Skip to contentSkip to conversation

OLPC: pci: fix up PCI stuff so that PCI_GOANY supports OLPC

PATCH In review Andres Salomon 7 participants 9 messages Jun 5, 2008 – Jun 6, 2008

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 OLPC: pci: fix up PCI stuff so that PCI_GOANY supports OLPC

Previously, one would have to specifically choose CONFIG_OLPC and CONFIG_PCI_GOOLPC in order to enable PCI_OLPC. That doesn't really work for distro kernels, so this patch allows one to choose CONFIG_OLPC and CONFIG_PCI_GOANY in order to build in OLPC support in a generic kernel (as requested by Robert Millan).

This also moves GOOLPC before GOANY in the menuconfig list.

Finally, make pci_access_init return early if we detect OLPC hardware. There's no need to continue probing stuff, and pci_pcbios_init specifically trashes our settings (we didn't run into that before because PCI_GOANY wasn't supported).

It would be great if we could sneak this in for 2.6.26.. *cough* :)

---
 arch/x86/Kconfig    |   11 +++++------
 arch/x86/pci/init.c |    3 ++-
 arch/x86/pci/olpc.c |    5 +++--
 arch/x86/pci/pci.h  |    2 +-
 4 files changed, 11 insertions(+), 10 deletions(-)
arch/x86/Kconfig +5 -6
index dcbec34..52e18e6 100644
@@ -1508,13 +1508,13 @@ config PCI_GOMMCONFIG
 config PCI_GODIRECT
 	bool "Direct"
 
-config PCI_GOANY
-	bool "Any"
-
 config PCI_GOOLPC
 	bool "OLPC"
 	depends on OLPC
 
+config PCI_GOANY
+	bool "Any"
+
 endchoice
 
 config PCI_BIOS
@@ -1531,9 +1531,8 @@ config PCI_MMCONFIG
 	depends on X86_32 && PCI && ACPI && (PCI_GOMMCONFIG || PCI_GOANY)
 
 config PCI_OLPC
-	bool
-	depends on PCI && PCI_GOOLPC
-	default y
+	def_bool y
+	depends on PCI && OLPC && (PCI_GOOLPC || PCI_GOANY)
 
 config PCI_DOMAINS
 	def_bool y
arch/x86/pci/init.c +2 -1
index e70b9c5..b821f44 100644
@@ -15,7 +15,8 @@ static __init int pci_access_init(void)
 	pci_mmcfg_early_init();
 
 #ifdef CONFIG_PCI_OLPC
-	pci_olpc_init();
+	if (!pci_olpc_init())
+		return 0;	/* skip additional checks if it's an XO */
 #endif
 #ifdef CONFIG_PCI_BIOS
 	pci_pcbios_init();
arch/x86/pci/olpc.c +3 -2
index 5e76365..e11e9e8 100644
@@ -302,12 +302,13 @@ static struct pci_raw_ops pci_olpc_conf = {
 	.write = pci_olpc_write,
 };
 
-void __init pci_olpc_init(void)
+int __init pci_olpc_init(void)
 {
 	if (!machine_is_olpc() || olpc_has_vsa())
-		return;
+		return -ENODEV;
 
 	printk(KERN_INFO "PCI: Using configuration type OLPC\n");
 	raw_pci_ops = &pci_olpc_conf;
 	is_lx = is_geode_lx();
+	return 0;
 }
arch/x86/pci/pci.h +1 -1
index f3972b1..720c4c5 100644
@@ -101,7 +101,7 @@ extern struct pci_raw_ops pci_direct_conf1;
 extern int pci_direct_probe(void);
 extern void pci_direct_init(int type);
 extern void pci_pcbios_init(void);
-extern void pci_olpc_init(void);
+extern int pci_olpc_init(void);
 
 /* pci-mmconfig.c */
 
Signature · 1 line
1.5.5.1
#1 lore raw
Quoted 6 lines · On Thu, 5 Jun 2008 15:29:14 -0400 Andres Salomon <[email protected]> wrote:

Previously, one would have to specifically choose CONFIG_OLPC and CONFIG_PCI_GOOLPC in order to enable PCI_OLPC. That doesn't really work for distro kernels, so this patch allows one to choose CONFIG_OLPC and CONFIG_PCI_GOANY in order to build in OLPC support in a generic kernel (as requested by Robert Millan).

Quoted 9 lines

This also moves GOOLPC before GOANY in the menuconfig list.

Finally, make pci_access_init return early if we detect OLPC hardware. There's no need to continue probing stuff, and pci_pcbios_init specifically trashes our settings (we didn't run into that before because PCI_GOANY wasn't supported).

It would be great if we could sneak this in for 2.6.26.. *cough* :)

It looks harmless enough.

