On Thu, Oct 04, 2012 at 11:05:07AM +0200, Rabin Vincent wrote:
> 2012/10/4 Mika Westerberg <mika.westerberg@xxxxxx>:
> > The unity-mapped region is mapped using 1MB pages. However, we actually have
> > (when using the Linux ARM 2-level translation scheme):
> >
> > see arch/arm/include/asm/pgtable-2level.h:
> >
> > #define PMD_SHIFT 21
> > #define PGDIR_SHIFT 21
> >
> > #define PTRS_PER_PGD 2048
> >
> > So we have 2048 entries in a PGD instead of 4096 making a PGD entry an array
> > of "two pointers".
> >
> > Anyway as you and Paawan suggested it looks like a bug - we always use the
> > first entry instead of the second given that bit 20 is set in the virtual
> > address.
>
> Isn't the problem actually that we read the section entry wrong?
> The following (and attached) is the fix I've been using for this.
Yeah, it looks like that part of the code is full of bugs :-/
>
> Rabin
>
> diff --git a/arm.c b/arm.c
> index 5930c02..4e7b6dc 100644
> --- a/arm.c
> +++ b/arm.c
> @@ -957,12 +957,14 @@ arm_vtop(ulong vaddr, ulong *pgd, physaddr_t
> *paddr, int verbose)
> MKSTR((ulong)page_middle)), pmd_pte);
>
> if ((pmd_pte & PMD_TYPE_MASK) == PMD_TYPE_SECT) {
> + ulong sectionbase = pmd_pte & _SECTION_PAGE_MASK;
> +
> if (verbose) {
> fprintf(fp, " PAGE: %s (1MB)\n\n",
> mkstring(buf, VADDR_PRLEN, RJUST | LONG_HEX,
> - MKSTR(PAGEBASE(pmd_pte))));
> + MKSTR(sectionbase)));
> }
> - *paddr = PAGEBASE(pmd_pte) + (vaddr & ~_SECTION_PAGE_MASK);
> + *paddr = sectionbase + (vaddr & ~_SECTION_PAGE_MASK);
> return TRUE;
> }
Ack for this patch.
--
Crash-utility mailing list
Crash-utility@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/crash-utility
