Messages in this thread | | | Date | Mon, 29 Nov 1999 07:46:43 -0600 (CST) | From | Jesse Pollard <> | Subject | Re: Linux needs flexible security |
| |
>> Callbacks: you send ioctl over network and if remote side needs to >> peek memory, sends "I want to peek at address" over network. > >yick. nasty.
The only way way to go, unless there is a local driver to interpret the IOCTL, collect the needed data, then transfer it (only once..) and then transfer the return value back.
>Hmm.. > > ssize_t write(int fd, const void *buf, size_t count); >vs > newioctl(fd, ptr, length);
or even ioctl(fd, ptr,...)
No difference. other than the inability to distinquish from the data transfer read/write...
IOCTLs (no matter what you call them) are there for out-of-band data. What the control data is happens to be is up to the filesystem/driver. Some of them have standardized interfaces, many more do not. As far as security goes, they still have to be optionally audited. Control over what is transferred doesn't have to be enforced - the open of the fd covers that.
The only time I can think of that would require more is when removable media has a mandatory lable - Users shouldn't be able to overwrite the label. I do think that this can be handled in a relatively simple way (of course I don't happen to have one handy...).
------------------------------------------------------------------------- Jesse I Pollard, II Email: pollard@navo.hpc.mil
Any opinions expressed are solely my own.
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |