On May 13, 2011, at 8:56 PM, Andrew Pinski wrote:
> On Fri, May 13, 2011 at 5:51 PM, Paul Koning wrote:
>> I was using 4.5.1 which has the bug. I thought later ones have it too, but
>> it looks like I misread the assembly output. (The fix makes a subtle
>> change: beq becomes beql.)
>
> A g
On Fri, May 13, 2011 at 5:51 PM, Paul Koning wrote:
> I was using 4.5.1 which has the bug. I thought later ones have it too, but
> it looks like I misread the assembly output. (The fix makes a subtle change:
> beq becomes beql.)
A good question is why does XLR have branch linkely enabled anyw
On May 13, 2011, at 5:29 PM, Eric Botcazou wrote:
>> rest_of_pass_free_cfg calls df_analyze but it doesn't call
>> df_note_add_problem. Is that the issue? I see that some other passes
>> (like regrename) do a sequence of df_xyz calls.
>
> It does now, you have outdated sources.
I was using 4.
> rest_of_pass_free_cfg calls df_analyze but it doesn't call
> df_note_add_problem. Is that the issue? I see that some other passes
> (like regrename) do a sequence of df_xyz calls.
It does now, you have outdated sources.
--
Eric Botcazou
On May 13, 2011, at 5:30 PM, Paul Koning wrote:
>
> On May 13, 2011, at 4:57 PM, Eric Botcazou wrote:
>
>>> Question: that code doesn't appear to touch REG_DEAD notes at all. Is that
>>> right? From the definition of REG_DEAD, it seems that rewriting a later
>>> insns to use register x means
> A later pass (machdep) relies on the notes, so the fact that the actual
> register liveness no longer matches what the REG_DEAD notes imply seems to
> be a problem.
Again, if a pass consumes these notes, it must df_note_add_problem() on entry.
For machine_reorg, this is too late because the CFG
On May 13, 2011, at 4:57 PM, Eric Botcazou wrote:
>> Question: that code doesn't appear to touch REG_DEAD notes at all. Is that
>> right? From the definition of REG_DEAD, it seems that rewriting a later
>> insns to use register x means that a preceding (REG_DEAD x) needs to be
>> deleted. Or m
On May 13, 2011, at 4:57 PM, Eric Botcazou wrote:
>> Question: that code doesn't appear to touch REG_DEAD notes at all. Is that
>> right? From the definition of REG_DEAD, it seems that rewriting a later
>> insns to use register x means that a preceding (REG_DEAD x) needs to be
>> deleted. Or m
> Question: that code doesn't appear to touch REG_DEAD notes at all. Is that
> right? From the definition of REG_DEAD, it seems that rewriting a later
> insns to use register x means that a preceding (REG_DEAD x) needs to be
> deleted. Or moved to the later insn, perhaps.
If the pass doesn't co
In attempting to find out what's wrong in PR 48990 (wrong code for MIPS N32
-O1) it looks to me like the issue is that a register is marked dead, then
still used afterwards. This causes the delay slot fill logic to make a wrong
decision (it brings a load instruction that overwrites the "dead" r
10 matches
Mail list logo
