The previous attempt (commit f77147cd4de) doesn't work as
expected, as we still have CONFIG_TCG=1 when using:
configure --disable-system --disable-user
Use Meson's get_option() instead to remove the softfloat tests
when building with --disable-tcg.
Suggested-by: Claudio Fontana <cfontana@suse.de>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
---
Cc: Richard Henderson <richard.henderson@linaro.org>
Cc: Alex Bennée <alex.bennee@linaro.org>
Cc: Emilio G. Cota <cota@braap.org>
---
tests/meson.build | 5 +----
1 file changed, 1 insertion(+), 4 deletions(-)
diff --git a/tests/meson.build b/tests/meson.build
index 55a7b082751..250fbc99a36 100644
--- a/tests/meson.build
+++ b/tests/meson.build
@@ -76,11 +76,8 @@
workdir: meson.current_source_dir() / 'decode',
suite: 'decodetree')
-if 'CONFIG_TCG' in config_all
- subdir('fp')
-endif
-
if not get_option('tcg').disabled()
+ subdir('fp')
if 'CONFIG_PLUGIN' in config_host
subdir('plugin')
endif
--
2.26.2
On 22/03/21 12:47, Philippe Mathieu-Daudé wrote: > The previous attempt (commit f77147cd4de) doesn't work as > expected, as we still have CONFIG_TCG=1 when using: > > configure --disable-system --disable-user But there's no change in behavior with this patch, right? I would rather first start removing CONFIG_TCG from target-dependent files (such as tests/qtest/bios-tables-test.c and tests/qtest/qmp-cmd-test.c), and then just remove the config_host += { 'CONFIG_TCG': 'y' } line that is not needed anymore. Paolo
On 3/22/21 1:57 PM, Paolo Bonzini wrote: > On 22/03/21 12:47, Philippe Mathieu-Daudé wrote: >> The previous attempt (commit f77147cd4de) doesn't work as >> expected, as we still have CONFIG_TCG=1 when using: >> >> configure --disable-system --disable-user > > But there's no change in behavior with this patch, right? I'm trying to save CI minutes from the build-tools-and-docs-debian job which uses: configure --disable-system --disable-user --disable-tools --enable-docs See https://gitlab.com/qemu-project/qemu/-/jobs/1117005759: [77/1268] Compiling C object libqemuutil.a.p/meson-generated_.._trace_trace-qapi.c.o [645/1268] Compiling C object tests/fp/libtestfloat.a.p/berkeley-testfloat-3_source_writeCase_a_f32.c.o [1079/1268] Compiling C object tests/unit/test-clone-visitor.p/test-clone-visitor.c.o [1132/1268] Compiling C object tests/unit/test-qdist.p/test-qdist.c.o > > I would rather first start removing CONFIG_TCG from target-dependent > files (such as tests/qtest/bios-tables-test.c and > tests/qtest/qmp-cmd-test.c), and then just remove the > > config_host += { 'CONFIG_TCG': 'y' } > > line that is not needed anymore. You seems right, I missed something, I'll revisit... Thanks.
On 3/22/21 2:09 PM, Philippe Mathieu-Daudé wrote: > On 3/22/21 1:57 PM, Paolo Bonzini wrote: >> On 22/03/21 12:47, Philippe Mathieu-Daudé wrote: >> I would rather first start removing CONFIG_TCG from target-dependent >> files (such as tests/qtest/bios-tables-test.c and >> tests/qtest/qmp-cmd-test.c), and then just remove the >> >> config_host += { 'CONFIG_TCG': 'y' } >> >> line that is not needed anymore. This looks the correct thing to do but not something I'm willing to do now, so please simply disregard this patch. Regards, Phil.
Philippe Mathieu-Daudé <f4bug@amsat.org> writes: > The previous attempt (commit f77147cd4de) doesn't work as > expected, as we still have CONFIG_TCG=1 when using: > > configure --disable-system --disable-user > > Use Meson's get_option() instead to remove the softfloat tests > when building with --disable-tcg. > > Suggested-by: Claudio Fontana <cfontana@suse.de> > Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Alex Bennée <alex.bennee@linaro.org> Tested-by: Alex Bennée <alex.bennee@linaro.org> -- Alex Bennée
© 2016 - 2024 Red Hat, Inc.