* [PATCH] PR target/89828 Inernal compiler error on -fno-omit-frame-pointer
@ 2023-01-05 13:05 Yoshinori Sato
2023-03-11 16:59 ` Jeff Law
0 siblings, 1 reply; 5+ messages in thread
From: Yoshinori Sato @ 2023-01-05 13:05 UTC (permalink / raw)
To: gcc-patches; +Cc: Yoshinori Sato
The problem was caused by an erroneous note about creating a stack frame,
which caused the cur_cfa reg to fail to assert with a value other than
the frame pointer.
This fix will generate notes that correctly update cur_cfa.
gcc/config/rx/
* rx.cc (add_pop_cfi_notes): Release the frame pointer if it is used.
(rx_expand_prologue): Redesigned stack pointer and frame pointer update process.
Signed-off-by: Yoshinori Sato <ysato@users.sourceforge.jp>
---
gcc/config/rx/rx.cc | 50 +++++++++++++++------------------------------
1 file changed, 17 insertions(+), 33 deletions(-)
diff --git a/gcc/config/rx/rx.cc b/gcc/config/rx/rx.cc
index 412a3a354b0..8c246d42cd8 100644
--- a/gcc/config/rx/rx.cc
+++ b/gcc/config/rx/rx.cc
@@ -1647,16 +1647,20 @@ mark_frame_related (rtx insn)
static void
add_pop_cfi_notes (rtx_insn *insn, unsigned int high, unsigned int low)
{
- rtx t = plus_constant (Pmode, stack_pointer_rtx,
- (high - low + 1) * UNITS_PER_WORD);
+ rtx src = stack_pointer_rtx;
+ rtx t;
+ for (unsigned int i = low; i <= high; i++)
+ {
+ add_reg_note (insn, REG_CFA_RESTORE, gen_rtx_REG (word_mode, i));
+ if (i == FRAME_POINTER_REGNUM)
+ src = frame_pointer_rtx;
+ }
+ t = plus_constant (Pmode, src, (high - low + 1) * UNITS_PER_WORD);
t = gen_rtx_SET (stack_pointer_rtx, t);
add_reg_note (insn, REG_CFA_ADJUST_CFA, t);
RTX_FRAME_RELATED_P (insn) = 1;
- for (unsigned int i = low; i <= high; i++)
- add_reg_note (insn, REG_CFA_RESTORE, gen_rtx_REG (word_mode, i));
}
-
static bool
ok_for_max_constant (HOST_WIDE_INT val)
{
@@ -1815,37 +1819,17 @@ rx_expand_prologue (void)
}
}
- /* If needed, set up the frame pointer. */
- if (frame_pointer_needed)
- gen_safe_add (frame_pointer_rtx, stack_pointer_rtx,
- GEN_INT (- (HOST_WIDE_INT) frame_size), true);
-
- /* Allocate space for the outgoing args.
