Messages in this thread | | | Date | Thu, 24 Nov 2011 13:23:49 +0900 | From | KAMEZAWA Hiroyuki <> | Subject | Re: [V3 PATCH 1/2] tmpfs: add fallocate support |
| |
On Thu, 24 Nov 2011 11:22:34 +0800 Cong Wang <amwang@redhat.com> wrote:
> 于 2011年11月24日 11:01, KAMEZAWA Hiroyuki 写道: > > On Wed, 23 Nov 2011 21:46:39 -0500 > > KOSAKI Motohiro<kosaki.motohiro@jp.fujitsu.com> wrote: > > > >>>> + while (index< end) { > >>>> + ret = shmem_getpage(inode, index,&page, SGP_WRITE, NULL); > >>> > >>> If the 'page' for index exists before this call, this will return the page without > >>> allocaton. > >>> > >>> Then, the page may not be zero-cleared. I think the page should be zero-cleared. > >> > >> No. fallocate shouldn't destroy existing data. It only ensure > >> subsequent file access don't make ENOSPC error. > >> > > FALLOC_FL_KEEP_SIZE > > This flag allocates and initializes to zero the disk space > > within the range specified by offset and len. .... > > > > just manual is unclear ? it seems that the range [offset, offset+len) is > > zero cleared after the call. > > I think we should fix the man page, because at least ext4 doesn't clear > the original contents, > > % echo hi > /tmp/foobar > % fallocate -n -l 1 -o 10 /tmp/foobar > % hexdump -Cv /tmp/foobar > 00000000 68 69 0a |hi.| > 00000003 >
thank you for checking. So, at failure path, original data should not be cleared, either.
Thanks, -Kame
| |