Quoted 5 lines
---
 arch/x86/Kconfig    |   11 +++++------
 arch/x86/pci/init.c |    3 ++-
 arch/x86/pci/olpc.c |    5 +++--
 arch/x86/pci/pci.h  |    2 +-

But I never know whether arch/x86/pci is a Jesse thing or an Ingo thing. Usual answer: it's an everyone thing and we all make a big mess ;)

#2 lore raw
Quoted 20 lines · Andrew Morton

On Thu, 5 Jun 2008 15:29:14 -0400

Quoted 6 lines · Andres Salomon

Previously, one would have to specifically choose CONFIG_OLPC and CONFIG_PCI_GOOLPC in order to enable PCI_OLPC. That doesn't really work for distro kernels, so this patch allows one to choose CONFIG_OLPC and CONFIG_PCI_GOANY in order to build in OLPC support in a generic kernel (as requested by Robert Millan).

Quoted 8 lines

This also moves GOOLPC before GOANY in the menuconfig list.

Finally, make pci_access_init return early if we detect OLPC hardware. There's no need to continue probing stuff, and pci_pcbios_init specifically trashes our settings (we didn't run into that before because PCI_GOANY wasn't supported).

It would be great if we could sneak this in for 2.6.26.. *cough* :)

It looks harmless enough.

Quoted 8 lines
Quoted 5 lines
---
 arch/x86/Kconfig    |   11 +++++------
 arch/x86/pci/init.c |    3 ++-
 arch/x86/pci/olpc.c |    5 +++--
 arch/x86/pci/pci.h  |    2 +-

But I never know whether arch/x86/pci is a Jesse thing or an Ingo thing. Usual answer: it's an everyone thing and we all make a big mess ;)

Heh, yeah I've been taking quite a few x86 PCI patches (Ingo actually has a queue of them for me now), so I can pick this one up and put it in what I hope will be my final 2.6.26 pull request.

Jesse

#3 lore raw
In reply to Andrew Morton
Quoted 21 lines · Andrew Morton
Quoted 6 lines · On Thu, 5 Jun 2008 15:29:14 -0400 Andres Salomon <[email protected]> wrote:

Previously, one would have to specifically choose CONFIG_OLPC and CONFIG_PCI_GOOLPC in order to enable PCI_OLPC. That doesn't really work for distro kernels, so this patch allows one to choose CONFIG_OLPC and CONFIG_PCI_GOANY in order to build in OLPC support in a generic kernel (as requested by Robert Millan).

Quoted 9 lines

This also moves GOOLPC before GOANY in the menuconfig list.

Finally, make pci_access_init return early if we detect OLPC hardware. There's no need to continue probing stuff, and pci_pcbios_init specifically trashes our settings (we didn't run into that before because PCI_GOANY wasn't supported).

It would be great if we could sneak this in for 2.6.26.. *cough* :)

It looks harmless enough.

Quoted 8 lines
Quoted 5 lines
---
 arch/x86/Kconfig    |   11 +++++------
 arch/x86/pci/init.c |    3 ++-
 arch/x86/pci/olpc.c |    5 +++--
 arch/x86/pci/pci.h  |    2 +-

But I never know whether arch/x86/pci is a Jesse thing or an Ingo thing. Usual answer: it's an everyone thing and we all make a big mess ;)

So the Salomonic solution is:

You pick it up and send it either to Jesse or [email protected] :)

I put it into my urgent bucket just to be on the safe side.

Thanks,
tglx

#4 lore raw
In reply to Andres Salomon
Quoted 7 lines · Andres Salomon

Previously, one would have to specifically choose CONFIG_OLPC and CONFIG_PCI_GOOLPC in order to enable PCI_OLPC. That doesn't really work for distro kernels, so this patch allows one to choose CONFIG_OLPC and CONFIG_PCI_GOANY in order to build in OLPC support in a generic kernel (as requested by Robert Millan).

This also moves GOOLPC before GOANY in the menuconfig list.

Please say why... (for above)

Quoted 96 lines

Finally, make pci_access_init return early if we detect OLPC hardware. There's no need to continue probing stuff, and pci_pcbios_init specifically trashes our settings (we didn't run into that before because PCI_GOANY wasn't supported).

It would be great if we could sneak this in for 2.6.26.. *cough* :)

---
 arch/x86/Kconfig    |   11 +++++------
 arch/x86/pci/init.c |    3 ++-
 arch/x86/pci/olpc.c |    5 +++--
 arch/x86/pci/pci.h  |    2 +-
 4 files changed, 11 insertions(+), 10 deletions(-)
arch/x86/Kconfig +5 -6
index dcbec34..52e18e6 100644
@@ -1508,13 +1508,13 @@ config PCI_GOMMCONFIG
 config PCI_GODIRECT
 	bool "Direct"
 
-config PCI_GOANY
-	bool "Any"
-
 config PCI_GOOLPC
 	bool "OLPC"
 	depends on OLPC
 
+config PCI_GOANY
+	bool "Any"
+
 endchoice
 
 config PCI_BIOS
@@ -1531,9 +1531,8 @@ config PCI_MMCONFIG
 	depends on X86_32 && PCI && ACPI && (PCI_GOMMCONFIG || PCI_GOANY)
 
 config PCI_OLPC
-	bool
-	depends on PCI && PCI_GOOLPC
-	default y
+	def_bool y
+	depends on PCI && OLPC && (PCI_GOOLPC || PCI_GOANY)
 
 config PCI_DOMAINS
 	def_bool y
arch/x86/pci/init.c +2 -1
index e70b9c5..b821f44 100644
@@ -15,7 +15,8 @@ static __init int pci_access_init(void)
 	pci_mmcfg_early_init();
 
