From: Mikhail Maltsev <maltsevm@gmail.com>
To: gcc-patches mailing list <gcc-patches@gcc.gnu.org>,
Uros Bizjak <ubizjak@gmail.com>, Jeff Law <law@redhat.com>
Subject: [PATCH, i386] PR68497. Fix ICE with -fno-checking
Date: Mon, 23 Nov 2015 23:20:00 -0000 [thread overview]
Message-ID: <56539C9F.9000903@gmail.com> (raw)
[-- Attachment #1: Type: text/plain, Size: 603 bytes --]
Hi!
The attached patch fixes a problem introduced in r229567: the assertion
gcc_assert (is_sse);
is checked if flag_checking is false, and this causes an ICE when compiling with
-fno-checking.
Bootstrapped and regtested on x86_64-pc-linux-gnu. OK for trunk?
--
Regards,
Mikhail Maltsev
gcc/ChangeLog:
2015-11-23 Mikhail Maltsev <maltsevm@gmail.com>
PR target/68497
* config/i386/i386.c (output_387_binary_op): Fix assertion for
-fno-checking case.
gcc/testsuite/ChangeLog:
2015-11-23 Mikhail Maltsev <maltsevm@gmail.com>
PR target/68497
* gcc.target/i386/pr68497.c: New test.
[-- Attachment #2: pr68497.patch --]
[-- Type: text/x-patch, Size: 1715 bytes --]
diff --git a/gcc/config/i386/i386.c b/gcc/config/i386/i386.c
index 83749d5..23dbb3a 100644
--- a/gcc/config/i386/i386.c
+++ b/gcc/config/i386/i386.c
@@ -17675,18 +17675,20 @@ output_387_binary_op (rtx insn, rtx *operands)
/* Even if we do not want to check the inputs, this documents input
constraints. Which helps in understanding the following code. */
- if (flag_checking
- && STACK_REG_P (operands[0])
- && ((REG_P (operands[1])
- && REGNO (operands[0]) == REGNO (operands[1])
- && (STACK_REG_P (operands[2]) || MEM_P (operands[2])))
- || (REG_P (operands[2])
- && REGNO (operands[0]) == REGNO (operands[2])
- && (STACK_REG_P (operands[1]) || MEM_P (operands[1]))))
- && (STACK_TOP_P (operands[1]) || STACK_TOP_P (operands[2])))
- ; /* ok */
- else
- gcc_checking_assert (is_sse);
+ if (flag_checking)
+ {
+ if (STACK_REG_P (operands[0])
+ && ((REG_P (operands[1])
+ && REGNO (operands[0]) == REGNO (operands[1])
+ && (STACK_REG_P (operands[2]) || MEM_P (operands[2])))
+ || (REG_P (operands[2])
+ && REGNO (operands[0]) == REGNO (operands[2])
+ && (STACK_REG_P (operands[1]) || MEM_P (operands[1]))))
+ && (STACK_TOP_P (operands[1]) || STACK_TOP_P (operands[2])))
+ ; /* ok */
+ else
+ gcc_assert (is_sse);
+ }
switch (GET_CODE (operands[3]))
{
diff --git a/gcc/testsuite/gcc.target/i386/pr68497.c b/gcc/testsuite/gcc.target/i386/pr68497.c
new file mode 100644
index 0000000..0135cda
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/pr68497.c
@@ -0,0 +1,8 @@
+/* { dg-do compile } */
+/* { dg-options "-fno-checking" } */
+
+long double
+foo (long double x, long double y)
+{
+ return x + y;
+}
next reply other threads:[~2015-11-23 23:09 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-11-23 23:20 Mikhail Maltsev [this message]
2015-11-24 0:04 ` Bernd Schmidt
2015-11-24 12:20 ` Mikhail Maltsev
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=56539C9F.9000903@gmail.com \
--to=maltsevm@gmail.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=law@redhat.com \
--cc=ubizjak@gmail.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
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).