[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Bug-tar] tar 1.23 Solaris regression
From: |
David Bartley |
Subject: |
Re: [Bug-tar] tar 1.23 Solaris regression |
Date: |
Tue, 10 Aug 2010 05:30:37 -0700 |
On Tue, Aug 10, 2010 at 5:00 AM, Petr Sumbera <address@hidden> wrote:
> On 08/02/10 22:36, Paul Eggert wrote:
>>
>> On 08/02/10 12:28, Eric Blake wrote:
>>>
>>> since tar does have the likelihood
>>> of creating children, yes, it should play nicely and restore privileges
>>> before exec()ing.
>>
>> Yes, that makes sense. However, the proposed patch isn't quite
>> right, since it restores PRIV_SYS_LINKDIR even if the user had
>> removed that privilege before invoking 'tar'.
>
> I don't think that PRIV_SYS_LINKDIR is added even if it wasn't there at
> program start. See priv_set_restore() which is checking whether the
> privilege was removed.
Correct.
>
> See the code from GNU tar 1.23:
> http://src.opensolaris.org/source/xref/sfw/usr/src/cmd/gtar/tar-1.23/gnu/priv-set.c#124
>
> David, I see this is your code in priv-set.c. I think your comments for
> priv_set_remove() and priv_set_restore() are not correct. The code looks to
> me that it actually returns -1 for:
Yes, the comments are wrong.
> 83 Returns 0 if ... or was not present in the effective set.
> 114 Returns 0 if ...
> 115 ... or if priv was already in the
> 116 effective set.
>
> Petr
>