Skip to contentSkip to conversation

SIS AGP clean up, blacklist and module options

PATCH Heilmann, Oliver 3 participants 9 messages Apr 26, 2004 – May 31, 2004 1 message not in the archive Agpgart driver

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

#1 lore raw
Quoted 7 lines

From: Andrew Morton <[email protected]>
To: Oliver Heilmann <[email protected]>
Subject: Re: [PATCH] SIS AGP clean up, blacklist and module options
Date: Sun, 25 Apr 2004 22:05:51 -0700

Quoted 1 line · Oliver Heilmann
Quoted 7 lines
Quoted 3 lines

Sorry, I'm not used to using yahoo webmail. It stupidly wrapped all the lines. Patch vs 2.6.5. is attached.

Generates several non-trivial rejects against the current kernel.

Please update the patch and resend both the patch and the description

of

Quoted 3 lines

the changes which it makes, thanks.

(Patches in `patch -p1' format are preferred, btw).

Mildly embarrassed. Sorry about wasting everybody's time. Btw. Documentation/SubmittingPatches seems to suggest -p0, At least for single file patches.

Anyway, here it is again (tested and -p1):

* renamed sis_648_enable to sis_delayed_enable and removed chipset references

* All chipsets that require the sis delay-hack are listed in sis_broken_chipsets (and no other place).

* There are two new module options to force agp3-spec compliant initialisation and/or the delay hack. As I only have a 648FX chipset to test on, I figured this might be useful to experiment with on other chipsets (i.e.746[FX]).

Basically, if you have an SiS chipset and your machine freezes when you start X, try the -agp_sis_force_delay=1 option. If this fixes your problem add your PCI ID to sis_broken_chipsets in sis-agp.c Note to 746[FX] people: I'm still not sure what the differences between the two 746 versions and the 648 series are. If this patch does not work for you try playing with the agp_sis_agp_spec module option. Any feedback is greatly appreciated.

Oliver

drivers/char/agp/sis-agp.c renamed from drivers/char/agp/sis-agp.c.orig +41 -18
@@ -13,6 +13,8 @@
 #define SIS_TLBCNTRL	0x97
 #define SIS_TLBFLUSH	0x98
 
+static int __devinitdata agp_sis_force_delay = 0;
+static int __devinitdata agp_sis_agp_spec = -1;
 
 static int sis_fetch_size(void)
 {
@@ -67,7 +69,7 @@ static void sis_cleanup(void)
 			      (previous_size->size_value & ~(0x03)));
 }
 
-static void sis_648_enable(u32 mode)
+static void sis_delayed_enable(u32 mode)
 {
 	struct pci_dev *device = NULL;
 	u32 command;
@@ -93,10 +95,13 @@ static void sis_648_enable(u32 mode)
 
 		pci_write_config_dword(device, agp + PCI_AGP_COMMAND, command);
 
-		if(device->device == PCI_DEVICE_ID_SI_648) {
-			// weird: on 648 and 648fx chipsets any rate change in the target command register
-			// triggers a 5ms screwup during which the master cannot be configured
-			printk(KERN_INFO PFX "sis 648 agp fix - giving bridge time to recover\n");
+		/*
+		 * Weird: on some sis chipsets any rate change in the target
+		 * command register triggers a 5ms screwup during which the master
+		 * cannot be configured		 
+		 */
+		if (device->device == agp_bridge->dev->device) {
+			printk(KERN_INFO PFX "SiS delay workaround: giving bridge time to recover.\n");
 			set_current_state(TASK_UNINTERRUPTIBLE);
 			schedule_timeout (1+(HZ*10)/1000);
 		}
@@ -219,21 +224,35 @@ static struct agp_device_ids sis_agp_dev
 };
 