- If the stack frame has not already been set up then handle this as well. */
- if (stack_size)
+ if (stack_size || frame_size)
{
- if (frame_size)
- {
- if (frame_pointer_needed)
- gen_safe_add (stack_pointer_rtx, frame_pointer_rtx,
- GEN_INT (- (HOST_WIDE_INT) stack_size), true);
- else
- gen_safe_add (stack_pointer_rtx, stack_pointer_rtx,
- GEN_INT (- (HOST_WIDE_INT) (frame_size + stack_size)),
- true);
- }
- else
- gen_safe_add (stack_pointer_rtx, stack_pointer_rtx,
- GEN_INT (- (HOST_WIDE_INT) stack_size), true);
+ gen_safe_add (stack_pointer_rtx, stack_pointer_rtx,
+ GEN_INT (- (HOST_WIDE_INT) (stack_size + frame_size)),
+ true);
}
- else if (frame_size)
+ if (frame_pointer_needed)
{
- if (! frame_pointer_needed)
- gen_safe_add (stack_pointer_rtx, stack_pointer_rtx,
- GEN_INT (- (HOST_WIDE_INT) frame_size), true);
- else
- gen_safe_add (stack_pointer_rtx, frame_pointer_rtx, NULL_RTX,
- false /* False because the epilogue will use the FP not the SP. */);
+ gen_safe_add (frame_pointer_rtx, stack_pointer_rtx,
+ GEN_INT ((HOST_WIDE_INT) stack_size),
+ true);
}
}
--
2.30.2
^ permalink raw reply [flat|nested] 5+ messages in thread* Re: [PATCH] PR target/89828 Inernal compiler error on -fno-omit-frame-pointer
2023-01-05 13:05 [PATCH] PR target/89828 Inernal compiler error on -fno-omit-frame-pointer Yoshinori Sato
@ 2023-03-11 16:59 ` Jeff Law
2023-03-15 7:51 ` [PATCH v2] " Yoshinori Sato
0 siblings, 1 reply; 5+ messages in thread
From: Jeff Law @ 2023-03-11 16:59 UTC (permalink / raw)
To: Yoshinori Sato, gcc-patches
On 1/5/23 06:05, Yoshinori Sato wrote:
> Subject:
> [PATCH] PR target/89828 Inernal compiler error on -fno-omit-frame-pointer
> From:
> Yoshinori Sato <ysato@users.sourceforge.jp>
> Date:
> 1/5/23, 06:05
>
> To:
> gcc-patches@gcc.gnu.org
> CC:
> Yoshinori Sato <ysato@users.sourceforge.jp>
>
>
> The problem was caused by an erroneous note about creating a stack frame,
> which caused the cur_cfa reg to fail to assert with a value other than
> the frame pointer.
>
> This fix will generate notes that correctly update cur_cfa.
>
> gcc/config/rx/
> * rx.cc (add_pop_cfi_notes): Release the frame pointer if it is used.
> (rx_expand_prologue): Redesigned stack pointer and frame pointer update process.
This seems to be taking a step backwards in that it's causing multiple
internal errors when compiling newlib.
The failures are here:
static void
dwarf2out_frame_debug_adjust_cfa (rtx pat)
{
rtx src, dest;
gcc_assert (GET_CODE (pat) == SET);
dest = XEXP (pat, 0);
src = XEXP (pat, 1);
switch (GET_CODE (src))
{
case PLUS:
gcc_assert (cur_cfa->reg == XEXP (src, 0));
cur_cfa->offset -= rtx_to_poly_int64 (XEXP (src, 1));
break;
case REG:
break;
default:
gcc_unreachable ();
}
cur_cfa->reg = dwf_cfa_reg (dest);
gcc_assert (cur_cfa->indirect == 0);
}
In particular it's tripping the gcc_assert in the PLUS case above.
Which is a pretty good indicator the contents of the note and the insn's
pattern are different WRT the expected registers.
jeff
^ permalink raw reply [flat|nested] 5+ messages in thread* [PATCH v2] PR target/89828 Inernal compiler error on -fno-omit-frame-pointer
2023-03-11 16:59 ` Jeff Law
@ 2023-03-15 7:51 ` Yoshinori Sato
2023-03-15 22:07 ` Jeff Law
2023-03-19 16:48 ` Jeff Law
0 siblings, 2 replies; 5+ messages in thread
From: Yoshinori Sato @ 2023-03-15 7:51 UTC (permalink / raw)
To: gcc-patches; +Cc: Yoshinori Sato
What about this?
It no longer occurs for me.
gcc/config/rx/
* rx.cc (add_pop_cfi_notes): Release the frame pointer if it is used.
(rx_expand_prologue): Redesigned stack pointer and frame pointer update process.
