On Thu, Feb 14, 2008 at 07:54:15AM -0600, Les Mikesell wrote: > >Because the device changes ownership > > Traditional unix behavior is that open file descriptors stay open and > working even if access permissions change. Actually no. The tty behaviour has been different since the earliest days for precisely these reasons > turn instead of rudely breaking a working process). I can see where > this might make sense on the VT keyboard since that device is > necessarily shared during the procedure. But it doesn't make any more > sense to interrupt a running phone or music player session because > someone else is temporarily using a certain keybord than it would to > break a running tape backup for the same circumstance. Or at least this > should be left as an easily chosen local policy. And local policy should defalt to security first. > the first session being interrupted has root access anyway and could > bypass the access restrictions the switch tries to impose. Wouldn't it > be better to kernel locking or some mechanism that can really ensure > exclusive access for situations like a phone session? VT switch locking policy is handled by X, and by X clients, the kernel just implements the rules. Your objections really make no rational sense anyway, you don't "accidentally" switch sessions to another user. Alan -- fedora-devel-list mailing list fedora-devel-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/fedora-devel-list
