| Age | Commit message (Collapse) | Author |
|---|
|
Using this unconditionally cuts the birds feet (;-p).
This reverts commit 491f263ebfb338c59abfbde6d4e0e7256a0150fa.
|
|
Commit 455bd4c430b0 ("ARM: 7668/1: fix memset-related crashes caused by
recent GCC (4.7.2) optimizations") attempted to fix a compliance issue
with the memset return value. However the memset itself became broken
by that patch for misaligned pointers.
This fixes the above by branching over the entry code from the
misaligned fixup code to avoid reloading the original pointer.
Also, because the function entry alignment is wrong in the Thumb mode
compilation, that fixup code is moved to the end.
While at it, the entry instructions are slightly reworked to help dual
issue pipelines.
Signed-off-by: Nicolas Pitre <nico@linaro.org>
Tested-by: Alexander Holler <holler@ahsoftware.de>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
|
|
optimizations
Recent GCC versions (e.g. GCC-4.7.2) perform optimizations based on
assumptions about the implementation of memset and similar functions.
The current ARM optimized memset code does not return the value of
its first argument, as is usually expected from standard implementations.
For instance in the following function:
void debug_mutex_lock_common(struct mutex *lock, struct mutex_waiter *waiter)
{
memset(waiter, MUTEX_DEBUG_INIT, sizeof(*waiter));
waiter->magic = waiter;
INIT_LIST_HEAD(&waiter->list);
}
compiled as:
800554d0 <debug_mutex_lock_common>:
800554d0: e92d4008 push {r3, lr}
800554d4: e1a00001 mov r0, r1
800554d8: e3a02010 mov r2, #16 ; 0x10
800554dc: e3a01011 mov r1, #17 ; 0x11
800554e0: eb04426e bl 80165ea0 <memset>
800554e4: e1a03000 mov r3, r0
800554e8: e583000c str r0, [r3, #12]
800554ec: e5830000 str r0, [r3]
800554f0: e5830004 str r0, [r3, #4]
800554f4: e8bd8008 pop {r3, pc}
GCC assumes memset returns the value of pointer 'waiter' in register r0; causing
register/memory corruptions.
This patch fixes the return value of the assembly version of memset.
It adds a 'mov' instruction and merges an additional load+store into
existing load/store instructions.
For ease of review, here is a breakdown of the patch into 4 simple steps:
Step 1
======
Perform the following substitutions:
ip -> r8, then
r0 -> ip,
and insert 'mov ip, r0' as the first statement of the function.
At this point, we have a memset() implementation returning the proper result,
but corrupting r8 on some paths (the ones that were using ip).
Step 2
======
Make sure r8 is saved and restored when (! CALGN(1)+0) == 1:
save r8:
- str lr, [sp, #-4]!
+ stmfd sp!, {r8, lr}
and restore r8 on both exit paths:
- ldmeqfd sp!, {pc} @ Now <64 bytes to go.
+ ldmeqfd sp!, {r8, pc} @ Now <64 bytes to go.
(...)
tst r2, #16
stmneia ip!, {r1, r3, r8, lr}
- ldr lr, [sp], #4
+ ldmfd sp!, {r8, lr}
Step 3
======
Make sure r8 is saved and restored when (! CALGN(1)+0) == 0:
save r8:
- stmfd sp!, {r4-r7, lr}
+ stmfd sp!, {r4-r8, lr}
and restore r8 on both exit paths:
bgt 3b
- ldmeqfd sp!, {r4-r7, pc}
+ ldmeqfd sp!, {r4-r8, pc}
(...)
tst r2, #16
stmneia ip!, {r4-r7}
- ldmfd sp!, {r4-r7, lr}
+ ldmfd sp!, {r4-r8, lr}
Step 4
======
Rewrite register list "r4-r7, r8" as "r4-r8".
