Skip to content

Commit

Permalink
uml: update address space affected by pud_clear
Browse files Browse the repository at this point in the history
pud_clear wasn't setting the _PAGE_NEWPAGE bit, fooling tlb_flush into
thinking that this area of the address space was up-to-date and not unmapping
whatever was covered by the pud.

This manifested itself as ldconfig on x86_64 complaining about the first
library it looked at not being a valid ELF file.  A config file is mapped at
0x4000000, as the only thing mapped under its pud, and unmapped.  The
unmapping caused a pud_clear, which, due to this bug, didn't actually unmap
the config file data on the host.  The first library is then mapped at the
same location, but is not actually mapped on the host because accesses to it
cause no page faults.  As a result, ldconfig sees the old config file data.

[akpm@linux-foundation.org: coding-style cleanups]
Signed-off-by: Jeff Dike <jdike@linux.intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
  • Loading branch information
Jeff Dike authored and Linus Torvalds committed Nov 15, 2007
1 parent d4d5d20 commit 8cd8fa5
Showing 1 changed file with 1 addition and 1 deletion.
2 changes: 1 addition & 1 deletion include/asm-um/pgtable-3level.h
Original file line number Diff line number Diff line change
Expand Up @@ -71,7 +71,7 @@ static inline pmd_t *pmd_alloc_one(struct mm_struct *mm, unsigned long address)

static inline void pud_clear (pud_t *pud)
{
set_pud(pud, __pud(0));
set_pud(pud, __pud(_PAGE_NEWPAGE));
}

#define pud_page(pud) phys_to_page(pud_val(pud) & PAGE_MASK)
Expand Down

0 comments on commit 8cd8fa5

Please sign in to comment.