Downplaying a serious issue

[Torbjorn Granlund]

Glen Barber <gjb at FreeBSD.org> writes:

  On Thu, Apr 03, 2014 at 07:36:05AM -0400, Glen Barber wrote:
  > On Thu, Apr 03, 2014 at 09:47:02AM +0200, Torbjorn Granlund wrote:
  > > Glen Barber <gjb at FreeBSD.org> writes:
  > >   The issue is specific to certain hardware configurations, and unless
  > >   anyone has made any breakthroughs that I am unaware of, the cause is
  > >   still unknown.
  > >   
  > > It happens on:
  > > 
  > > AMD piledriver running Linux+KVM
  > > AMD piledriver running Linux+Xen
  > > Intel Nehalem running NetBSD+Xen
  > > Intel Sandybridge running NetBSD+Xen
  > > Intel Haswell running NetBSD+Xen
  > > AMD K10 Barcelona running NetBSD+Xen
  > > AMD Bulldozer running NetBSD+Xen
  > > 
  > 
  > We need more specifics.
  > 
  
  Here is the type of information we need:
   CPU information
   RAM
   ACPI information
   Drive controller information
  
The CPU list is above.

I have used RAM from 192 MiB to 512 MiB.  Your machine where you have
not observed problems clearly has an amount of memory in a quite
different range.  Perhaps this bug is related to the amount of RAM?

For all Xen runs, I have had 'apic = 1' in the guest config files.  I
don't know about KVM, I used the default.  (I cannot easily retest this,
since the machine in question now runs Linux+Xen instead of Linux+KVM.
I have no other KVM machine.)

Not sure about emulated drive controller, and how to control that.  I've
used file:/xen/freebsd/32/v100/disk.img,hda,w' mostly, but also tried
xvda instead of hda.

I have been persistent, trying at least 30 installs of 10.0-foo for
several values of foo (BETA, RC, RELEASE, STABLE) and on the various
systems above.

It seems quite likely that this is very easy to confirm, given that
every one of my attempts has been a failure.

Note that the actual installation procedure apparently works.  After
reboot some simple repeated filesystem command, say

  sha1 /bin/cat
  sha1 /bin/cat

will almost certainly yield different results.

I've tested vfs.unmapped_buf_allowed=0 only for one machine (since I saw
that only recently) and it works in that case.  It is the AMD piledriver
running Linux+Xen, 512 MiB RAM.


Torbjörn
Please encrypt, key id 0xC8601622