On Wed, 25 Feb 2009 10:55:25 +0200 Boaz Harrosh <bharrosh@xxxxxxxxxxx> wrote: > FUJITA Tomonori wrote: > > On Mon, 23 Feb 2009 19:51:25 +0200 > > Boaz Harrosh <bharrosh@xxxxxxxxxxx> wrote: > > > >> usr/util.h redefines sync_file_range (as __sync_file_range) > >> for old systems that don't have this defintion. Make these > >> definitions conditional on __CONF_TGT_NO_SYNC_FILE_RANGE which > >> should be set by a Makefile on very old systems. > >> (Or a ./configure script can detect that automatically) > >> > >> Normally sync_file_range() definition is now taken from gcc > >> headers. > >> > >> Signed-off-by: Boaz Harrosh <bharrosh@xxxxxxxxxxx> > >> --- > >> usr/bs_mmap.c | 2 +- > >> usr/bs_rdwr.c | 2 +- > >> usr/util.h | 6 ++++-- > >> 3 files changed, 6 insertions(+), 4 deletions(-) > > > > Using "sync_file_range" does not work on a box having glibc supporting > > sync_file_range properly, I think. > > > > I don't understand? Are you saying that sync_file_range() is > broken in glibc? On my new FC10 the source looks the same > as yours. Do you have any info on what systems are broken and > how? > > I will try to investigate this, any pointers will help > Thanks. I guess that you should test your patch on different environments. You patch wrongly assumes that glibc has sync_file_range. With debian etch, I got the following error: bs_rdwr.o: In function `bs_sync_sync_range': /home/fujita/git/tgt/usr/bs_rdwr.c:54: undefined reference to `sync_file_range' collect2: ld returned 1 exit status The code need to work both with and without sync_file_range in glibc. That's what I pointed out in the previous mail. -- To unsubscribe from this list: send the line "unsubscribe stgt" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html