Mark Shinwell wrote:
Bernd Schmidt wrote:
Mark Shinwell wrote:
Do you think it should be the case that, at the point below, _any_
reload
with reg_rtx corresponding to a hard register should have the relevant
bit set in reload_spill_index?
I think so. I'm attaching a patch below. It appears
Bernd Schmidt wrote:
Mark Shinwell wrote:
Do you think it should be the case that, at the point below, _any_ reload
with reg_rtx corresponding to a hard register should have the relevant
bit set in reload_spill_index?
I think so. I'm attaching a patch below. It appears to have no effect
on a
Mark Mitchell wrote:
Bernd Schmidt wrote:
Mark Shinwell wrote:
Do you think it should be the case that, at the point below, _any_ reload
with reg_rtx corresponding to a hard register should have the relevant
bit set in reload_spill_index?
I think so. I'm attaching a patch below. It appears t
Bernd Schmidt wrote:
> Mark Shinwell wrote:
>> Do you think it should be the case that, at the point below, _any_ reload
>> with reg_rtx corresponding to a hard register should have the relevant
>> bit set in reload_spill_index?
>
> I think so. I'm attaching a patch below. It appears to have no
Mark Shinwell wrote:
> Do you think it should be the case that, at the point below, _any_ reload
> with reg_rtx corresponding to a hard register should have the relevant
> bit set in reload_spill_index?
I think so. I'm attaching a patch below. It appears to have no effect
on all code I've tried
On Mon, Jun 11, 2007 at 09:22:24AM +0100, Mark Shinwell wrote:
> + if (rld[r].when_needed == RELOAD_FOR_INPUT
> + && rld[r].reg_rtx
> + && REGNO (rld[r].reg_rtx) < FIRST_PSEUDO_REGISTER)
> + {
$ grep -F -e HARD_REGISTER gcc/rtl.h
#define HARD_REGISTER_P(R
Bernd Schmidt wrote:
Mark Shinwell wrote:
and the code following in emit_reload_insns? Perhaps if spill_reg_index
took account of registers selected during find_reloads then this could
be simplified too.
So what do you think is the best approach to fix all of this? :-)
Sounds like you gave
Bernd Schmidt wrote:
Mark Shinwell wrote:
and the code following in emit_reload_insns? Perhaps if spill_reg_index
took account of registers selected during find_reloads then this could
be simplified too.
So what do you think is the best approach to fix all of this? :-)
Sounds like you gave
Mark Shinwell wrote:
> and the code following in emit_reload_insns? Perhaps if spill_reg_index
> took account of registers selected during find_reloads then this could
> be simplified too.
>
> So what do you think is the best approach to fix all of this? :-)
Sounds like you gave the answer your
Bernd Schmidt wrote:
It appears that spill_reg_index is only set up for registers that go
through the choose_reload_regs process, not for the ones selected during
find_reloads. That's probably a bad thing, as a reg_rtx chosen in
find_reloads could be used as a spill reg in a previous insn (as in
Mark Shinwell wrote:
> This code is guarded by
>
> /* I is nonneg if this reload used a register.
> If rld[r].reg_rtx is 0, this is an optional reload
> that we opted to ignore. */
>
> if (i >= 0 && rld[r].reg_rtx != 0)
>
> and in this [my original] case, i is negative (se
Bernd Schmidt wrote:
My gut feeling is that this example will work as a consequence.
... note that I inserted which could conceivably use
R9 as an input reload, as the hard reg is dead. Where would we
invalidate previous information about R9? I assume it would be the loop
at the end of emit
Mark Shinwell wrote:
> As you say, one unusual thing about this situation must be the fact
> that the reload register is getting chosen by the code in
> push_reload heralded by "If this is an input reload and the operand
> contains a register that dies in this insn and is used nowhere else,
> see i
Bernd Schmidt wrote:
insn 5301: (set (reg/f:SI 4082) (reg/f:SI 3275))
insn 5291 (set (reg:DF 4078])
(mem/s:DF (plus:SI (reg/f:SI 3275) (reg:SI 3812
REG_DEAD 3275
insn 5314 (set (reg:DF 4096)
(mem/s:DF (plus:SI (reg/f:SI 4082) (reg:SI 4084
After reload we end up
Mark Shinwell wrote:
> The relevant RTL instructions before reload are as follows. These
> correspond to points A, B and C respectively in my previous email.
I must admit I'm still stumbling in the dark a bit - this would be so
much easier to digest with a testcase.
