On Wed, 2005-06-01 at 01:32 +0100, Dave Airlie wrote:
> In radeon_irq.c:radeon_wait_irq
>
> There is a comment
> /* This is a hack to work around mysterious freezes on certain
> * systems:
> */
> radeon_acknowledge_irqs(dev_priv);
>
>
> On my 7500 system I've been seeing mysterious hangs after a week or so
> running, and I've removed this "hack" from my system and my hangs seem to
> have gone away, I've talked to Keith about it and he can't remember
> exactly what it fixed but he remembered never producing it on his own
> system, and it is a bad idea IMHO..
>
> I'm mainly sending this mail as I can guarantee about 5 mins after I send
> one of my machines wil hang :-)
It looks quite wrong to "ack" the interrupt that way indeed... It's a
good way to lose interrupts. Besides, it causes the wait on SWI irq to
also "ack" (and thus clear possible pending) vblanks. Bad.
Also, I don't understand why radeon_driver_irq_handler() is ack'ing
interrupts _after_ doing work. That may also cause a loss of irqs. It
should ack them first, and then do work.
I suggest this (untested) patch:
Index: linux-work/drivers/char/drm/radeon_irq.c
===================================================================
--- linux-work.orig/drivers/char/drm/radeon_irq.c 2005-05-02
10:48:09.000000000 +1000
+++ linux-work/drivers/char/drm/radeon_irq.c 2005-06-01 17:20:31.000000000
+1000
@@ -35,6 +35,16 @@
#include "radeon_drm.h"
#include "radeon_drv.h"
+static __inline__ u32 radeon_acknowledge_irqs(drm_radeon_private_t *dev_priv)
+{
+ u32 irqq = RADEON_READ( RADEON_GEN_INT_STATUS )
+ & (RADEON_SW_INT_TEST_ACK | RADEON_CRTC_VBLANK_STAT);
+ if (irqs)
+ RADEON_WRITE( RADEON_GEN_INT_STATUS, irqs );
+ return irqs;
+}
+
+
/* Interrupts - Used for device synchronization and flushing in the
* following circumstances:
*
@@ -63,8 +73,7 @@
/* Only consider the bits we're interested in - others could be used
* outside the DRM
*/
- stat = RADEON_READ(RADEON_GEN_INT_STATUS)
- & (RADEON_SW_INT_TEST | RADEON_CRTC_VBLANK_STAT);
+ stat = radeon_acknowledge_irqs(dev_priv);
if (!stat)
return IRQ_NONE;
@@ -80,19 +89,9 @@
drm_vbl_send_signals( dev );
}
- /* Acknowledge interrupts we handle */
- RADEON_WRITE(RADEON_GEN_INT_STATUS, stat);
return IRQ_HANDLED;
}
-static __inline__ void radeon_acknowledge_irqs(drm_radeon_private_t *dev_priv)
-{
- u32 tmp = RADEON_READ( RADEON_GEN_INT_STATUS )
- & (RADEON_SW_INT_TEST_ACK | RADEON_CRTC_VBLANK_STAT);
- if (tmp)
- RADEON_WRITE( RADEON_GEN_INT_STATUS, tmp );
-}
-
static int radeon_emit_irq(drm_device_t *dev)
{
drm_radeon_private_t *dev_priv = dev->dev_private;
@@ -123,11 +122,6 @@
dev_priv->stats.boxes |= RADEON_BOX_WAIT_IDLE;
- /* This is a hack to work around mysterious freezes on certain
- * systems:
- */
- radeon_acknowledge_irqs( dev_priv );
-
DRM_WAIT_ON( ret, dev_priv->swi_queue, 3 * DRM_HZ,
RADEON_READ( RADEON_LAST_SWI_REG ) >= swi_nr );
-------------------------------------------------------
This SF.Net email is sponsored by Yahoo.
Introducing Yahoo! Search Developer Network - Create apps using Yahoo!
Search APIs Find out how you can build Yahoo! directly into your own
Applications - visit http://developer.yahoo.net/?fr=offad-ysdn-ostg-q22005
--
_______________________________________________
Dri-devel mailing list
Dri-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/dri-devel
