Skip to contentSkip to conversation

i386,2.6 cyrix.c cann't found companion chip

DISCUSSION TAKADA 3 participants 5 messages Jan 7, 2007 – Jan 9, 2007

lore
Share & tools
lore

TAKADA

author
#0 lore raw

Hi. I use MediaGX with kernel 2.6.19.
cirix.c try to find companion chip (CS5510 and CS5520) with pci_devPresent().
However, cyrix.c cannot find a companion chip because a list of pci_devices is not yet initialized when __cpuinit is called. Therefore, Search functions such as the 2.4 kernel which pci_devices list is needless is necessary.

How will it be good?

Signature · 1 line
TAKADA <[email protected]>
#1 lore raw

Hi Takada-san,

It is obviously bad.
These part is added several years ago by my post.
A cyrix.c try to find chip because of chip hardware bug affected to timer which has started early.

Now, these chips have already been obsolete.
There are 2 options. One is simply remove these functionality. The other is to move it to compile time ifdef that is off by default.

For user who use in embbeded environment,
I wanna change it to ifdef.

Thank you for report!

Hiroshi

Quoted 10 lines · TAKADA

Hi. I use MediaGX with kernel 2.6.19.
cirix.c try to find companion chip (CS5510 and CS5520) with pci_devPresent().
However, cyrix.c cannot find a companion chip because a list of pci_devices is not yet initialized when __cpuinit is called. Therefore, Search functions such as the 2.4 kernel which pci_devices list is needless is necessary.

How will it be good?

#2 lore raw
Quoted 3 lines

Now, these chips have already been obsolete.
There are 2 options. One is simply remove these functionality. The other is to move it to compile time ifdef that is off by default.

Fortunately we have pci functions for early pci accesses. I will take a look at this as I still have a CS5520 board.

#3 lore raw
In reply to Hiroshi Miura
Quoted 11 lines · On Mon, 08 Jan 2007 01:00:00 +0900 Hiroshi Miura <[email protected]> wrote:

Hi Takada-san,

It is obviously bad.
These part is added several years ago by my post.
A cyrix.c try to find chip because of chip hardware bug affected to timer which has started early.

Now, these chips have already been obsolete.
There are 2 options. One is simply remove these functionality. The other is to move it to compile time ifdef that is off by default.

Quoted 6 lines

For user who use in embbeded environment,
I wanna change it to ifdef.

Thank you for report!

Hiroshi

What do the you think of this as a solution ?

arch/i386/kernel/cpu/cyrix.c +12 -8
@@ -196,6 +196,7 @@
 	unsigned char dir0, dir0_msn, dir0_lsn, dir1 = 0;
 	char *buf = c->x86_model_id;
 	const char *p = NULL;
+	u16 vendor, device;
 
 	/* Bit 31 in normal CPUID used for nonstandard 3DNow ID;
 	   3DNow is IDd by bit 31 in extended CPUID (1*32+31) anyway */
@@ -257,7 +258,6 @@
 		break;
 
 	case 4: /* MediaGX/GXm or Geode GXM/GXLV/GX1 */
-#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
 		   SB emulation. It thows away the fifo on disable_dma() which
@@ -265,22 +265,26 @@
 
 		   Bug2: VSA1 has a wrap bug so that using maximum sized DMA 
 		   causes bad things. According to NatSemi VSA2 has another
-		   bug to do with 'hlt'. I've not seen any boards using VSA2
-		   and X doesn't seem to support it either so who cares 8).
-		   VSA1 we work around however.
+		   bug to do with 'hlt'.
 		*/
 
 		printk(KERN_INFO "Working around Cyrix MediaGX virtual DMA bugs.\n");
 		isa_dma_bridge_buggy = 2;
-
+		
+		/* We do this before the PCI layer is running. However we 
+		   are safe here as we know the bridge must be a Cyrix 
+		   companion and must be present */
+		   
+		pci_conf1_read(0, 0, PCI_DEVFN(0x12, 0), 0, 4, &vendor);
+		pci_conf1_read(0, 0, PCI_DEVFN(0x12, 0), 4, 4, &device);
 
 		/*
 		 *  The 5510/5520 companion chips have a funky PIT.
 		 */  
-		if (pci_dev_present(cyrix_55x0))
+		if (vendor == PCI_VENDOR_ID_CYRIX && (device == PCI_DEVICE_ID_CYRIX_5510 || device == PCI_DEVICE_ID_CYRIX_5520))
 			pit_latch_buggy = 1;
-#endif
-		c->x86_cache_size=16;	/* Yep 16K integrated cache thats it */
+
+		c->x86_cache_size=16;	/* Yep 16K integrated cache thats it (12K in X) */
 
 		/* GXm supports extended cpuid levels 'ala' AMD */
 		if (c->cpuid_level == 2) {
author
#4 lore raw

Thanks for your reply.
Your patch dosen't work on 2.6.19.1.
pci_conf1_read() requires u32 pointer for a value of a register. I replaced u16 to u32.
The vendor-ID and dvice ID that pci_conf1_read() got were:
vender: 0x00021078
device: 0x0280000f.

My result of lcpci -vn is follow:

00:12.0 0601: 1078:0002

	Flags: bus master, medium devsel, latency 64
	I/O ports at 3000 [size=128]
	I/O ports at 4000 [size=32]
	I/O ports at 5000 [size=16]
	Memory at 40010000 (32-bit, non-prefetchable) [size=4K]

The low word in vender-id that pci_conf1_reqd() store to u32 is correct. But, device-id is wrong. I attach modified patch. It just work on my machine.
BTW, This patch is assuming that the PCI slot number is 0x12, Is it by Cyrix's spec?

arch/i386/kernel/cpu/cyrix.c renamed from arch/i386/kernel/cpu/cyrix.c.orig +8 -9
@@ -8,6 +8,7 @@
 #include <asm/timer.h>
 
 #include "cpu.h"
+#include "../../pci/pci.h"
 
 /*
  * Read NSC/Cyrix DEVID registers (DIR) to get more detailed info. about the CPU
@@ -183,19 +184,12 @@ static void __cpuinit geode_configure(vo
 }
 
 
-#ifdef CONFIG_PCI
-static struct pci_device_id __cpuinitdata cyrix_55x0[] = {
-	{ PCI_DEVICE(PCI_VENDOR_ID_CYRIX, PCI_DEVICE_ID_CYRIX_5510) },
-	{ PCI_DEVICE(PCI_VENDOR_ID_CYRIX, PCI_DEVICE_ID_CYRIX_5520) },
-	{ },
-};
-#endif
-
 static void __cpuinit init_cyrix(struct cpuinfo_x86 *c)
 {
 	unsigned char dir0, dir0_msn, dir0_lsn, dir1 = 0;
 	char *buf = c->x86_model_id;
 	const char *p = NULL;
+	u32 vendor, device;
 
 	/* Bit 31 in normal CPUID used for nonstandard 3DNow ID;
 	   3DNow is IDd by bit 31 in extended CPUID (1*32+31) anyway */
@@ -273,11 +267,16 @@ static void __cpuinit init_cyrix(struct 
 		printk(KERN_INFO "Working around Cyrix MediaGX virtual DMA bugs.\n");
 		isa_dma_bridge_buggy = 2;
 
+		/* We do this before the PCI layer is running. However we 
+		   are safe here as we know the bridge must be a Cyrix 
+		   companion and must be present */
+		pci_conf1_read(0, 0, PCI_DEVFN(0x12, 0), PCI_VENDOR_ID, 2, &vendor);
+		pci_conf1_read(0, 0, PCI_DEVFN(0x12, 0), PCI_DEVICE_ID, 2, &device);
 
 		/*
 		 *  The 5510/5520 companion chips have a funky PIT.
 		 */  
-		if (pci_dev_present(cyrix_55x0))
+		if (vendor == PCI_VENDOR_ID_CYRIX && (device == PCI_DEVICE_ID_CYRIX_5510 || device == PCI_DEVICE_ID_CYRIX_5520))
 			pit_latch_buggy = 1;
 #endif
 		c->x86_cache_size=16;	/* Yep 16K integrated cache thats it */

From: Alan <[email protected]>
Subject: Re: i386,2.6 cyrix.c cann't found companion chip
Date: Mon, 8 Jan 2007 17:06:54 +0000

Quoted 86 lines
Quoted 11 lines · On Mon, 08 Jan 2007 01:00:00 +0900 Hiroshi Miura <[email protected]> wrote:

Hi Takada-san,

It is obviously bad.
These part is added several years ago by my post.
A cyrix.c try to find chip because of chip hardware bug affected to timer which has started early.

Now, these chips have already been obsolete.
There are 2 options. One is simply remove these functionality. The other is to move it to compile time ifdef that is off by default.

Quoted 6 lines

For user who use in embbeded environment,
I wanna change it to ifdef.

Thank you for report!

Hiroshi

What do the you think of this as a solution ?

arch/i386/kernel/cpu/cyrix.c +12 -9
@@ -196,6 +196,7 @@
 	unsigned char dir0, dir0_msn, dir0_lsn, dir1 = 0;
 	char *buf = c->x86_model_id;
 	const char *p = NULL;
+	u16 vendor, device;
 
 	/* Bit 31 in normal CPUID used for nonstandard 3DNow ID;
 	   3DNow is IDd by bit 31 in extended CPUID (1*32+31) anyway */
@@ -257,7 +258,6 @@
 		break;
 
 	case 4: /* MediaGX/GXm or Geode GXM/GXLV/GX1 */
-#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
 		   SB emulation. It thows away the fifo on disable_dma() which
@@ -265,22 +265,26 @@
 
 		   Bug2: VSA1 has a wrap bug so that using maximum sized DMA 
 		   causes bad things. According to NatSemi VSA2 has another
-		   bug to do with 'hlt'. I've not seen any boards using VSA2
-		   and X doesn't seem to support it either so who cares 8).
-		   VSA1 we work around however.
+		   bug to do with 'hlt'.
 		*/
 
 		printk(KERN_INFO "Working around Cyrix MediaGX virtual DMA bugs.\n");
 		isa_dma_bridge_buggy = 2;
-
+		
+		/* We do this before the PCI layer is running. However we 
+		   are safe here as we know the bridge must be a Cyrix 
+		   companion and must be present */
+		   
+		pci_conf1_read(0, 0, PCI_DEVFN(0x12, 0), 0, 4, &vendor);
+		pci_conf1_read(0, 0, PCI_DEVFN(0x12, 0), 4, 4, &device);
 
 		/*
 		 *  The 5510/5520 companion chips have a funky PIT.
 		 */  
-		if (pci_dev_present(cyrix_55x0))
+		if (vendor == PCI_VENDOR_ID_CYRIX && (device == PCI_DEVICE_ID_CYRIX_5510 || device == PCI_DEVICE_ID_CYRIX_5520))
 			pit_latch_buggy = 1;
-#endif
-		c->x86_cache_size=16;	/* Yep 16K integrated cache thats it */
+
+		c->x86_cache_size=16;	/* Yep 16K integrated cache thats it (12K in X) */
 
 		/* GXm supports extended cpuid levels 'ala' AMD */
 		if (c->cpuid_level == 2) {
-

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/