Signed-off-by: Yoshinori Sato <ysato@users.sourceforge.jp>
---
gcc/config/rx/rx.cc | 50 +++++++++++++++------------------------------
1 file changed, 17 insertions(+), 33 deletions(-)
diff --git a/gcc/config/rx/rx.cc b/gcc/config/rx/rx.cc
index 412a3a354b0..51ef4a6be34 100644
--- a/gcc/config/rx/rx.cc
+++ b/gcc/config/rx/rx.cc
@@ -1647,16 +1647,20 @@ mark_frame_related (rtx insn)
static void
add_pop_cfi_notes (rtx_insn *insn, unsigned int high, unsigned int low)
{
- rtx t = plus_constant (Pmode, stack_pointer_rtx,
- (high - low + 1) * UNITS_PER_WORD);
+ rtx src = stack_pointer_rtx;
+ rtx t;
+ for (unsigned int i = low; i <= high; i++)
+ {
+ add_reg_note (insn, REG_CFA_RESTORE, gen_rtx_REG (word_mode, i));
+ if (i == FRAME_POINTER_REGNUM && frame_pointer_needed)
+ src = frame_pointer_rtx;
+ }
+ t = plus_constant (Pmode, src, (high - low + 1) * UNITS_PER_WORD);
t = gen_rtx_SET (stack_pointer_rtx, t);
add_reg_note (insn, REG_CFA_ADJUST_CFA, t);
RTX_FRAME_RELATED_P (insn) = 1;
- for (unsigned int i = low; i <= high; i++)
- add_reg_note (insn, REG_CFA_RESTORE, gen_rtx_REG (word_mode, i));
}
-
static bool
ok_for_max_constant (HOST_WIDE_INT val)
{
@@ -1815,37 +1819,17 @@ rx_expand_prologue (void)
}
}
- /* If needed, set up the frame pointer. */
- if (frame_pointer_needed)
- gen_safe_add (frame_pointer_rtx, stack_pointer_rtx,
- GEN_INT (- (HOST_WIDE_INT) frame_size), true);
-
- /* Allocate space for the outgoing args.
- If the stack frame has not already been set up then handle this as well. */
- if (stack_size)
+ if (stack_size || frame_size)
{
- if (frame_size)
- {
- if (frame_pointer_needed)
- gen_safe_add (stack_pointer_rtx, frame_pointer_rtx,
- GEN_INT (- (HOST_WIDE_INT) stack_size), true);
- else
- gen_safe_add (stack_pointer_rtx, stack_pointer_rtx,
- GEN_INT (- (HOST_WIDE_INT) (frame_size + stack_size)),
- true);
- }
- else
- gen_safe_add (stack_pointer_rtx, stack_pointer_rtx,
- GEN_INT (- (HOST_WIDE_INT) stack_size), true);
+ gen_safe_add (stack_pointer_rtx, stack_pointer_rtx,
+ GEN_INT (- (HOST_WIDE_INT) (stack_size + frame_size)),
+ true);
}
- else if (frame_size)
+ if (frame_pointer_needed)
{
- if (! frame_pointer_needed)
- gen_safe_add (stack_pointer_rtx, stack_pointer_rtx,
- GEN_INT (- (HOST_WIDE_INT) frame_size), true);
- else
- gen_safe_add (stack_pointer_rtx, frame_pointer_rtx, NULL_RTX,
- false /* False because the epilogue will use the FP not the SP. */);
+ gen_safe_add (frame_pointer_rtx, stack_pointer_rtx,
+ GEN_INT ((HOST_WIDE_INT) stack_size),
+ true);
}
}
--
2.30.2
^ permalink raw reply [flat|nested] 5+ messages in thread* Re: [PATCH v2] PR target/89828 Inernal compiler error on -fno-omit-frame-pointer
2023-03-15 7:51 ` [PATCH v2] " Yoshinori Sato
@ 2023-03-15 22:07 ` Jeff Law
2023-03-19 16:48 ` Jeff Law
1 sibling, 0 replies; 5+ messages in thread
From: Jeff Law @ 2023-03-15 22:07 UTC (permalink / raw)
To: Yoshinori Sato, gcc-patches
On 3/15/23 01:51, Yoshinori Sato wrote:
> What about this?
> It no longer occurs for me.
>
> gcc/config/rx/
> * rx.cc (add_pop_cfi_notes): Release the frame pointer if it is used.
> (rx_expand_prologue): Redesigned stack pointer and frame pointer update process.
