[3.10] [AM335x-EVMSK] BUG: spinlock bad magic on CPU#0

2013-07-02 Thread Ambresh K
Hi, While booting 3.10 on AM355x-evmsk with following steps seeing race condition leading to SPIN BUG: Attached log for reference. Config: omap2plus_defconfig. make ARCH=arm CROSS_COMPILE=arm-none-linux-gnueabi- uImage -j 8 LOADADDR=0x80008000 make ARCH=arm

Re: [PATCH V2 3/3] ARM: OMAP2+: clk: Fix return type of callbacks

2013-06-23 Thread Ambresh K
On Friday 21 June 2013 11:43 PM, Mike Turquette wrote: Quoting Ambresh K (2013-06-17 01:39:44) From: Ambresh K ambr...@ti.com clk_ops's .get_parent member data return's signed value. Signed-off-by: Ambresh K ambr...@ti.com Unsurprisingly I get the following errors when building

[PATCH V2 0/3] Fix to clk framework while handling orphan clks

2013-06-17 Thread Ambresh K
From: Ambresh K ambr...@ti.com On a possible HW bug or in-correct configuration of MUX register's in bootloader; might return a value greater than available parent clocks for a MUX clk. Sensing invalid parent index, clk_mux_get_parent returns -EINVALID. Due to function's u8 return type

[PATCH V2 2/3] clk: skip re-parenting orphan clk

2013-06-17 Thread Ambresh K
From: Ambresh K ambr...@ti.com If clk is same as orphan clk then skip the iteration, there by avoiding unnecessary look-up. Signed-off-by: Ambresh K ambr...@ti.com --- drivers/clk/clk.c |4 1 files changed, 4 insertions(+), 0 deletions(-) diff --git a/drivers/clk/clk.c b/drivers/clk

[PATCH V2 1/3] clk: fix clk_mux_get_parent return's signed value

2013-06-17 Thread Ambresh K
From: Ambresh K ambr...@ti.com clk_mux_get_parent should return an error if the value read from the register is erroneous. Currently if the value read is greater than the number of available parents clk_mux_get_parent return's signed error which will result in NULL pointer dereferencing

[PATCH V2 3/3] ARM: OMAP2+: clk: Fix return type of callbacks

2013-06-17 Thread Ambresh K
From: Ambresh K ambr...@ti.com clk_ops's .get_parent member data return's signed value. Signed-off-by: Ambresh K ambr...@ti.com --- arch/arm/mach-omap2/clkt_clksel.c |2 +- arch/arm/mach-omap2/clkt_dpll.c |2 +- arch/arm/mach-omap2/clock.h |4 ++-- 3 files changed, 4

Re: [Patch 3/3] clk: Avoid re-parenting orphan clk's having invalid parent index.

2013-06-13 Thread Ambresh K
Sorry for not being descriptive in commit message. a) Avoids unnecessary re-parenting cycle for orphan clock's with invalid parent for every clock True, but this is a minor optimisation. If this is a big optimization for you then you really need to fix your bootloader. We

Re: [Patch 1/3] clk: fix clk_mux_get_parent return's signed value

2013-06-04 Thread Ambresh K
clk_mux_get_parent return's signed error which will result in NULL pointer dereferencing in the calling functions. Signed-off-by: Ambresh K ambr...@ti.com --- drivers/clk/clk-mux.c|2 +- drivers/clk/clk.c| 12 +++- include/linux/clk-provider.h |4 ++-- 3 files

Re: [Patch 3/3] clk: Avoid re-parenting orphan clk's having invalid parent index.

2013-06-04 Thread Ambresh K
On Wednesday 29 May 2013 12:48 PM, Mike Turquette wrote: Quoting Ambresh K (2013-05-01 23:25:29) From: Ambresh K ambr...@ti.com Add orhan clk nodes having invalid parent index to list and use the list to skip re-parenting orphan clk having invalid parents. Signed-off-by: Ambresh K ambr

[Patch 0/3] Fix to clk framework while handling orphan clks

2013-05-02 Thread Ambresh K
From: Ambresh K ambr...@ti.com On a possible HW bug or in-correct configuration of clk_sel bits in bootloaders, there are high probablity of returning a value greater than available parent clocks for a MUX clk. Sensing invalid parent index, clk_mux_get_parent returns -EINVALID. Due

[Patch 1/3] clk: fix clk_mux_get_parent return's signed value

2013-05-02 Thread Ambresh K
From: Ambresh K ambr...@ti.com If for some reason, the value read from clksel field return erroneous due to HW bug or improper configuration, then clk_mux_get_parent should return appropriate error's. Currently if the value read is greater than the number of available parents clk_mux_get_parent

[Patch 2/3] clk: skip re-parenting orphan clk

2013-05-02 Thread Ambresh K
From: Ambresh K ambr...@ti.com If clk is same as orphan clk than skip the iteration, there by avoiding unnecessary look-up. Signed-off-by: Ambresh K ambr...@ti.com --- drivers/clk/clk.c |4 1 files changed, 4 insertions(+), 0 deletions(-) diff --git a/drivers/clk/clk.c b/drivers/clk

[Patch 3/3] clk: Avoid re-parenting orphan clk's having invalid parent index.

2013-05-02 Thread Ambresh K
From: Ambresh K ambr...@ti.com Add orhan clk nodes having invalid parent index to list and use the list to skip re-parenting orphan clk having invalid parents. Signed-off-by: Ambresh K ambr...@ti.com --- drivers/clk/clk.c | 21 +++-- 1 files changed, 19 insertions(+), 2

[PATCH] OMAP4: GPIO: Fix missing interrupts during device wakeup due to IOPAD.

2011-06-15 Thread Ambresh K
From: Ambresh K ambr...@ti.com If gpio pins from bank[2-5] are marked as wakeup enable and if the wake is through gpio IO pad wakeup, then that wakeup gpio interrupt is lost. In the current implementation, GPIO driver stores the context of DATAIN of all the gpio in the bank. During GPIO resuming

[PATCH] OMAP4: DMTIMER: enable autoidle mode

2011-06-15 Thread Ambresh K
From: Ambresh K ambr...@ti.com Currently autoidle is only enabled for OMAP2/3; by enabling autoidle, the automatic L4 clock gating strategy is applied based on L4 activity, otherwise L4 clock to module will be a free running. Signed-off-by: Ambresh K ambr...@ti.com --- arch/arm/plat-omap

[PATCH] OMAP3: CLK: FIX: warning SR fclk clk domain not found

2010-04-22 Thread Ambresh K
Added clock domain name to sr[1,2] clk structure to fix warning while initialising sr hwmod during boot-up. Signed-off-by: Ambresh K ambr...@ti.com --- arch/arm/mach-omap2/clock3xxx_data.c |2 ++ 1 files changed, 2 insertions(+), 0 deletions(-) diff --git a/arch/arm/mach-omap2