[PATCH] implement early_io{re,un}map for ia64

[PATCH] implement early_io{re,un}map for ia64

[Luck, Tony]

drivers/pci/dmar.c uses these functions, so provide them for ia64
Date: Mon, 14 Dec 2009 12:00:36 -0800
Message-Id: <4b26996416535305c4@agluck-desktop.sc.intel.com>

Signed-off-by: Tony Luck <tony.luck@intel.com>

---

linux-next hasn't been building for ia64 for a few days. This fixes
the build problem (but hasn't been tested on a system that has a working
iommu).  Need to synchronize so this gets to mainline before Chris'
"intel-iommu: Detect DMAR in hyperspace at probe time" patch.

David: Do you want to run it through your iommu tree, or should
I just pre-emptively push it in via the ia64 tree?

diff --git a/arch/ia64/include/asm/io.h b/arch/ia64/include/asm/io.h
index 0d9d16e..cc8335e 100644
--- a/arch/ia64/include/asm/io.h
+++ b/arch/ia64/include/asm/io.h
@@ -424,6 +424,8 @@ __writeq (unsigned long val, volatile void __iomem *addr)
 extern void __iomem * ioremap(unsigned long offset, unsigned long size);
 extern void __iomem * ioremap_nocache (unsigned long offset, unsigned long size);
 extern void iounmap (volatile void __iomem *addr);
+extern void __iomem * early_ioremap (unsigned long phys_addr, unsigned long size);
+extern void early_iounmap (volatile void __iomem *addr, unsigned long size);
 
 /*
  * String version of IO memory access ops:
diff --git a/arch/ia64/mm/ioremap.c b/arch/ia64/mm/ioremap.c
index 2a14062..3dccdd8 100644
--- a/arch/ia64/mm/ioremap.c
+++ b/arch/ia64/mm/ioremap.c
@@ -22,6 +22,12 @@ __ioremap (unsigned long phys_addr)
 }
 
 void __iomem *
+early_ioremap (unsigned long phys_addr, unsigned long size)
+{
+	return __ioremap(phys_addr);
+}
+
+void __iomem *
 ioremap (unsigned long phys_addr, unsigned long size)
 {
 	void __iomem *addr;
@@ -102,6 +108,11 @@ ioremap_nocache (unsigned long phys_addr, unsigned long size)
 EXPORT_SYMBOL(ioremap_nocache);
 
 void
+early_iounmap (volatile void __iomem *addr, unsigned long size)
+{
+}
+
+void
 iounmap (volatile void __iomem *addr)
 {
 	if (REGION_NUMBER(addr) = RGN_GATE)

Re: [PATCH] implement early_io{re,un}map for ia64

[Chris Wright]

* Luck, Tony (tony.luck@intel.com) wrote:
> drivers/pci/dmar.c uses these functions, so provide them for ia64
> Date: Mon, 14 Dec 2009 12:00:36 -0800
> Message-Id: <4b26996416535305c4@agluck-desktop.sc.intel.com>
> 
> Signed-off-by: Tony Luck <tony.luck@intel.com>

Sorry about breaking the build Tony.

Ack

Re: [PATCH] implement early_io{re,un}map for ia64

[David Woodhouse]

On Mon, 2009-12-14 at 20:00 +0000, Luck, Tony wrote:
> 
> David: Do you want to run it through your iommu tree, or should
> I just pre-emptively push it in via the ia64 tree? 

I'll push it, thanks. That keeps it very close to the commit which broke
it, so bisection isn't hard.

-- 
David Woodhouse                            Open Source Technology Centre
David.Woodhouse@intel.com                              Intel Corporation

RE: [PATCH] implement early_io{re,un}map for ia64

[Luck, Tony]

> I'll push it, thanks. That keeps it very close to the commit which broke
> it, so bisection isn't hard.

Too late ... I bundled it with a pull request to Linus yesterday ... so
he will probably do the pull and get it into 2.6.33-rc1 (which will keep
it ahead of when Chris' iommu fix hits mainline).

-Tony

RE: [PATCH] implement early_io{re,un}map for ia64

[David Woodhouse]

On Wed, 2009-12-16 at 17:18 +0000, Luck, Tony wrote:
> > I'll push it, thanks. That keeps it very close to the commit which broke
> > it, so bisection isn't hard.
> 
> Too late ... I bundled it with a pull request to Linus yesterday ...

So did I :)

>  so he will probably do the pull and get it into 2.6.33-rc1 (which
> will keep it ahead of when Chris' iommu fix hits mainline).

That doesn't necessarily help for bisection.

-- 
David Woodhouse                            Open Source Technology Centre
David.Woodhouse@intel.com                              Intel Corporation