xen-devel.lists.xenproject.org archive mirror
 help / color / mirror / Atom feed
* [PATCH for-next] configure: probe for gcc -m32 integer sizes
@ 2021-02-24  1:08 Stefano Stabellini
  2021-02-24  7:26 ` Jan Beulich
  2021-02-24 10:20 ` Roger Pau Monné
  0 siblings, 2 replies; 3+ messages in thread
From: Stefano Stabellini @ 2021-02-24  1:08 UTC (permalink / raw)
  To: xen-devel; +Cc: sstabellini, jbeulich, andrew.cooper3, wl, iwj, anthony.perard

The hvmloader build on Alpine Linux x86_64 currenly fails:


hvmloader.c: In function 'init_vm86_tss':
hvmloader.c:202:39: error: left shift count >= width of type [-Werror=shift-count-overflow]
  202 |                   ((uint64_t)TSS_SIZE << 32) | virt_to_phys(tss));

util.c: In function 'get_cpu_mhz':
util.c:824:15: error: conversion from 'long long unsigned int' to 'uint64_t' {aka 'long unsigned int'} changes value from
'4294967296000000' to '0' [-Werror=overflow]
  824 |     cpu_khz = 1000000ull << 32;


The root cause of the issue is that gcc -m32 picks up headers meant for
64-bit builds.

The failures are currently causing problems to the xen-project
gitlab-ci pipeline.

This patch introduces code to detect this kind of errors in the
configure script, and disables with a warning the compilation of
hvmloader if problems are detected.

This patch also updates tools/configure. It has been done by calling
autoreconf -fi.

Signed-off-by: Stefano Stabellini <stefano.stabellini@xilinx.com>


diff --git a/config/Tools.mk.in b/config/Tools.mk.in
index d47936686b..395ed2a6d2 100644
--- a/config/Tools.mk.in
+++ b/config/Tools.mk.in
@@ -51,6 +51,7 @@ CONFIG_OVMF         := @ovmf@
 CONFIG_ROMBIOS      := @rombios@
 CONFIG_SEABIOS      := @seabios@
 CONFIG_IPXE         := @ipxe@
+CONFIG_HVMLOADER    := @hvmloader@
 CONFIG_QEMU_TRAD    := @qemu_traditional@
 CONFIG_QEMU_XEN     := @qemu_xen@
 CONFIG_QEMUU_EXTRA_ARGS:= @EXTRA_QEMUU_CONFIGURE_ARGS@
diff --git a/tools/configure b/tools/configure
index bb5acf9d43..f23a3bb8aa 100755
--- a/tools/configure
+++ b/tools/configure
@@ -687,6 +687,7 @@ INSTALL_DATA
 INSTALL_SCRIPT
 INSTALL_PROGRAM
 SET_MAKE
+hvmloader
 AWK
 IASL
 XGETTEXT
@@ -5279,6 +5280,25 @@ ac_compile='$CC -c $CFLAGS $CPPFLAGS conftest.$ac_ext >&5'
 ac_link='$CC -o conftest$ac_exeext $CFLAGS $CPPFLAGS $LDFLAGS conftest.$ac_ext $LIBS >&5'
 ac_compiler_gnu=$ac_cv_c_compiler_gnu
 
+ac_ext=c
+ac_cpp='$CPP $CPPFLAGS'
+ac_compile='$CC -c $CFLAGS $CPPFLAGS conftest.$ac_ext >&5'
+ac_link='$CC -o conftest$ac_exeext $CFLAGS $CPPFLAGS $LDFLAGS conftest.$ac_ext $LIBS >&5'
+ac_compiler_gnu=$ac_cv_c_compiler_gnu
+
+cat confdefs.h - <<_ACEOF >conftest.$ac_ext
+/* end confdefs.h.  */
+#include <stdint.h>
+#define BUILD_BUG_ON(p) ((void)sizeof(char[1 - 2 * !!(p)]))
+int main() { BUILD_BUG_ON(sizeof(uint64_t) != 8); }
+_ACEOF
+if gcc -m32 -c conftest.c -o /dev/null 2>/dev/null; then :
+  hvmloader=y
+else
+  { $as_echo "$as_me:${as_lineno-$LINENO}: WARNING: hvmloader build disabled as the compiler cannot build 32bit binaries" >&5
+$as_echo "$as_me: WARNING: hvmloader build disabled as the compiler cannot build 32bit binaries" >&2;}
+fi
+
 { $as_echo "$as_me:${as_lineno-$LINENO}: checking whether ${MAKE-make} sets \$(MAKE)" >&5
 $as_echo_n "checking whether ${MAKE-make} sets \$(MAKE)... " >&6; }
 set x ${MAKE-make}
diff --git a/tools/configure.ac b/tools/configure.ac
index 636e7077be..706c162322 100644
--- a/tools/configure.ac
+++ b/tools/configure.ac
@@ -307,6 +307,12 @@ AC_ARG_VAR([AWK], [Path to awk tool])
 