 
+// chipsets that require the 'delay hack'
+static int sis_broken_chipsets[] __devinitdata = {
+	PCI_DEVICE_ID_SI_648,
+	PCI_DEVICE_ID_SI_746,
+	0 // terminator
+};
+
 static void __devinit sis_get_driver(struct agp_bridge_data *bridge)
 {
-	if (bridge->dev->device == PCI_DEVICE_ID_SI_648) {
-		if (agp_bridge->major_version == 3 && agp_bridge->minor_version < 5) {
-			sis_driver.agp_enable=sis_648_enable;
-		} else {
-			sis_driver.agp_enable			= sis_648_enable;
-			sis_driver.aperture_sizes		= agp3_generic_sizes;
-			sis_driver.size_type			= U16_APER_SIZE;
-			sis_driver.num_aperture_sizes	= AGP_GENERIC_SIZES_ENTRIES;
-			sis_driver.configure			= agp3_generic_configure;
-			sis_driver.fetch_size			= agp3_generic_fetch_size;
-			sis_driver.cleanup				= agp3_generic_cleanup;
-			sis_driver.tlb_flush			= agp3_generic_tlbflush;
-		}
+	int i;
+
+	for(i=0; sis_broken_chipsets[i]!=0; ++i)
+		if(bridge->dev->device==sis_broken_chipsets[i])
+			break;
+	
+	if(sis_broken_chipsets[i] || agp_sis_force_delay)
+		sis_driver.agp_enable=sis_delayed_enable;
+
+	// sis chipsets that indicate less than agp3.5
+	// are not actually fully agp3 compliant
+	if ((agp_bridge->major_version == 3 && agp_bridge->minor_version >= 5
+	     && agp_sis_agp_spec!=0) || agp_sis_agp_spec==1) {
+		sis_driver.aperture_sizes = agp3_generic_sizes;
+		sis_driver.size_type = U16_APER_SIZE;
+		sis_driver.num_aperture_sizes = AGP_GENERIC_SIZES_ENTRIES;
+		sis_driver.configure = agp3_generic_configure;
+		sis_driver.fetch_size = agp3_generic_fetch_size;
+		sis_driver.cleanup = agp3_generic_cleanup;
+		sis_driver.tlb_flush = agp3_generic_tlbflush;
 	}
 }
 
@@ -324,4 +343,8 @@ static void __exit agp_sis_cleanup(void)
 module_init(agp_sis_init);
 module_exit(agp_sis_cleanup);
 
+MODULE_PARM(agp_sis_force_delay,"i");
+MODULE_PARM_DESC(agp_sis_force_delay,"forces sis delay hack");
+MODULE_PARM(agp_sis_agp_spec,"i");
+MODULE_PARM_DESC(agp_sis_agp_spec,"0=force sis init, 1=force generic agp3 init, default: autodetect");
 MODULE_LICENSE("GPL and additional rights");

--------------------------------------------------------------------------------
The information contained herein is confidential and is intended solely for the addressee. Access by any other party is unauthorised without the express written permission of the sender. If you are not the intended recipient, please contact the sender either via the company switchboard on +44 (0)20 7623 8000, or via e-mail return. If you have received this e-mail in error or wish to read our e-mail disclaimer statement and monitoring policy, please refer to
http://www.drkw.com/disc/email/ or contact the sender.
--------------------------------------------------------------------------------

#2 lore raw

Re: [PATCH] SIS AGP vs 2.6.6-rc2

And here it is yet again this time diffed against 2.6.6-rc2 plus latest cset (patches cleanly onto "vanilla" 2.6.6.-rc2 too). I've included the description again:

* renamed sis_648_enable to sis_delayed_enable and removed chipset references

* All chipsets that require the sis delay-hack are listed in sis_broken_chipsets (and no other place).

* There are two new module options to force agp3-spec compliant initialisation and/or the delay hack. As I only have a 648FX chipset to test on, I figured this might be useful to experiment with on other chipsets (i.e.746[FX]).

Basically, if you have an SiS chipset and your machine freezes when you start X, try the -agp_sis_force_delay=1 option. If this fixes your problem add your PCI ID to sis_broken_chipsets in sis-agp.c Note to 746[FX] people: I'm still not sure what the differences between the two 746 versions and the 648 series are. If this patch does not work for you try playing with the agp_sis_agp_spec module option. Any feedback is greatly appreciated.

Oliver

drivers/char/agp/sis-agp.c renamed from drivers/char/agp/sis-agp.c.orig +36 -24
@@ -13,6 +13,8 @@
 #define SIS_TLBCNTRL	0x97
 #define SIS_TLBFLUSH	0x98
 
+static int __devinitdata agp_sis_force_delay = 0;
+static int __devinitdata agp_sis_agp_spec = -1;
 
 static int sis_fetch_size(void)
 {
@@ -67,7 +69,7 @@ static void sis_cleanup(void)
 			      (previous_size->size_value & ~(0x03)));
 }
 
-static void sis_648_enable(u32 mode)
+static void sis_delayed_enable(u32 mode)
 {
 	struct pci_dev *device = NULL;
 	u32 command;
@@ -94,13 +96,12 @@ static void sis_648_enable(u32 mode)
 		pci_write_config_dword(device, agp + PCI_AGP_COMMAND, command);
 
 		/*
-		 * Weird: on 648(fx) and 746(fx) chipsets any rate change in the target
+		 * Weird: on some sis chipsets any rate change in the target
 		 * command register triggers a 5ms screwup during which the master
 		 * cannot be configured		 
 		 */
-		if (device->device == PCI_DEVICE_ID_SI_648 ||
-		    device->device == PCI_DEVICE_ID_SI_746) {
-			printk(KERN_INFO PFX "SiS chipset with AGP problems detected. Giving bridge time to recover.\n");
+		if (device->device == agp_bridge->dev->device) {
+			printk(KERN_INFO PFX "SiS delay workaround: giving bridge time to recover.\n");
 			set_current_state(TASK_UNINTERRUPTIBLE);
 			schedule_timeout (1+(HZ*10)/1000);
 		}
@@ -223,28 +224,35 @@ static struct agp_device_ids sis_agp_dev
 };
 
 
+// chipsets that require the 'delay hack'
+static int sis_broken_chipsets[] __devinitdata = {
+	PCI_DEVICE_ID_SI_648,
+	PCI_DEVICE_ID_SI_746,
+	0 // terminator
+};
+
 static void __devinit sis_get_driver(struct agp_bridge_data *bridge)
 {
-	if (bridge->dev->device == PCI_DEVICE_ID_SI_648) { 
-		sis_driver.agp_enable=sis_648_enable;
-		if (agp_bridge->major_version == 3) {
-			sis_driver.aperture_sizes = agp3_generic_sizes;
-			sis_driver.size_type = U16_APER_SIZE;
-			sis_driver.num_aperture_sizes = AGP_GENERIC_SIZES_ENTRIES;
-			sis_driver.configure = agp3_generic_configure;
-			sis_driver.fetch_size = agp3_generic_fetch_size;
-			sis_driver.cleanup = agp3_generic_cleanup;
-			sis_driver.tlb_flush = agp3_generic_tlbflush;
-		}
-	}
+	int i;
 
-	if (bridge->dev->device == PCI_DEVICE_ID_SI_746) {
-		/*
-		 * We don't know enough about the 746 to enable it properly.
-		 * Though we do know that it needs the 'delay' hack to settle
-		 * after changing modes.
-		 */
-		sis_driver.agp_enable=sis_648_enable;
+	for(i=0; sis_broken_chipsets[i]!=0; ++i)
+		if(bridge->dev->device==sis_broken_chipsets[i])
+			break;
+	
+	if(sis_broken_chipsets[i] || agp_sis_force_delay)
+		sis_driver.agp_enable=sis_delayed_enable;
+
+	// sis chipsets that indicate less than agp3.5
+	// are not actually fully agp3 compliant
+	if ((agp_bridge->major_version == 3 && agp_bridge->minor_version >= 5
+	     && agp_sis_agp_spec!=0) || agp_sis_agp_spec==1) {
+		sis_driver.aperture_sizes = agp3_generic_sizes;
+		sis_driver.size_type = U16_APER_SIZE;
+		sis_driver.num_aperture_sizes = AGP_GENERIC_SIZES_ENTRIES;
+		sis_driver.configure = agp3_generic_configure;
+		sis_driver.fetch_size = agp3_generic_fetch_size;
+		sis_driver.cleanup = agp3_generic_cleanup;
+		sis_driver.tlb_flush = agp3_generic_tlbflush;
 	}
 }
 
@@ -335,4 +343,8 @@ static void __exit agp_sis_cleanup(void)
 module_init(agp_sis_init);
 module_exit(agp_sis_cleanup);
 
+MODULE_PARM(agp_sis_force_delay,"i");
+MODULE_PARM_DESC(agp_sis_force_delay,"forces sis delay hack");
+MODULE_PARM(agp_sis_agp_spec,"i");
+MODULE_PARM_DESC(agp_sis_agp_spec,"0=force sis init, 1=force generic agp3 init, default: autodetect");
 MODULE_LICENSE("GPL and additional rights");

