LKML Archive on lore.kernel.org help / color / mirror / Atom feed
* fix typo in geode_configre()@cyrix.c @ 2007-01-09 9:41 takada 2007-01-09 15:43 ` Lennart Sorensen 2007-01-09 17:33 ` Lennart Sorensen 0 siblings, 2 replies; 11+ messages in thread From: takada @ 2007-01-09 9:41 UTC (permalink / raw) To: LKML In kernel 2.6, write back wrong register when configure Geode processor. Instead of storing to CCR4, it stores to CCR3. --- linux-2.6.19/arch/i386/kernel/cpu/cyrix.c.orig 2007-01-09 16:45:21.000000000 +0900 +++ linux-2.6.19/arch/i386/kernel/cpu/cyrix.c 2007-01-09 17:10:13.000000000 +0900 @@ -173,7 +173,7 @@ static void __cpuinit geode_configure(vo ccr4 = getCx86(CX86_CCR4); ccr4 |= 0x38; /* FPU fast, DTE cache, Mem bypass */ - setCx86(CX86_CCR3, ccr3); + setCx86(CX86_CCR4, ccr4); set_cx86_memwb(); set_cx86_reorder(); ^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: fix typo in geode_configre()@cyrix.c 2007-01-09 9:41 fix typo in geode_configre()@cyrix.c takada @ 2007-01-09 15:43 ` Lennart Sorensen 2007-01-17 0:15 ` Juergen Beisert 2007-01-09 17:33 ` Lennart Sorensen 1 sibling, 1 reply; 11+ messages in thread From: Lennart Sorensen @ 2007-01-09 15:43 UTC (permalink / raw) To: takada; +Cc: LKML On Tue, Jan 09, 2007 at 06:41:56PM +0900, takada wrote: > In kernel 2.6, write back wrong register when configure Geode processor. > Instead of storing to CCR4, it stores to CCR3. > > --- linux-2.6.19/arch/i386/kernel/cpu/cyrix.c.orig 2007-01-09 16:45:21.000000000 +0900 > +++ linux-2.6.19/arch/i386/kernel/cpu/cyrix.c 2007-01-09 17:10:13.000000000 +0900 > @@ -173,7 +173,7 @@ static void __cpuinit geode_configure(vo > ccr4 = getCx86(CX86_CCR4); > ccr4 |= 0x38; /* FPU fast, DTE cache, Mem bypass */ > > - setCx86(CX86_CCR3, ccr3); > + setCx86(CX86_CCR4, ccr4); > > set_cx86_memwb(); > set_cx86_reorder(); Any idea what the consequence of this would be? Any chance that while fixing this file anyhow, adding a missing variant could be done? The patch I currently run with is this (To cyrix.c): @@ -257,6 +257,8 @@ break; case 4: /* MediaGX/GXm or Geode GXM/GXLV/GX1 */ + case 11: /* SC1200 seems to return 11 or something but it is a geode gx1 too as far as I know */ + dir0_msn = 4; #ifdef CONFIG_PCI /* It isn't really a PCI quirk directly, but the cure is the same. The MediaGX has deep magic SMM stuff that handles the After the patch I get this: # cat /proc/cpuinfo processor : 0 vendor_id : CyrixInstead cpu family : 5 model : 9 model name : Geode(TM) Integrated Processor by National Semi stepping : 1 cpu MHz : 266.759 cache size : 16 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 2 wp : yes flags : fpu tsc msr cx8 cmov mmx cxmmx bogomips : 535.09 Before that the cpu type was unknown, the cache size wasn't there, and a number of fixups were not being applied as far as I could tell. I know I have had to fix the jsm serial driver which uses memcpy_toio, and I got characters out of order on transmit on the geode, while on a pentium 4 or athlon it works perfectly fine with the same card and driver. My workaround (which of course