* [vm 0/76] convert remap_page_range() to remap_pfn_range() vs. 2.6.9-rc2-mm3
@ 2004-09-25 6:53 William Lee Irwin III
2004-09-25 6:58 ` [vm 1/76] introduce remap_pfn_range() William Lee Irwin III
0 siblings, 1 reply; 7+ messages in thread
From: William Lee Irwin III @ 2004-09-25 6:53 UTC (permalink / raw)
To: Andrew Morton; +Cc: linux-kernel
This series resolves a physical address overflow issue in the
remap_page_range() API by replacing it with remap_pfn_range(), which
accepts its physical address argument as a pfn, hence allowing the use
of a single-precision physical address argument without the risk of
overflow at the API boundary. The above issue has hobbled support for
various 32-bit architectures, including some embedded systems (ppc440
IIRC), caused persistent portability issues for sound drivers for
legacy systems (sparc32; unfortunately this patch alone does not fully
resolve those), and according to John Fusco's reports, made drivers for
some PCI-X hardware infeasible to port to recent ia32 PAE enterprise
systems. With this patch series applied, physical address overflows on
32-bit systems caused directly by remap_page_range() are gone forever,
and ca. 100LOC of cut-and-waste driver code are swept out of existence
alongside them.
vs. 2.6.9-rc2-mm3 and the scheduler header cleanups. The parts touching
mm.h should apply with just offsets if those aren't applied beforehand.
Successfully tested on x86-64.
-- wli
P.S.: The existing solution to the sparc32 issue was to pass a double
precision representation of the physical address as 2 single-
precision arguments in an API (io_remap_page_range()) whose
argument corresponding to those two was a single single-
precision argument on most/all other architectures. The
sparc32-specific issue requires more work beyond these patches
to rectify. The most apparent consequence of the API skew is
that drivers don't compile on sparc32 when they use
io_remap_page_range() due to passing insufficient arguments,
or vice-versa for drivers originally written for sparc32.
^ permalink raw reply [flat|nested] 7+ messages in thread
* [vm 1/76] introduce remap_pfn_range()
2004-09-25 6:53 [vm 0/76] convert remap_page_range() to remap_pfn_range() vs. 2.6.9-rc2-mm3 William Lee Irwin III
@ 2004-09-25 6:58 ` William Lee Irwin III
2004-09-25 6:59 ` [vm 2/76] convert atyfb.c to use remap_pfn_range() William Lee Irwin III
0 siblings, 1 reply; 7+ messages in thread
From: William Lee Irwin III @ 2004-09-25 6:58 UTC (permalink / raw)
To: Andrew Morton; +Cc: linux-kernel
This patch introduces remap_pfn_range(), destined to replace
remap_page_range(). In the sequel, the callers of remap_page_range()
will be converted one at a time.
By using a pfn to specify its physical address argument,
remap_pfn_range() resolves a longstanding physical address overflow
issue.
Index: mm3-2.6.9-rc2/include/linux/mm.h
===================================================================
--- mm3-2.6.9-rc2.orig/include/linux/mm.h 2004-09-24 19:17:00.863688360 -0700
+++ mm3-2.6.9-rc2/include/linux/mm.h 2004-09-24 22:01:48.366559448 -0700
@@ -856,8 +856,15 @@
extern struct page * vmalloc_to_page(void *addr);
extern struct page * follow_page(struct mm_struct *mm, unsigned long address,
int write);
-extern int remap_page_range(struct vm_area_struct *vma, unsigned long from,
- unsigned long to, unsigned long size, pgprot_t prot);
+int remap_pfn_range(struct vm_area_struct *, unsigned long,
+ unsigned long, unsigned long, pgprot_t);
+
+static inline
+int remap_page_range(struct vm_area_struct *vma, unsigned long uvaddr,
+ unsigned long paddr, unsigned long size, pgprot_t prot)
+{
+ return remap_pfn_range(vma, uvaddr, paddr >> PAGE_SHIFT, size, prot);
+}
#ifdef CONFIG_PROC_FS
void __vm_stat_account(struct mm_struct *, unsigned long, struct file *, long);
Index: mm3-2.6.9-rc2/mm/memory.c
===================================================================
--- mm3-2.6.9-rc2.orig/mm/memory.c 2004-09-24 17:37:15.000000000 -0700
+++ mm3-2.6.9-rc2/mm/memory.c 2004-09-24 21:58:50.300629584 -0700
@@ -945,16 +945,14 @@
* in null mappings (currently treated as "copy-on-access")
*/
static inline void remap_pte_range(pte_t * pte, unsigned long address, unsigned long size,
- unsigned long phys_addr, pgprot_t prot)
+ unsigned long pfn, pgprot_t prot)
{
unsigned long end;
- unsigned long pfn;
address &= ~PMD_MASK;
end = address + size;
if (end > PMD_SIZE)
end = PMD_SIZE;
- pfn = phys_addr >> PAGE_SHIFT;
do {
BUG_ON(!pte_none(*pte));
if (!pfn_valid(pfn) || PageReserved(pfn_to_page(pfn)))
@@ -966,7 +964,7 @@
}
static inline int remap_pmd_range(struct mm_struct *mm, pmd_t * pmd, unsigned long address, unsigned long size,
- unsigned long phys_addr, pgprot_t prot)
+ unsigned long pfn, pgprot_t prot)
{
unsigned long base, end;
@@ -975,12 +973,12 @@
end = address + size;
if (end > PGDIR_SIZE)
end = PGDIR_SIZE;
- phys_addr -= address;
+ pfn -= address >> PAGE_SHIFT;
do {
pte_t * pte = pte_alloc_map(mm, pmd, base + address);
if (!pte)
return -ENOMEM;
- remap_pte_range(pte, base + address, end - address, address + phys_addr, prot);
+ remap_pte_range(pte, base + address, end - address, pfn + (address >> PAGE_SHIFT), prot);
pte_unmap(pte);
address = (address + PMD_SIZE) & PMD_MASK;
pmd++;
@@ -989,7 +987,7 @@
}
/* Note: this is only safe if the mm semaphore is held when called. */
-int remap_page_range(struct vm_area_struct *vma, unsigned long from, unsigned long phys_addr, unsigned long size, pgprot_t prot)
+int remap_pfn_range(struct vm_area_struct *vma, unsigned long from, unsigned long pfn, unsigned long size, pgprot_t prot)
{
int error = 0;
pgd_t * dir;
@@ -997,7 +995,7 @@
unsigned long end = from + size;
struct mm_struct *mm = vma->vm_mm;
- phys_addr -= from;
+ pfn -= from >> PAGE_SHIFT;
dir = pgd_offset(mm, from);
flush_cache_range(vma, beg, end);
if (from >= end)
@@ -1009,7 +1007,7 @@
error = -ENOMEM;
if (!pmd)
break;
- error = remap_pmd_range(mm, pmd, from, end - from, phys_addr + from, prot);
+ error = remap_pmd_range(mm, pmd, from, end - from, pfn + (from >> PAGE_SHIFT), prot);
if (error)
break;
from = (from + PGDIR_SIZE) & PGDIR_MASK;
@@ -1022,8 +1020,7 @@
spin_unlock(&mm->page_table_lock);
return error;
}
-
-EXPORT_SYMBOL(remap_page_range);
+EXPORT_SYMBOL(remap_pfn_range);
/*
* Do pte_mkwrite, but only if the vma says VM_WRITE. We do this when
Index: mm3-2.6.9-rc2/mm/nommu.c
===================================================================
--- mm3-2.6.9-rc2.orig/mm/nommu.c 2004-09-24 02:10:30.000000000 -0700
+++ mm3-2.6.9-rc2/mm/nommu.c 2004-09-24 22:06:16.400812048 -0700
@@ -560,7 +560,7 @@
return NULL;
}
-int remap_page_range(struct vm_area_struct *vma, unsigned long from,
+int remap_pfn_range(struct vm_area_struct *vma, unsigned long from,
unsigned long to, unsigned long size, pgprot_t prot)
{
return -EPERM;
^ permalink raw reply [flat|nested] 7+ messages in thread
* [vm 2/76] convert atyfb.c to use remap_pfn_range()
2004-09-25 6:58 ` [vm 1/76] introduce remap_pfn_range() William Lee Irwin III
@ 2004-09-25 6:59 ` William Lee Irwin III
2004-09-25 7:01 ` [vm 3/76] convert gbefb.c " William Lee Irwin III
0 siblings, 1 reply; 7+ messages in thread
From: William Lee Irwin III @ 2004-09-25 6:59 UTC (permalink / raw)
To: Andrew Morton; +Cc: linux-kernel
On Fri, Sep 24, 2004 at 11:58:16PM -0700, William Lee Irwin III wrote:
> This patch introduces remap_pfn_range(), destined to replace
> remap_page_range(). In the sequel, the callers of remap_page_range()
> will be converted one at a time.