--------------------------------------------------------------------------------
The information contained herein is confidential and is intended solely for the addressee. Access by any other party is unauthorised without the express written permission of the sender. If you are not the intended recipient, please contact the sender either via the company switchboard on +44 (0)20 7623 8000, or via e-mail return. If you have received this e-mail in error or wish to read our e-mail disclaimer statement and monitoring policy, please refer to
http://www.drkw.com/disc/email/ or contact the sender.
--------------------------------------------------------------------------------

#3 lore raw

Re: [PATCH] SIS AGP vs 2.6.6-rc2

Hi,
I tried your patch and is has rejects again:
energy linux # patch -p1 < sis.agp.patch
patching file drivers/char/agp/sis-agp.c
Hunk #1 succeeded at 13 with fuzz 2.
Hunk #2 succeeded at 69 with fuzz 2.
Hunk #3 FAILED at 96.
Hunk #4 FAILED at 224.
2 out of 5 hunks FAILED -- saving rejects to file
drivers/char/agp/sis-agp.c.rej

Attached is the .rej file.

Glück Auf
Volker

Signature · 4 lines
Conclusions
 In a straight-up fight, the Empire squashes the Federation like a bug. Even 
with its numerical advantage removed, the Empire would still squash the 
Federation like a bug. Accept it. -Michael Wong

***************
*** 94,106 ****

                 pci_write_config_dword(device, agp + PCI_AGP_COMMAND, command);
  
                 /*

- * Weird: on 648(fx) and 746(fx) chipsets any rate change in the target
* command register triggers a 5ms screwup during which the master
* cannot be configured
*/
- if (device->device == PCI_DEVICE_ID_SI_648 ||
- device->device == PCI_DEVICE_ID_SI_746) {
- printk(KERN_INFO PFX "SiS chipset with AGP problems detected. Giving bridge time to recover.\n");

                         set_current_state(TASK_UNINTERRUPTIBLE);
                         schedule_timeout (1+(HZ*10)/1000);
                 }
--- 96,107 ----
                 pci_write_config_dword(device, agp + PCI_AGP_COMMAND, command);
  
                 /*

+ * Weird: on some sis chipsets any rate change in the target
* command register triggers a 5ms screwup during which the master
* cannot be configured
*/
+ if (device->device == agp_bridge->dev->device) {
+ printk(KERN_INFO PFX "SiS delay workaround: giving bridge time to recover.\n");

                         set_current_state(TASK_UNINTERRUPTIBLE);
                         schedule_timeout (1+(HZ*10)/1000);
                 }

***************
*** 223,250 ****
};

static void __devinit sis_get_driver(struct agp_bridge_data *bridge) {
- if (bridge->dev->device == PCI_DEVICE_ID_SI_648) {
- sis_driver.agp_enable=sis_648_enable;
- if (agp_bridge->major_version == 3) {

-                        sis_driver.aperture_sizes = agp3_generic_sizes;
-                        sis_driver.size_type = U16_APER_SIZE;
-                        sis_driver.num_aperture_sizes = AGP_GENERIC_SIZES_ENTRIES;
-                        sis_driver.configure = agp3_generic_configure;
-                        sis_driver.fetch_size = agp3_generic_fetch_size;
-                        sis_driver.cleanup = agp3_generic_cleanup;
-                        sis_driver.tlb_flush = agp3_generic_tlbflush;

- }
- }

- if (bridge->dev->device == PCI_DEVICE_ID_SI_746) {
- /*

-                 * We don't know enough about the 746 to enable it properly.
-                 * Though we do know that it needs the 'delay' hack to settle
-                 * after changing modes.
-                 */
-                sis_driver.agp_enable=sis_648_enable;
         }

}

--- 224,258 ----
};

+ // chipsets that require the 'delay hack'
+ static int sis_broken_chipsets[] __devinitdata = {

+        PCI_DEVICE_ID_SI_648,
+        PCI_DEVICE_ID_SI_746,
+        0 // terminator

+ };
+
static void __devinit sis_get_driver(struct agp_bridge_data *bridge) {
+ int i;

