a pedantic observation, you can do with it what you wish (i'm not a list subscriber). in drivers/gpu/drm/drm_vm.c, we read (line 629):
#if !defined(__arm__) if (io_remap_pfn_range(vma, vma->vm_start, (map->offset + offset) >> PAGE_SHIFT, vma->vm_end - vma->vm_start, vma->vm_page_prot)) return -EAGAIN; #else if (remap_pfn_range(vma, vma->vm_start, (map->offset + offset) >> PAGE_SHIFT, vma->vm_end - vma->vm_start, vma->vm_page_prot)) return -EAGAIN; #endif
suggesting that only with ARM need we distinguish between those two routines. but for various architectures, in pgtable.h, you can see:
#define io_remap_pfn_range remap_pfn_range
even the arm file contains:
#define io_remap_pfn_range(vma,from,pfn,size,prot) \ remap_pfn_range(vma, from, pfn, size, prot)
so it's not clear what the preprocessor test is for. admittedly, it should work, it just seems unnecessary. and it makes the code more confusing than it needs to be, but perhaps i've just misread something.
rday
On Wed, Aug 29, 2012 at 06:29:53PM -0400, Robert P. J. Day wrote:
a pedantic observation, you can do with it what you wish (i'm not a list subscriber). in drivers/gpu/drm/drm_vm.c, we read (line 629):
#if !defined(__arm__) if (io_remap_pfn_range(vma, vma->vm_start, (map->offset + offset) >> PAGE_SHIFT, vma->vm_end - vma->vm_start, vma->vm_page_prot)) return -EAGAIN; #else if (remap_pfn_range(vma, vma->vm_start, (map->offset + offset) >> PAGE_SHIFT, vma->vm_end - vma->vm_start, vma->vm_page_prot)) return -EAGAIN; #endif
This code goes down to:
commit 4b7fb9b5746554d460e7bc986341d4b2db01e0b6 Author: Jordan Crouse jcrouse@codeaurora.org Date: Thu May 27 13:40:26 2010 -0600
drm: Add __arm defines to DRM
Add __arm defines to specify behavior specific for an ARM processor.
Signed-off-by: Jordan Crouse jcrouse@codeaurora.org Signed-off-by: Dave Airlie airlied@redhat.com
Even at that time io_remap_pfn_range on arm was defined as:
#define io_remap_pfn_range(vma,from,pfn,size,prot) \ remap_pfn_range(vma, from, pfn, size, prot)
so it's not clear what the preprocessor test is for. admittedly, it should work, it just seems unnecessary. and it makes the code more confusing than it needs to be, but perhaps i've just misread something.
I agree that this shouldn't be necessary.
Sascha
On Thu, 30 Aug 2012, Sascha Hauer wrote:
On Wed, Aug 29, 2012 at 06:29:53PM -0400, Robert P. J. Day wrote:
a pedantic observation, you can do with it what you wish (i'm not a list subscriber). in drivers/gpu/drm/drm_vm.c, we read (line 629):
#if !defined(__arm__) if (io_remap_pfn_range(vma, vma->vm_start, (map->offset + offset) >> PAGE_SHIFT, vma->vm_end - vma->vm_start, vma->vm_page_prot)) return -EAGAIN; #else if (remap_pfn_range(vma, vma->vm_start, (map->offset + offset) >> PAGE_SHIFT, vma->vm_end - vma->vm_start, vma->vm_page_prot)) return -EAGAIN; #endif
This code goes down to:
commit 4b7fb9b5746554d460e7bc986341d4b2db01e0b6 Author: Jordan Crouse jcrouse@codeaurora.org Date: Thu May 27 13:40:26 2010 -0600
drm: Add __arm defines to DRM Add __arm defines to specify behavior specific for an ARM processor. Signed-off-by: Jordan Crouse <jcrouse@codeaurora.org> Signed-off-by: Dave Airlie <airlied@redhat.com>Even at that time io_remap_pfn_range on arm was defined as:
#define io_remap_pfn_range(vma,from,pfn,size,prot) \ remap_pfn_range(vma, from, pfn, size, prot)
so it's not clear what the preprocessor test is for. admittedly, it should work, it just seems unnecessary. and it makes the code more confusing than it needs to be, but perhaps i've just misread something.
I agree that this shouldn't be necessary.
for simplicity, i can submit a patch that just does a straight call to remap_pfn_range(), unless someone can recognize something really subtle happening here.
rday
On Thu, 30 Aug 2012, Robert P. J. Day wrote:
On Thu, 30 Aug 2012, Sascha Hauer wrote:
On Wed, Aug 29, 2012 at 06:29:53PM -0400, Robert P. J. Day wrote:
a pedantic observation, you can do with it what you wish (i'm not a list subscriber). in drivers/gpu/drm/drm_vm.c, we read (line 629):
#if !defined(__arm__) if (io_remap_pfn_range(vma, vma->vm_start, (map->offset + offset) >> PAGE_SHIFT, vma->vm_end - vma->vm_start, vma->vm_page_prot)) return -EAGAIN; #else if (remap_pfn_range(vma, vma->vm_start, (map->offset + offset) >> PAGE_SHIFT, vma->vm_end - vma->vm_start, vma->vm_page_prot)) return -EAGAIN; #endif
This code goes down to:
commit 4b7fb9b5746554d460e7bc986341d4b2db01e0b6 Author: Jordan Crouse jcrouse@codeaurora.org Date: Thu May 27 13:40:26 2010 -0600
drm: Add __arm defines to DRM Add __arm defines to specify behavior specific for an ARM processor. Signed-off-by: Jordan Crouse <jcrouse@codeaurora.org> Signed-off-by: Dave Airlie <airlied@redhat.com>Even at that time io_remap_pfn_range on arm was defined as:
#define io_remap_pfn_range(vma,from,pfn,size,prot) \ remap_pfn_range(vma, from, pfn, size, prot)
so it's not clear what the preprocessor test is for. admittedly, it should work, it just seems unnecessary. and it makes the code more confusing than it needs to be, but perhaps i've just misread something.
I agree that this shouldn't be necessary.
for simplicity, i can submit a patch that just does a straight call to remap_pfn_range(), unless someone can recognize something really subtle happening here.
sorry, that should have been a straight call to io_remap_pfn_range() since that is the call invoked for all but ARM. i'll post a patch shortly and someone else can decide whether or not to apply it.
rday
dri-devel@lists.freedesktop.org
-
Robert P. J. Day -
Sascha Hauer