public inbox for gdb-cvs@sourceware.org
help / color / mirror / Atom feed
* [binutils-gdb] Fix SVE-related failure in gdb.arch/aarch64-fp.exp
@ 2020-02-28 15:25 Luis Machado
  0 siblings, 0 replies; only message in thread
From: Luis Machado @ 2020-02-28 15:25 UTC (permalink / raw)
  To: gdb-cvs

https://sourceware.org/git/gitweb.cgi?p=binutils-gdb.git;h=f7a7000d486f6fc456e9332214b89d01c3639fb1

commit f7a7000d486f6fc456e9332214b89d01c3639fb1
Author: Luis Machado <luis.machado@linaro.org>
Date:   Wed Feb 26 10:37:28 2020 -0300

    Fix SVE-related failure in gdb.arch/aarch64-fp.exp
    
    The gdb.arch/aarch64-fp.exp test assumes it is dealing with a regular SIMD
    target that exposes the V registers as raw registers.  SVE-enabled targets
    turn the V registers into pseudo-registers.
    
    That is all fine, but the testcase uses the "info registers" command, which
    prints pseudo-register's contents twice. One for the hex format and another
    for the natural format of the type.
    
    (gdb) info registers v0
    v0             {d = {f = {0x0, 0x0}, u = {0x1716151413121110, 0x1f1e1d1c1b1a1918}, s = {0x1716151413121110, 0x1f1e1d1c1b1a1918}}, s = {f = {0x0, 0x0, 0x0, 0x0}, u = {0x13121110, 0x17161514, 0x1b1a1918, 0x1f1e1d1c}, s = {0x13121110, 0x17161514, 0x1b1a1918, 0x1f1e1d1c}}, h = {f = {0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}, u = {0x1110, 0x1312, 0x1514, 0x1716, 0x1918, 0x1b1a, 0x1d1c, 0x1f1e}, s = {0x1110, 0x1312, 0x1514, 0x1716, 0x1918, 0x1b1a, 0x1d1c, 0x1f1e}}, b = {u = {0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17, 0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f}, s = {0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17, 0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f}}, q = {u = {0x1f1e1d1c1b1a19181716151413121110}, s = {0x1f1e1d1c1b1a19181716151413121110}}} {d = {f = {1.846323925681849e-197, 8.5677456166123577e-159}, u = {1663540288323457296, 2242261671028070680}, s = {1663540288323457296, 2242261671028070680}}, s = {f = {1.84362032e-27, 4.84942184e-25, 1.27466897e-22, 3.34818801e-20}, u = {319951120, 387323156, 454695192, 522067228}, s = {319951120, 387323156, 454695192, 522067228}}, h = {f = {0.00061798, 0.00086308, 0.0012398, 0.00173, 0.0024872, 0.0034676, 0.0049896, 0.0069504}, u = {4368, 4882, 5396, 5910, 6424, 6938, 7452, 7966}, s = {4368, 4882, 5396, 5910, 6424, 6938, 7452, 7966}}, b = {u = {16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31}, s = {16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31}}, q = {u = {41362427191743139026751447860679676176}, s = {41362427191743139026751447860679676176}}}
    
    (gdb) p/x $v0
    $1 = {d = {f = {0x0, 0x0}, u = {0x1716151413121110, 0x1f1e1d1c1b1a1918}, s = {0x1716151413121110, 0x1f1e1d1c1b1a1918}}, s = {f = {0x0, 0x0, 0x0, 0x0}, u = {0x13121110, 0x17161514, 0x1b1a1918, 0x1f1e1d1c}, s = {0x13121110, 0x17161514, 0x1b1a1918, 0x1f1e1d1c}}, h = {f = {0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}, u = {0x1110, 0x1312, 0x1514, 0x1716, 0x1918, 0x1b1a, 0x1d1c, 0x1f1e}, s = {0x1110, 0x1312, 0x1514, 0x1716, 0x1918, 0x1b1a, 0x1d1c, 0x1f1e}}, b = {u = {0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17, 0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f}, s = {0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17, 0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f}}, q = {u = {0x1f1e1d1c1b1a19181716151413121110}, s = {0x1f1e1d1c1b1a19181716151413121110}}}
    
    Since the testcase is not expecting that, we run into a couple failures:
    
    FAIL: gdb.arch/aarch64-fp.exp: check register v0 value
    FAIL: gdb.arch/aarch64-fp.exp: check register v1 value
    
    The following patch switches to using "p/x" for printing register values, which
    prints the values once with the hex format, instead of twice.
    
    gdb/testsuite/ChangeLog
    
    2020-02-28  Luis Machado  <luis.machado@linaro.org>
    
    	* gdb.arch/aarch64-fp.exp: Switch from "info registers" command
    	to "p/x".

Diff:
---
 gdb/testsuite/ChangeLog               |  5 +++++
 gdb/testsuite/gdb.arch/aarch64-fp.exp | 12 ++++++------
 2 files changed, 11 insertions(+), 6 deletions(-)

diff --git a/gdb/testsuite/ChangeLog b/gdb/testsuite/ChangeLog
index c3f49e0..5248d59 100644
--- a/gdb/testsuite/ChangeLog
+++ b/gdb/testsuite/ChangeLog
@@ -1,5 +1,10 @@
 2020-02-28  Luis Machado  <luis.machado@linaro.org>
 
+	* gdb.arch/aarch64-fp.exp: Switch from "info registers" command
+	to "p/x".
+
+2020-02-28  Luis Machado  <luis.machado@linaro.org>
+
 	* gdb.arch/aarch64-dbreg-contents.c: Include stdlib.h, unistd.h,
 	asm/ptrace.h and error.h.
 
diff --git a/gdb/testsuite/gdb.arch/aarch64-fp.exp b/gdb/testsuite/gdb.arch/aarch64-fp.exp
index 8ba3377..74224d4 100644
--- a/gdb/testsuite/gdb.arch/aarch64-fp.exp
+++ b/gdb/testsuite/gdb.arch/aarch64-fp.exp
@@ -51,27 +51,27 @@ if {$endianness == "little"} {
     set reg_value1 "0x202122232425262728292a2b2c2d2e2f"
 }
 
-gdb_test "info registers q0" \
+gdb_test "p/x \$q0" \
     "q0.*{u = $reg_value0, s = $reg_value0.*" \
     "check register q0 value"
 
-gdb_test "info registers q1" \
+gdb_test "p/x \$q1" \
     "q1.*{u = $reg_value1, s = $reg_value1.*" \
     "check register q1 value"
 
-gdb_test "info registers v0" \
+gdb_test "p/x \$v0" \
     "v0.*$reg_value0}}}" \
     "check register v0 value"
 
-gdb_test "info registers v1" \
+gdb_test "p/x \$v1" \
     "v1.*$reg_value1}}}" \
     "check register v1 value"
 
-gdb_test "info registers fpsr" \
+gdb_test "p/x \$fpsr" \
     "fpsr.*0x\[0-9a-fA-F\].*" \
     "check register fpsr value"
 
-gdb_test "info registers fpcr" \
+gdb_test "p/x \$fpcr" \
     "fpcr.*0x\[0-9a-fA-F\].*" \
     "check register fpcr value"


^ permalink raw reply	[flat|nested] only message in thread

only message in thread, other threads:[~2020-02-28 15:25 UTC | newest]

Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-02-28 15:25 [binutils-gdb] Fix SVE-related failure in gdb.arch/aarch64-fp.exp Luis Machado

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).