+ for(i=0; sis_broken_chipsets[i]!=0; ++i)
+ if(bridge->dev->device==sis_broken_chipsets[i])
+ break;
+
+ if(sis_broken_chipsets[i] || agp_sis_force_delay)
+ sis_driver.agp_enable=sis_delayed_enable;
+

+        // sis chipsets that indicate less than agp3.5
+        // are not actually fully agp3 compliant
+        if ((agp_bridge->major_version == 3 && agp_bridge->minor_version >= 5
+             && agp_sis_agp_spec!=0) || agp_sis_agp_spec==1) {
+                sis_driver.aperture_sizes = agp3_generic_sizes;
+                sis_driver.size_type = U16_APER_SIZE;
+                sis_driver.num_aperture_sizes = AGP_GENERIC_SIZES_ENTRIES;
+                sis_driver.configure = agp3_generic_configure;
+                sis_driver.fetch_size = agp3_generic_fetch_size;
+                sis_driver.cleanup = agp3_generic_cleanup;
+                sis_driver.tlb_flush = agp3_generic_tlbflush;
         }

}

#4 lore raw

Re: [PATCH] SIS AGP vs 2.6.6-rc2

The patch works! (Just to be sure I tried it again with the posting I received back from vger.). Make sure you don't screw up the tabs when you copy and paste. I always save the entire mail to a patch file.

Quoted 11 lines · Volker Armin

Hi,
I tried your patch and is has rejects again:
energy linux # patch -p1 < sis.agp.patch
patching file drivers/char/agp/sis-agp.c
Hunk #1 succeeded at 13 with fuzz 2.
Hunk #2 succeeded at 69 with fuzz 2.
Hunk #3 FAILED at 96.
Hunk #4 FAILED at 224.
2 out of 5 hunks FAILED -- saving rejects to file
drivers/char/agp/sis-agp.c.rej

Quoted 4 lines

Attached is the .rej file.

Glück Auf
Volker

--------------------------------------------------------------------------------
The information contained herein is confidential and is intended solely for the addressee. Access by any other party is unauthorised without the express written permission of the sender. If you are not the intended recipient, please contact the sender either via the company switchboard on +44 (0)20 7623 8000, or via e-mail return. If you have received this e-mail in error or wish to read our e-mail disclaimer statement and monitoring policy, please refer to
http://www.drkw.com/disc/email/ or contact the sender.
--------------------------------------------------------------------------------

#5 lore raw

Re: [PATCH] SIS AGP vs 2.6.6-rc2

In reply to Hemmann, Volker Armin

I forgot, this (-l) will probably do the trick for you:
patch -p1 -l < sis.agp.patch

Quoted 11 lines · Volker Armin

Hi,
I tried your patch and is has rejects again:
energy linux # patch -p1 < sis.agp.patch
patching file drivers/char/agp/sis-agp.c
Hunk #1 succeeded at 13 with fuzz 2.
Hunk #2 succeeded at 69 with fuzz 2.
Hunk #3 FAILED at 96.
Hunk #4 FAILED at 224.
2 out of 5 hunks FAILED -- saving rejects to file
drivers/char/agp/sis-agp.c.rej

Quoted 4 lines

Attached is the .rej file.

Glück Auf
Volker

--------------------------------------------------------------------------------
The information contained herein is confidential and is intended solely for the addressee. Access by any other party is unauthorised without the express written permission of the sender. If you are not the intended recipient, please contact the sender either via the company switchboard on +44 (0)20 7623 8000, or via e-mail return. If you have received this e-mail in error or wish to read our e-mail disclaimer statement and monitoring policy, please refer to
http://www.drkw.com/disc/email/ or contact the sender.
--------------------------------------------------------------------------------

#6 lore raw

Re: [PATCH] SIS AGP vs 2.6.6-rc2

Hi,

I tried the gzip'ed patch and it applied cleanly.
testgart works and gives the usual results, dmesg is ok, showing the warning.

Nvidia is not working with 2.6.6-rc2, so I do not know exactly that everything is ok, but if testgart is fine, X had no problems in the past.

Glück Auf
Volker

Signature · 4 lines
Conclusions
 In a straight-up fight, the Empire squashes the Federation like a bug. Even 
with its numerical advantage removed, the Empire would still squash the 
Federation like a bug. Accept it. -Michael Wong
#7 lore raw

Crash in 2.6.6-rc3-mm1 (SiS/Radeon)

In reply to Heilmann, Oliver

Hi Oliver,

2.6.6-rc3-mm1 contains your patch.

