Message ID | 1392339605-20691-12-git-send-email-laurent.pinchart+renesas@ideasonboard.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
On Fri, 14 Feb 2014, Laurent Pinchart wrote: > CMT hardware devices can support multiple channels, with global > registers and per-channel registers. The sh_cmt driver currently models > the hardware with one Linux device per channel. This model makes it > difficult to handle global registers in a clean way. > > Add support for a new model that uses one Linux device per timer with > multiple channels per device. This requires changes to platform data, > add new channel configuration fields. > > Support for the legacy model is kept and will be removed after all > platforms switch to the new model. > > Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com> > --- > drivers/clocksource/sh_cmt.c | 299 +++++++++++++++++++++++++++++++++---------- > include/linux/sh_timer.h | 9 ++ > 2 files changed, 239 insertions(+), 69 deletions(-) > > diff --git a/drivers/clocksource/sh_cmt.c b/drivers/clocksource/sh_cmt.c > index 5280231..8390f0f 100644 > --- a/drivers/clocksource/sh_cmt.c > +++ b/drivers/clocksource/sh_cmt.c > @@ -53,7 +53,16 @@ struct sh_cmt_device; > * channel registers block. All other versions have a shared start/stop register > * located in the global space. > * > - * Note that CMT0 on r8a73a4, r8a7790 and r8a7791, while implementing 32-bit > + * Channels are indexed from 0 to N-1 in the documentation. The channel index > + * infers the start/stop bit position in the control register and the channel > + * registers block address. Some CMT instances have a subset of channels > + * available, in which case the index in the documentation doesn't match the > + * "real" index as implemented in hardware. This is for instance the case with > + * CMT0 on r8a7740, which is a 32-bit variant with a single channel numbered 0 > + * in the documentation but using start/stop bit 5 and having its registers > + * block at 0x60. > + * > + * Similarly CMT0 on r8a73a4, r8a7790 and r8a7791, while implementing 32-bit > * channels only, is a 48-bit gen2 CMT with the 48-bit channels unavailable. > */ > > @@ -85,11 +94,15 @@ struct sh_cmt_info { > > struct sh_cmt_channel { > struct sh_cmt_device *cmt; > - unsigned int index; > > - void __iomem *base; > + unsigned int index; /* Index in the documentation */ > + unsigned int hwidx; /* Real hardware index */ > + > + void __iomem *iostart; > + void __iomem *ioctrl; > struct irqaction irqaction; While you are at it, can you please get rid of that irqaction and use request_irq() ? Thanks, tglx
Hi Thomas, On Saturday 15 February 2014 13:46:54 Thomas Gleixner wrote: > On Fri, 14 Feb 2014, Laurent Pinchart wrote: > > CMT hardware devices can support multiple channels, with global > > registers and per-channel registers. The sh_cmt driver currently models > > the hardware with one Linux device per channel. This model makes it > > difficult to handle global registers in a clean way. > > > > Add support for a new model that uses one Linux device per timer with > > multiple channels per device. This requires changes to platform data, > > add new channel configuration fields. > > > > Support for the legacy model is kept and will be removed after all > > platforms switch to the new model. > > > > Signed-off-by: Laurent Pinchart > > <laurent.pinchart+renesas@ideasonboard.com> > > --- > > > > drivers/clocksource/sh_cmt.c | 299 +++++++++++++++++++++++++++++--------- > > include/linux/sh_timer.h | 9 ++ > > 2 files changed, 239 insertions(+), 69 deletions(-) > > > > diff --git a/drivers/clocksource/sh_cmt.c b/drivers/clocksource/sh_cmt.c > > index 5280231..8390f0f 100644 > > --- a/drivers/clocksource/sh_cmt.c > > +++ b/drivers/clocksource/sh_cmt.c > > @@ -53,7 +53,16 @@ struct sh_cmt_device; > > > > * channel registers block. All other versions have a shared start/stop > > register * located in the global space. > > * > > > > - * Note that CMT0 on r8a73a4, r8a7790 and r8a7791, while implementing > > 32-bit + * Channels are indexed from 0 to N-1 in the documentation. The > > channel index + * infers the start/stop bit position in the control > > register and the channel + * registers block address. Some CMT instances > > have a subset of channels + * available, in which case the index in the > > documentation doesn't match the + * "real" index as implemented in > > hardware. This is for instance the case with + * CMT0 on r8a7740, which > > is a 32-bit variant with a single channel numbered 0 + * in the > > documentation but using start/stop bit 5 and having its registers + * > > block at 0x60. > > + * > > + * Similarly CMT0 on r8a73a4, r8a7790 and r8a7791, while implementing > > 32-bit> > > * channels only, is a 48-bit gen2 CMT with the 48-bit channels > > unavailable. > > */ > > > > @@ -85,11 +94,15 @@ struct sh_cmt_info { > > > > struct sh_cmt_channel { > > > > struct sh_cmt_device *cmt; > > > > - unsigned int index; > > > > - void __iomem *base; > > + unsigned int index; /* Index in the documentation */ > > + unsigned int hwidx; /* Real hardware index */ > > + > > + void __iomem *iostart; > > + void __iomem *ioctrl; > > > > struct irqaction irqaction; > > While you are at it, can you please get rid of that irqaction and use > request_irq() ? The driver claims it can't use request_irq() because the function is not available for early platform devices. If the situation has changed I'd gladly get rid of irqaction.