Convert atyfb.c to use remap_pfn_range(). Here (as in numerous other
cases) shifting the argument right by PAGE_SHIFT suffices.
Index: mm3-2.6.9-rc2/drivers/video/aty/atyfb_base.c
===================================================================
--- mm3-2.6.9-rc2.orig/drivers/video/aty/atyfb_base.c 2004-09-24 17:37:14.000000000 -0700
+++ mm3-2.6.9-rc2/drivers/video/aty/atyfb_base.c 2004-09-24 22:07:40.573015928 -0700
@@ -1174,8 +1174,8 @@
~(par->mmap_map[i].prot_mask);
pgprot_val(vma->vm_page_prot) |= par->mmap_map[i].prot_flag;
- if (remap_page_range(vma, vma->vm_start + page, map_offset,
- map_size, vma->vm_page_prot))
+ if (remap_pfn_range(vma, vma->vm_start + page,
+ map_offset >> PAGE_SHIFT, map_size, vma->vm_page_prot))
return -EAGAIN;
page += map_size;
^ permalink raw reply [flat|nested] 7+ messages in thread
* [vm 3/76] convert gbefb.c to use remap_pfn_range()
2004-09-25 6:59 ` [vm 2/76] convert atyfb.c to use remap_pfn_range() William Lee Irwin III
@ 2004-09-25 7:01 ` William Lee Irwin III
2004-09-25 7:03 ` [vm 4/76] convert sgivwfb.c " William Lee Irwin III
0 siblings, 1 reply; 7+ messages in thread
From: William Lee Irwin III @ 2004-09-25 7:01 UTC (permalink / raw)
To: Andrew Morton; +Cc: linux-kernel
On Fri, Sep 24, 2004 at 11:59:59PM -0700, William Lee Irwin III wrote:
> Convert atyfb.c to use remap_pfn_range(). Here (as in numerous other
> cases) shifting the argument right by PAGE_SHIFT suffices.
Conver the SGI GBE framebuffer driver to use remap_pfn_range().
Index: mm3-2.6.9-rc2/drivers/video/gbefb.c
===================================================================
--- mm3-2.6.9-rc2.orig/drivers/video/gbefb.c 2004-09-24 02:10:27.000000000 -0700
+++ mm3-2.6.9-rc2/drivers/video/gbefb.c 2004-09-24 22:08:19.215141432 -0700
@@ -1018,8 +1018,8 @@
else
phys_size = TILE_SIZE - offset;
- if (remap_page_range
- (vma, addr, phys_addr, phys_size, vma->vm_page_prot))
+ if (remap_pfn_range(vma, addr, phys_addr >> PAGE_SHIFT,
+ phys_size, vma->vm_page_prot))
return -EAGAIN;
offset = 0;
^ permalink raw reply [flat|nested] 7+ messages in thread
* [vm 4/76] convert sgivwfb.c to use remap_pfn_range()
2004-09-25 7:01 ` [vm 3/76] convert gbefb.c " William Lee Irwin III
@ 2004-09-25 7:03 ` William Lee Irwin III
2004-09-25 7:05 ` [vm 5/76] convert igafb.c " William Lee Irwin III
0 siblings, 1 reply; 7+ messages in thread
From: William Lee Irwin III @ 2004-09-25 7:03 UTC (permalink / raw)
To: Andrew Morton; +Cc: linux-kernel
On Sat, Sep 25, 2004 at 12:01:51AM -0700, William Lee Irwin III wrote:
> Conver the SGI GBE framebuffer driver to use remap_pfn_range().
Convert the SGI DBE framebuffer driver to use remap_pfn_range().
Index: mm3-2.6.9-rc2/drivers/video/sgivwfb.c
===================================================================
--- mm3-2.6.9-rc2.orig/drivers/video/sgivwfb.c 2004-09-24 02:10:27.000000000 -0700
+++ mm3-2.6.9-rc2/drivers/video/sgivwfb.c 2004-09-24 22:08:57.675294600 -0700
@@ -719,8 +719,8 @@
pgprot_val(vma->vm_page_prot) =
pgprot_val(vma->vm_page_prot) | _PAGE_PCD;
vma->vm_flags |= VM_IO;
- if (remap_page_range
- (vma, vma->vm_start, offset, size, vma->vm_page_prot))
+ if (remap_pfn_range(vma, vma->vm_start, offset >> PAGE_SHIFT,
+ size, vma->vm_page_prot))
return -EAGAIN;
vma->vm_file = file;
printk(KERN_DEBUG "sgivwfb: mmap framebuffer P(%lx)->V(%lx)\n",
^ permalink raw reply [flat|nested] 7+ messages in thread
* [vm 5/76] convert igafb.c to use remap_pfn_range()
2004-09-25 7:03 ` [vm 4/76] convert sgivwfb.c " William Lee Irwin III
@ 2004-09-25 7:05 ` William Lee Irwin III
2004-09-25 7:08 ` [vm 6/76] convert zr36120.c " William Lee Irwin III
0 siblings, 1 reply; 7+ messages in thread
From: William Lee Irwin III @ 2004-09-25 7:05 UTC (permalink / raw)
To: Andrew Morton; +Cc: linux-kernel
On Sat, Sep 25, 2004 at 12:03:46AM -0700, William Lee Irwin III wrote:
> Convert the SGI DBE framebuffer driver to use remap_pfn_range().
