qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Thomas Huth <thuth@redhat.com>
To: qemu-devel@nongnu.org
Cc: Peter Maydell <peter.maydell@linaro.org>,
	Richard Henderson <richard.henderson@linaro.org>,
	David Hildenbrand <david@redhat.com>
Subject: [PULL 2/7] tests/tcg/s390x: Fix the exrl-trt* tests with Clang
Date: Mon,  7 Mar 2022 19:26:04 +0100	[thread overview]
Message-ID: <20220307182609.94466-3-thuth@redhat.com> (raw)
In-Reply-To: <20220307182609.94466-1-thuth@redhat.com>

The exrl-trt* tests use two pre-initialized variables for the
results of the assembly code:

    uint64_t r1 = 0xffffffffffffffffull;
    uint64_t r2 = 0xffffffffffffffffull;

But then the assembly code copies over the full contents
of the register into the output variable, without taking
care of this pre-initialized values:

        "    lgr %[r1],%%r1\n"
        "    lgr %[r2],%%r2\n"

The code then finally compares the register contents to
a value that apparently depends on the pre-initialized values:

    if (r2 != 0xffffffffffffffaaull) {
        write(1, "bad r2\n", 7);
        return 1;
    }

This all works with GCC, since the 0xffffffffffffffff got into
the r2 register there by accident, but it fails completely with
Clang.

Let's fix this by declaring the r1 and r2 variables as proper
register variables instead, so the pre-initialized values get
correctly passed into the inline assembly code.

Message-Id: <20220301092431.1448419-1-thuth@redhat.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Thomas Huth <thuth@redhat.com>
---
 tests/tcg/s390x/exrl-trt.c  | 8 +++-----
 tests/tcg/s390x/exrl-trtr.c | 8 +++-----
 2 files changed, 6 insertions(+), 10 deletions(-)

diff --git a/tests/tcg/s390x/exrl-trt.c b/tests/tcg/s390x/exrl-trt.c
index 16711a3181..451f777b9d 100644
--- a/tests/tcg/s390x/exrl-trt.c
+++ b/tests/tcg/s390x/exrl-trt.c
@@ -5,8 +5,8 @@ int main(void)
 {
     char op1[] = "hello";
     char op2[256];
-    uint64_t r1 = 0xffffffffffffffffull;
-    uint64_t r2 = 0xffffffffffffffffull;
+    register uint64_t r1 asm("r1") = 0xffffffffffffffffull;
+    register uint64_t r2 asm("r2") = 0xffffffffffffffffull;
     uint64_t cc;
     int i;
 
@@ -21,8 +21,6 @@ int main(void)
         "    j 2f\n"
         "1:  trt 0(1,%[op1]),%[op2]\n"
         "2:  exrl %[op1_len],1b\n"
-        "    lgr %[r1],%%r1\n"
-        "    lgr %[r2],%%r2\n"
         "    ipm %[cc]\n"
         : [r1] "+r" (r1),
           [r2] "+r" (r2),
@@ -30,7 +28,7 @@ int main(void)
         : [op1] "a" (&op1),
           [op1_len] "a" (5),
           [op2] "Q" (op2)
-        : "r1", "r2", "cc");
+        : "cc");
     cc = (cc >> 28) & 3;
     if (cc != 2) {
         write(1, "bad cc\n", 7);
diff --git a/tests/tcg/s390x/exrl-trtr.c b/tests/tcg/s390x/exrl-trtr.c
index 5f30cda6bd..422f7f385a 100644
--- a/tests/tcg/s390x/exrl-trtr.c
+++ b/tests/tcg/s390x/exrl-trtr.c
@@ -5,8 +5,8 @@ int main(void)
 {
     char op1[] = {0, 1, 2, 3};
     char op2[256];
-    uint64_t r1 = 0xffffffffffffffffull;
-    uint64_t r2 = 0xffffffffffffffffull;
+    register uint64_t r1 asm("r1") = 0xffffffffffffffffull;
+    register uint64_t r2 asm("r2") = 0xffffffffffffffffull;
     uint64_t cc;
     int i;
 
@@ -21,8 +21,6 @@ int main(void)
         "    j 2f\n"
         "1:  trtr 3(1,%[op1]),%[op2]\n"
         "2:  exrl %[op1_len],1b\n"
-        "    lgr %[r1],%%r1\n"
-        "    lgr %[r2],%%r2\n"
         "    ipm %[cc]\n"
         : [r1] "+r" (r1),
           [r2] "+r" (r2),
@@ -30,7 +28,7 @@ int main(void)
         : [op1] "a" (&op1),
           [op1_len] "a" (3),
           [op2] "Q" (op2)
-        : "r1", "r2", "cc");
+        : "cc");
     cc = (cc >> 28) & 3;
     if (cc != 1) {
         write(1, "bad cc\n", 7);
-- 
2.27.0



  parent reply	other threads:[~2022-03-07 18:50 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-03-07 18:26 [PULL 0/7] s390x, tests and misc patches Thomas Huth
2022-03-07 18:26 ` [PULL 1/7] tests/tcg/s390x: Fix mvc, mvo and pack tests with Clang Thomas Huth
2022-03-07 18:26 ` Thomas Huth [this message]
2022-03-07 18:26 ` [PULL 3/7] tests/tcg/s390x: Cleanup of mie3 tests Thomas Huth
2022-03-07 18:26 ` [PULL 4/7] MAINTAINERS: Update the files in the FreeBSD section Thomas Huth
2022-03-07 18:26 ` [PULL 5/7] tests/avocado: Cancel BootLinux tests in case there is no free port Thomas Huth
2022-03-07 18:26 ` [PULL 6/7] tests/vm: Update haiku test vm to R1/Beta3 Thomas Huth
2022-03-07 18:26 ` [PULL 7/7] Check and report for incomplete 'global' option format Thomas Huth
2022-03-08 19:30 ` [PULL 0/7] s390x, tests and misc patches Peter Maydell

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=20220307182609.94466-3-thuth@redhat.com \
    --to=thuth@redhat.com \
    --cc=david@redhat.com \
    --cc=peter.maydell@linaro.org \
    --cc=qemu-devel@nongnu.org \
    --cc=richard.henderson@linaro.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).