This is an archived copy of the Xen.org mailing list, which we have preserved to ensure that existing links to archives are not broken. The live archive, which contains the latest emails, can be found at http://lists.xen.org/
Home Products Support Community News


Re: [Xen-devel] [PATCH] Fixing iopl and ioperm

On 14 Jun 2005, at 01:22, Li, Xin B wrote:

On xeno32, test case iopl02 of LTP will fail, pls see below:

My guess is that LTP tries to change iopl to 1. This doesn;t fail on xenlinux because the iopl is *already* 1 (because the kernel runs at ring 1; not ring 0 as on native). If this is the case, your patch 'fixes' LTP by breaking our implementation of sys_iopl. By reading the old iopl out of eflags, you will always read zero. This would certainly fix this LTP testcase but it is the wrong thing to do.

Frankly, fixing this LTP test case will be a pain. If I have correctly described what's going on then let's ignore it at least for now -- it is not indicative of a real bug and it is entirely harmless. LTP is a measure of stability for us, but passing every LTP test is not an end in itself (we should understand the failures though).

 -- Keir

Xen-devel mailing list