On 2019/7/31 11:27, Xiaojun Wang wrote: > This patch fix bug that variable dentries going > out of scope leaks the storage it points to.
Normally, we use title like "sload.f2fs: xxx" to indicate this is a patch of f2fs-tools, otherwise it looks like a kernel one. It's local filesystem change, so we'd better get rid of fsdevel mailing list submission to avoid in-there folks' complaint which has happened previously... So just f2fs mailing list is okay. For f2fs change in kernel side, scripts/get_maintainer.pl can indicate which list we should send to or cc. :) > > Signed-off-by: Xiaojun Wang<[email protected]> > --- > fsck/sload.c | 7 +++++-- > 1 file changed, 5 insertions(+), 2 deletions(-) > > diff --git a/fsck/sload.c b/fsck/sload.c > index f5a4651..e5de2e1 100644 > --- a/fsck/sload.c > +++ b/fsck/sload.c > @@ -240,15 +240,18 @@ static int build_directory(struct f2fs_sb_info *sbi, > const char *full_path, > ret = set_selinux_xattr(sbi, dentries[i].path, > dentries[i].ino, dentries[i].mode); > if (ret) > - return ret; > + goto out; > + } > > +out: I'd like to suggest to release each entry in the end of cycle to save memory in time, and in error path, we can traverse and release dentries[i, entries]. How do you think? Thanks, > + for (i = 0; i < entries; i++) { > free(dentries[i].path); > free(dentries[i].full_path); > free((void *)dentries[i].name); > } > > free(dentries); > - return 0; > + return ret; > } > > static int configure_files(void) > _______________________________________________ Linux-f2fs-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