Hi Laurent, On Mon, Feb 17, 2014 at 3:18 AM, Laurent Pinchart <laurent.pinchart@ideasonboard.com> wrote: > Hi Thomas, > > On Saturday 15 February 2014 13:46:54 Thomas Gleixner wrote: >> On Fri, 14 Feb 2014, Laurent Pinchart wrote: >> > CMT hardware devices can support multiple channels, with global >> > registers and per-channel registers. The sh_cmt driver currently models >> > the hardware with one Linux device per channel. This model makes it >> > difficult to handle global registers in a clean way. >> > >> > Add support for a new model that uses one Linux device per timer with >> > multiple channels per device. This requires changes to platform data, >> > add new channel configuration fields. >> > >> > Support for the legacy model is kept and will be removed after all >> > platforms switch to the new model. >> > >> > Signed-off-by: Laurent Pinchart >> > <laurent.pinchart+renesas@ideasonboard.com> >> > --- >> > >> > drivers/clocksource/sh_cmt.c | 299 +++++++++++++++++++++++++++++--------- >> > include/linux/sh_timer.h | 9 ++ >> > 2 files changed, 239 insertions(+), 69 deletions(-) >> > >> > diff --git a/drivers/clocksource/sh_cmt.c b/drivers/clocksource/sh_cmt.c >> > index 5280231..8390f0f 100644 >> > --- a/drivers/clocksource/sh_cmt.c >> > +++ b/drivers/clocksource/sh_cmt.c >> > @@ -53,7 +53,16 @@ struct sh_cmt_device; >> > >> > * channel registers block. All other versions have a shared start/stop >> > register * located in the global space. >> > * >> > >> > - * Note that CMT0 on r8a73a4, r8a7790 and r8a7791, while implementing >> > 32-bit + * Channels are indexed from 0 to N-1 in the documentation. The >> > channel index + * infers the start/stop bit position in the control >> > register and the channel + * registers block address. Some CMT instances >> > have a subset of channels + * available, in which case the index in the >> > documentation doesn't match the + * "real" index as implemented in >> > hardware. This is for instance the case with + * CMT0 on r8a7740, which >> > is a 32-bit variant with a single channel numbered 0 + * in the >> > documentation but using start/stop bit 5 and having its registers + * >> > block at 0x60. >> > + * >> > + * Similarly CMT0 on r8a73a4, r8a7790 and r8a7791, while implementing >> > 32-bit> >> > * channels only, is a 48-bit gen2 CMT with the 48-bit channels >> > unavailable. >> > */ >> > >> > @@ -85,11 +94,15 @@ struct sh_cmt_info { >> > >> > struct sh_cmt_channel { >> > >> > struct sh_cmt_device *cmt; >> > >> > - unsigned int index; >> > >> > - void __iomem *base; >> > + unsigned int index; /* Index in the documentation */ >> > + unsigned int hwidx; /* Real hardware index */ >> > + >> > + void __iomem *iostart; >> > + void __iomem *ioctrl; >> > >> > struct irqaction irqaction; >> >> While you are at it, can you please get rid of that irqaction and use >> request_irq() ? > > The driver claims it can't use request_irq() because the function is not > available for early platform devices. If the situation has changed I'd gladly > get rid of irqaction. With the risk of stating the obvious, this depends on how early the early platform device stuff is being run. The actual location may not be the same on SH and ARM for instance. On ARM we are doing all we can to initialize these devices as late as ever possible in the MULTIPLAFORM (DT reference) case. Once we manage to remove the legacy ARM case then there is one less user of early platform timers. One perhaps reasonable way forward could be to use request_irq() in case of DT and leave the legacy platform data case to rely on irqaction. Cheers, / magnus
Hi Magnus, On Monday 17 February 2014 10:41:31 Magnus Damm wrote: > On Mon, Feb 17, 2014 at 3:18 AM, Laurent Pinchart wrote: > > On Saturday 15 February 2014 13:46:54 Thomas Gleixner wrote: > >> On Fri, 14 Feb 2014, Laurent Pinchart wrote: > >> > CMT hardware devices can support multiple channels, with global > >> > registers and per-channel registers. The sh_cmt driver currently models > >> > the hardware with one Linux device per channel. This model makes it > >> > difficult to handle global registers in a clean way. > >> > > >> > Add support for a new model that uses one Linux device per timer with > >> > multiple channels per device. This requires changes to platform data, > >> > add new channel configuration fields. > >> > > >> > Support for the legacy model is kept and will be removed after all > >> > platforms switch to the new model. > >> > > >> > Signed-off-by: Laurent Pinchart > >> > <laurent.pinchart+renesas@ideasonboard.com> > >> > --- > >> > > >> > drivers/clocksource/sh_cmt.c | 299 ++++++++++++++++++++++++++--------- > >> > include/linux/sh_timer.h | 9 ++ > >> > 2 files changed, 239 insertions(+), 69 deletions(-) > >> > > >> > diff --git a/drivers/clocksource/sh_cmt.c > >> > b/drivers/clocksource/sh_cmt.c > >> > index 5280231..8390f0f 100644 > >> > --- a/drivers/clocksource/sh_cmt.c > >> > +++ b/drivers/clocksource/sh_cmt.c [snip] > >> > @@ -85,11 +94,15 @@ struct sh_cmt_info { > >> > > >> > struct sh_cmt_channel { > >> > struct