 # Checks for programs.
 AC_PROG_CC
+AC_LANG(C)
+AC_LANG_CONFTEST([AC_LANG_SOURCE([[#include <stdint.h>
+#define BUILD_BUG_ON(p) ((void)sizeof(char[1 - 2 * !!(p)]))
+int main() { BUILD_BUG_ON(sizeof(uint64_t) != 8); }]])])
+AS_IF([gcc -m32 -c conftest.c -o /dev/null 2>/dev/null], [hvmloader=y], [AC_MSG_WARN(hvmloader build disabled due to headers mismatch)])
+AC_SUBST(hvmloader)
 AC_PROG_MAKE_SET
 AC_PROG_INSTALL
 AC_PATH_PROG([FLEX], [flex])
diff --git a/tools/firmware/Makefile b/tools/firmware/Makefile
index 1f27117794..5c395ad738 100644
--- a/tools/firmware/Makefile
+++ b/tools/firmware/Makefile
@@ -13,7 +13,7 @@ SUBDIRS-$(CONFIG_ROMBIOS) += rombios
 SUBDIRS-$(CONFIG_ROMBIOS) += vgabios
 SUBDIRS-$(CONFIG_IPXE) += etherboot
 SUBDIRS-$(CONFIG_PV_SHIM) += xen-dir
-SUBDIRS-y += hvmloader
+SUBDIRS-$(CONFIG_HVMLOADER) += hvmloader
 
 SEABIOSCC ?= $(CC)
 SEABIOSLD ?= $(LD)


^ permalink raw reply related	[flat|nested] 3+ messages in thread

* Re: [PATCH for-next] configure: probe for gcc -m32 integer sizes
  2021-02-24  1:08 [PATCH for-next] configure: probe for gcc -m32 integer sizes Stefano Stabellini
@ 2021-02-24  7:26 ` Jan Beulich
  2021-02-24 10:20 ` Roger Pau Monné
  1 sibling, 0 replies; 3+ messages in thread
From: Jan Beulich @ 2021-02-24  7:26 UTC (permalink / raw)
  To: Stefano Stabellini; +Cc: andrew.cooper3, wl, iwj, anthony.perard, xen-devel

On 24.02.2021 02:08, Stefano Stabellini wrote:
> --- a/tools/configure.ac
> +++ b/tools/configure.ac
> @@ -307,6 +307,12 @@ AC_ARG_VAR([AWK], [Path to awk tool])
>  
>  # Checks for programs.
>  AC_PROG_CC
> +AC_LANG(C)
> +AC_LANG_CONFTEST([AC_LANG_SOURCE([[#include <stdint.h>
> +#define BUILD_BUG_ON(p) ((void)sizeof(char[1 - 2 * !!(p)]))
> +int main() { BUILD_BUG_ON(sizeof(uint64_t) != 8); }]])])
> +AS_IF([gcc -m32 -c conftest.c -o /dev/null 2>/dev/null], [hvmloader=y], [AC_MSG_WARN(hvmloader build disabled due to headers mismatch)])
> +AC_SUBST(hvmloader)
>  AC_PROG_MAKE_SET
>  AC_PROG_INSTALL
>  AC_PATH_PROG([FLEX], [flex])

I'm fine with the approach now, but I'm rather uncertain about
the insertion point you've selected (in the middle of the
"Checks for programs" section). It'll need to be the tools
maintainers to judge about this.

Jan


^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [PATCH for-next] configure: probe for gcc -m32 integer sizes
  2021-02-24  1:08 [PATCH for-next] configure: probe for gcc -m32 integer sizes Stefano Stabellini
  2021-02-24  7:26 ` Jan Beulich
@ 2021-02-24 10:20 ` Roger Pau Monné
  1 sibling, 0 replies; 3+ messages in thread
From: Roger Pau Monné @ 2021-02-24 10:20 UTC (permalink / raw)
  To: Stefano Stabellini
  Cc: xen-devel, jbeulich, andrew.cooper3, wl, iwj, anthony.perard

On Tue, Feb 23, 2021 at 05:08:43PM -0800, Stefano Stabellini wrote:
> The hvmloader build on Alpine Linux x86_64 currenly fails:
> 
> 
> hvmloader.c: In function 'init_vm86_tss':
> hvmloader.c:202:39: error: left shift count >= width of type [-Werror=shift-count-overflow]
>   202 |                   ((uint64_t)TSS_SIZE << 32) | virt_to_phys(tss));
> 
> util.c: In function 'get_cpu_mhz':
> util.c:824:15: error: conversion from 'long long unsigned int' to 'uint64_t' {aka 'long unsigned int'} changes value from
> '4294967296000000' to '0' [-Werror=overflow]
>   824 |     cpu_khz = 1000000ull << 32;
> 
> 
> The root cause of the issue is that gcc -m32 picks up headers meant for
> 64-bit builds.

I'm working on getting hvmloader to build standalone without using any
system headers, which I think it's a worthwhile change to do rather
than this configure bodge. Will post the series now.

Thanks, Roger.


^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2021-02-24 10:21 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-02-24  1:08 [PATCH for-next] configure: probe for gcc -m32 integer sizes Stefano Stabellini
2021-02-24  7:26 ` Jan Beulich
2021-02-24 10:20 ` Roger Pau Monné

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).