Convert the IGA 1682 framebuffer driver to use remap_pfn_range().
Index: mm3-2.6.9-rc2/drivers/video/igafb.c
===================================================================
--- mm3-2.6.9-rc2.orig/drivers/video/igafb.c 2004-09-24 17:37:14.000000000 -0700
+++ mm3-2.6.9-rc2/drivers/video/igafb.c 2004-09-24 22:09:34.845643848 -0700
@@ -262,8 +262,8 @@
pgprot_val(vma->vm_page_prot) &= ~(par->mmap_map[i].prot_mask);
pgprot_val(vma->vm_page_prot) |= par->mmap_map[i].prot_flag;
- if (remap_page_range(vma, vma->vm_start + page, map_offset,
- map_size, vma->vm_page_prot))
+ if (remap_pfn_range(vma, vma->vm_start + page,
+ map_offset >> PAGE_SHIFT, map_size, vma->vm_page_prot))
return -EAGAIN;
page += map_size;
^ permalink raw reply [flat|nested] 7+ messages in thread
* [vm 6/76] convert zr36120.c to use remap_pfn_range()
2004-09-25 7:05 ` [vm 5/76] convert igafb.c " William Lee Irwin III
@ 2004-09-25 7:08 ` William Lee Irwin III
0 siblings, 0 replies; 7+ messages in thread
From: William Lee Irwin III @ 2004-09-25 7:08 UTC (permalink / raw)
To: Andrew Morton; +Cc: linux-kernel
On Sat, Sep 25, 2004 at 12:05:36AM -0700, William Lee Irwin III wrote:
> Convert the IGA 1682 framebuffer driver to use remap_pfn_range().
Conver the Zoran 36120/36125 framegrabber driver to use remap_pfn_range().
Index: mm3-2.6.9-rc2/drivers/media/video/zr36120.c
===================================================================
--- mm3-2.6.9-rc2.orig/drivers/media/video/zr36120.c 2004-09-24 17:37:19.000000000 -0700
+++ mm3-2.6.9-rc2/drivers/media/video/zr36120.c 2004-09-24 22:10:34.377593616 -0700
@@ -1474,8 +1474,8 @@
/* start mapping the whole shabang to user memory */
pos = (unsigned long)ztv->fbuffer;
while (size>0) {
- unsigned long page = virt_to_phys((void*)pos);
- if (remap_page_range(vma, start, page, PAGE_SIZE, PAGE_SHARED))
+ unsigned long pfn = virt_to_phys((void*)pos) >> PAGE_SHIFT;
+ if (remap_pfn_range(vma, start, pfn, PAGE_SIZE, PAGE_SHARED))
return -EAGAIN;
start += PAGE_SIZE;
pos += PAGE_SIZE;
^ permalink raw reply [flat|nested] 7+ messages in thread
end of thread, other threads:[~2004-09-25 7:08 UTC | newest]
Thread overview: 7+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2004-09-25 6:53 [vm 0/76] convert remap_page_range() to remap_pfn_range() vs. 2.6.9-rc2-mm3 William Lee Irwin III
2004-09-25 6:58 ` [vm 1/76] introduce remap_pfn_range() William Lee Irwin III
2004-09-25 6:59 ` [vm 2/76] convert atyfb.c to use remap_pfn_range() William Lee Irwin III
2004-09-25 7:01 ` [vm 3/76] convert gbefb.c " William Lee Irwin III
2004-09-25 7:03 ` [vm 4/76] convert sgivwfb.c " William Lee Irwin III
2004-09-25 7:05 ` [vm 5/76] convert igafb.c " William Lee Irwin III
2004-09-25 7:08 ` [vm 6/76] convert zr36120.c " William Lee Irwin III
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.