sh_cmt_device *cmt; > >> > - unsigned int index; > >> > - void __iomem *base; > >> > + unsigned int index; /* Index in the documentation */ > >> > + unsigned int hwidx; /* Real hardware index */ > >> > + > >> > + void __iomem *iostart; > >> > + void __iomem *ioctrl; > >> > > >> > struct irqaction irqaction; > >> > >> While you are at it, can you please get rid of that irqaction and use > >> request_irq() ? > > > > The driver claims it can't use request_irq() because the function is not > > available for early platform devices. If the situation has changed I'd > > gladly get rid of irqaction. > > With the risk of stating the obvious, this depends on how early the > early platform device stuff is being run. The actual location may not > be the same on SH and ARM for instance. > > On ARM we are doing all we can to initialize these devices as late as > ever possible in the MULTIPLAFORM (DT reference) case. Once we manage > to remove the legacy ARM case then there is one less user of early > platform timers. > > One perhaps reasonable way forward could be to use request_irq() in > case of DT and leave the legacy platform data case to rely on > irqaction. The whole point of switching from setup_irq() to request_irq() is to simplify the code. Adding request_irq() as an option would go in the opposite direction.
Hi Laurent, On Mon, Feb 17, 2014 at 10:48 AM, Laurent Pinchart <laurent.pinchart@ideasonboard.com> wrote: > Hi Magnus, > > On Monday 17 February 2014 10:41:31 Magnus Damm wrote: >> On Mon, Feb 17, 2014 at 3:18 AM, Laurent Pinchart wrote: >> > On Saturday 15 February 2014 13:46:54 Thomas Gleixner wrote: >> >> On Fri, 14 Feb 2014, Laurent Pinchart wrote: >> >> > CMT hardware devices can support multiple channels, with global >> >> > registers and per-channel registers. The sh_cmt driver currently models >> >> > the hardware with one Linux device per channel. This model makes it >> >> > difficult to handle global registers in a clean way. >> >> > >> >> > Add support for a new model that uses one Linux device per timer with >> >> > multiple channels per device. This requires changes to platform data, >> >> > add new channel configuration fields. >> >> > >> >> > Support for the legacy model is kept and will be removed after all >> >> > platforms switch to the new model. >> >> > >> >> > Signed-off-by: Laurent Pinchart >> >> > <laurent.pinchart+renesas@ideasonboard.com> >> >> > --- >> >> > >> >> > drivers/clocksource/sh_cmt.c | 299 ++++++++++++++++++++++++++--------- >> >> > include/linux/sh_timer.h | 9 ++ >> >> > 2 files changed, 239 insertions(+), 69 deletions(-) >> >> > >> >> > diff --git a/drivers/clocksource/sh_cmt.c >> >> > b/drivers/clocksource/sh_cmt.c >> >> > index 5280231..8390f0f 100644 >> >> > --- a/drivers/clocksource/sh_cmt.c >> >> > +++ b/drivers/clocksource/sh_cmt.c > > [snip] > >> >> > @@ -85,11 +94,15 @@ struct sh_cmt_info { >> >> > >> >> > struct sh_cmt_channel { >> >> > struct sh_cmt_device *cmt; >> >> > - unsigned int index; >> >> > - void __iomem *base; >> >> > + unsigned int index; /* Index in the documentation */ >> >> > + unsigned int hwidx; /* Real hardware index */ >> >> > + >> >> > + void __iomem *iostart; >> >> > + void __iomem *ioctrl; >> >> > >> >> > struct irqaction irqaction; >> >> >> >> While you are at it, can you please get rid of that irqaction and use >> >> request_irq() ? >> > >> > The driver claims it can't use request_irq() because the function is not >> > available for early platform devices. If the situation has changed I'd >> > gladly get rid of irqaction. >> >> With the risk of stating the obvious, this depends on how early the >> early platform device stuff is being run. The actual location may not >> be the same on SH and ARM for instance. >> >> On ARM we are doing all we can to initialize these devices as late as >> ever possible in the MULTIPLAFORM (DT reference) case. Once we manage >> to remove the legacy ARM case then there is one less user of early >> platform timers. >> >> One perhaps reasonable way forward could be to use request_irq() in >> case of DT and leave the legacy platform data case to rely on >> irqaction. > > The whole point of switching from setup_irq() to request_irq() is to simplify > the code. Adding request_irq() as an option would go in the opposite > direction. The point of switching to setup_irq() was not very clear to me. I think it is fine to switch to using it over time, and I'm open to any alternative ways forward. Usually moving in incremental steps means more crap in the short term, but if someone could come up with a way that involves little crap then I'm all for that! =) Thanks, / magnus
Hi Magnus, On Monday 17 February 2014 11:07:06 Magnus Damm wrote: > On Mon, Feb 17, 2014 at 10:48 AM, Laurent Pinchart wrote: > > On Monday 17 February 2014 10:41:31 Magnus Damm wrote: > >> On Mon, Feb 17, 2014 at 3:18 AM, Laurent Pinchart wrote: > >> > On Saturday 15 February 2014 13:46:54 Thomas Gleixner wrote: > >> >> On Fri, 14 Feb 2014, Laurent Pinchart wrote: > >> >> > CMT hardware devices can support multiple channels, with global > >> >> > registers and per-channel registers. The sh_cmt driver currently > >> >> > models the hardware with one Linux device per channel. This model > >> >> > makes it difficult to handle global registers in a clean way. > >> >> > > >> >> > Add support for a new model that uses one Linux device per timer > >> >> > with multiple channels per device. This requires changes to platform > >> >> > data, add new channel configuration fields. > >> >> > > >> >> > Support for the legacy model is kept and will be removed after all > >> >> > platforms switch to the new model. > >> >> > > >> >> > Signed-off-by: Laurent Pinchart > >> >> > <laurent.pinchart+renesas@ideasonboard.com> > >> >> > --- > >> >> > > >> >> > drivers/clocksource/sh_cmt.c | 299 ++++++++++++++++++++++++-------- > >> >> > include/linux/sh_timer.h | 9 ++ > >> >> > 2 files changed, 239 insertions(+), 69 deletions(-) > >> >> > > >> >> > diff --git a/drivers/clocksource/sh_cmt.c > >> >> > b/drivers/clocksource/sh_cmt.c > >> >> > index 5280231..8390f0f 100644 > >> >> > --- a/drivers/clocksource/sh_cmt.c > >> >> > +++ b/drivers/clocksource/sh_cmt.c > > > > [snip] > > > >> >> > @@ -85,11 +94,15 @@ struct sh_cmt_info { > >> >> > > >> >> > struct sh_cmt_channel { > >> >> > struct sh_cmt_device *cmt; > >> >> > > >> >> > - unsigned int index; > >> >> > - void __iomem *base; > >> >> > + unsigned int index; /* Index in the documentation */ > >> >> > + unsigned int hwidx; /* Real hardware index */ > >> >> > + > >> >> > + void __iomem *iostart; > >> >> > + void __iomem *ioctrl; > >> >> > > >> >> > struct irqaction irqaction; > >> >> > >> >> While you are at it, can you please get rid of that irqaction and use > >> >> request_irq() ? > >> > > >> > The driver claims it can't use request_irq() because the function is > >> > not available for early platform devices. If the situation has changed > >> > I'd gladly get rid of irqaction. > >> > >> With the risk of stating the obvious, this depends on how early the > >> early platform device stuff is being run. The actual location may not > >> be the same on SH and ARM for instance. > >> > >> On ARM we are doing all we can to initialize these devices as late as > >> ever possible in the MULTIPLAFORM (DT reference) case. Once we manage > >> to remove the legacy ARM case then there is one less user of early > >> platform timers. > >> > >> One perhaps reasonable way forward could be to use request_irq() in > >> case of DT and leave the legacy platform data case to rely on > >> irqaction. > > > > The whole point of switching from setup_irq() to request_irq() is to > > simplify the code. Adding request_irq() as an option would go in the > > opposite direction. > > The point of switching to setup_irq() was not very clear to me. I > think it is fine to switch to using it over time, and I'm open to any > alternative ways forward. Usually moving in incremental steps means > more crap in the short term, but if someone could come up with a way > that involves little crap then I'm all for that! =) I've had a quick look at request_irq(). The function is defined as a static inline in include/linux/interrupt.h: static inline int __must_check request_irq(unsigned int irq, irq_handler_t handler, unsigned long flags, const char *name, void *dev) { return request_threaded_irq(irq, handler, NULL, flags, name, dev); } request_threaded_irq() is implemented in kernel/irq/manage.c. I've removed the code paths that would never be taken when called by the sh_cmt driver, due to - handler not being NULL - thread_fn being NULL - CONFIG_DEBUG_SHIRQ_FIXME not being set - IRQF_SHARED not being set int request_threaded_irq(unsigned int irq, irq_handler_t handler, irq_handler_t thread_fn, unsigned long irqflags, const char *devname, void *dev_id) { struct irqaction *action; struct irq_desc *desc; int retval; desc = irq_to_desc(irq); if (!desc) return -EINVAL; if (!irq_settings_can_request(desc) || WARN_ON(irq_settings_is_per_cpu_devid(desc))) return -EINVAL; action = kzalloc(sizeof(struct irqaction), GFP_KERNEL); if (!action) return -ENOMEM; action->handler = handler; action->thread_fn = thread_fn; action->flags = irqflags; action->name = devname; action->dev_id = dev_id; chip_bus_lock(desc); retval = __setup_irq(irq, desc, action); chip_bus_sync_unlock(desc); if (retval) kfree(action); return retval; } setup_irq() is implemented in the same file as int setup_irq(unsigned int irq, struct irqaction *act) { int retval; struct irq_desc *desc = irq_to_desc(irq); if (WARN_ON(irq_settings_is_per_cpu_devid(desc))) return -EINVAL; chip_bus_lock(desc); retval = __setup_irq(irq, desc, act); chip_bus_sync_unlock(desc); return retval; } The only two differences between request_irq() and setup_irq() would thus be - the extra !irq_settings_can_request(desc) check - the extra kzalloc() call The former should only be true for chained IRQ handlers, which isn't the case here. The latter is just a kmalloc + kmap_atomic + memset as far as I can tell. It thus seems safe to replace setup_irq() with request_irq().
diff --git a/drivers/clocksource/sh_cmt.c b/drivers/clocksource/sh_cmt.c index 5280231..8390f0f 100644 --- a/drivers/clocksource/sh_cmt.c +++ b/drivers/clocksource/sh_cmt.c @@ -53,7 +53,16 @@ struct sh_cmt_device; * channel registers block. All other versions have a shared start/stop register * located in the global space. * - * Note that CMT0 on r8a73a4, r8a7790 and r8a7791, while implementing 32-bit + * Channels are indexed from 0 to N-1 in the documentation. The channel index + * infers the start/stop bit position in the control register and the channel + * registers block address. Some CMT instances have a subset of channels + * available, in which case the index in the documentation doesn't match the + * "real" index as implemented in hardware. This is for instance the case with + * CMT0 on r8a7740, which is a 32-bit variant with a single channel numbered 0 + * in the documentation but using start/stop bit 5 and having its registers + * block at 0x60. + * + * Similarly CMT0 on r8a73a4, r8a7790 and r8a7791, while implementing 32-bit * channels only, is a 48-bit gen2 CMT with the 48-bit channels unavailable. */ @@ -85,11 +94,15 @@ struct sh_cmt_info { struct sh_cmt_channel { struct sh_cmt_device *cmt; - unsigned int index; - void __iomem *base; + unsigned int index; /* Index in the documentation */ + unsigned int hwidx; /* Real hardware index */ + + void __iomem *iostart; + void __iomem *ioctrl; struct irqaction irqaction; + unsigned int timer_bit; unsigned long flags; unsigned long match_value; unsigned long next_match_value; @@ -106,6 +119,7 @@ struct sh_cmt_device { struct platform_device *pdev; const struct sh_cmt_info *info; + bool legacy; void __iomem *mapbase_ch; void __iomem *mapbase; @@ -113,6 +127,10 @@ struct sh_cmt_device { struct sh_cmt_channel *channels; unsigned int num_channels; + unsigned int hw_channels; + + bool has_clockevent; + bool has_clocksource; }; #define SH_CMT16_CMCSR_CMF (1 << 7) @@ -224,41 +242,47 @@ static const struct sh_cmt_info sh_cmt_info[] = { static inline unsigned long sh_cmt_read_cmstr(struct sh_cmt_channel *ch) { - return ch->cmt->info->read_control(ch->cmt->mapbase, 0); + if (ch->iostart) + return ch->cmt->info->read_control(ch->iostart, 0); + else + return ch->cmt->info->read_control(ch->cmt->mapbase, 0); } -static inline unsigned long sh_cmt_read_cmcsr(struct sh_cmt_channel *ch) +static inline void sh_cmt_write_cmstr(struct sh_cmt_channel *ch, + unsigned long value) { - return ch->cmt->info->read_control(ch->base, CMCSR); + if (ch->iostart) + ch->cmt->info->write_control(ch->iostart, 0, value); + else + ch->cmt->info->write_control(ch->cmt->mapbase, 0, value); } -static inline unsigned long sh_cmt_read_cmcnt(struct sh_cmt_channel *ch) +static inline unsigned long sh_cmt_read_cmcsr(struct sh_cmt_channel *ch) { - return ch->cmt->info->read_count(ch->base, CMCNT); + return ch->cmt->info->read_control(ch->ioctrl, CMCSR); } -static inline void sh_cmt_write_cmstr(struct sh_cmt_channel *ch, +static inline void sh_cmt_write_cmcsr(struct sh_cmt_channel *ch, unsigned long value) { - ch->cmt->info->write_control(ch->cmt->mapbase, 0, value); + ch->cmt->info->write_control(ch->ioctrl, CMCSR, value); } -static inline void sh_cmt_write_cmcsr(struct sh_cmt_channel *ch, - unsigned long value) +static inline unsigned long sh_cmt_read_cmcnt(struct sh_cmt_channel *ch) { - ch->cmt->info->write_control(ch->base, CMCSR, value); + return ch->cmt->info->read_count(ch->ioctrl, CMCNT); } static inline void sh_cmt_write_cmcnt(struct sh_cmt_channel *ch, unsigned long value) { - ch->cmt->info->write_count(ch->base, CMCNT, value); + ch->cmt->info->write_count(ch->ioctrl, CMCNT, value); } static inline void sh_cmt_write_cmcor(struct sh_cmt_channel *ch, unsigned long value) { - ch->cmt->info->write_count(ch->base, CMCOR, value); + ch->cmt->info->write_count(ch->ioctrl, CMCOR, value); } static