Hi again,
Niek Green wrote:
Hoi,
Thank you for testing. It confirms what I have found in the
meantime, and I have even found the cause.
I do not need the logs I asked, you can drop them.
Can you please try the attached patch over the latest
version 2013.1.13AR.4 ?
Done, no errors anymore.
Great !
I have attached a different variant of the patch, which
I prefer because it should have less overhead.
Can you try it also ?
And I would also like to know whether there is a difference in
the execution times between 2013.1.13AR.3 and 2013.1.13AR.4
(with the new patch applied - if successful of course).
Just mount, make a rather big torrent download, and note the
execution time of the ntfs mounting process, before unmounting
(just do "ps -ef | grep ntfs")
Sorry I don't understand what to do.
ps -ef has a column time but that only says 1 second, is that what you
are looking for?
Too bad, this is probably what I wanted, but your download
does not lead to measurable times. Good news for ntfs-3g
performance anyway. So forget about it.
Regards
Jean-Pierre
--- libntfs-3g/attrib.c.ref 2014-04-06 19:50:17.000000000 +0200
+++ libntfs-3g/attrib.c 2014-04-06 19:48:19.000000000 +0200
@@ -1274,9 +1274,10 @@
goto err_out;
} else {
/* make sure the run ahead of hole is mapped */
- if (((*rl)->lcn == LCN_HOLE) && cur_vcn) {
- if (ntfs_attr_map_partial_runlist(na, cur_vcn - 1))
- goto err_out;
+ if ((*rl)->lcn == LCN_HOLE) {
+ if (ntfs_attr_map_partial_runlist(na,
+ (cur_vcn ? cur_vcn - 1 : cur_vcn)))
+ goto err_out;
}
}
#else