That fixes the problems building newlib. It also fixes roughly 300
failures in the testsuite (primarily fixing ICEs in
dwarf2out_frame_debug_adjust_cfa).
So the next step is to sit down with the code and review it :-)
jeff
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH v2] PR target/89828 Inernal compiler error on -fno-omit-frame-pointer
2023-03-15 7:51 ` [PATCH v2] " Yoshinori Sato
2023-03-15 22:07 ` Jeff Law
@ 2023-03-19 16:48 ` Jeff Law
1 sibling, 0 replies; 5+ messages in thread
From: Jeff Law @ 2023-03-19 16:48 UTC (permalink / raw)
To: Yoshinori Sato, gcc-patches
On 3/15/23 01:51, Yoshinori Sato wrote:
> What about this?
> It no longer occurs for me.
>
> gcc/config/rx/
> * rx.cc (add_pop_cfi_notes): Release the frame pointer if it is used.
> (rx_expand_prologue): Redesigned stack pointer and frame pointer update process.
So I think the ChangeLog entry needs a bit of work. I don't see how
the ChangeLog entry for add_pop_cfi_notes relates to the changes at all.
This might be better:
* config/rx/rx.cc (add_pop_cfi_notes): Attach CFA_RESTORE notes
first, then the CFA_ADJUST_CFA note. If restoring the frame
pointer, use (fp + offset) for the CFA_ADJUST_CFA note.
> @@ -1815,37 +1819,17 @@ rx_expand_prologue (void)
> }
> }
>
> - /* If needed, set up the frame pointer. */
> - if (frame_pointer_needed)
> - gen_safe_add (frame_pointer_rtx, stack_pointer_rtx,
> - GEN_INT (- (HOST_WIDE_INT) frame_size), true);
> -
> - /* Allocate space for the outgoing args.
> - If the stack frame has not already been set up then handle this as well. */
> - if (stack_size)
> + if (stack_size || frame_size)
> {
> - if (frame_size)
> - {
> - if (frame_pointer_needed)
> - gen_safe_add (stack_pointer_rtx, frame_pointer_rtx,
> - GEN_INT (- (HOST_WIDE_INT) stack_size), true);
> - else
> - gen_safe_add (stack_pointer_rtx, stack_pointer_rtx,
> - GEN_INT (- (HOST_WIDE_INT) (frame_size + stack_size)),
> - true);
> - }
> - else
> - gen_safe_add (stack_pointer_rtx, stack_pointer_rtx,
> - GEN_INT (- (HOST_WIDE_INT) stack_size), true);
> + gen_safe_add (stack_pointer_rtx, stack_pointer_rtx,
> + GEN_INT (- (HOST_WIDE_INT) (stack_size + frame_size)),
> + true);
> }
> - else if (frame_size)
> + if (frame_pointer_needed)
> {
> - if (! frame_pointer_needed)
> - gen_safe_add (stack_pointer_rtx, stack_pointer_rtx,
> - GEN_INT (- (HOST_WIDE_INT) frame_size), true);
> - else
> - gen_safe_add (stack_pointer_rtx, frame_pointer_rtx, NULL_RTX,
> - false /* False because the epilogue will use the FP not the SP. */);
> + gen_safe_add (frame_pointer_rtx, stack_pointer_rtx,
> + GEN_INT ((HOST_WIDE_INT) stack_size),
> + true);
It looks like we're emitting;
(set (sp) (plus (sp) (stack_size + frame_size)
Then we emit
(set (fp) (plus (sp) (stack_size))
Unless I missing something important, that seems wrong and results in
stack_size being added to FP twice.
jeff
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2023-03-19 16:48 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-01-05 13:05 [PATCH] PR target/89828 Inernal compiler error on -fno-omit-frame-pointer Yoshinori Sato
2023-03-11 16:59 ` Jeff Law
2023-03-15 7:51 ` [PATCH v2] " Yoshinori Sato
2023-03-15 22:07 ` Jeff Law
2023-03-19 16:48 ` Jeff Law
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).