| Message ID | 1363618756-15851-3-git-send-email-eduardo.valentin@ti.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Mon, Mar 18, 2013 at 10:59:10AM -0400, Eduardo Valentin wrote: > Even if the IRQ is not firing because it is ONE_SHOT and disable > at INTC level, the IRQ handler must use spin_lock_irqsave. > It is necessary to disable IRQs from the current > CPU while it is holding a spin_lock which is need. > Gar... I think I was just totally wrong on this. I think your original code was fine. Sorry Eduardo and Greg. This is a threaded IRQ so the regular spin_lock is fine or even the mutex would have been. IRQ_ONESHOT is about triggering a second IRQ before the first one has been finished, btw. I am an idiot. regards, dan carpenter -- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On 18-03-2013 15:16, Dan Carpenter wrote: > On Mon, Mar 18, 2013 at 10:59:10AM -0400, Eduardo Valentin wrote: >> Even if the IRQ is not firing because it is ONE_SHOT and disable >> at INTC level, the IRQ handler must use spin_lock_irqsave. >> It is necessary to disable IRQs from the current >> CPU while it is holding a spin_lock which is need. >> > > Gar... I think I was just totally wrong on this. I think your > original code was fine. Sorry Eduardo and Greg. > > This is a threaded IRQ so the regular spin_lock is fine or even the > mutex would have been. In fact it is. But I rather prefer to use spinlocks there, just to keep the irq handler sane, even if it is moved to non-threaded IRQ. > > IRQ_ONESHOT is about triggering a second IRQ before the first one > has been finished, btw. It is, and that gets done by masking the IRQ at INTC level. > > I am an idiot. Not really. Thanks for your time reviewing the driver. I will resend this series. Drop this one and split patch 4/8 into two I think (one for moving files, one for renaming functions) > > regards, > dan carpenter > > > -- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Mon, Mar 18, 2013 at 03:38:38PM -0400, Eduardo Valentin wrote: > On 18-03-2013 15:16, Dan Carpenter wrote: > >On Mon, Mar 18, 2013 at 10:59:10AM -0400, Eduardo Valentin wrote: > >>Even if the IRQ is not firing because it is ONE_SHOT and disable > >>at INTC level, the IRQ handler must use spin_lock_irqsave. > >>It is necessary to disable IRQs from the current > >>CPU while it is holding a spin_lock which is need. > >> > > > >Gar... I think I was just totally wrong on this. I think your > >original code was fine. Sorry Eduardo and Greg. > > > >This is a threaded IRQ so the regular spin_lock is fine or even the > >mutex would have been. > > In fact it is. But I rather prefer to use spinlocks there, just to > keep the irq handler sane, even if it is moved to non-threaded IRQ. Yep. I'd agree there. > > > > >IRQ_ONESHOT is about triggering a second IRQ before the first one > >has been finished, btw. > > It is, and that gets done by masking the IRQ at INTC level. > > > > >I am an idiot. > > > Not really. Thanks for your time reviewing the driver. > > I will resend this series. Drop this one and split patch 4/8 into > two I think (one for moving files, one for renaming functions) Great. Much appreciated. regards, dan carpenter -- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/drivers/staging/omap-thermal/omap-bandgap.c b/drivers/staging/omap-thermal/omap-bandgap.c index cb7aa35..a4ac06c 100644 --- a/drivers/staging/omap-thermal/omap-bandgap.c +++ b/drivers/staging/omap-thermal/omap-bandgap.c @@ -168,9 +168,10 @@ static irqreturn_t omap_bandgap_talert_irq_handler(int irq, void *data) struct omap_bandgap *bg_ptr = data; struct temp_sensor_registers *tsr; u32 t_hot = 0, t_cold = 0, ctrl; + unsigned long flags; int i; - spin_lock(&bg_ptr->lock); + spin_lock_irqsave(&bg_ptr->lock, flags); for (i = 0; i < bg_ptr->conf->sensor_count; i++) { tsr = bg_ptr->conf->sensors[i].registers; ctrl = omap_bandgap_readl(bg_ptr, tsr->bgap_status); @@ -209,7 +210,7 @@ static irqreturn_t omap_bandgap_talert_irq_handler(int irq, void *data) if (bg_ptr->conf->report_temperature) bg_ptr->conf->report_temperature(bg_ptr, i); } - spin_unlock(&bg_ptr->lock); + spin_unlock_irqrestore(&bg_ptr->lock, flags); return IRQ_HANDLED; }
Even if the IRQ is not firing because it is ONE_SHOT and disable at INTC level, the IRQ handler must use spin_lock_irqsave. It is necessary to disable IRQs from the current CPU while it is holding a spin_lock which is need. Cc: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Eduardo Valentin <eduardo.valentin@ti.com>