linux-sparse.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Luc Van Oostenryck <luc.vanoostenryck@gmail.com>
To: linux-sparse@vger.kernel.org
Cc: Luc Van Oostenryck <luc.vanoostenryck@gmail.com>
Subject: [PATCH 1/4] ssa: add some testcases for mismatched memops
Date: Wed, 10 Mar 2021 00:42:29 +0100	[thread overview]
Message-ID: <20210309234232.50889-2-luc.vanoostenryck@gmail.com> (raw)
In-Reply-To: <20210309234232.50889-1-luc.vanoostenryck@gmail.com>

The SSA conversion is incorrect when the size or offset of the
memory operations doesn't match. It shouldn't be done at all.

So, add a few testcases for this.

Signed-off-by: Luc Van Oostenryck <luc.vanoostenryck@gmail.com>
---
 validation/mem2reg/not-same-memop0.c | 49 ++++++++++++++++++++++++++++
 validation/mem2reg/packed-bitfield.c | 36 ++++++++++++++++++++
 2 files changed, 85 insertions(+)
 create mode 100644 validation/mem2reg/not-same-memop0.c
 create mode 100644 validation/mem2reg/packed-bitfield.c

diff --git a/validation/mem2reg/not-same-memop0.c b/validation/mem2reg/not-same-memop0.c
new file mode 100644
index 000000000000..7a84ddce4efc
--- /dev/null
+++ b/validation/mem2reg/not-same-memop0.c
@@ -0,0 +1,49 @@
+struct s {
+	int:16;
+	short f:6;
+};
+
+static short local(struct s s)
+{
+	return s.f;
+}
+
+static void foo(struct s s)
+{
+	while (s.f) ;
+}
+
+/*
+ * check-name: not-same-memop0
+ * check-command: test-linearize -Wno-decl -fdump-ir=mem2reg $file
+ * check-known-to-fail
+ *
+ * check-output-start
+local:
+.L0:
+	<entry-point>
+	store.32    %arg1 -> 0[s]
+	load.16     %r1 <- 2[s]
+	trunc.6     %r2 <- (16) %r1
+	sext.16     %r3 <- (6) %r2
+	ret.16      %r3
+
+
+foo:
+.L2:
+	<entry-point>
+	store.32    %arg1 -> 0[s]
+	br          .L6
+
+.L6:
+	load.16     %r5 <- 2[s]
+	trunc.6     %r6 <- (16) %r5
+	setne.1     %r7 <- %r6, $0
+	cbr         %r7, .L6, .L5
+
+.L5:
+	ret
+
+
+ * check-output-end
+ */
diff --git a/validation/mem2reg/packed-bitfield.c b/validation/mem2reg/packed-bitfield.c
new file mode 100644
index 000000000000..4eaf0befeaf5
--- /dev/null
+++ b/validation/mem2reg/packed-bitfield.c
@@ -0,0 +1,36 @@
+struct s {
+	int:16;
+	int f:16;
+} __attribute__((__packed__));
+
+static void foo(struct s s)
+{
+	while (s.f)
+		;
+}
+
+/*
+ * check-name: packed-bitfield
+ * check-command: test-linearize -fmem2reg $file
+ * check-known-to-fail
+ *
+ * check-output-contains: store.32
+ * check-output-contains: load.16
+ *
+ * check-output-start
+foo:
+.L0:
+	<entry-point>
+	store.32    %arg1 -> 0[s]
+	br          .L4
+
+.L4:
+	load.16     %r1 <- 2[s]
+	cbr         %r1, .L4, .L3
+
+.L3:
+	ret
+
+
+ * check-output-end
+ */
-- 
2.30.0


  reply	other threads:[~2021-03-09 23:43 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-03-09 23:42 [PATCH 0/4] fix SSA conversion of mismatched memops Luc Van Oostenryck
2021-03-09 23:42 ` Luc Van Oostenryck [this message]
2021-03-09 23:42 ` [PATCH 2/4] ssa: the sparse set is not needed Luc Van Oostenryck
2021-03-09 23:42 ` [PATCH 3/4] ssa: avoid SSA conversion of packed bitfields Luc Van Oostenryck
2021-03-09 23:42 ` [PATCH 4/4] ssa: fix conversion with mismatched size or offset Luc Van Oostenryck

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=20210309234232.50889-2-luc.vanoostenryck@gmail.com \
    --to=luc.vanoostenryck@gmail.com \
    --cc=linux-sparse@vger.kernel.org \
    /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 NNTP newsgroup(s).