Messages in this thread | | | Subject | Re: PATCH: cdrecord: avoiding scsi device numbering for ide devices | From | Alan Cox <> | Date | Fri, 20 Aug 2004 12:22:14 +0100 |
| |
On Iau, 2004-08-19 at 23:57, Bartlomiej Zolnierkiewicz wrote: > Alan's example is invalid because IDE driver requires CAP_SYS_ADMIN and > CAP_SYS_RAWIO so if there is some security risk involved - it is in the user > apps not in the kernel. Also Linus first fixed SG_IO correctly with > requiring CAP_SYS_RAWIO but then (under Alan's influence?) he added filtering > which broke cd writing and which is just unmaintainable.
SG_IO prior to 2.6.8 doesn't do any checks on any path into and through the IDE driver. Thus I sent Linus a patch for 2.6.8 that just added capable(CAP_SYS_RAWIO) to the raw command path.
Filtering was something Jens and I were talking about. I was a little suprised when Linus added filters just before release. We do now have good traces of what various apps want. Some of those commands may be problematic without sg_io knowing the target class however.
> Also filtering cannot work in all cases because there are vendor specific > opcodes, some devices redefines some opcodes etc. - this should be left to > user space.
Possibly. Vendor commands are not in themselves a problems. The answer to those is "no" 8)
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| |