unsigned long sh_cmt_get_counter(struct sh_cmt_channel *ch, @@ -287,7 +311,6 @@ static DEFINE_RAW_SPINLOCK(sh_cmt_lock); static void sh_cmt_start_stop_ch(struct sh_cmt_channel *ch, int start) { - struct sh_timer_config *cfg = ch->cmt->pdev->dev.platform_data; unsigned long flags, value; /* start stop register shared by multiple timer channels */ @@ -295,9 +318,9 @@ static void sh_cmt_start_stop_ch(struct sh_cmt_channel *ch, int start) value = sh_cmt_read_cmstr(ch); if (start) - value |= 1 << cfg->timer_bit; + value |= 1 << ch->timer_bit; else - value &= ~(1 << cfg->timer_bit); + value &= ~(1 << ch->timer_bit); sh_cmt_write_cmstr(ch, value); raw_spin_unlock_irqrestore(&sh_cmt_lock, flags); @@ -792,28 +815,71 @@ static int sh_cmt_register(struct sh_cmt_channel *ch, char *name, unsigned long clockevent_rating, unsigned long clocksource_rating) { - if (clockevent_rating) + if (clockevent_rating) { + ch->cmt->has_clockevent = true; sh_cmt_register_clockevent(ch, name, clockevent_rating); + } - if (clocksource_rating) + if (clocksource_rating) { + ch->cmt->has_clocksource = true; sh_cmt_register_clocksource(ch, name, clocksource_rating); + } return 0; } static int sh_cmt_setup_channel(struct sh_cmt_channel *ch, unsigned int index, - struct sh_cmt_device *cmt) + const struct sh_timer_channel_config *ch_cfg, + unsigned int hwidx, struct sh_cmt_device *cmt) { - struct sh_timer_config *cfg = cmt->pdev->dev.platform_data; + unsigned long clockevent_rating; + unsigned long clocksource_rating; int irq; int ret; ch->cmt = cmt; - ch->base = cmt->mapbase_ch; - ch->index = index; + ch->index = ch_cfg ? ch_cfg->index : index; + ch->hwidx = hwidx; + + /* + * Compute the address of the channel control register block. For the + * timers with a per-channel start/stop register, compute its address + * as well. + * + * For legacy configuration the address has been mapped explicitly. + */ + if (cmt->legacy) { + ch->ioctrl = cmt->mapbase_ch; + } else { + switch (cmt->info->model) { + case SH_CMT_16BIT: + ch->ioctrl = cmt->mapbase + 2 + ch->hwidx * 6; + break; + case SH_CMT_32BIT: + case SH_CMT_48BIT: + ch->ioctrl = cmt->mapbase + 0x10 + ch->hwidx * 0x10; + break; + case SH_CMT_32BIT_FAST: + /* + * The 32-bit "fast" timer has a single channel at hwidx + * 5 but is located at offset 0x40 instead of 0x60 for + * some reason. + */ + ch->ioctrl = cmt->mapbase + 0x40; + break; + case SH_CMT_48BIT_GEN2: + ch->iostart = cmt->mapbase + ch->hwidx * 0x100; + ch->ioctrl = ch->iostart + 0x10; + break; + } + } /* Request irq using setup_irq() (too early for request_irq()). */ - irq = platform_get_irq(cmt->pdev, 0); + if (cmt->legacy) + irq = platform_get_irq(cmt->pdev, 0); + else + irq = platform_get_irq(cmt->pdev, ch->index); + if (irq < 0) { dev_err(&cmt->pdev->dev, "failed to get irq\n"); return irq; @@ -832,9 +898,20 @@ static int sh_cmt_setup_channel(struct sh_cmt_channel *ch, unsigned int index, ch->match_value = ch->max_match_value; raw_spin_lock_init(&ch->lock); + if (ch_cfg) { + ch->timer_bit = cmt->info->model == SH_CMT_48BIT_GEN2 + ? 0 : ch->hwidx; + clockevent_rating = ch_cfg->clockevent_rating; + clocksource_rating = ch_cfg->clocksource_rating; + } else { + struct sh_timer_config *cfg = cmt->pdev->dev.platform_data; + ch->timer_bit = cfg->timer_bit; + clockevent_rating = cfg->clockevent_rating; + clocksource_rating = cfg->clocksource_rating; + } + ret = sh_cmt_register(ch, (char *)dev_name(&cmt->pdev->dev), - cfg->clockevent_rating, - cfg->clocksource_rating); + clockevent_rating, clocksource_rating); if (ret) { dev_err(&cmt->pdev->dev, "registration failed\n"); return ret; @@ -850,97 +927,169 @@ static int sh_cmt_setup_channel(struct sh_cmt_channel *ch, unsigned int index, return 0; } -static int sh_cmt_setup(struct sh_cmt_device *cmt, struct platform_device *pdev) +static int sh_cmt_map_memory(struct sh_cmt_device *cmt) { - struct sh_timer_config *cfg = pdev->dev.platform_data; - struct resource *res, *res2; - int ret; - ret = -ENXIO; + struct resource *mem; - cmt->pdev = pdev; + mem = platform_get_resource(cmt->pdev, IORESOURCE_MEM, 0); + if (!mem) { + dev_err(&cmt->pdev->dev, "failed to get I/O memory\n"); + return -ENXIO; + } - if (!cfg) { - dev_err(&cmt->pdev->dev, "missing platform data\n"); - goto err0; + cmt->mapbase = ioremap_nocache(mem->start, resource_size(mem)); + if (cmt->mapbase == NULL) { + dev_err(&cmt->pdev->dev, "failed to remap I/O memory\n"); + return -ENXIO; } + return 0; +} + +static int sh_cmt_map_memory_legacy(struct sh_cmt_device *cmt) +{ + struct sh_timer_config *cfg = cmt->pdev->dev.platform_data; + struct resource *res, *res2; + + /* map memory, let mapbase_ch point to our channel */ res = platform_get_resource(cmt->pdev, IORESOURCE_MEM, 0); if (!res) { dev_err(&cmt->pdev->dev, "failed to get I/O memory\n"); - goto err0; + return -ENXIO; } - /* optional resource for the shared timer start/stop register */ - res2 = platform_get_resource(cmt->pdev, IORESOURCE_MEM, 1); - - /* map memory, let mapbase_ch point to our channel */ cmt->mapbase_ch = ioremap_nocache(res->start, resource_size(res)); if (cmt->mapbase_ch == NULL) { dev_err(&cmt->pdev->dev, "failed to remap I/O memory\n"); - goto err0; + return -ENXIO; } + /* optional resource for the shared timer start/stop register */ + res2 = platform_get_resource(cmt->pdev, IORESOURCE_MEM, 1); + /* map second resource for CMSTR */ cmt->mapbase = ioremap_nocache(res2 ? res2->start : res->start - cfg->channel_offset, res2 ? resource_size(res2) : 2); if (cmt->mapbase == NULL) { dev_err(&cmt->pdev->dev, "failed to remap I/O second memory\n"); - goto err1; + iounmap(cmt->mapbase_ch); + return -ENXIO; } - /* get hold of clock */ + /* identify the model based on the resources */ + if (resource_size(res) == 6) + cmt->info = &sh_cmt_info[SH_CMT_16BIT]; + else if (res2 && (resource_size(res2) == 4)) + cmt->info = &sh_cmt_info[SH_CMT_48BIT_GEN2]; + else + cmt->info = &sh_cmt_info[SH_CMT_32BIT]; + + return 0; +} + +static void sh_cmt_unmap_memory(struct sh_cmt_device *cmt) +{ + iounmap(cmt->mapbase); + if (cmt->mapbase_ch) + iounmap(cmt->mapbase_ch); +} + +static int sh_cmt_setup(struct sh_cmt_device *cmt, struct platform_device *pdev) +{ + struct sh_timer_config *cfg = pdev->dev.platform_data; + const struct platform_device_id *id = pdev->id_entry; + int ret; + + memset(cmt, 0, sizeof(*cmt)); + cmt->pdev = pdev; + + if (!cfg) { + dev_err(&cmt->pdev->dev, "missing platform data\n"); + return -ENXIO; + } + + cmt->info = (const struct sh_cmt_info *)id->driver_data; + cmt->legacy = cmt->info ? false : true; + + /* Get hold of clock. */ cmt->clk = clk_get(&cmt->pdev->dev, "cmt_fck"); if (IS_ERR(cmt->clk)) { dev_err(&cmt->pdev->dev, "cannot get clock\n"); - ret = PTR_ERR(cmt->clk); - goto err2; + return PTR_ERR(cmt->clk); } ret = clk_prepare(cmt->clk); if (ret < 0) - goto err3; + goto err_clk_put; - /* identify the model based on the resources */ - if (resource_size(res) == 6) - cmt->info = &sh_cmt_info[SH_CMT_16BIT]; - else if (res2 && (resource_size(res2) == 4)) - cmt->info = &sh_cmt_info[SH_CMT_48BIT_GEN2]; + /* + * Map the memory resource(s). We need to support both the legacy + * platform device configuration (with one device per channel) and the + * new version (with multiple channels per device). + */ + if (cmt->legacy) + ret = sh_cmt_map_memory_legacy(cmt); else - cmt->info = &sh_cmt_info[SH_CMT_32BIT]; + ret = sh_cmt_map_memory(cmt); - cmt->channels = kzalloc(sizeof(*cmt->channels), GFP_KERNEL); + if (ret < 0) + goto err_clk_unprepare; + + /* Allocate and setup the channels. */ + if (cmt->legacy) { + cmt->num_channels = 1; + } else { + cmt->num_channels = cfg->num_channels; + cmt->hw_channels = ~cfg->channels_mask; + } + + cmt->channels = kzalloc(cmt->num_channels * sizeof(*cmt->channels), + GFP_KERNEL); if (cmt->channels == NULL) { ret = -ENOMEM; - goto err4; + goto err_unmap; } - cmt->num_channels = 1; + if (cmt->legacy) { + ret = sh_cmt_setup_channel(&cmt->channels[0], cfg->timer_bit, + NULL, cfg->timer_bit, cmt); + if (ret < 0) + goto err_unmap; + } else { + const struct sh_timer_channel_config *ch_cfg = cfg->channels; + unsigned int mask = cmt->hw_channels; + unsigned int i; + + for (i = 0; i < cmt->num_channels; ++i) { + unsigned int hwidx = ffs(mask) - 1; - ret = sh_cmt_setup_channel(&cmt->channels[0], cfg->timer_bit, cmt); - if (ret < 0) - goto err4; + ret = sh_cmt_setup_channel(&cmt->channels[i], 0, ch_cfg, + hwidx, cmt); + if (ret < 0) + goto err_unmap; + + mask &= ~(1 << hwidx); + } + } platform_set_drvdata(pdev, cmt); return 0; -err4: + +err_unmap: kfree(cmt->channels); + sh_cmt_unmap_memory(cmt); +err_clk_unprepare: clk_unprepare(cmt->clk); -err3: +err_clk_put: clk_put(cmt->clk); -err2: - iounmap(cmt->mapbase); -err1: - iounmap(cmt->mapbase_ch); -err0: return ret; } static int sh_cmt_probe(struct platform_device *pdev) { struct sh_cmt_device *cmt = platform_get_drvdata(pdev); - struct sh_timer_config *cfg = pdev->dev.platform_data; int ret; if (!is_early_platform_device(pdev)) { @@ -969,7 +1118,7 @@ static int sh_cmt_probe(struct platform_device *pdev) return 0; out: - if (cfg->clockevent_rating || cfg->clocksource_rating) + if (cmt->has_clockevent || cmt->has_clocksource) pm_runtime_irq_safe(&pdev->dev); else pm_runtime_idle(&pdev->dev); @@ -982,12 +1131,24 @@ static int sh_cmt_remove(struct platform_device *pdev) return -EBUSY; /* cannot unregister clockevent and clocksource */ } +static const struct platform_device_id sh_cmt_id_table[] = { + { "sh_cmt", 0 }, + { "sh-cmt-16", (kernel_ulong_t)&sh_cmt_info[SH_CMT_16BIT] }, + { "sh-cmt-32", (kernel_ulong_t)&sh_cmt_info[SH_CMT_32BIT] }, + { "sh-cmt-32-fast", (kernel_ulong_t)&sh_cmt_info[SH_CMT_32BIT_FAST] }, + { "sh-cmt-48", (kernel_ulong_t)&sh_cmt_info[SH_CMT_48BIT] }, + { "sh-cmt-48-gen2", (kernel_ulong_t)&sh_cmt_info[SH_CMT_48BIT_GEN2] }, + { } +}; +MODULE_DEVICE_TABLE(platform, sh_cmt_id_table); + static struct platform_driver sh_cmt_device_driver = { .probe = sh_cmt_probe, .remove = sh_cmt_remove, .driver = { .name = "sh_cmt", - } + }, + .id_table = sh_cmt_id_table, }; static int __init sh_cmt_init(void) diff --git a/include/linux/sh_timer.h b/include/linux/sh_timer.h index 4d9dcd1..b8273e2 100644 --- a/include/linux/sh_timer.h +++ b/include/linux/sh_timer.h @@ -1,12 +1,21 @@ #ifndef __SH_TIMER_H__ #define __SH_TIMER_H__ +struct sh_timer_channel_config { + unsigned int index; + unsigned long clockevent_rating; + unsigned long clocksource_rating; +}; + struct sh_timer_config { char *name; long channel_offset; int timer_bit; unsigned long clockevent_rating; unsigned long clocksource_rating; + const struct sh_timer_channel_config *channels; + unsigned int num_channels; + unsigned int channels_mask; }; #endif /* __SH_TIMER_H__ */
CMT hardware devices can support multiple channels, with global registers and per-channel registers. The sh_cmt driver currently models the hardware with one Linux device per channel. This model makes it difficult to handle global registers in a clean way. Add support for a new model that uses one Linux device per timer with multiple channels per device. This requires changes to platform data, add new channel configuration fields. Support for the legacy model is kept and will be removed after all platforms switch to the new model. Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com> --- drivers/clocksource/sh_cmt.c | 299 +++++++++++++++++++++++++++++++++---------- include/linux/sh_timer.h | 9 ++ 2 files changed, 239 insertions(+), 69 deletions(-)