| To: | yyyyyyyyyyyyyy@xxxxxxxxxxxxx |
|---|---|
| Subject: | Re: nohup with stdout closed (was: Defect in XCU nohup) |
| From: | Gunnar Ritter <yyyyyyy@xxxxxxx> |
| Date: | Thu, 14 Jul 2005 13:51:37 +0200 |
| Organization: | Privat. |
| References: | <200507132133.WAA13330@xoneweb.opengroup.org><20050714101704.GB6585@squonk.masqnet><20050714113409.GA3966@excession.spiral-arm.org> |
yyy@xxxxxxxxxxxxxxxxxxxxxxxx (James Youngman) wrote:
> It can be worse than this, too. For example, 'ping' is set-user-id on
> many systems, and by this mechanism it can be persuaded to send
> arbitrary data of an attacker's choice out of its raw socket. Hence
> the wording of execve() which I quote below.
The correct way to handle this problem is to fix that implementation
of the "ping" program, and to fix it _now_. A program that relies on
a "may" condition in the standard is not a secure one and should not
be run set-used-id anyway, regardless of possible future changes.
Gunnar
|
| Previous by Date: | Re: nohup with stdout closed (was: Defect in XCU nohup), Gunnar Ritter |
|---|---|
| Next by Date: | Re: nohup with stdout closed (was: Defect in XCU nohup), James Youngman |
| Previous by Thread: | Re: nohup with stdout closed (was: Defect in XCU nohup), James Youngman |
| Next by Thread: | Re: nohup with stdout closed (was: Defect in XCU nohup), James Youngman |
| Indexes: | [Date] [Thread] [All Lists] |