drivers/char/pc110pad.c
PATCH 1/1 drivers/char/pc110pad.c
pc110pad.c doesn't compile if SMP is enabled. This patch is needed:
<<<<<<
drivers/char/pc110pad.c Sun Dec 30 14:02:32 2001 renamed from drivers/char/pc110pad.c Sat Dec 29 19:13:41 2001 +0 -0
@@ -590,7 +590,7 @@ spin_lock_irqsave(&pc110_lock, flags); if (!--active_count) outb(0x30, current_params.io+2); /* switch off
digitiser */
- spin_unlock_irqrestore(&active_lock, flags);
+ spin_unlock_irqrestore(&pc110_lock, flags);
return 0;}
<<<<<<<<<
The bug was introduced between 2.4.17 and 2.5.1: someone added spinlocks instead of cli(), without adding his name to the changelog.
--
ManfredP.S.: Netscape eats space, I've attached is as MIME, too.
drivers/char/pc110pad.c +1 -1
@@ -590,7 +590,7 @@spin_lock_irqsave (&pc110_lock, flags);if (!--active_count)outb (0x30 , current_params.io+2 );/* switch off digitiser */ -spin_unlock_irqrestore (&active_lock, flags); +spin_unlock_irqrestore (&pc110_lock, flags);return 0 ; }
Quoted 2 lines
The bug was introduced between 2.4.17 and 2.5.1: someone added spinlocks instead of
cli(), without adding his name to the changelog.
It was broken as part of the weird cli -> spinlock patch that someone added. I'd fixed it but there is no SMP PC110 so I never tested that case.
Alan
Quoted 4 lines · Alan Cox
Quoted 2 lines
The bug was introduced between 2.4.17 and 2.5.1: someone added spinlocks instead of
cli(), without adding his name to the changelog.It was broken as part of the weird cli -> spinlock patch that someone added. I'd fixed it but there is no SMP PC110 so I never tested that case.
So someone improved SMP scalability on a driver only used on hardware which can't do SMP ?
*boggle*
Dave.
Signature · 2 lines
| Dave Jones. http://www.codemonkey.org.uk | SuSE Labs