Signed-off-by: Ivan Djelic <ivan.djelic@parrot.com>
Reviewed-by: Nicolas Pitre <nico@linaro.org>
Signed-off-by: Dirk Behme <dirk.behme@gmail.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
|
|
Enable VLAN networking functionality.
While at it also enable some more kernel modules for Apalis T30 PCIe
based devices like the Intel E1000E NICs, Intel IWLAGN Wi-Fi cards,
PCIe serial cards and XHCI based USB 3.0 controllers.
|
|
This reverts commit f162e88fc6aac0fef03c49c1ac596995a03ff134.
Somehow NVIDIA's media player does not like it to be fixed as
it starts stuttering like crazy at times.
Still keep the dangerous time out handling clean-up.
While at it get rid of the excessive dev_info() logging as well.
|
|
These 2 syncronize_rcu()s make attaching a task to a cgroup
quite slow, and it can't be ignored in some situations.
A real case from Colin Cross: Android uses cgroups heavily to
manage thread priorities, putting threads in a background group
with reduced cpu.shares when they are not visible to the user,
and in a foreground group when they are. Some RPCs from foreground
threads to background threads will temporarily move the background
thread into the foreground group for the duration of the RPC.
This results in many calls to cgroup_attach_task.
In cgroup_attach_task() it's task->cgroups that is protected by RCU,
and put_css_set() calls kfree_rcu() to free it.
If we remove this synchronize_rcu(), there can be threads in RCU-read
sections accessing their old cgroup via current->cgroups with
concurrent rmdir operation, but this is safe.
# time for ((i=0; i<50; i++)) { echo $$ > /mnt/sub/tasks; echo $$ > /mnt/tasks; }
real 0m2.524s
user 0m0.008s
sys 0m0.004s
With this patch:
real 0m0.004s
user 0m0.004s
sys 0m0.000s
tj: These synchronize_rcu()s are utterly confused. synchornize_rcu()
necessarily has to come between two operations to guarantee that
the changes made by the former operation are visible to all rcu
readers before proceeding to the latter operation. Here,
synchornize_rcu() are at the end of attach operations with nothing
beyond it. Its only effect would be delaying completion of
write(2) to sysfs tasks/procs files until all rcu readers see the
change, which doesn't mean anything.
Signed-off-by: Li Zefan <lizefan@huawei.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: Colin Cross <ccross@google.com>
Conflicts:
kernel/cgroup.c
|
|
Use HDA HDMI audio now on Colibri T30 as well allowing for up to 7.1
multi-channel playback. Split board sound initialisation between Apalis
and Colibri T30 in order to distinguish better between default SPDIF
being supported out-of-the-box on Apalis vs. this being an optional
feature for the Colibri. Therefore remove SPDIF for the Colibri T30.
While at it add csus clock required for vi_sensor camera master clock
on Apalis T30 and clk_out_2/extern2 for Colibri T30, fix debug UART1
initialisation, get rid of spurious CONFIG_SND_USB enable and further
clean-up both board platform data files.
|
|
Hack HDA driver to make it the 2nd rather than the first sound card as
the first one should be the SGTL5000 which is always present on our T30
modules. Especially on Ixora where there is no HDA codec on the carrier
board HDA being the first sound card does not make much sense even more
though as there was no primary device (e.g. device=hw:0,0).
|
|
Enable CONFIG_FHANDLE required for systemd >= 209 according to
lists.freedesktop.org/archives/systemd-devel/2014-February/017362.html
While at it put them all 3 on par concerning ARM Errata configuration.
|
|
Fix issue with slow SPDIF clock by just doubling desired rate prior to
doing any further calculations on it.
BTW: A proper fix might involve overall clock doubler configuration.
|
|
Fix (Colibri) resp. introduce (Apalis) SPI device controller data to
properly use HW based chip select with one clock of setup and hold time
each for both MCP2515 CAN controller(s) as well as spidev.
|
|
This works around a issue we see when using eMMC 4.5 on tegra
SDHCI host controller.
|
|
The 1.8V quirk also affected the internal eMMC which disabled newer
modes such as SDR50, SDR104 and DDR50. This in turn lead to an
out of spec usage since the clock was still 50MHz.
By creating a no_1v8 field in the platform data we can now enable
this work around on a per-instance basis. Hence we enable the
quirk only on the controllers which are connected to the external
SD-slots.
|
|
PCIe Reset line must be asserted for at least 100us after clock is enabled.
PEX 2 Controller Register fix, offsets are not at constant intervals.
Bug 1521306
Reviewed-on: http://git-master/r/225399
(cherry picked from commit df0760bf515236bed2e87e590509642ab72a01b5)
Change-Id: I7b44ea51e7e02f2bca93cfc75ed85e01ab91fe03
Signed-off-by: Shreshtha Sahu <ssahu@nvidia.com>
|
|
Force round down in division calculating required logo height as
preceding code did round up as follows:
logo_lines = DIV_ROUND_UP(logo_height, vc->vc_font.height);
Which resulted in no boot-logo shown at all due to not enough lines
being available.
|
|
The AC97 sampling rate is really fixed to 48 kHz at all times.
While at it clean-up dangerous time out handling.
|
|
|
|
|
|
This commit resolves an issue of non-working vi camera driver
on Colibri T20 occured after a l4t-r16-r4 merge.
|
|
|
|
Add SPI controller platform data required for use of hardware chip
selects.
While at it pull-up SPI chip select pin due to NVIDIA's designers
taking the term chip select a little too personal: they indeed only
select a chip otherwise all the chip select pins are just left
floating!
|
|
The MCP2515 on the Colibri EvalBoard V3.1a actually uses regular system
reset aka RESET_OUT# on SODIMM 87 to reset the MCP2515.
Therefore clean-up resp. remove any MECS Tellurium legacy in that
respect.
While at it adjust copyright year, re-order some include, clean-up some
struct spacing, re-order clocks and improve some comments.
|
|
|
|
Don't initialize the whole display controller (dc) when framebuffer
mode is set to off.
This avoids creation of the /dev/tegra_dc_x device. The tegra X
driver requires for each tegra_dc_x device an framebuffer device
with the same id. This patch avoids creation of both, which makes
the tegra X driver load correctly.
Note: Turning off tegrafb0 still leads to kernel crash due to
hard coded device enumeration.
|
|
Symptoms:
The interface stops receiving CAN frames when reception of CAN frame
and call to the introduced get_berr_counter function coincide.
Reproduction:
Ip requests, among others, the berr field for the can interfaces. NOTE
that it does so even if called as 'ip link show lo', as interface-based
filtering is only done after the system calls to receive interface
information.
Run candump on a MCP-based can interface that has a lot of traffic (via
'cangen -g1' on another can node).
On the system with the MCP chip, run
while /bin/true; do ip link > /dev/null ; done;
Reception of CAN-frames (as shown by candump) will cease within
seconds.
This issue got introduced by commit
1b59090a47101b4bf379d2c35986ad3f0bcb2aa1 as access to the SPI registers
of the MCP251x were not locked.
Issue report and patch courtesy David R. Piegdon of Indurad GmbH.
|
|
Driver features:
1. Supported formats:
- Full HD, 1920x1080, 8.5 FPS, YUYV
- VGA, 640x480, 30 FPS, YUYV
2. Available controls with enabled AE (auto exposure):
- brightness [0...255],
- contrast [16...64],
- saturation [0...256],
- hue [-2200...2200],
- gamma [100...280].
3. Works only in parallell interface mode (CSI in progress).
4. No external tools needed (for i2c).
|
|
This commit restores the full framerate (25 FPS), after a drop to
16 FPS caused by a l4t-r16-r4 merge.
|
|
This commit adds settings such as:
- values of horizontal/vertical active start,
- enabling/disabling internal sync,
to the private data of decoders: MAX9526, ADV7180, TVP5150,
OV7670 and AS0260.
The feature is available through *_tegra_camera_platform_data
struct initialised in board-*.c files.
|
|
This commit fixes the crash that occures while loading tegra_v4l2_camera
module caused by its new features (private data per sensor).
updated files:
arch/arm/mach-tegra/board-apalis_t30.c
arch/arm/mach-tegra/board-colibri_t20.c
arch/arm/mach-tegra/board-colibri_t30.c
|
|
This commit fixes the kernel compilation error with the lack
of the export of tegra_powergate_partition symbol in
arch/arm/mach-tegra/powergate.c
|
|
|
|
Bug 1369083
Change-Id: I1a81bcb62e8f6bb654ffbebba09661187ab4b512
Signed-off-by: Bryan Wu <pengw@nvidia.com>
Reviewed-on: http://git-master/r/309536
Tested-by: Vikram Fugro <vfugro@nvidia.com>
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Kaustubh Purandare <kpurandare@nvidia.com>
Reviewed-by: Winnie Hsu <whsu@nvidia.com>
Reviewed-by: Matthew Pedro <mapedro@nvidia.com>
|
|
Add support for dual cameras from both CSI-A and CSI-B:
- move all the CSI settings into video buffer struct
- queue the video buffer struct to a dedicated queue
- process one video buffer struct from the queue at one time
Bug 1369083
Change-Id: Ie64d69282ab991b66e97327e288a2bacde088bd6
Signed-off-by: Bryan Wu <pengw@nvidia.com>
Reviewed-on: http://git-master/r/246269
(cherry picked from commit 228b0c2d9ae3fa1121f88836626d654ae0fc4ff0)
Reviewed-on: http://git-master/r/279987
Reviewed-by: Matthew Pedro <mapedro@nvidia.com>
Tested-by: Matthew Pedro <mapedro@nvidia.com>
Conflicts:
drivers/media/video/tegra_v4l2_camera.c
|
|
soc_camera_link supports passing power on/off control callback to
soc_camera stack. So the power control can be handled by soc_camera
stack instead of our Tegra V4L2 host driver.
Also pass other platform_data fields via soc_camera_link instead of
a hacking nvhost_device_data struct.
Bug 1240806
Bug 1369083
Change-Id: I443a7d28196cc8292805da70d2d5ff1c3cd50a5d
Signed-off-by: Bryan Wu <pengw@nvidia.com>
Reviewed-on: http://git-master/r/246267
(cherry picked from commit 9083d270bf93b583cd5bf5151a52ea250f8541a3)
Reviewed-on: http://git-master/r/279986
Reviewed-by: Matthew Pedro <mapedro@nvidia.com>
Tested-by: Matthew Pedro <mapedro@nvidia.com>
Conflicts:
drivers/media/video/tegra_v4l2_camera.c
|
|
CSI host controller of Tegra can support Bayer Raw8 and Raw 10 data
input, it also output data as extended 16-bit data format in memory
directly.
Raw data output should be handled by the second output channel.
This patch add supporting for Raw data input/output and the second
output channel.
All were tested on Cardhu board.
Change-Id: Ifd1d8c94671a1d4571dcf176774e2d3cffafc399
Signed-off-by: Bryan Wu <pengw@nvidia.com>
Reviewed-on: http://git-master/r/200225
Reviewed-by: Automatic_Commit_Validation_User
GVS: Gerrit_Virtual_Submit
Reviewed-by: Matthew Pedro <mapedro@nvidia.com>
Conflicts:
drivers/media/video/tegra_v4l2_camera.c
|
|
Enable IEEE 1588 precision time protocol (PTP) support.
|
|
Springville/i211 with a blank Flash/iNVM use different PCI IDs. Extend
the driver to load despite i210/i211 data sheets claiming tools only,
not for driver.
Only warn rather than fail on NVM validation failures on Apalis T30.
