On Fri, 2005-02-25 at 10:37 -0800, David S. Miller wrote:
> On Fri, 25 Feb 2005 18:20:55 +1100
> Benjamin Herrenschmidt <[EMAIL PROTECTED]> wrote:
>
> > Ok, LTP dies on POWER5, investigation in progress ;)
>
> I bet the address arg is incorrect in some case.
Yes, probably. That's the same symptoms we had when zeromap_pud_range
had the bug getting the address wrong, which means we fail to properly
flush the hash & TLB for this PTE. I'm not at work (it's sat. already
here :) but I'll have a look asap.
> It is ea<sy
> to add debugging for this, for example in your set_pte_at()
> implementation you can do something like:
>
> static inline void set_pte_at(struct mm_struct *mm, unsigned long addr,
> pte_t *ptep, pte_t pte)
> {
> #ifdef DEBUG_SET_PTE_AT
> {
> pgd_t *pgd_check = pgd_offset(mm, addr);
> pud_t *pud_check = pud_offset(pgd_check, addr);
> pmd_t *pmd_check = pmd_offset(pud_check, addr);
> pte_t *pte_check = pte_offset(pmd_check, addr);
>
> BUG_ON(pte_check != ptep);
> }
> #endif
> }
Good idea.
> BUG_ON() may hinder debugging if you hit a range of such
> incorrect calculations, so maybe a single printk which
> just prints __builtin_return_address(0) or current_text_addr()
> as well.
Nah, just xmon(NULL); :=)
Ben.
