[PATCH v7 00/26] gpio/omap: driver cleanup and fixes

Thu Sep 22 11:27:34 EDT 2011

On Tue, Sep 13, 2011 at 6:32 PM, Tarun Kanti DebBarma
<tarun.kanti at ti.com> wrote:
> This series is continuation of cleanup of OMAP GPIO driver and fixes.
> The cleanup include getting rid of cpu_is_* checks wherever possible,
> use of gpio_bank list instead of static array, use of unique platform
> specific value associated data member to OMAP platforms to avoid
> cpu_is_* checks. The series also include PM runtime support.*
>
> Baseline: git://gitorious.org/khilman/linux-omap-pm.git
> Branch: for_3.2/gpio-cleanup
> Commit: 8323374
Thanks Tony for ack'ing the patches.
Kevin,
Since Tony's has acked the patches, can you please pull them?
Thanks.
--
Tarun
>
> Test Details:
> - Compile tested for omap1_defconfig and omap2plus_defconfig.
> - OMAP1710-H3: Bootup test.
> - OMAP2430/SDP, OMAP3430/SDP, OMAP4430/SDP: Functional testing.
> - PM Testing on OMAP3430-SDP: retention, off_mode, system_wide
>  suspend and gpio wakeup.
>
> v7:
> - Use pm_runtime_put() instead of pm_runtime_put_sync_suspend()
>
> - Keep *_runtime_get/put*()  outside spinlock
>
> - Remove additional checking of conditions in _restore_context()
>  From:
>  if (bank->regs->set_dataout && bank->regs->clear_dataout)
>  ...
>  To:
>  if (bank->regs->set_dataout)
>  ...
>
> - Use SET_RUNTIME_PM_OPS and SET_SYSTEM_SLEEP_PM_OPS macros
>
> - In [PATCH 19/25] gpio/omap: cleanup prepare_for_idle and resume_after_idle,
>  protect the bank data elements and register access using spinlock in
>  runtime_suspend/resume() callbacks.
>  This is because these callbacks run with interrupts enabled.
>
> - Add dbclk aliases for all GPIO modules. Without this, GPIO modules were not
>  getting the correct clock handle to enable/disable debounec clock.
>
> - Fix log comments on the following patches:
>  [PATCH 19/25] gpio/omap: cleanup prepare_for_idle and resume_after_idle
>  [PATCH 20/25] gpio/omap: skip operations in runtime callbacks
>  [PATCH 24/25] gpio/omap: restore OE only after setting the output level
>
> v6:
> - Save and restore debounce registers for proper driver operation.
> - Restore interrupt enable after all configuration to avoid spurious interrupts.
> - Restore dataout register before oe register.
> - Restore dataout into dataout_set or dataout based upon the OMAP version.
> - Change register name from wkup_status to wkup_en.
> - Remove wrapper around omap_pm_get_dev_context_loss_count(). Use it directly.
>  Also, changed the signature of get_context_loss_count in pdata and bank structure
>  from int to u32.
>
> - Use 'context' instead of 'ctx' for clarity wherever it is used.
> - Merged two patches into one which are related to bank_is_mpuio() modification.
> - Use shift operator instead of following:
> +       .irqctrl        = OMAP_MPUIO_GPIO_INT_EDGE / 2,
>
> - Remove redundant check from the following
> +       if (bank_is_mpuio(bank)) {
> +               if (bank->regs->wkup_status) <--- redundant check
> +                       mpuio_init(bank);
>
> - Change subject of following patch
>  [PATCH v5 15/22] gpio/omap: use readl in irq_handler for all access
>  into
>  [PATCH 14/25] gpio/omap: remove unnecessary bit-masking for read access
>
> - Fix multi-line comments in
>  [PATCH v5 20/22] gpio/omap: cleanup prepare_for_idle and resume_after_idle
>
> v5:
> - Reduce runtime callback overhead when *_get/put_sync() called from probe()
>  and *_gpio_request/free().
>
> - Dynamic context save within functions where context is modified instead of
>  saving all context within a common function.
>
> - Removed call to mpuio_init() from omap_gpio_mod_init(). Both the functions are
>  called once during initialization in *_gpio_probe().
>  Call to omap_gpio_mod_init() has been removed from omap_gpio_request() on the
>  first access to gpio bank. One time initialization looks sufficient.
>
> - In *_gpio_irq_handler() use *_put_sync_suspend() instead of *_put_sync().
>
> - Removed hardcoding of OMAP16xx sysconfig register value and instead defined an
>  associated constant.
>
> - Removed *_get_sync() call from *_gpio_suspend() and *_put_sync() call from
>  *_gpio_resume(). They got wrongly slipped into the code.
>
> - Removed following redundant zero allocated initialization from mach-omap2/gpio.c
> +       pdata->regs->irqctrl = 0;
> +       pdata->regs->edgectrl1 = 0;
> +       pdata->regs->edgectrl2 = 0;
>
> - Removed following redundant code in gpio-omap.c
>  -#define bank_is_mpuio(bank)  ((bank)->method == METHOD_MPUIO)
>
> v4:
> - since all accesses to registers are 4-byte aligned, removing special
>  checks and handling of 16 and 32-bit wide bank registers and instead
>  use 32-bit read/write access consistently.
>
> - redundant usage of MOD_REG_BIT has been corrected and replaced with
>  _gpio_rmw().
>
> - omap_gpio_mod_init() function has been simplified further using _gpio_rmw().
>
> - sysconfig register offset specific to omap16xx has been removed along
>  with its usage.
>
> - additional logic to skip from suspend/resume:
>
>  if (!bank->regs->wkup_status || !bank->suspend_wakeup)
>        return 0;
>
>  if (!bank->regs->wkup_status || !bank->saved_wakeup)
>        return 0;
>
> - separated mpuio related changes into a different patch from the patch where
>  wakeup status register related changes are done.
>
> - Incorrect replacement of !cpu_class_is_omap2() in gpio_irq_type()
>  corrected:
> +       if (!bank->regs->leveldetect0 &&
> +               (type & (IRQ_TYPE_LEVEL_LOW|IRQ_TYPE_LEVEL_HIGH)))
>                return -EINVAL;
>
> v3:
> - Avoid use of wkup_set and wkup_clear registers. Instead use wkup_status
>  register for all platforms. This is because on OMAP4 it is recommended
>  not to use them.
>
> - Remove duplicate code in omap_gpio_mod_init() for handling the same for
>  32-bit and 16-bit GPIO bank widths. This is accomplished by having two
>  functions to handle each case while assiging a common function pointer
>  during initialization.
>
> - Remove OMAP16xx specific one time initialization from omap_gpio_mod_init().
>  Move it inside omap16xx_gpio_init().
>
> - Avoid usage of USHRT_MAX to indicate undefined values. Use 0 instead.
>
> - In omap_gpio_suspend()/resume() functions remove code that checks
>  if the feature is supported. Instead, assign these functions to
>  struct platform_driver's suspend & resume function pointers for those
>  OMAP platforms whcih support this feature.
>
> - Remove 'suspend_support' flag because it is redundant. Instead use
>  wkup_* registers to decode the same information.
>
> - Restore context also when we don't know if the context is lost.
>
> - Make omap_gpio_save_context() and omap_gpio_restore_context()
>  static.
>
> v2:
> - Do special handling of non-wakeup GPIOs only on OMAP2420. Avoid this
>  handling on OMAP3430.
> - Isolate cleanups and fixes into separate set of patches. Keep the cleanup
>  first followed by the fixes.
> - Avoid calling omap_gpio_get_context_loss() directly and instead call it
>  through function pointer in pdata initialized during init.
> - workaround_enabled flag is not longer needed and is removed.
> - Call pwrdm_post_transition() before calling omap_gpio_resume_after_idle().
> - In omap2_gpio_resume_after_idle() do context restore before handling
>  workaround.
> - Use PM runtime framework.
> - Modify register offset names to : wkup_status, wkup_clear, wkup_set.
>  Also use 'base + offset' for readibility in all relevant places.
> - Remove unwanted messages from commit section like TODO, etc.
>
>
> Charulatha V (10):
>  gpio/omap: remove dependency on gpio_bank_count
>  gpio/omap: use flag to identify wakeup domain
>  gpio/omap: make gpio_context part of gpio_bank structure
>  gpio/omap: fix pwrdm_post_transition call sequence
>  gpio/omap: handle save/restore context in GPIO driver
>  gpio/omap: make non-wakeup GPIO part of pdata
>  gpio/omap: avoid cpu checks during module ena/disable
>  gpio/omap: use pinctrl offset instead of macro
>  gpio/omap: remove bank->method & METHOD_* macros
>  gpio/omap: fix bankwidth for OMAP7xx MPUIO
>
> Nishanth Menon (4):
>  gpio/omap: save and restore debounce registers
>  gpio/omap: enable irq at the end of all configuration in restore
>  gpio/omap: restore OE only after setting the output level
>  gpio/omap: handle set_dataout reg capable IP on restore
>
> Tarun Kanti DebBarma (12):
>  gpio/omap: further cleanup using wkup_en register
>  gpio/omap: use level/edge detect reg offsets
>  gpio/omap: remove hardcoded offsets in context save/restore
>  gpio/omap: cleanup set_gpio_triggering function
>  gpio/omap: cleanup omap_gpio_mod_init function
>  gpio/omap: remove unnecessary bit-masking for read access
>  gpio/omap: use pm-runtime framework
>  gpio/omap: optimize suspend and resume functions
>  gpio/omap: cleanup prepare_for_idle and resume_after_idle
>  gpio/omap: skip operations in runtime callbacks
>  gpio/omap: remove omap_gpio_save_context overhead
>  gpio/omap: add dbclk aliases for all gpio modules
>
>  arch/arm/mach-omap1/gpio15xx.c         |    7 +-
>  arch/arm/mach-omap1/gpio16xx.c         |   47 ++-
>  arch/arm/mach-omap1/gpio7xx.c          |   14 +-
>  arch/arm/mach-omap2/clock3xxx_data.c   |    6 +
>  arch/arm/mach-omap2/clock44xx_data.c   |    6 +
>  arch/arm/mach-omap2/gpio.c             |   36 +-
>  arch/arm/mach-omap2/pm34xx.c           |   21 +-
>  arch/arm/plat-omap/include/plat/gpio.h |   29 +-
>  drivers/gpio/gpio-omap.c               | 1051 +++++++++++++-------------------
>  9 files changed, 524 insertions(+), 693 deletions(-)
>
>