metafs-devel — General MetaFS Development

Re: [MetaFS-devel] Is this a bug in metafs version 0.1.5?

[Joshua J. B. <de...@co...>]

On Thursday 22 February 2007 18:28:40 汤斯亮 wrote:
> When I was using metafs, I found something a little unusual that may be a
> bug in metafs.

Thanks for your note.  Sorry for not getting back to you sooner.

> In the operation rename, uid changes to root automatically.
>
>  I look over the code in file metafsd/fuse_ll.cc, function "static void
> _rename(fuse_req_t req, fuse_ino_t parent, const char *name, fuse_ino_t
> newparent, const char *newname)", line 503,
> "destdir->rename(&Context::root, srcdir, name, newname);"
>
> Why metafs pass parameter "&Context::root" instead of "&ctx"?

Probably because I wasn't paying attention. ;)  I've pushed a fix for this 
into the Darcs repository.  Thanks for the bug report.

> And in the root-plugin realfs, file plugins/realfs/xattr_db.cc.
>
> When operating Berkeley db, code, such as "if (! strncmp((char
> *)oldvalue.get_data(), name.c_str(), oldvalue.get_size()))", seems not
> work in it's originally intention.
>
> Metafs compares data "oldvalue" and "name", if they have same prefix?
>
> An "oldvalue" "aaa" will not be replaced if "name" string is "aaaaa".
> Should metafs check the string length before comparing?

hmmm, yes.  There was another problem I saw while I was reading through, so 
I reworked that code a bit.  Please try the attached patch and let me know 
if it fixes the problem.

I don't use xattr_db myself anymore, so I don't have any easy way of testing 
it, unfortunately.

Thanks for the feedback!

-- Josh


-- 
Joshua J. Berry

"I haven't lost my mind -- it's backed up on tape somewhere."
    -- /usr/games/fortune

[MetaFS-devel] Is this a bug in metafs version 0.1.5?

[<sil...@gm...>]

Greetings,



When I was using metafs, I found something a little unusual that may be a
bug in metafs.

In the operation rename, uid changes to root automatically.

 I look over the code in file metafsd/fuse_ll.cc, function "static void
_rename(fuse_req_t req, fuse_ino_t parent, const char *name, fuse_ino_t
newparent, const char *newname)", line 503, "destdir->rename(&Context::root,
srcdir, name, newname);"

Why metafs pass parameter "&Context::root" instead of "&ctx"?



And in the root-plugin realfs, file plugins/realfs/xattr_db.cc.

When operating Berkeley db, code, such as "if (! strncmp((char
*)oldvalue.get_data(), name.c_str(), oldvalue.get_size()))", seems not work
in it's originally intention.

Metafs compares data "oldvalue" and "name", if they have same prefix?

An "oldvalue" "aaa" will not be replaced if "name" string is "aaaaa". Should
metafs check the string length before comparing?



TANG Siliang.