Re: [PATCH] Posix file attribute support on VFAT

["Machida, Hiroyuki" <machida@sm.sony.co.jp>]

Bodo Eggert wrote:
> Hiroyuki Machida <machida@sm.sony.co.jp> wrote:
> 
> 
>>For newly created and/or modified files/dirs, system can utilize
>>full posix attributes, because memory resident inode storage can
>>hold those. After umount-mount cycle, system may lose some
>>attributes to preserve VFAT format.
> 
> 
> Inodes may be reclaimed, therefore you might also lose some attributes at
> runtime. For your users, that will look like a heisenbug. A similar bug
> has been reported for procfs. Is your implementation affected?

Yes.
Thank you, pointed it out. So I need to select
	1) restric attribute even in mamory
or
	2) pinn inode in memory during mount
		(I think it's not good idea)

I'll revise the patch with 1).

>>- Special file
>>        To distinguish special files, look if this fat dir entry 
>>        has ATTR_SYS, first. If it has ATTR_SYS, then check
>>        1st. LSB bit in ctime_cs, refered as "special file flag".
>>        If set,  this file is created under VFAT with "posix_attr". 
>>        Look up TYPE field to decide special file type.
>>        This spcial file detection mothod has some flaw to make
>>        potential confusion. E.g. some system file created under
>>        dos/win may be treated as special file.  However in most case,
>>        user don't create system file under dos/win.
> 
> 
> You can add additional magic, e.g.: nodes must be empty, except for symlinks
> which must be not larger than 4KB (current PATH_MAX?). This will get rid
> of io.sys, logo.sys etc.\. If you want to be really sure, prepend a magic
> code to the on-disk representation of symlinks.

Please confirm my understanding.
You sugessted that symlink should not have ATTR_SYS, to prevent
some over 4KB files created in DOS/WIN to be treated as symlinks?

-- 
Hiroyuki Machida		machida@sm.sony.co.jp		
SSW Dept. HENC, Sony Corp.