Age | Commit message (Collapse) | Author |
|
Signed-off-by: Sebastien Jan <s-jan@ti.com>
|
|
CONFIG_OPROFILE breaks the PM feature: when cpu1 is cut, the platform
enters a weird state (not really crashed: the terminal used to cut
cpu1 still echos the strings typed, but the console does not respond.
Other previously open terminal on the board LCD can still be used but
with limited capabilities).
Deactivating OPROFILE 'fixes' these issues.
Signed-off-by: Sebastien Jan <s-jan@ti.com>
|
|
Signed-off-by: Sebastien Jan <s-jan@ti.com>
|
|
WLAN is not supported on blaze board.
If the WLAN driver is loaded, it generates a crash.
Deactivating WLAN permits to avoid this crash.
Long-term solution is to fix the WLAN driver (or the
way it was integrated into mmc/sdio/board file), so
that the same kernel imagete can support Blaze without
WLAN and pandaboard with WLAN.
Signed-off-by: Sebastien Jan <s-jan@ti.com>
|
|
Signed-off-by: Sebastien Jan <s-jan@ti.com>
|
|
Signed-off-by: Sebastien Jan <s-jan@ti.com>
|
|
Signed-off-by: Sebastien Jan <s-jan@ti.com>
|
|
the bounced page needs to be flushed after data is copied into it,
to ensure that architecture implementations can synchronize
instruction and data caches if necessary.
Signed-off-by: Gary King <gking@nvidia.com>
Signed-off-by: Bryan Wu <bryan.wu@canonical.com>
|
|
BugLink: http://bugs.launchpad.net/bugs/632235
Signed-off-by: Andy Whitcroft <apw@canonical.com>
|
|
Signed-off-by: Sebastien Jan <s-jan@ti.com>
|
|
BugLink: http://bugs.launchpad.net/bugs/632235
Ensure the flavour header depends on the correctly named common header.
Also turn off those packages a derivative kernel is not permitted to
generate.
Signed-off-by: Andy Whitcroft <apw@canonical.com>
|
|
- define some components as modules instead of static
- enable USB DMA
Signed-off-by: Sebastien Jan <s-jan@ti.com>
|
|
This patch masks the Secure interrupt at wakeupgen level. This interrupt
is controlled only from secure SW and non-secure SW should be immune to it.
Additinaly the L3 error interupts are masked since they are gating the
lowpower states. This hack will be removed once the L3 error-handling
is in place
Signed-off-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
|
|
This patch fixes two issues.
1. Barriers in the cpudile path
2. Restoration of CPUx state, to avoid accidental transitions
of CPUx power domain to low power states while executing
WFI outside the idle thread. Secure software and hotplog
path are examples of this problems
Signed-off-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
|
|
This patch fixes the lopw power powerdomain, clockdomain counters
One can mount the debug fs and check the low-power state
e.g
$cat /proc/sys/debug/pm_debug/count | grep mpu
$mpu_pwrdm (ON),OFF:87186,RET:2605909,INA:0,ON:2693096,RET-LOGIC-OFF:2605909
Additionaly MPU OFF is programmed to OFF instead of RET in suspend
path
Signed-off-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
|
|
PCI dependency is removed in the DRM makefile. DRM stub files from
DDK are added to drm source tree in the kernel.
Signed-off-by: Gowtham Tammana <g-tammana@ti.com>
[s-jan@ti.com: defconfig changes to debian folder + checkpatch fixes]
Signed-off-by: Sebastien Jan <s-jan@ti.com>
|
|
Signed-off-by: Sebastien Jan <s-jan@ti.com>
|
|
It's possible that the mem struct assignment inside of
the kernel list doesn't assigned properly (e.g. NULL), which
causes a deref crash during the free call.
Signed-off-by: David Sin <davidsin@ti.com>
Signed-off-by: Sebastien Jan <s-jan@ti.com>
|
|
If free_list is empty the element won't be assigned
to NULL. Causing issues when freeing the pages.
Signed-off-by: Fernando Guzman Lugo <fernando.lugo@ti.com>
Signed-off-by: Bhavin Shah <bshah@ti.com>
Signed-off-by: David Sin <davidsin@ti.com>
|
|
This patch fixes the following kernel-hacking compile errors that are
present when the kernel-hacking for mutexes and spinlocks is enabled:
drivers/dsp/syslink/multicore_ipc/ipc_drv.c:76: error: braced-group
within expression allowed only inside a function
drivers/dsp/syslink/multicore_ipc/ipc_drv.c:77: error: braced-group
within expression allowed only inside a function
The fix is to change the DECLARE_COMPLETION_ON_STACK declarations to
DECLARE_COMPLETION.
Signed-off-by: Angela Stegmaier <angelabaker@ti.com>
|
|
If there has been an MPU-side app failure and the message queue has
been deleted, then obj could be NULL in the local put part of
messageq_put. So, after setting obj, obj should be checked for
NULL-ness before being de-referenced.
Signed-off-by: Angela Stegmaier <angelabaker@ti.com>
|
|
Fix this build error with CONFIG_OMAP2_DSS_HDMI disabled:
Signed-off-by: Anand Gadiyar <gadiyar@ti.com>
|
|
various includes were left from the original based of the 4430sdp.
remove these for cleaner board file.
Signed-off-by: David Anders <x0132446@ti.com>
|
|
clean up various coding format issues
Signed-off-by: David Anders <x0132446@ti.com>
|
|
The trigger heartbeat is used by default for Beagleboard and Pandaboard,
but it's only activated when the module is loaded. Changing to built-in
to be always used and activated during boot time.
Signed-off-by: Ricardo Salveti de Araujo <ricardo.salveti@canonical.com>
|
|
At Pandaboard we have 2 status leds, so adding them with similar usage as
we have for Beagleboard (heartbeat and mmc0).
Signed-off-by: Ricardo Salveti de Araujo <ricardo.salveti@canonical.com>
|
|
Add support to determine which GPIO to use for HUB_NRESET based on
the board revision of the PandaBoard.
Signed-off-by: David Anders <x0132446@ti.com>
|
|
Add initial support for determining the board revision of the
PandaBoard and reports the version on boot. This will assist in
determining which board is being used when providing techincal
support. The board revision will also be used to determine which
GPIO and other configuration changes are needed between different
hardware revisions of the PandaBoard.
Signed-off-by: David Anders <x0132446@ti.com>
|
|
(renamed from:
RFC: don't handle DISCONNECT if we've already got a CONNECT)
Change the irq status to be a bitmask, so multiple irq's can be
handled at one time in work-queue function. This avoids handling
a disconnect if a connect already comes by the time the work
queue function runs.
Also, avoid powering off PHY/PLL to avoid getting inconsistant
connect/disconnect interrupts and inconsistant HPD status.
Signed-off-by: Rob Clark <rob@ti.com>
|
|
This reverts commit 560118846ca54b81e6498433e6d4398f628866d2.
Signed-off-by: Sebastien Jan <s-jan@ti.com>
|
|
Signed-off-by: Mythri P K <mythripk@ti.com>
checkpatch fixes
Signed-off-by: Sebastien Jan <s-jan@ti.com>
|
|
Signed-off-by: Mythri P K <mythripk@ti.com>
|
|
Integration of L24.9-p2 kernel updates.
Fix for blaze boot (WLAN).
Signed-off-by: Sebastien Jan <s-jan@ti.com>
|
|
Disable auto gating for PER M3 divider
Signed-off-by: Rajendra Nayak <rnayak@ti.com>
Signed-off-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
|
|
Changing 755 file permissions to 644 proper ones
Signed-off-by: Ricardo Perez Olivares <x0081762@ti.com>
|
|
IPC uses the first 4 notify events for internal notifications. The Notify
reserved events should therefore be 4, and not 3. Otherwise, it would allow
the users to get notified and send a wrong value back using event 3.
Signed-off-by: Suman Anna <s-anna@ti.com>
|
|
This patch reverts back Russell King's patch to do the userspace
buffer cache flush. This allows unconditional cache operations and this
patch is considered as a hack patch until the new API is alinged with MM
team
Signed-off-by: Hari Kanigeri <h-kanigeri2@ti.com>
|
|
In ipu_pm_notifications the PID_DEATH/SUSPEND/RESUME
event is sent to IPU and a semaphore is used to wait
for IPU to respond. In the case when the send_event
is failing, there is no need to called the semaphore
and wait.
Also if one of the proc (SYS/APP) is not responding
an error is returned without sending the notification
to the other proc.
Signed-off-by: Miguel Vadillo <vadillo@ti.com>
Signed-off-by: Juan Gutierrez <jgutierrez@ti.com>
Signed-off-by: Angela Stegmaier <angelabaker@ti.com>
Signed-off-by: Paul Hunt <hunt@ti.com>
|
|
Exclude Tiler buffers from getting added to runtime mapped information
since they are already pre-mapped in TLB.
By including them in runtime leads to MMU fault during the process
shutdown.
Signed-off-by: Hari Kanigeri <h-kanigeri2@ti.com>
|
|
Signed-off-by: Hemant Hariyani <hemanthariyani@ti.com>
|
|
disabled.
CC arch/arm/mach-omap2/board-4430sdp.o
arch/arm/mach-omap2/board-4430sdp.c:459: warning: 'sdp4430_hdmi_device' defined but not used
Signed-off-by: Anand Gadiyar <gadiyar@ti.com>
Cc: Mythri P K <mythripk@ti.com>
|
|
CC arch/arm/mach-omap2/board-4430sdp.o
arch/arm/mach-omap2/board-4430sdp.c:506: warning: 'sdp4430_picoDLP_device' defined but not used
Signed-off-by: Anand Gadiyar <gadiyar@ti.com>
CC: Mythri P K <mythripk@ti.com>
CC: David Anders <x0132446@ti.com>
|
|
Wrong adress was getting assigned to vma struct
So a dump was observed in FB with TILER buffers.
Signed-off-by: Mayuresh Janorkar <mayur@ti.com>
|
|
The following SW errata are not required anymore on OMAP4 ES2.0:
- AXI2COP
- PL310_ERRATA_588369
Signed-off-by: Sebastien Jan <s-jan@ti.com>
|
|
Kernel crashes at board init time with TIWLAN_SDIO activated.
This is caused by a wrong memory access.
Signed-off-by: Sebastien Jan <s-jan@ti.com>
|
|
Local change, not to be merged into Ubuntu tree.
Signed-off-by: Sebastien Jan <s-jan@ti.com>
|
|
Ignore: yes
Signed-off-by: Bryan Wu <bryan.wu@canonical.com>
|
|
ES1.0 support breaks ES2.0, so revert this option.
Note: if we wanna test Ubuntu 2.6.35 kernel on ES1.0 hardware, we need
to revert this commit and also tweak some USB EHCI configs.
Signed-off-by: Sebastien Jan <s-jan@ti.com>
Signed-off-by: Bryan Wu <bryan.wu@canonical.com>
|
|
A new random value for the canary is stored in the task struct whenever
a new task is forked. This is meant to allow for different canary values
per task. On ARM, GCC expects the canary value to be found in a global
variable called __stack_chk_guard. So this variable has to be updated
with the value stored in the task struct whenever a task switch occurs.
Because the variable GCC expects is global, this cannot work on SMP
unfortunately. So, on SMP, the same initial canary value is kept
throughout, making this feature a bit less effective although it is still
useful.
One way to overcome this GCC limitation would be to locate the
__stack_chk_guard variable into a memory page of its own for each CPU,
and then use TLB locking to have each CPU see its own page at the same
virtual address for each of them.
Signed-off-by: Nicolas Pitre <nicolas.pitre@linaro.org>
(cherry picked from commit df0698be14c6683606d5df2d83e3ae40f85ed0d9
bryan.wu@canonical.com: modified the code to make it apply to ti-omap4)
Signed-off-by: Bryan Wu <bryan.wu@canonical.com>
|
|
dirty modules
After they cleanup their dirty modules, we can turn it on.
Signed-off-by: Bryan Wu <bryan.wu@canonical.com>
|