isn't good for efficiency) has been: - memcpy_toio(&ch->ch_neo_uart->txrxburst, ch->ch_wqueue + tail, s); + for(i=0;i<s;i++) { + memcpy_toio(&(ch->ch_neo_uart->txrxburst[i]), ch->ch_wqueue + tail + i, 1); + } + /*memcpy_toio(&ch->ch_neo_uart->txrxburst, ch->ch_wqueue + tail, s);*/ Could the wrong register being saved have anything to do with that? Without my fix a burst transfer of the data 0123456789ABCDEF results in the other end receiving: 123056749AB8DEFC -- Len Sorensen ^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: fix typo in geode_configre()@cyrix.c 2007-01-09 15:43 ` Lennart Sorensen @ 2007-01-17 0:15 ` Juergen Beisert 0 siblings, 0 replies; 11+ messages in thread From: Juergen Beisert @ 2007-01-17 0:15 UTC (permalink / raw) To: linux-kernel; +Cc: Lennart Sorensen, takada On Tuesday 09 January 2007 16:43, Lennart Sorensen wrote: > On Tue, Jan 09, 2007 at 06:41:56PM +0900, takada wrote: > > In kernel 2.6, write back wrong register when configure Geode processor. > > Instead of storing to CCR4, it stores to CCR3. > > > > --- linux-2.6.19/arch/i386/kernel/cpu/cyrix.c.orig 2007-01-09 > > 16:45:21.000000000 +0900 +++ > > linux-2.6.19/arch/i386/kernel/cpu/cyrix.c 2007-01-09 17:10:13.000000000 > > +0900 @@ -173,7 +173,7 @@ static void __cpuinit geode_configure(vo > > ccr4 = getCx86(CX86_CCR4); > > ccr4 |= 0x38; /* FPU fast, DTE cache, Mem bypass */ > > > > - setCx86(CX86_CCR3, ccr3); > > + setCx86(CX86_CCR4, ccr4); > > > > set_cx86_memwb(); > > set_cx86_reorder(); > > Any idea what the consequence of this would be? Any chance that while > fixing this file anyhow, adding a missing variant could be done? Writing back of ccr4 should be intented here, but also writing back the ccr3 to disable the MAPEN again. So both are required. But the ccr4 first: setCx86(CX86_CCR4, ccr4); setCx86(CX86_CCR3, ccr3); Juergen ^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: fix typo in geode_configre()@cyrix.c 2007-01-09 9:41 fix typo in geode_configre()@cyrix.c takada 2007-01-09 15:43 ` Lennart Sorensen @ 2007-01-09 17:33 ` Lennart Sorensen 2007-01-16 16:38 ` takada ` (2 more replies) 1 sibling, 3 replies; 11+ messages in thread From: Lennart Sorensen @ 2007-01-09 17:33 UTC (permalink / raw) To: takada; +Cc: LKML On Tue, Jan 09, 2007 at 06:41:56PM +0900, takada wrote: > In kernel 2.6, write back wrong register when configure Geode processor. > Instead of storing to CCR4, it stores to CCR3. > > --- linux-2.6.19/arch/i386/kernel/cpu/cyrix.c.orig 2007-01-09 16:45:21.000000000 +0900 > +++ linux-2.6.19/arch/i386/kernel/cpu/cyrix.c 2007-01-09 17:10:13.000000000 +0900 > @@ -173,7 +173,7 @@ static void __cpuinit geode_configure(vo > ccr4 = getCx86(CX86_CCR4); > ccr4 |= 0x38; /* FPU fast, DTE cache, Mem bypass */ > > - setCx86(CX86_CCR3, ccr3); > + setCx86(CX86_CCR4, ccr4); > > set_cx86_memwb(); > set_cx86_reorder(); > - One more comment on this: Why is this function using 3 different styles to do the same thing to 3 different registers? First it does CCR2 by doing: setCx86(CX86_CCR2, getCx86(CX86_CCR2) | 0x88); Nice, no temp values, and it is obvious this is a permanent change. Then for the next one it does: ccr3 = GetCx86(CX86_CCR3); setCx86(CX86_CCR3, (ccr3 & 0x0f) | 0x10); Couldn't that have been: setCx86(CX86_CCR3, (getCx86(CX86_CCR3) & 0x0f) | 0x10); No temp variable, and again it clearly does not intend to restore the value again later (even though the bug actually did cause the value to be restored by accident). Then the last case does: ccr4 = GetCx86(CX86_CCR4); ccr4 |= 0x38; setCx86(CX86_CCR4, ccr4); (after fixing the typo you found of course). Why did this one have to modify the variable first before using it? Maybe the ccr3 and ccr4 cases would have made the line too long and needed wrapping, but at least those two cases should be done the same way. It makes it look like it was written by 3 different people who didn't look at the lines around the changes they were making. -- Len Sorensen ^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: fix typo in geode_configre()@cyrix.c 2007-01-09 17:33 ` Lennart Sorensen @ 2007-01-16 16:38 ` takada 2007-01-16 16:50 ` Lennart Sorensen 2007-01-16 22:47 ` Lennart Sorensen 2007-01-17 0:24 ` Juergen Beisert 2007-01-17 15:26 ` takada 2 siblings, 2 replies; 11+ messages in thread From: takada @ 2007-01-16 16:38 UTC (permalink / raw) To: lsorense; +Cc: linux-kernel From: lsorense@csclub.uwaterloo.ca (Lennart Sorensen) Subject: Re: fix typo in geode_configre()@cyrix.c Date: Tue, 9 Jan 2007 12:33:48 -0500 Thank you for comments. > On Tue, Jan 09, 2007 at 06:41:56PM +0900, takada wrote: > > In kernel 2.6, write back wrong register when configure Geode processor. > > Instead of storing to CCR4, it stores to CCR3. > > > > --- linux-2.6.19/arch/i386/kernel/cpu/cyrix.c.orig 2007-01-09 16:45:21.000000000 +0900 > > +++ linux-2.6.19/arch/i386/kernel/cpu/cyrix.c 2007-01-09 17:10:13.000000000 +0900 > > @@ -173,7 +173,7 @@ static void __cpuinit geode_configure(vo > > ccr4 = getCx86(CX86_CCR4); > > ccr4 |= 0x38; /* FPU fast, DTE cache, Mem bypass */ > > > > - setCx86(CX86_CCR3, ccr3); > > + setCx86(CX86_CCR4, ccr4); > > > > set_cx86_memwb(); > > set_cx86_reorder(); > > - > > One more comment on this: > > Why is this function using 3 different styles to do the same thing to 3 > different registers? > > First it does CCR2 by doing: > setCx86(CX86_CCR2, getCx86(CX86_CCR2) | 0x88); > > Nice, no temp values, and it is obvious this is a permanent change. > > Then for the next one it does: > ccr3 = GetCx86(CX86_CCR3); > setCx86(CX86_CCR3, (ccr3 & 0x0f) | 0x10); > > Couldn't that have been: > setCx86(CX86_CCR3, (getCx86(CX86_CCR3) & 0x0f) | 0x10); > > No temp variable, and again it clearly does not intend to restore the > value again later (even though the bug actually did cause the value to > be restored by accident). > > Then the last case does: > ccr4 = GetCx86(CX86_CCR4); > ccr4 |= 0x38; > setCx86(CX86_CCR4, ccr4); (after fixing the typo you found of course). > > Why did this one have to modify the variable first before using it? > Maybe the ccr3 and ccr4 cases would have made the line too long and > needed wrapping, but at least those two cases should be done the same > way. It makes it look like it was written by 3 different people who > didn't look at the lines around the changes they were making. You are right. I agree to your comment. These variables are needless. I made a patch again. diff -Narup linux-2.6.19.orig/arch/i386/kernel/cpu/cyrix.c linux-2.6.19/arch/i386/kernel/cpu/cyrix.c --- linux-2.6.19.orig/arch/i386/kernel/cpu/cyrix.c 2006-11-30 06:57:37.000000000 +0900 +++ linux-2.6.19/arch/i386/kernel/cpu/cyrix.c 2007-01-16 19:55:05.000000000 +0900 @@ -161,19 +161,15 @@ static void __cpuinit set_cx86_inc(void) static void __cpuinit geode_configure(void) { unsigned long flags; - u8 ccr3, ccr4; local_irq_save(flags); /* Suspend on halt power saving and enable #SUSP pin */ setCx86(CX86_CCR2, getCx86(CX86_CCR2) | 0x88); - ccr3 = getCx86(CX86_CCR3); - setCx86(CX86_CCR3, (ccr3 & 0x0f) | 0x10); /* Enable */ + setCx86(CX86_CCR3, (getCx86(CX86_CCR3) & 0x0f) | 0x10); /* Enable */ - ccr4 = getCx86(CX86_CCR4); - ccr4 |= 0x38; /* FPU fast, DTE cache, Mem bypass */ - - setCx86(CX86_CCR3, ccr3); + /* FPU fast, DTE cache, Mem bypass */ + setCx86(CX86_CCR4, getCx86(CX86_CCR4) | 0x30); set_cx86_memwb(); set_cx86_reorder(); @@ -415,15 +411,14 @@ static void __cpuinit cyrix_identify(str if (dir0 == 5 || dir0 == 3) { - unsigned char ccr3, ccr4; + unsigned char ccr3; unsigned long flags; printk(KERN_INFO "Enabling CPUID on Cyrix processor.\n"); local_irq_save(flags); ccr3 = getCx86(CX86_CCR3); - setCx86(CX86_CCR3, (ccr3 & 0x0f) | 0x10); /* enable MAPEN */ - ccr4 = getCx86(CX86_CCR4); - setCx86(CX86_CCR4, ccr4 | 0x80); /* enable cpuid */ - setCx86(CX86_CCR3, ccr3); /* disable MAPEN */ + setCx86(CX86_CCR3, (ccr3 & 0x0f) | 0x10); /* enable MAPEN */ + setCx86(CX86_CCR4, getCx86(CX86_CCR4) | 0x80); /* enable cpuid */ + setCx86(CX86_CCR3, ccr3); /* disable MAPEN */ local_irq_restore(flags); } } > > -- > Len Sorensen > - > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html > Please read the FAQ at http://www.tux.org/lkml/ > ^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: fix typo in geode_configre()@cyrix.c 2007-01-16 16:38 ` takada @ 2007-01-16 16:50 ` Lennart Sorensen 2007-01-16 17:12 ` takada 2007-01-16 22:47 ` Lennart Sorensen 1 sibling, 1 reply; 11+ messages in thread From: Lennart Sorensen @ 2007-01-16 16:50 UTC (permalink / raw) To: takada; +Cc: linux-kernel On Wed, Jan 17, 2007 at 01:38:35AM +0900, takada wrote: > You are right. I agree to your comment. These variables are needless. > I made a patch again. Of course there are also lots of "magic numbers" around, but I must admit I don't personally really feel like going through the data sheet and naming all of them. > diff -Narup linux-2.6.19.orig/arch/i386/kernel/cpu/cyrix.c linux-2.6.19/arch/i386/kernel/cpu/cyrix.c > --- linux-2.6.19.orig/arch/i386/kernel/cpu/cyrix.c 2006-11-30 06:57:37.000000000 +0900 > +++ linux-2.6.19/arch/i386/kernel/cpu/cyrix.c 2007-01-16 19:55:05.000000000 +0900 > @@ -161,19 +161,15 @@ static void __cpuinit set_cx86_inc(void) > static void __cpuinit geode_configure(void) > { > unsigned long flags; > - u8 ccr3, ccr4; > local_irq_save(flags); > > /* Suspend on halt power saving and enable #SUSP pin */ > setCx86(CX86_CCR2, getCx86(CX86_CCR2) | 0x88); > > - ccr3 = getCx86(CX86_CCR3); > - setCx86(CX86_CCR3, (ccr3 & 0x0f) | 0x10); /* Enable */ > + setCx86(CX86_CCR3, (getCx86(CX86_CCR3) & 0x0f) | 0x10); /* Enable */ > > - ccr4 = getCx86(CX86_CCR4); > - ccr4 |= 0x38; /* FPU fast, DTE cache, Mem bypass */ > - > - setCx86(CX86_CCR3, ccr3); > + /* FPU fast, DTE cache, Mem bypass */ > + setCx86(CX86_CCR4, getCx86(CX86_CCR4) | 0x30); Why did that change from 0x38 to 0x30? > set_cx86_memwb(); > set_cx86_reorder(); > @@ -415,15 +411,14 @@ static void __cpuinit cyrix_identify(str > > if (dir0 == 5 || dir0 == 3) > { > - unsigned char ccr3, ccr4; > + unsigned char ccr3; > unsigned long flags; > printk(KERN_INFO "Enabling CPUID on Cyrix processor.\n"); > local_irq_save(flags); > ccr3 = getCx86(CX86_CCR3); > - setCx86(CX86_CCR3, (ccr3 & 0x0f) | 0x10); /* enable MAPEN */ > - ccr4 = getCx86(CX86_CCR4); > - setCx86(CX86_CCR4, ccr4 | 0x80); /* enable cpuid */ > - setCx86(CX86_CCR3, ccr3); /* disable MAPEN */ > + setCx86(CX86_CCR3, (ccr3 & 0x0f) | 0x10); /* enable MAPEN */ > + setCx86(CX86_CCR4, getCx86(CX86_CCR4) | 0x80); /* enable cpuid */ > + setCx86(CX86_CCR3, ccr3); /* disable MAPEN */ > local_irq_restore(flags); > } > } -- Len Sorensen ^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: fix typo in geode_configre()@cyrix.c 2007-01-16 16:50 ` Lennart Sorensen @ 2007-01-16 17:12 ` takada 2007-01-24 4:04 ` Andrew Morton 0 siblings, 1 reply; 11+ messages in thread From: takada @ 2007-01-16 17:12 UTC (permalink / raw) To: lsorense; +Cc: linux-kernel From: lsorense@csclub.uwaterloo.ca (Lennart Sorensen) Subject: Re: fix typo in geode_configre()@cyrix.c Date: Tue, 16 Jan 2007 11:50:07 -0500 > On Wed, Jan 17, 2007 at 01:38:35AM +0900, takada wrote: > > You are right. I agree to your comment. These variables are needless. > > I made a patch again. > > Of course there are also lots of "magic numbers" around, but I must > admit I don't personally really feel like going through the data sheet > and naming all of them. Me too. > > > diff -Narup linux-2.6.19.orig/arch/i386/kernel/cpu/cyrix.c linux-2.6.19/arch/i386/kernel/cpu/cyrix.c > > --- linux-2.6.19.orig/arch/i386/kernel/cpu/cyrix.c 2006-11-30 06:57:37.000000000 +0900 > > +++ linux-2.6.19/arch/i386/kernel/cpu/cyrix.c 2007-01-16 19:55:05.000000000 +0900 > > @@ -161,19 +161,15 @@ static void __cpuinit set_cx86_inc(void) > > static void __cpuinit geode_configure(void) > > { > > unsigned long flags; > > - u8 ccr3, ccr4; > > local_irq_save(flags); > > > > /* Suspend on halt power saving and enable #SUSP pin */ > > setCx86(CX86_CCR2, getCx86(CX86_CCR2) | 0x88); > > > > - ccr3 = getCx86(CX86_CCR3); > > - setCx86(CX86_CCR3, (ccr3 & 0x0f) | 0x10); /* Enable */ > > + setCx86(CX86_CCR3, (getCx86(CX86_CCR3) & 0x0f) | 0x10); /* Enable */ > > > > - ccr4 = getCx86(CX86_CCR4); > > - ccr4 |= 0x38; /* FPU fast, DTE cache, Mem bypass */ > > - > > - setCx86(CX86_CCR3, ccr3); > > + /* FPU fast, DTE cache, Mem bypass */ > > + setCx86(CX86_CCR4, getCx86(CX86_CCR4) | 0x30); > > Why did that change from 0x38 to 0x30? ah... I made mistake. My eye is tired. Correctry, 0x38. > > > set_cx86_memwb(); > > set_cx86_reorder(); > > @@ -415,15 +411,14 @@ static void __cpuinit cyrix_identify(str > > > > if (dir0 == 5 || dir0 == 3) > > { > > - unsigned char ccr3, ccr4; > > + unsigned char ccr3; > > unsigned long flags; > > printk(KERN_INFO "Enabling CPUID on Cyrix processor.\n"); > > local_irq_save(flags); > > ccr3 = getCx86(CX86_CCR3); > > - setCx86(CX86_CCR3, (ccr3 & 0x0f) | 0x10); /* enable MAPEN */ > > - ccr4 = getCx86(CX86_CCR4); > > - setCx86(CX86_CCR4, ccr4 | 0x80); /* enable cpuid */ > > - setCx86(CX86_CCR3, ccr3); /* disable MAPEN */ > > + setCx86(CX86_CCR3, (ccr3 & 0x0f) | 0x10); /* enable MAPEN */ > > + setCx86(CX86_CCR4, getCx86(CX86_CCR4) | 0x80); /* enable cpuid */ > > + setCx86(CX86_CCR3, ccr3); /* disable MAPEN */ > > local_irq_restore(flags); > > } > > } > > -- > Len Sorensen > ^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: fix typo in geode_configre()@cyrix.c 2007-01-16 17:12 ` takada @ 2007-01-24 4:04 ` Andrew Morton 0 siblings, 0 replies; 11+ messages in thread From: Andrew Morton @ 2007-01-24 4:04 UTC (permalink / raw) To: takada; +Cc: lsorense, linux-kernel On Wed, 17 Jan 2007 02:12:42 +0900 (JST) takada <takada@mbf.nifty.com> wrote: > From: lsorense@csclub.uwaterloo.ca (Lennart Sorensen) > Subject: Re: fix typo in geode_configre()@cyrix.c > Date: Tue, 16 Jan 2007 11:50:07 -0500 > > > On Wed, Jan 17, 2007 at 01:38:35AM +0900, takada wrote: > > > You are right. I agree to your comment. These variables are needless. > > > I made a patch again. > > > > Of course there are also lots of "magic numbers" around, but I must > > admit I don't personally really feel like going through the data sheet > > and naming all of them. > > Me too. > > > > > > diff -Narup linux-2.6.19.orig/arch/i386/kernel/cpu/cyrix.c linux-2.6.19/arch/i386/kernel/cpu/cyrix.c > > > --- linux-2.6.19.orig/arch/i386/kernel/cpu/cyrix.c 2006-11-30 06:57:37.000000000 +0900 > > > +++ linux-2.6.19/arch/i386/kernel/cpu/cyrix.c 2007-01-16 19:55:05.000000000 +0900 > > > @@ -161,19 +161,15 @@ static void __cpuinit set_cx86_inc(void) > > > static void __cpuinit geode_configure(void) > > > { > > > unsigned long flags; > > > - u8 ccr3, ccr4; > > > local_irq_save(flags); > > > > > > /* Suspend on halt power saving and enable #SUSP pin */ > > > setCx86(CX86_CCR2, getCx86(CX86_CCR2) | 0x88); > > > > > > - ccr3 = getCx86(CX86_CCR3); > > > - setCx86(CX86_CCR3, (ccr3 & 0x0f) | 0x10); /* Enable */ > > > + setCx86(CX86_CCR3, (getCx86(CX86_CCR3) & 0x0f) | 0x10); /* Enable */ > > > > > > - ccr4 = getCx86(CX86_CCR4); > > > - ccr4 |= 0x38; /* FPU fast, DTE cache, Mem bypass */ > > > - > > > - setCx86(CX86_CCR3, ccr3); > > > + /* FPU fast, DTE cache, Mem bypass */ > > > + setCx86(CX86_CCR4, getCx86(CX86_CCR4) | 0x30); > > > > Why did that change from 0x38 to 0x30? > > ah... I made mistake. My eye is tired. > Correctry, 0x38. > I'm lost. Could someone please prepare a complete new patch against the latest -linus tree? Thanks. ^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: fix typo in geode_configre()@cyrix.c 2007-01-16 16:38 ` takada 2007-01-16 16:50 ` Lennart Sorensen @ 2007-01-16 22:47 ` Lennart Sorensen 1 sibling, 0 replies; 11+ messages in thread From: Lennart Sorensen @ 2007-01-16 22:47 UTC (permalink / raw) To: takada; +Cc: linux-kernel On Wed, Jan 17, 2007 at 01:38:35AM +0900, takada wrote: > You are right. I agree to your comment. These variables are needless. > I made a patch again. > > diff -Narup linux-2.6.19.orig/arch/i386/kernel/cpu/cyrix.c linux-2.6.19/arch/i386/kernel/cpu/cyrix.c > --- linux-2.6.19.orig/arch/i386/kernel/cpu/cyrix.c 2006-11-30 06:57:37.000000000 +0900 > +++ linux-2.6.19/arch/i386/kernel/cpu/cyrix.c 2007-01-16 19:55:05.000000000 +0900 > @@ -161,19 +161,15 @@ static void __cpuinit set_cx86_inc(void) > static void __cpuinit geode_configure(void) > { > unsigned long flags; > - u8 ccr3, ccr4; > local_irq_save(flags); > > /* Suspend on halt power saving and enable #SUSP pin */ > setCx86(CX86_CCR2, getCx86(CX86_CCR2) | 0x88); > > - ccr3 = getCx86(CX86_CCR3); > - setCx86(CX86_CCR3, (ccr3 & 0x0f) | 0x10); /* Enable */ > + setCx86(CX86_CCR3, (getCx86(CX86_CCR3) & 0x0f) | 0x10); /* Enable */ > > - ccr4 = getCx86(CX86_CCR4); > - ccr4 |= 0x38; /* FPU fast, DTE cache, Mem bypass */ > - > - setCx86(CX86_CCR3, ccr3); > + /* FPU fast, DTE cache, Mem bypass */ > + setCx86(CX86_CCR4, getCx86(CX86_CCR4) | 0x30); Actually is it possible that the original intent was: ccr3 = getCx86(CX86_CCR3); setCx86(CX86_CCR3, (ccr3 & 0x0f) | 0x10); /* Enable */ /* enable advanced register access? */ ccr4 = getCx86(CX86_CCR4); ccr4 |= 0x38; /* FPU fast, DTE cache, Mem bypass */ setCx86(CX86_CCR4, ccr4); setCx86(CX86_CCR3, ccr3); /* restore ccr3 register */ Seems something similar with ccr3 was taking place elsewhere in the function. -- Len Sorensen ^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: fix typo in geode_configre()@cyrix.c 2007-01-09 17:33 ` Lennart Sorensen 2007-01-16 16:38 ` takada @ 2007-01-17 0:24 ` Juergen Beisert 2007-01-17 15:26 ` takada 2 siblings, 0 replies; 11+ messages in thread From: Juergen Beisert @ 2007-01-17 0:24 UTC (permalink / raw) To: linux-kernel; +Cc: Lennart Sorensen, takada On Tuesday 09 January 2007 18:33, Lennart Sorensen wrote: > Then for the next one it does: > ccr3 = GetCx86(CX86_CCR3); > setCx86(CX86_CCR3, (ccr3 & 0x0f) | 0x10); > > Couldn't that have been: > setCx86(CX86_CCR3, (getCx86(CX86_CCR3) & 0x0f) | 0x10); > > No temp variable, and again it clearly does not intend to restore the > value again later (even though the bug actually did cause the value to > be restored by accident). No, ccr3 should be restored to protect some registers (or at least bit 4 should be cleared in ccr3). BTW: In function set_cx86_inc() [...] /* PCR1 -- Performance Control */ /* Incrementor on, whatever that is */ setCx86(CX86_PCR1, getCx86(CX86_PCR1) | 0x02); /* PCR0 -- Performance Control */ /* Incrementor Margin 10 */ setCx86(CX86_PCR0, getCx86(CX86_PCR0) | 0x04); [...] This setting is only valid for 200MHz...266MHz CPUs, for 300MHz and 333MHz CPUs the Incrementor Margin should be 1-1. There is an AppNote about this setting: AMD Geode GX1 Processor Memory Timings for Maximum Performance. Juergen ^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: fix typo in geode_configre()@cyrix.c 2007-01-09 17:33 ` Lennart Sorensen 2007-01-16 16:38 ` takada 2007-01-17 0:24 ` Juergen Beisert @ 2007-01-17 15:26 ` takada 2 siblings, 0 replies; 11+ messages in thread From: takada @ 2007-01-17 15:26 UTC (permalink / raw) To: linux-kernel; +Cc: Lennart Sorensen, takada Hi. Thanks Sorensen and Juergen. I overlooked the restored to CCR3. The bit4(0x10, MAPEN) of CCR3 is necessary to access advanced configuration registers. I'll repost patch. - restore CCR3 - fix use 0x30 instead of 0x38. --- linux-2.6.19.orig/arch/i386/kernel/cpu/cyrix.c 2006-11-30 06:57:37.000000000 +0900 +++ linux-2.6.19/arch/i386/kernel/cpu/cyrix.c 2007-01-17 16:15:40.000000000 +0900 @@ -161,19 +161,18 @@ static void __cpuinit set_cx86_inc(void) static void __cpuinit geode_configure(void) { unsigned long flags; - u8 ccr3, ccr4; + u8 ccr3; local_irq_save(flags); /* Suspend on halt power saving and enable #SUSP pin */ setCx86(CX86_CCR2, getCx86(CX86_CCR2) | 0x88); ccr3 = getCx86(CX86_CCR3); - setCx86(CX86_CCR3, (ccr3 & 0x0f) | 0x10); /* Enable */ + setCx86(CX86_CCR3, (ccr3 & 0x0f) | 0x10); /* enable MAPEN */ - ccr4 = getCx86(CX86_CCR4); - ccr4 |= 0x38; /* FPU fast, DTE cache, Mem bypass */ - - setCx86(CX86_CCR3, ccr3); + /* FPU fast, DTE cache, Mem bypass */ + setCx86(CX86_CCR4, getCx86(CX86_CCR4) | 0x38); + setCx86(CX86_CCR3, ccr3); /* disable MAPEN */ set_cx86_memwb(); set_cx86_reorder(); @@ -415,15 +414,14 @@ static void __cpuinit cyrix_identify(str if (dir0 == 5 || dir0 == 3) { - unsigned char ccr3, ccr4; + unsigned char ccr3; unsigned long flags; printk(KERN_INFO "Enabling CPUID on Cyrix processor.\n"); local_irq_save(flags); ccr3 = getCx86(CX86_CCR3); - setCx86(CX86_CCR3, (ccr3 & 0x0f) | 0x10); /* enable MAPEN */ - ccr4 = getCx86(CX86_CCR4); - setCx86(CX86_CCR4, ccr4 | 0x80); /* enable cpuid */ - setCx86(CX86_CCR3, ccr3); /* disable MAPEN */ + setCx86(CX86_CCR3, (ccr3 & 0x0f) | 0x10); /* enable MAPEN */ + setCx86(CX86_CCR4, getCx86(CX86_CCR4) | 0x80); /* enable cpuid */ + setCx86(CX86_CCR3, ccr3); /* disable MAPEN */ local_irq_restore(flags); } } ^ permalink raw reply [flat|nested] 11+ messages in thread
end of thread, other threads:[~2007-01-24 4:04 UTC | newest] Thread overview: 11+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2007-01-09 9:41 fix typo in geode_configre()@cyrix.c takada 2007-01-09 15:43 ` Lennart Sorensen 2007-01-17 0:15 ` Juergen Beisert 2007-01-09 17:33 ` Lennart Sorensen 2007-01-16 16:38 ` takada 2007-01-16 16:50 ` Lennart Sorensen 2007-01-16 17:12 ` takada 2007-01-24 4:04 ` Andrew Morton 2007-01-16 22:47 ` Lennart Sorensen 2007-01-17 0:24 ` Juergen Beisert 2007-01-17 15:26 ` takada
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).