Revise Ethernet MAC address assignment: should now handle up to two
instances of custom user MACs (2nd one with a 0x100000 offset). This
way customer does not have to worry about NVM on a secondary Ethernet
on the carrier board and still gets a valid official MAC address from
us (e.g. analogous to how we did it on our Protea carrier board).
Use the Toradex OUI as default MAC address if no valid one is
encountered.
Tested on samples of Apalis T30 2GB V1.0B, V1.0C and Apalis T30 1GB
V1.0A both with blank NVMs as well as iNVMs programmed with Intel's
defaults.
|
|
Springville/i211 with a blank Flash/iNVM use a different PCI ID (tools
only, not for driver) than properly programmed ones.
|
|
Integrate latest igb driver version 5.1.2
(igb-5.1.2.tar.gz from e1000.sf.net).
|
|
This commit adds a method that MAC drivers may call in order to find out
the device number of their associated PTP Hardware Clock.
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
This commit adds a new ethtool ioctl that exposes the SO_TIMESTAMPING
capabilities of a network interface. In addition, user space programs
can use this ioctl to discover the PTP Hardware Clock (PHC) device
associated with the interface.
Since software receive time stamps are handled by the stack, the generic
ethtool code can answer the query correctly in case the MAC or PHY
drivers lack special time stamping features.
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
Reviewed-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Add a new ethtool operation (get_rxfh_indir_size) to get the
indirectional table size. Use this to validate the user buffer size
before calling get_rxfh_indir or set_rxfh_indir. Use get_rxnfc to get
the number of RX rings, and validate the contents of the new
indirection table before calling set_rxfh_indir. Remove this
validation from drivers.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
Acked-by: Dimitris Michailidis <dm@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
|
|
|
|
|
|
Tegra20 and Tegra30 do not support gen2 PCIe, so correct the
register setting to disable it.
Signed-off-by: Eric Brower <ebrower@nvidia.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Thierry Reding <treding@nvidia.com>
|
|
Conflicts:
drivers/media/video/tegra_v4l2_camera.c
reverted to current driver supporting ACM rather than CSI2
drivers/media/video/videobuf2-dma-nvmap.c
drivers/video/tegra/host/Makefile
|
|
While setup, there is a time the serial port is running on a
predefined speed (115200). When data arrived during this phase at
a lower baudrate (e.g. 9600), a lot of overrun and break error
messages were shown. Due to those messages, the terminal setup
never finished, and thus never reached the selected baudrate (e.g.
9600). By choosing a lower baudrate at the startup (9600), a lot
less (if any at all) messages are shown and those initialization
always succeeds.
Another solution would be to disable those messages completely.
However, while normal mode its good to have those UART error
messages.
|
|
Hack to avoid 24 Hz mode in X resulting in no display at all. This has
been encountered using e.g. a Samsung SyncMaster F2380 if connected by
HDMI.
root@colibri-t20:~# xrandr
Screen 0: minimum 256 x 128, current 1920 x 1080, maximum 2048 x 2048
LVDS-1 connected 1920x1080+0+0 (normal left inverted right x axis y
axis) 0mm x 0mm
1920x1080 60.0*+
HDMI-1 connected 1920x1080+0+0 (normal left inverted right x axis y
axis) 160mm x 90mm
1920x1080 60.0 + 24.0*
1680x1050 59.9
1280x1024 75.0 60.0
1440x900 75.0 59.9
1280x960 60.0
1280x800 59.9
1152x864 75.0
1280x720 60.0 50.0
1024x768 75.0 70.1 60.0
832x624 74.6
800x600 72.2 75.0 60.3 56.2
720x576 50.0
720x480 59.9
640x480 75.0 72.8 66.7 59.9
720x400 70.1
This is really just enabling the same hack from Apalis/Colibri T30
which is now relevant on the Colibri T20 as well due to the modedb
changes:
3a6fc5303b9e798295caa01d2cfe75b05fdccf55
|
|
Before updating modes on HDMI hotplug, check if framebuffer is
registered at all. This allows to disable framebuffers completely.
|
