Hi,
Do we really want these two filenames as they are right now?
We have:
Documentation/IO-mapping.txt
Documentation/io-mapping.txt
kees@plankie:~/src/linux-2.6$ git blame Documentation/IO-mapping.txt|head
^1da177e (Linus Torvalds 2005-04-16 15:20:36 -0700 1) [ NOTE: The virt_to_bus() and bus_to_virt() functions have been
5872fb94 (Randy Dunlap 2009-01-29 16:28:02 -0800 2) superseded by the functionality provided by the PCI DMA interface
5872fb94 (Randy Dunlap 2009-01-29 16:28:02 -0800 3) (see Documentation/PCI/PCI-DMA-mapping.txt). They continue
^1da177e (Linus Torvalds 2005-04-16 15:20:36 -0700 4) to be documented below for historical purposes, but new code
^1da177e (Linus Torvalds 2005-04-16 15:20:36 -0700 5) must not use them. --davidm 00/12/12 ]
^1da177e (Linus Torvalds 2005-04-16 15:20:36 -0700 6)
^1da177e (Linus Torvalds 2005-04-16 15:20:36 -0700 7) [ This is a mail message in response to a query on IO mapping, thus the
^1da177e (Linus Torvalds 2005-04-16 15:20:36 -0700 8) strange format for a "document" ]
^1da177e (Linus Torvalds 2005-04-16 15:20:36 -0700 9)
^1da177e (Linus Torvalds 2005-04-16 15:20:36 -0700 10) The AHA-1542 is a bus-master device, and your patch makes the driver give the
kees@plankie:~/src/linux-2.6$ git blame Documentation/io-mapping.txt|head
9663f2e6 (Keith Packard 2008-10-30 19:38:18 -0700 1) The io_mapping functions in linux/io-mapping.h provide an abstraction for
9663f2e6 (Keith Packard 2008-10-30 19:38:18 -0700 2) efficiently mapping small regions of an I/O device to the CPU. The initial
9663f2e6 (Keith Packard 2008-10-30 19:38:18 -0700 3) usage is to support the large graphics aperture on 32-bit processors where
9663f2e6 (Keith Packard 2008-10-30 19:38:18 -0700 4) ioremap_wc cannot be used to statically map the entire aperture to the CPU
9663f2e6 (Keith Packard 2008-10-30 19:38:18 -0700 5) as it would consume too much of the kernel address space.
9663f2e6 (Keith Packard 2008-10-30 19:38:18 -0700 6)
9663f2e6 (Keith Packard 2008-10-30 19:38:18 -0700 7) A mapping object is created during driver initialization using
9663f2e6 (Keith Packard 2008-10-30 19:38:18 -0700 8)
9663f2e6 (Keith Packard 2008-10-30 19:38:18 -0700 9) struct io_mapping *io_mapping_create_wc(unsigned long base,
9663f2e6 (Keith Packard 2008-10-30 19:38:18 -0700 10) unsigned long size)
So the IO-mapping.txt was there first. Keith, can't we rename your
file? For example,
Documentation/io-mapping.txt => Documentation/io-mapping.h.txt
--
Kees
On Wed, 30 Jun 2010 22:58:21 +0200 Kees Bakker wrote:
> Hi,
>
> Do we really want these two filenames as they are right now?
> We have:
> Documentation/IO-mapping.txt
> Documentation/io-mapping.txt
>
> kees@plankie:~/src/linux-2.6$ git blame Documentation/IO-mapping.txt|head
> ^1da177e (Linus Torvalds 2005-04-16 15:20:36 -0700 1) [ NOTE: The virt_to_bus() and bus_to_virt() functions have been
> 5872fb94 (Randy Dunlap 2009-01-29 16:28:02 -0800 2) superseded by the functionality provided by the PCI DMA interface
> 5872fb94 (Randy Dunlap 2009-01-29 16:28:02 -0800 3) (see Documentation/PCI/PCI-DMA-mapping.txt). They continue
> ^1da177e (Linus Torvalds 2005-04-16 15:20:36 -0700 4) to be documented below for historical purposes, but new code
> ^1da177e (Linus Torvalds 2005-04-16 15:20:36 -0700 5) must not use them. --davidm 00/12/12 ]
> ^1da177e (Linus Torvalds 2005-04-16 15:20:36 -0700 6)
> ^1da177e (Linus Torvalds 2005-04-16 15:20:36 -0700 7) [ This is a mail message in response to a query on IO mapping, thus the
> ^1da177e (Linus Torvalds 2005-04-16 15:20:36 -0700 8) strange format for a "document" ]
> ^1da177e (Linus Torvalds 2005-04-16 15:20:36 -0700 9)
> ^1da177e (Linus Torvalds 2005-04-16 15:20:36 -0700 10) The AHA-1542 is a bus-master device, and your patch makes the driver give the
>
> kees@plankie:~/src/linux-2.6$ git blame Documentation/io-mapping.txt|head
> 9663f2e6 (Keith Packard 2008-10-30 19:38:18 -0700 1) The io_mapping functions in linux/io-mapping.h provide an abstraction for
> 9663f2e6 (Keith Packard 2008-10-30 19:38:18 -0700 2) efficiently mapping small regions of an I/O device to the CPU. The initial
> 9663f2e6 (Keith Packard 2008-10-30 19:38:18 -0700 3) usage is to support the large graphics aperture on 32-bit processors where
> 9663f2e6 (Keith Packard 2008-10-30 19:38:18 -0700 4) ioremap_wc cannot be used to statically map the entire aperture to the CPU
> 9663f2e6 (Keith Packard 2008-10-30 19:38:18 -0700 5) as it would consume too much of the kernel address space.
> 9663f2e6 (Keith Packard 2008-10-30 19:38:18 -0700 6)
> 9663f2e6 (Keith Packard 2008-10-30 19:38:18 -0700 7) A mapping object is created during driver initialization using
> 9663f2e6 (Keith Packard 2008-10-30 19:38:18 -0700 8)
> 9663f2e6 (Keith Packard 2008-10-30 19:38:18 -0700 9) struct io_mapping *io_mapping_create_wc(unsigned long base,
> 9663f2e6 (Keith Packard 2008-10-30 19:38:18 -0700 10) unsigned long size)
>
> So the IO-mapping.txt was there first. Keith, can't we rename your
> file? For example,
>
> Documentation/io-mapping.txt => Documentation/io-mapping.h.txt
> --
Alternatively:
$ git mv Documentation/IO-mapping.txt Documentation/bus-virt-phys-mapping.txt
---
~Randy
*** Remember to use Documentation/SubmitChecklist when testing your code ***
On Thursday 01 July 2010, Randy Dunlap wrote:
> On Wed, 30 Jun 2010 22:58:21 +0200 Kees Bakker wrote:
>
> > Hi,
> >
> > Do we really want these two filenames as they are right now?
> > We have:
> > Documentation/IO-mapping.txt
> > Documentation/io-mapping.txt
> >
[...]
> >
> > So the IO-mapping.txt was there first. Keith, can't we rename your
> > file? For example,
> >
> > Documentation/io-mapping.txt => Documentation/io-mapping.h.txt
> > --
>
> Alternatively:
>
> $ git mv Documentation/IO-mapping.txt Documentation/bus-virt-phys-mapping.txt
Sounds good to me.
--
Kees