 #ifdef CONFIG_PCI_OLPC
-	pci_olpc_init();
+	if (!pci_olpc_init())
+		return 0;	/* skip additional checks if it's an XO */
 #endif
 #ifdef CONFIG_PCI_BIOS
 	pci_pcbios_init();
arch/x86/pci/olpc.c +3 -2
index 5e76365..e11e9e8 100644
@@ -302,12 +302,13 @@ static struct pci_raw_ops pci_olpc_conf = {
 	.write = pci_olpc_write,
 };
 
-void __init pci_olpc_init(void)
+int __init pci_olpc_init(void)
 {
 	if (!machine_is_olpc() || olpc_has_vsa())
-		return;
+		return -ENODEV;
 
 	printk(KERN_INFO "PCI: Using configuration type OLPC\n");
 	raw_pci_ops = &pci_olpc_conf;
 	is_lx = is_geode_lx();
+	return 0;
 }
arch/x86/pci/pci.h +1 -1
index f3972b1..720c4c5 100644
@@ -101,7 +101,7 @@ extern struct pci_raw_ops pci_direct_conf1;
 extern int pci_direct_probe(void);
 extern void pci_direct_init(int type);
 extern void pci_pcbios_init(void);
-extern void pci_olpc_init(void);
+extern int pci_olpc_init(void);
 
 /* pci-mmconfig.c */
 
Signature · 1 line
~Randy
#5 lore raw
Quoted 8 lines · On Thu, 5 Jun 2008 13:07:35 -0700 (PDT) "Randy.Dunlap" <[email protected]>…
Quoted 6 lines · Andres Salomon

Previously, one would have to specifically choose CONFIG_OLPC and CONFIG_PCI_GOOLPC in order to enable PCI_OLPC. That doesn't really work for distro kernels, so this patch allows one to choose CONFIG_OLPC and CONFIG_PCI_GOANY in order to build in OLPC support in a generic kernel (as requested by Robert Millan).

Quoted 4 lines
Quoted 1 line

This also moves GOOLPC before GOANY in the menuconfig list.

Please say why... (for above)

The options above GOANY are included in GOANY.
Previously, GOANY did not support GOOLPC, so GOOLPC was listed after it. Now that GOANY supports GOOLPC, it can be listed above w/ the others.

Plus, _I_ think it looks nicer. :) That hunk can easily be removed from the patch if others disagree.

#6 lore raw
In reply to Andres Salomon
Quoted 3 lines · Andres Salomon

Previously, one would have to specifically choose CONFIG_OLPC and CONFIG_PCI_GOOLPC in order to enable PCI_OLPC. That doesn't really work for distro kernels, so this patch allows one to choose

How do distros work on OLPC, anyway? It lacks normal BIOS, so you can't really use GRUB, right?
Pavel

Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
#7 lore raw
Quoted 8 lines · On Fri, 6 Jun 2008 13:05:49 +0200 Pavel Machek <[email protected]> wrote:
Quoted 3 lines · Andres Salomon

Previously, one would have to specifically choose CONFIG_OLPC and CONFIG_PCI_GOOLPC in order to enable PCI_OLPC. That doesn't really work for distro kernels, so this patch allows one to choose

How do distros work on OLPC, anyway? It lacks normal BIOS, so you can't really use GRUB, right?
Pavel

I believe Robert was actually working on grub support. I thought it made little sense within the context of OFW, but if it is a requirement to make installers happy, so be it.

#8 lore raw
Quoted 8 lines · Andres Salomon
Quoted 5 lines · On Fri, 6 Jun 2008 13:05:49 +0200 Pavel Machek <[email protected]> wrote:
Quoted 3 lines · Andres Salomon

Previously, one would have to specifically choose CONFIG_OLPC and CONFIG_PCI_GOOLPC in order to enable PCI_OLPC. That doesn't really work for distro kernels, so this patch allows one to choose

Quoted 5 lines
Quoted 3 lines

How do distros work on OLPC, anyway? It lacks normal BIOS, so you can't really use GRUB, right?
Pavel

I believe Robert was actually working on grub support.

The OLPC port of GRUB is mostly complete now. There's a problem with finding the boot drive though (I'm working on that atm). Other details at http://grub.enbug.org/OLPC

As for the BIOS dependency, GRUB 2 is very portable so it wasn't a problem (we used the existing keyboard driver from the Coreboot port, and the OFW client framework we already had for powerpc).

Quoted 3 lines

I thought it made
little sense within the context of OFW, but if it is a requirement to make installers happy, so be it.

Some might find it practical to have the GRUB menu & update-grub to autobuild it :-)

Signature · 5 lines
Robert Millan

<GPLv2> I know my rights; I want my phone call!
<DRM> What good is a phone call… if you are unable to speak?
(as seen on /.)