Re: [Xen-devel] [PATCH v2 2/2] libxl: fix build (missing CLONE_NEWIPC) on astonishingly old systems

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 14.01.19 at 15:22, <ian.jackson@xxxxxxxxxx> wrote:
> Wei Liu writes ("Re: [PATCH v2 2/2] libxl: fix build (missing CLONE_NEWIPC) 
> on astonishingly old systems"):
>> On Mon, Jan 14, 2019 at 02:47:58AM -0700, Jan Beulich wrote:
>> > On 11.01.19 at 20:23, <ian.jackson@xxxxxxxxxxxxx> wrote:
>> > > CLONE_NEWIPC was introduced in Linux 2.6.19, on the 29th of November
>> > > 2006, which was 12 years, 1 month, and 14 days ago.
>> > 
>> > Thanks for the very precise counting, the latter part which will be
>> > wrong - even if just slightly - by the time you commit it ;-)
> ...
>> > Sadly the situation is more complicated: The check to disallow
>> > unknown flags was introduced only in 2.6.17 [1], and apparently
>> > never backported to 2.6.16 or older stable trees despite the
>> > description talking about it going into 2.6.16. Since it didn't
>> > matter in my variant of the workaround, I didn't mention this.
> 
> Good grief.
> 
>> > Of course a pretty reasonable position to take would be to
>> > consider the 2.6.18-based XenoLinux tree a "baseline", beyond
>> > which we don't care about undesirable behavior here.
>> 
>> I think using 2.6.18 as baseline is very reasonable. 
> 
> I guess we need to write this in the SUPPORT.md statement for
> dm_restrict.

Ah yes, we should.

> TBH how about writing somewhere general in SUPPORT.md that "all bets
> are off if you use Linux before 2.6.18" ?  Do we even have a limit
> anywhere for security supported Linux versions ?

I don't think so, and leaving this specific case aside it's also
unclear to me why we should.

Jan



_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel