Re: [Qemu-devel] Minor MMU fixes for PowerPC 40x emulation

[Edgar E. Iglesias]

On Sat, Oct 02, 2010 at 06:55:36PM +0200, Alexander Graf wrote:
> 
> On 02.10.2010, at 18:49, John Clark wrote:
> 
> >>>            /* Check from TLB entry */
> >>> -            /* XXX: there is a problem here or in the TLB fill code... */
> >>> +            /* There is no longer a need to force PAGE_EXEC permission 
> >>> here */
> >>> +            /* because of the tlb->attr fix in helper_4xx_tlbwe_lo() */
> >> 
> >> I guess that comment is superfluous, as readers several years from now 
> >> don't care what was broken back in the day :).
> > 
> > Yes, I suppose so :)
> > 
> >>> @@ -3939,7 +3939,7 @@ target_ulong helper_4xx_tlbre_lo (target_ulong 
> >>> entry)
> >>>    tlb = &env->tlb[entry].tlbe;
> >>>    ret = tlb->EPN;
> >>>    if (tlb->prot & PAGE_VALID)
> >>> -        ret |= 0x400;
> >>> +        ret |= 0x40;    /* V bit is 0x40, not 0x400 */
> >> 
> >> Ouch. Mind to make it a define?
> > 
> > Sure, I was surprised that there wasn't a define for that when I found it.
> 
> The ppc emulation code lacks a lot of defines. In fact, the same goes for x86 
> emulation too ;). But that doesn't mean we have to keep it that way!
> 
> > 
> >>>    size = booke_page_size_to_tlb(tlb->size);
> >>>    if (size < 0 || size > 0x7)
> >>>        size = 1;
> >>> @@ -3948,7 +3948,7 @@ target_ulong helper_4xx_tlbre_lo (target_ulong 
> >>> entry)
> >>>    return ret;
> >>> }
> >>> 
> >>> -target_ulong helper_4xx_tlbre_hi (target_ulong entry)
> >>> +target_ulong helper_4xx_tlbre_lo (target_ulong entry)
> >> 
> >> Huh?
> > 
> > To summarize, 'tlbre' has two forms: one to retrieve the high bits of
> > a TLB entry (TLBHI), and one to retrieve the low bits (TLBLO) of a TLB
> > entry.  This code had the TLBLO form returning the bits corresponding
> > to TLBHI and vice versa, hence the name change.  You can verify this
> > if you like with this IBM PowerPC 405 core user manual on page 362:
> 
> Well the thing that strikes me as weird is mostly that you're changing a 
> function name, but no callers to it. So is this function never used? Or was 
> tlbre_lo defined before already and is now defined twice?

Hi,

Alex:
I think you've missed the part of the patch that renames the _lo -> _hi.
As John says, qemu had the hi/lo parts reversed when reading 4xx TLB regs.

Except for the comments and the define, the patch looks good to me.
John, please also add a Signed-off-by line.

Cheers