Pádraig Brady wrote: > On 13/05/10 15:25, jeff.liu wrote: >> >> diff --git a/src/copy.c b/src/copy.c >> index c16cef6..960e5fb 100644 >> --- a/src/copy.c >> +++ b/src/copy.c >> @@ -63,6 +63,10 @@ >> >> #include <sys/ioctl.h> >> >> +#ifndef HAVE_FIEMAP >> +# include "fiemap.h" >> +#endif > > Is HAVE_FIEMAP ever defined anywhere? > In future will we use this to check for <linux/fiemap.h> ?
I'll look. > On 20/05/10 20:23, Jim Meyering wrote: >> >> diff --git a/tests/cp/sparse-fiemap b/tests/cp/sparse-fiemap >> old mode 100644 >> new mode 100755 >> index f9d3a94..814d537 >> --- a/tests/cp/sparse-fiemap >> +++ b/tests/cp/sparse-fiemap >> @@ -28,8 +28,7 @@ cwd=`pwd` >> cleanup_() { cd /; umount "$cwd/mnt"; } >> >> # Create an ext4 loopback file system >> -dd if=/dev/zero of=blob bs=8192 count=1000 > /dev/null 2>&1 \ >> - || skip=1 >> +dd if=/dev/zero of=blob bs=8192 count=1000 || skip=1 >> mkdir mnt >> mkfs -t ext4 -F blob || >> skip_test_ "failed to create ext4 file system" > > There is the unlikely combination of ext4 without fiemap support I think? > If so then that dependency is worth a comment. I don't know off hand. Is there a shell-level way to test for that? >> @@ -42,20 +41,15 @@ test $skip = 1 && >> >> rm -f mnt/f >> >> -# Create a 2gb sparse file >> -dd if=/dev/zero of=mnt/sparse bs=1k count=1 seek=2096128 > /dev/null 2>&1 >> || framework_failure >> +# Create a 2TiB sparse file >> +dd if=/dev/zero of=mnt/sparse bs=1k count=1 seek=2G || framework_failure > > If we don't need any actual data in the files then one could use: > truncate -s 2TB mnt/sparse > > For my reference, I used TB rather than TiB because on ext3 > the limit is 0x1FEFF7FC000 (2194719883264) > (0x1FF7FFFD000 (2196875759616) before 2.6.25) Thanks. For now I'll limit it to 1GiB using dd. That gives a slightly less uniform input. # Create a 1TiB sparse file dd if=/dev/zero of=mnt/sparse bs=1k count=1 seek=1G || framework_failure