Message ID | m3iqh5l2bi.fsf@neno.mitica |
---|---|
State | Superseded |
Headers | show |
On Mon, Aug 3, 2009 at 10:38 AM, Juan Quintela<quintela@redhat.com> wrote: > "Sebastian Herbszt" <herbszt@gmx.de> wrote: >> Juan Quintela wrote: > > Hi > > Thanks for the report, on my next series, I add this patch that fix the problem: > >> I think this patch or a related one breaks MinGW/Win32: >> >> CC posix-aio-compat.o >> posix-aio-compat.c:14:23: warning: sys/ioctl.h: No such file or directory >> posix-aio-compat.c: In function `handle_aiocb_ioctl': >> posix-aio-compat.c:90: warning: implicit declaration of function `ioctl' >> posix-aio-compat.c: In function `handle_aiocb_rw_linear': >> posix-aio-compat.c:182: warning: implicit declaration of function `pwrite' >> posix-aio-compat.c:187: warning: implicit declaration of function `pread' >> posix-aio-compat.c: In function `aio_thread': >> posix-aio-compat.c:284: warning: implicit declaration of function `sigfillset' >> posix-aio-compat.c:285: warning: implicit declaration of function `sigprocmask' >> posix-aio-compat.c:332: warning: implicit declaration of function `kill' >> posix-aio-compat.c: In function `qemu_paio_submit': >> posix-aio-compat.c:367: error: `EINPROGRESS' undeclared (first use in this function) >> posix-aio-compat.c:367: error: (Each undeclared identifier is reported only once >> posix-aio-compat.c:367: error: for each function it appears in.) >> posix-aio-compat.c: In function `qemu_paio_cancel': >> posix-aio-compat.c:424: error: `ECANCELED' undeclared (first use in this function) >> posix-aio-compat.c:426: error: `EINPROGRESS' undeclared (first use in this function) >> make: *** [posix-aio-compat.o] Error 1 >> >> v0.11.0-rc0-182-g28e738d on gcc version 3.4.5 (mingw32 special) >> >> - Sebastian > > From 53e6f57abb0534b5f654dd1b18183c18443ecb5b Mon Sep 17 00:00:00 2001 > From: Juan Quintela <quintela@redhat.com> > Date: Mon, 3 Aug 2009 10:36:09 +0200 > Subject: [PATCH] disable aio with mingw32 > > > Signed-off-by: Juan Quintela <quintela@redhat.com> > --- > configure | 1 + > 1 files changed, 1 insertions(+), 0 deletions(-) > > diff --git a/configure b/configure > index 58a32a7..a25343a 100755 > --- a/configure > +++ b/configure > @@ -359,6 +359,7 @@ if test "$mingw32" = "yes" ; then > EXESUF=".exe" > CFLAGS="-DWIN32_LEAN_AND_MEAN -DWINVER=0x501 $CFLAGS" > LIBS="-lwinmm -lws2_32 -liphlpapi $LIBS" > + aio=no > fi > > # find source path I don't think this is the correct way to fix it in long term or at least the variables have to be renamed to something like CONFIG_POSIX_AIO / posix_aio. There's no reason why AIO can't be added to Win32 port and it surely won't use the posix-aio-compat file. Best regards, Filip Navara
Filip Navara <filip.navara@gmail.com> wrote: > I don't think this is the correct way to fix it in long term or at > least the variables have to be renamed to something like > CONFIG_POSIX_AIO / posix_aio. There's no reason why AIO can't be added > to Win32 port and it surely won't use the posix-aio-compat file. If we agree on that, I can change the names in my cleanups. I was just doing with the current names. /me add another patch to his already very long series. Later, Juan.
diff --git a/configure b/configure index 58a32a7..a25343a 100755 --- a/configure +++ b/configure @@ -359,6 +359,7 @@ if test "$mingw32" = "yes" ; then EXESUF=".exe" CFLAGS="-DWIN32_LEAN_AND_MEAN -DWINVER=0x501 $CFLAGS" LIBS="-lwinmm -lws2_32 -liphlpapi $LIBS" + aio=no fi # find source path