much saner
as there can be other users who need it. Probably this did not cause
any problems so far as we only recently started dual fb support and
otherwise there won't be other users and LCD is preferred assigned to
second display channel.
Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de>
Cc: Joseph Chan <JosephChan@via.com.tw>
Move all output path routing directly in the viafb_setmode. This
gives a better overview and allows to factor similar parts out.
Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de>
Cc: Joseph Chan <JosephChan@via.com.tw>
This patch moves the power on/off sequences to lcd.c as they are only
used there. This allows removing lcdtbl.h as the other stuff is not
used.
Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de>
Cc: Joseph Chan <JosephChan@via.com.tw>
Looks like we did reset the PLL of the (whatever) engine instead of
the PLL of the secondary display (IGA2, LCDCK). This patch fixes it.
Signed-off-by: Florian Tobias Schandinat <FlorianSchandinat@gmx.de>
Cc: Joseph Chan <JosephChan@via.com.tw>
Correct function being needlessly visible outside compilation unit
when the only users are internal.
Signed-off-by: Daniel J Blueman <daniel.blueman@gmail.com>
Reviewed-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Fix string interpreted as trigraph and typo.
Signed-off-by: Daniel J Blueman <daniel.blueman@gmail.com>
Reviewed-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Correct function storage class, and correct assignment type.
Signed-off-by: Daniel J Blueman <daniel.blueman@gmail.com>
Reviewed-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Correct allocation flags type and function prototype for ANSI C compliance.
[airlied: whitespace fixed]
Signed-off-by: Daniel J Blueman <daniel.blueman@gmail.com>
Reviewed-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
nouveau_bios_fp_mode() zeroes the mode struct before filling in relevant
entries. This nukes the mode id initialised by drm_mode_create(), and
causes warnings from idr when we try to remove the mode.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
A single driver shouldn't be overriding the kernel-wide -W options.
This removes them from the Makefile.
Bonus is that the code at least now will build on a 64bit platform.
Problem is that both drivers can't be built at the same time right now
or bad things happen when linking.
Cc: Henry Ptasinski <henryp@broadcom.com>
Cc: Nohee Ko <noheek@broadcom.com>
Cc: jason <jason@lakedaemon.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Change to use the proper ccflags-y option, as well as splitting the
options out one-per-line so that we can see what is needed to be cleaned
up and removed over time.
Cc: Henry Ptasinski <henryp@broadcom.com>
Cc: Nohee Ko <noheek@broadcom.com>
Cc: jason <jason@lakedaemon.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
E1000 can benefit from calling the GRO receive functions.
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Jeff Pieper <jeffrey.e.pieper@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Net drivers in general have an issue where timers fired
by mod_timer or work threads with schedule_work are running
outside of the rtnl_lock.
With no other lock protection these routines are vulnerable
to races with driver unload or reset paths.
The longer term solution to this might be a redesign with
safer locks being taken in the driver to guarantee no
reentrance, but for now a safe and effective fix is
to take the rtnl_lock in these routines.
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Jeff Pieper <jeffrey.e.pieper@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
E1000 is using several timers that in a follow on patch
will need to acquire the rtnl_lock in order to be safe.
This patch moves the timer bodies into work queues which
will allow the next patch to add rtnl_lock.
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Jeff Pieper <jeffrey.e.pieper@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When using simultaneously the two DMA channels on a same engine, some
transfers are never completed. For example, an endless lock can occur
while writing heavily on a RAID5 array (with async-tx offload support
enabled).
Note that this issue can also be reproduced by using the DMA test
client.
On a same engine, the interrupt cause register is shared between two
DMA channels. This patch make sure that the cause bit is only cleared
for the requested channel.
Signed-off-by: Simon Guinot <sguinot@lacie.com>
Tested-by: Luc Saillard <luc@saillard.org>
Acked-by: saeed bishara <saeed.bishara@gmail.com>
Cc: <stable@kernel.org>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Change the code so that it will use the correct size for keymap entries.
Do it in a way that makes it harder to screw it up in the future.
Reported-by: Jaime Velasco Juan <jsagarribay@gmail.com>
Signed-off-by: Henrique de Moraes Holschuh <hmh@hmh.eng.br>
Signed-off-by: Matthew Garrett <mjg@redhat.com>
In the event that the external chipset doesn't implement the
GET_SUPPORTED_ENHANCEMENTS commands, gracefully treat it as having no
enhancments rather than bailing.
Reported-and-tested-by: Woody Suwalski <terraluna977@gmail.com>
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=18342
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
We weren't unlinking the freed connector from the drm lists, and so
hit some use-after-free if we failed to initialise the connector.
Reported-and-tested-by: Woody Suwalski <terraluna977@gmail.com>
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=18342
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Currently the kernel is screaming about slowpath at me for the
wp/cd callbacks. Switch to the _cansleep variants so as to silence
this.
Signed-off-by: Linus Walleij <linus.walleij@stericsson.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
The card may be always present on the board, and for these cases neither
a status callback nor a card detect GPIO is required, and card detection
polling can be disabled.
Acked-by: Linus Walleij <linus.walleij@stericsson.com>
Signed-off-by: Rabin Vincent <rabin.vincent@stericsson.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
If an IRQ can be requested on the card detected GPIO, use it instead of
polling.
Acked-by: Linus Walleij <linus.walleij@stericsson.com>
Signed-off-by: Rabin Vincent <rabin.vincent@stericsson.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Fix this Kconfig warning on allmodconfig for s390:
warning: (BATTERY_DS2760 && POWER_SUPPLY) selects W1 which has unmet
direct dependencies (HAS_IOMEM)
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Anton Vorontsov <cbouatmailru@gmail.com>
We need to wait for the PLLs to settle prior to detecting the state
changes. The BIOS writers guide suggests waiting for the next vblank.
Reported-by: Carlos R. Mafra <crmafra2@gmail.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
This is already performed with the pipelined flush, so by the time we
schedule the flush in the page-flip, the ring is NULL and we OOPs
instead.
Reported-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
A minor typo caused a single fence register to be incorrectly
programmed, resulting in occassional tiling corruption.
Reported-and-tested-by: Hans de Bruin <bruinjm@xs4all.nl>
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=18962
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: stable@kernel.org