My computer is usually very stable, but I had the following strange crash [1]:

- running XFree86 4.3.0 (from Debian unstable) with FVWM and 3x3 virtual terminals
- in one xterm a kernel compile that was most likely at the final linking (IOW: medium cpu and IO load - I heard my HD working)

I suddenly saw in my virtual terminal the outlines of a xclock in a different virtual terminal (but not the content).

After a Ctrl-Alt-F5 and Alt-F5 I'm also in 2.6.6-rc3-mm1 seeing for a short time the following:
- in the lower half of the window the lower half of the X window
- in the left upper half and the right upper half of the window two copies of tty5

But in the case of the crash, this didn't vanish any more.

At boot time, the kernel said:

May  1 12:07:01 r063144 kernel: agpgart: Found an AGP 2.0 compliant device at 0000:00:00.0.
May  1 12:07:01 r063144 kernel: agpgart: Putting AGP V3 device at 0000:00:00.0 into 4x mode
May  1 12:07:01 r063144 kernel: agpgart: SiS delay workaround: giving bridge time to recover.
May  1 12:07:01 r063144 kernel: agpgart: Putting AGP V3 device at 0000:01:00.0 into 4x mode

My hardware is:

0000:00:00.0 Host bridge: Silicon Integrated Systems [SiS] 746 Host (rev 10)

        Subsystem: Unknown device 1849:0746
        Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B-
        Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort+ >SERR- <PERR-
        Latency: 0
        Region 0: Memory at d0000000 (32-bit, non-prefetchable)
        Capabilities: [c0] AGP version 2.0
                Status: RQ=32 Iso- ArqSz=0 Cal=0 SBA+ ITACoh- GART64- HTrans- 64bit- FW+ AGP3- Rate=x1,x2,x4
                Command: RQ=1 ArqSz=0 Cal=0 SBA+ AGP+ GART64- 64bit- FW- Rate=x1

0000:01:00.0 VGA compatible controller: ATI Technologies Inc Radeon
RV100 QY [Radeon 7000/VE] (prog-if 00 [VGA])
Subsystem: C.P. Technology Co. Ltd: Unknown device 2072
Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping+ SERR+ FastB2B-
Status: Cap+ 66MHz+ UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR-
Latency: 32 (2000ns min), Cache Line Size: 0x08 (32 bytes)
Interrupt: pin A routed to IRQ 11
Region 0: Memory at c0000000 (32-bit, prefetchable)
Region 1: I/O ports at b800 [size=256]
Region 2: Memory at cfef0000 (32-bit, non-prefetchable) [size=64K]
Capabilities: [58] AGP version 2.0
Status: RQ=48 Iso- ArqSz=0 Cal=0 SBA+ ITACoh- GART64- HTrans- 64bit- FW- AGP3- Rate=x1,x2,x4
Command: RQ=32 ArqSz=0 Cal=0 SBA+ AGP+ GART64- 64bit- FW- Rate=x1
Capabilities: [50] Power Management version 2
Flags: PMEClk- DSI- D1+ D2+ AuxCurrent=0mA PME(D0-,D1-,D2-,D3hot-,D3cold-)
Status: D0 PME-Enable- DSel=0 DScale=0 PME-

Radeon framebuffer support is compiled into the kernel, but X is configured not to use it.

cu
Adrian

[1] the SiS AGP patch is first possible suspect, but it might be unrelated

Signature · 4 lines
       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed
#8 lore raw

Re: Crash in 2.6.6-rc3-mm1 (SiS/Radeon)

It's possible that this is indeed an AGP problem (the 746 hasn't received a lot of testing yet).
However, I don't think it's related to the patch. The only change it introduces for your chip is the delay-hack, which I don't think could adversely affect stability.
_If_ it is an AGP problem it might be a GART screw-up. Here's a bunch of things you can do/try:

1. Run testgart.

2. Tell me which kernel version you were running when you thought your system was stable. (Are you sure you were even using agp?)

3. Send me the output of lspci -s0 -vvv -xxx and lspci -s1 -vvv -xxx preferably when X is up.

Thanks for testing,

Oliver

Quoted 63 lines · Adrian Bunk

Hi Oliver,

2.6.6-rc3-mm1 contains your patch.

