[PATCH] PR target/106877: Robustify reg-stack to malformed asm.

[PATCH] PR target/106877: Robustify reg-stack to malformed asm.

[Roger Sayle]

[-- Attachment #1: Type: text/plain, Size: 866 bytes --]


This patch resolves PR target/106877 an ICE-on-invalid inline-asm
regression.  An innocent upstream change means that the test case
from PR inline-asm/84683 now hits a different assert in reg-stack.cc's
move_for_stack_reg.  Fixed by duplicating Jakub's solution to PR 84683
https://gcc.gnu.org/pipermail/gcc-patches/2018-March/495193.html
at this second (similar) gcc_assert.

This patch has been tested on x86_64-pc-linux-gnu with make bootstrap
and make -k check, both with and without --target_board=unix{-m32},
with no new failures.  Ok for mainline?


2022-09-13  Roger Sayle  <roger@nextmovesoftware.com>

gcc/ChangeLog
        PR target/106877
        * reg-stack.cc (move_for_stack_reg): Check for any_malformed_asm
        in gcc_assert.

gcc/testsuite/ChangeLog
        PR target/106877
        * g++.dg/ext/pr106877.C: New test case.


Thanks,
Roger
--


[-- Attachment #2: patchrs.txt --]
[-- Type: text/plain, Size: 1185 bytes --]

diff --git a/gcc/reg-stack.cc b/gcc/reg-stack.cc
index fd03250..95e0e61 100644
--- a/gcc/reg-stack.cc
+++ b/gcc/reg-stack.cc
@@ -1073,7 +1073,8 @@ move_for_stack_reg (rtx_insn *insn, stack_ptr regstack, rtx pat)
 	      break;
 
 	  /* The destination must be dead, or life analysis is borked.  */
-	  gcc_assert (get_hard_regnum (regstack, dest) < FIRST_STACK_REG);
+	  gcc_assert (get_hard_regnum (regstack, dest) < FIRST_STACK_REG
+		      || any_malformed_asm);
 
 	  /* If the source is not live, this is yet another case of
 	     uninitialized variables.  Load up a NaN instead.  */
diff --git a/gcc/testsuite/g++.dg/ext/pr106877.C b/gcc/testsuite/g++.dg/ext/pr106877.C
new file mode 100644
index 0000000..6bffed9
--- /dev/null
+++ b/gcc/testsuite/g++.dg/ext/pr106877.C
@@ -0,0 +1,13 @@
+// PR target/106877
+// { dg-do compile { target i?86-*-* x86_64-*-* } }
+// { dg-options "-O1 -m16 -mtune=sandybridge -flive-range-shrinkage -fno-dce" }
+
+void
+foo (float b, double c)
+{
+  for (int e = 0; e < 2; e++)
+    {
+      asm volatile ("" : "+f" (c));	// { dg-error "must specify a single register" }
+      asm ("" : "+rm" (c = b));
+    }
+}

Re: [PATCH] PR target/106877: Robustify reg-stack to malformed asm.

[Uros Bizjak]

On Tue, Sep 13, 2022 at 11:57 AM Roger Sayle <roger@nextmovesoftware.com> wrote:
>
>
> This patch resolves PR target/106877 an ICE-on-invalid inline-asm
> regression.  An innocent upstream change means that the test case
> from PR inline-asm/84683 now hits a different assert in reg-stack.cc's
> move_for_stack_reg.  Fixed by duplicating Jakub's solution to PR 84683
> https://gcc.gnu.org/pipermail/gcc-patches/2018-March/495193.html
> at this second (similar) gcc_assert.
>
> This patch has been tested on x86_64-pc-linux-gnu with make bootstrap
> and make -k check, both with and without --target_board=unix{-m32},
> with no new failures.  Ok for mainline?
>
>
> 2022-09-13  Roger Sayle  <roger@nextmovesoftware.com>
>
> gcc/ChangeLog
>         PR target/106877
>         * reg-stack.cc (move_for_stack_reg): Check for any_malformed_asm
>         in gcc_assert.
>
> gcc/testsuite/ChangeLog
>         PR target/106877
>         * g++.dg/ext/pr106877.C: New test case.

OK.

Thanks,
Uros.