The question I'm trying to an
On Tue, May 15, 2007 at 04:27:21PM +0100, Mark Shinwell wrote:
> Part of the reason for starting this thread was that I was concerned
> about invalidating reloads that could be re-used later. However, it
> seems to me that in every circumstance where the reload register is a
> hard register and t
Bernd Schmidt wrote:
Mark Shinwell wrote:
These dumps are of course taken before the application of my patch.
Hope that helps,
Thanks. I may have missed it in the previous mails, but which piece of
code exactly decides to use R9 for reload 0 of insn 5314?
I'm afraid I'm not sure exactly --
Rask Ingemann Lambertsen wrote:
On Mon, May 14, 2007 at 10:47:13PM +0100, Mark Shinwell wrote:
I'm fairly certain that this is the correct approach to fix this
issue, but I'm less certain that I have correctly guarded the call
to forget_old_reloads_1,
[snip]
Index: reload1.c
On Mon, May 14, 2007 at 10:47:13PM +0100, Mark Shinwell wrote:
> I'm fairly certain that this is the correct approach to fix this
> issue, but I'm less certain that I have correctly guarded the call
> to forget_old_reloads_1,
[snip]
> Index: reload1.c
> ===
Mark Shinwell wrote:
> These dumps are of course taken before the application of my patch.
>
> Hope that helps,
Thanks. I may have missed it in the previous mails, but which piece of
code exactly decides to use R9 for reload 0 of insn 5314?
Bernd
--
This footer brought to you by insane German
Bernd Schmidt wrote:
Mark Shinwell wrote:
The bug is currently only exhibiting itself on a proprietary testcase
when compiled for an ARM target and is extremely sensitive to the
particular code involved. It arises as follows, using the same notation
as in Richard's mail:
If you can't post the
Rask Ingemann Lambertsen wrote:
On Mon, May 14, 2007 at 10:47:13PM +0100, Mark Shinwell wrote:
I'm fairly certain that this is the correct approach to fix this
issue, but I'm less certain that I have correctly guarded the call
to forget_old_reloads_1, and indeed that I've done everything to
era
On Mon, May 14, 2007 at 10:47:13PM +0100, Mark Shinwell wrote:
> I'm fairly certain that this is the correct approach to fix this
> issue, but I'm less certain that I have correctly guarded the call
> to forget_old_reloads_1, and indeed that I've done everything to
> eradicate the previous reloads
Mark Shinwell wrote:
> The bug is currently only exhibiting itself on a proprietary testcase
> when compiled for an ARM target and is extremely sensitive to the
> particular code involved. It arises as follows, using the same notation
> as in Richard's mail:
If you can't post the testcase, the be
Rask Ingemann Lambertsen wrote:
On Tue, May 15, 2007 at 09:31:10AM +0100, Mark Shinwell wrote:
Rask Ingemann Lambertsen wrote:
On Mon, May 14, 2007 at 10:47:13PM +0100, Mark Shinwell wrote:
[snip]
- the last use of reg2 (in B) is inside a matched input operand;
[snip]
The reload used for
On Tue, May 15, 2007 at 09:31:10AM +0100, Mark Shinwell wrote:
> Rask Ingemann Lambertsen wrote:
> >On Mon, May 14, 2007 at 10:47:13PM +0100, Mark Shinwell wrote:
> >
> >[snip]
> >>- the last use of reg2 (in B) is inside a matched input operand;
> >[snip]
> >>The reload used for the instruction
Rask Ingemann Lambertsen wrote:
On Mon, May 14, 2007 at 10:47:13PM +0100, Mark Shinwell wrote:
[snip]
- the last use of reg2 (in B) is inside a matched input operand;
[snip]
The reload used for the instruction at B looks like this:
Reload 0: reload_in (SI) = (plus:SI (reg/f:SI 9 r9 [3275
On Mon, May 14, 2007 at 10:47:13PM +0100, Mark Shinwell wrote:
[snip]
> - the last use of reg2 (in B) is inside a matched input operand;
[snip]
>
> The reload used for the instruction at B looks like this:
>
> Reload 0: reload_in (SI) = (plus:SI (reg/f:SI 9 r9 [3275])
>
> In order to fix this I think that reload1.c:emit_reload_insns should,
> at the point of discovery of an input reload whose reload register is a
> non-spill hard register H, invalidate all previous reloads involving
> that hard register.
If the reload is inherited, and there is no non-input part
29 matches
Mail list logo
