[patch 10/11] iomap: sort out the broken address reporting caused by the iomap layer

[patch 10/11] iomap: sort out the broken address reporting caused by the iomap layer

[akpm]

From: Alan Cox <alan@lxorguk.ukuu.org.uk>

Add an iomap_name() function which translates an I/O map into a string to
print.

Use it for the Libata layer

For now we use 0xXXXX for I/O and 0xXXXXXXXX for MMIO.  I'm assuming that
eventually some other platforms will want to use their own iomap_name() and
we can add ARCH_HAVE_IOMAP_NAME later as such a platform needs it.

Signed-off-by: Alan Cox <alan@redhat.com>
Cc: Jeff Garzik <jeff@garzik.org>
Cc: Tejun Heo <htejun@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
---

 include/asm-generic/iomap.h |    5 +++++
 lib/iomap.c                 |   12 ++++++++++++
 2 files changed, 17 insertions(+)

diff -puN include/asm-generic/iomap.h~iomap-sort-out-the-broken-address-reporting-caused-by-the-iomap-layer include/asm-generic/iomap.h
--- a/include/asm-generic/iomap.h~iomap-sort-out-the-broken-address-reporting-caused-by-the-iomap-layer
+++ a/include/asm-generic/iomap.h
@@ -65,4 +65,9 @@ struct pci_dev;
 extern void __iomem *pci_iomap(struct pci_dev *dev, int bar, unsigned long max);
 extern void pci_iounmap(struct pci_dev *dev, void __iomem *);
 
+/* Convert an iomap to text for this platform */
+extern char *iomap_name(void __iomem *addr, char *buf, size_t len);
+#define IOMAP_NAMELEN	12
+
+
 #endif
diff -puN lib/iomap.c~iomap-sort-out-the-broken-address-reporting-caused-by-the-iomap-layer lib/iomap.c
--- a/lib/iomap.c~iomap-sort-out-the-broken-address-reporting-caused-by-the-iomap-layer
+++ a/lib/iomap.c
@@ -268,3 +268,15 @@ void pci_iounmap(struct pci_dev *dev, vo
 }
 EXPORT_SYMBOL(pci_iomap);
 EXPORT_SYMBOL(pci_iounmap);
+
+char *iomap_name(void __iomem *port, char *buf, size_t len)
+{
+	if (port == NULL) {
+		snprintf(buf, len, "None");
+		return buf;
+	}
+	IO_COND(port, snprintf(buf, len, "0x%04lx", port),
+		snprintf(buf, len, "%08lx", port));
+	return buf;
+}
+EXPORT_SYMBOL(iomap_name);
_

Re: [patch 10/11] iomap: sort out the broken address reporting caused by the iomap layer

[Jeff Garzik]

akpm@linux-foundation.org wrote:
> From: Alan Cox <alan@lxorguk.ukuu.org.uk>
> 
> Add an iomap_name() function which translates an I/O map into a string to
> print.
> 
> Use it for the Libata layer
> 
> For now we use 0xXXXX for I/O and 0xXXXXXXXX for MMIO.  I'm assuming that
> eventually some other platforms will want to use their own iomap_name() and
> we can add ARCH_HAVE_IOMAP_NAME later as such a platform needs it.
> 
> Signed-off-by: Alan Cox <alan@redhat.com>
> Cc: Jeff Garzik <jeff@garzik.org>
> Cc: Tejun Heo <htejun@gmail.com>
> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
> ---
> 
>  include/asm-generic/iomap.h |    5 +++++
>  lib/iomap.c                 |   12 ++++++++++++
>  2 files changed, 17 insertions(+)
> 
> diff -puN include/asm-generic/iomap.h~iomap-sort-out-the-broken-address-reporting-caused-by-the-iomap-layer include/asm-generic/iomap.h
> --- a/include/asm-generic/iomap.h~iomap-sort-out-the-broken-address-reporting-caused-by-the-iomap-layer
> +++ a/include/asm-generic/iomap.h
> @@ -65,4 +65,9 @@ struct pci_dev;
>  extern void __iomem *pci_iomap(struct pci_dev *dev, int bar, unsigned long max);
>  extern void pci_iounmap(struct pci_dev *dev, void __iomem *);
>  
> +/* Convert an iomap to text for this platform */
> +extern char *iomap_name(void __iomem *addr, char *buf, size_t len);
> +#define IOMAP_NAMELEN	12

this will truncate on 64-bit addresses

otherwise OK

Re: [patch 10/11] iomap: sort out the broken address reporting caused by the iomap layer

[Alan Cox]

> > +/* Convert an iomap to text for this platform */
> > +extern char *iomap_name(void __iomem *addr, char *buf, size_t len);
> > +#define IOMAP_NAMELEN	12
> 
> this will truncate on 64-bit addresses

Yes I suppose we could have 64bit MMIO mappings. So make it 20 8)

Shall I send Andrew/you another diff to do this in the next batch of
updates I'll send out ?

Alan

Re: [patch 10/11] iomap: sort out the broken address reporting caused by the iomap layer

[Jeff Garzik]

Alan Cox wrote:
>>> +/* Convert an iomap to text for this platform */
>>> +extern char *iomap_name(void __iomem *addr, char *buf, size_t len);
>>> +#define IOMAP_NAMELEN	12
>> this will truncate on 64-bit addresses
> 
> Yes I suppose we could have 64bit MMIO mappings. So make it 20 8)
> 
> Shall I send Andrew/you another diff to do this in the next batch of
> updates I'll send out ?

Please.  In general, I was running through patches from you sent by 
Andrew (his periodic libata patchbomb).  IIRC, a few of those were for 
testing and did want immediate upstreaming anyway?