My computer is usually very stable, but I had the following strange crash [1]:

- running XFree86 4.3.0 (from Debian unstable) with FVWM and 3x3 virtual terminals
- in one xterm a kernel compile that was most likely at the final linking (IOW: medium cpu and IO load - I heard my HD working)

I suddenly saw in my virtual terminal the outlines of a xclock in a different virtual terminal (but not the content).

After a Ctrl-Alt-F5 and Alt-F5 I'm also in 2.6.6-rc3-mm1 seeing for a short time the following:
- in the lower half of the window the lower half of the X window
- in the left upper half and the right upper half of the window two copies of tty5

But in the case of the crash, this didn't vanish any more.

At boot time, the kernel said:

May  1 12:07:01 r063144 kernel: agpgart: Found an AGP 2.0 compliant device at 0000:00:00.0.
May  1 12:07:01 r063144 kernel: agpgart: Putting AGP V3 device at 0000:00:00.0 into 4x mode
May  1 12:07:01 r063144 kernel: agpgart: SiS delay workaround: giving bridge time to recover.
May  1 12:07:01 r063144 kernel: agpgart: Putting AGP V3 device at 0000:01:00.0 into 4x mode

My hardware is:

0000:00:00.0 Host bridge: Silicon Integrated Systems [SiS] 746 Host (rev 10)

        Subsystem: Unknown device 1849:0746
        Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B-
        Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort+ >SERR- <PERR-
        Latency: 0
        Region 0: Memory at d0000000 (32-bit, non-prefetchable)
        Capabilities: [c0] AGP version 2.0
                Status: RQ=32 Iso- ArqSz=0 Cal=0 SBA+ ITACoh- GART64- HTrans- 64bit- FW+ AGP3- Rate=x1,x2,x4
                Command: RQ=1 ArqSz=0 Cal=0 SBA+ AGP+ GART64- 64bit- FW- Rate=x1

0000:01:00.0 VGA compatible controller: ATI Technologies Inc Radeon
RV100 QY [Radeon 7000/VE] (prog-if 00 [VGA])
Subsystem: C.P. Technology Co. Ltd: Unknown device 2072
Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping+ SERR+ FastB2B-
Status: Cap+ 66MHz+ UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR-
Latency: 32 (2000ns min), Cache Line Size: 0x08 (32 bytes)
Interrupt: pin A routed to IRQ 11
Region 0: Memory at c0000000 (32-bit, prefetchable)
Region 1: I/O ports at b800 [size=256]
Region 2: Memory at cfef0000 (32-bit, non-prefetchable) [size=64K]
Capabilities: [58] AGP version 2.0
Status: RQ=48 Iso- ArqSz=0 Cal=0 SBA+ ITACoh- GART64- HTrans- 64bit- FW- AGP3- Rate=x1,x2,x4
Command: RQ=32 ArqSz=0 Cal=0 SBA+ AGP+ GART64- 64bit- FW- Rate=x1
Capabilities: [50] Power Management version 2
Flags: PMEClk- DSI- D1+ D2+ AuxCurrent=0mA PME(D0-,D1-,D2-,D3hot-,D3cold-)
Status: D0 PME-Enable- DSel=0 DScale=0 PME-

Radeon framebuffer support is compiled into the kernel, but X is configured not to use it.

Quoted 5 lines

cu
Adrian

[1] the SiS AGP patch is first possible suspect, but it might be unrelated

--------------------------------------------------------------------------------
The information contained herein is confidential and is intended solely for the addressee. Access by any other party is unauthorised without the express written permission of the sender. If you are not the intended recipient, please contact the sender either via the company switchboard on +44 (0)20 7623 8000, or via e-mail return. If you have received this e-mail in error or wish to read our e-mail disclaimer statement and monitoring policy, please refer to
http://www.drkw.com/disc/email/ or contact the sender.
--------------------------------------------------------------------------------

#9 lore raw

Re: Crash in 2.6.6-rc3-mm1 (SiS/Radeon)

Hi Oliver,

first of all sorry for my very late answer.

It seems the crash I experienced was caused by something else, and not by your AGP patch.

cu
Adrian

Signature · 4 lines
       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed