[Xen-ia64-devel] Re: [Xen-devel] Help? Red Hat fails, Suse/Debian both work fine

2006-03-01 Thread Tristan Gingold
Le Mercredi 01 Mars 2006 02:30, Xu, Anthony a écrit :
 It is likely some subtle difference (or bug), perhaps in mmap?
Hi all,

yesterday I got the same bug as Akio.  Today it is working again.  I have just 
reinstalled tools+xen+kernel.
Akio, could you try to reinstall xen+tools+kernel using a clean repository ?
(make clean may be not enough).

It seems *really* strange that only a few of us got the bug because it seems 
to be related only to Xen+Xenstore+Linux.  Therefore it should not depend on 
the distribution.

Tristan.



___
Xen-ia64-devel mailing list
Xen-ia64-devel@lists.xensource.com
http://lists.xensource.com/xen-ia64-devel


RE: [Xen-ia64-devel] Why is cpl read as 0 in para-virtualization ?

2006-03-01 Thread Magenheimer, Dan (HP Labs Fort Collins)
 Subject: [Xen-ia64-devel] Why is cpl read as 0 in 
 para-virtualization ?
 
 while I was modifying xen_early_setup, I tried to replace dcr 
 kludge using 
 cpl.  Unfortunatly, it doesn't work since cpl is read as 0 
 (see vcpu.c).
 So my question is simple: why ?

I think some places in Linux/ia64 determine whether code
is running in kernel or user mode by checking psr.cpl.

___
Xen-ia64-devel mailing list
Xen-ia64-devel@lists.xensource.com
http://lists.xensource.com/xen-ia64-devel


Re: [Xen-ia64-devel] Why is cpl read as 0 in para-virtualization ?

2006-03-01 Thread Tristan Gingold
Le Mercredi 01 Mars 2006 16:44, Magenheimer, Dan (HP Labs Fort Collins) a 
écrit :
  Subject: [Xen-ia64-devel] Why is cpl read as 0 in
  para-virtualization ?
 
  while I was modifying xen_early_setup, I tried to replace dcr
  kludge using
  cpl.  Unfortunatly, it doesn't work since cpl is read as 0
  (see vcpu.c).
  So my question is simple: why ?

 I think some places in Linux/ia64 determine whether code
 is running in kernel or user mode by checking psr.cpl.
Correct.
One place is user_mode() is ptrace.h.
Currently it is true if cpl != 0.  I have made a simple test: I have changed 
the code to cpl == 3, and cpl reads as 2.  Linux is booting (dom0+domU).

Tristan.


___
Xen-ia64-devel mailing list
Xen-ia64-devel@lists.xensource.com
http://lists.xensource.com/xen-ia64-devel


[Xen-ia64-devel] RE: Please pull xen-ia64-unstable

2006-03-01 Thread Magenheimer, Dan (HP Labs Fort Collins)
Hi Keir --

Please pull

http://xenbits.xensource.com/ext/xen-ia64-unstable.hg

Includes catchup with various xen-unstable changes through
9023, various bug fixes and cleanups, some preparatory work
for moving domain0 from p==m to VP, and default to SMP host.

Thanks,
Dan

___
Xen-ia64-devel mailing list
Xen-ia64-devel@lists.xensource.com
http://lists.xensource.com/xen-ia64-devel


[Xen-ia64-devel] Re: [Xen-devel] Help? Red Hat fails, Suse/Debian both work fine

2006-03-01 Thread Akio Takebe
Hi, Tristan

Thank you. I'll try to reinstall soon.
If reinstall is a root cause, it is mean some tools are not overwrite 
when we run make install-tools.

Best Regards,

Akio Takebe

Le Mercredi 01 Mars 2006 02:30, Xu, Anthony a ィヲcrit :
 It is likely some subtle difference (or bug), perhaps in mmap?
Hi all,

yesterday I got the same bug as Akio.  Today it is working again.  I have 
just 
reinstalled tools+xen+kernel.
Akio, could you try to reinstall xen+tools+kernel using a clean repository ?
(make clean may be not enough).

It seems *really* strange that only a few of us got the bug because it seems 
to be related only to Xen+Xenstore+Linux.  Therefore it should not depend on 
the distribution.

Tristan.




___
Xen-ia64-devel mailing list
Xen-ia64-devel@lists.xensource.com
http://lists.xensource.com/xen-ia64-devel


[Xen-ia64-devel] RE: [Xen-devel] Help? Red Hat fails, Suse/Debian both work fine

2006-03-01 Thread Xu, Anthony
From: Akio Takebe 
Sent: 2006年3月2日 8:44
Thank you for your advice.
I checked retun value of mmap(), and it is not NULL.
I'll check vcpu_translate().

Sorry for not clear, we need to check if the return address of mmap is in 
region0,
Current vcpu_translate may not handle all scenarios when fault address
is in region 0.

Thanks,
Anthony

___
Xen-ia64-devel mailing list
Xen-ia64-devel@lists.xensource.com
http://lists.xensource.com/xen-ia64-devel


RE: [Xen-ia64-devel] Why is cpl read as 0 in para-virtualization ?

2006-03-01 Thread Tian, Kevin
From: Magenheimer, Dan (HP Labs Fort Collins)
Sent: 2006年3月2日 1:01
 
  I think some places in Linux/ia64 determine whether code
  is running in kernel or user mode by checking psr.cpl.
 Correct.
 One place is user_mode() is ptrace.h.
 Currently it is true if cpl != 0.  I have made a simple test:
 I have changed
 the code to cpl == 3, and cpl reads as 2.  Linux is booting
 (dom0+domU).

IIRC, there was at least one place in assembly code which
checked psr.cpl==0 (maybe mca code?) but this was a longtime
ago (possibly 2.4.x) so the code may be gone.

Yes, MCA recover assembly code has that check. Also alt dtlb miss
handler has similar check to prevent user space speculation from 
polluting the TLB. Temporarily hacking one macro doesn't solve 
anything. You need to capture every places (rare or not) to eliminate 
all knowledges based on 4 ring levels. However, I didn't see necessity 
by far.

Thanks,
Kevin

There is Itanium architecture work going on to establish a
mechanism for determining whether an OS is running native or
as a (paravirtualized or fully-virtualized) guest.  I don't
know the status of this but am pretty sure it was checking
a currently reserved bit, possibly in cr.dcr.

Dan

___
Xen-ia64-devel mailing list
Xen-ia64-devel@lists.xensource.com
http://lists.xensource.com/xen-ia64-devel

___
Xen-ia64-devel mailing list
Xen-ia64-devel@lists.xensource.com
http://lists.xensource.com/xen-ia64-devel


[Xen-ia64-devel] [PATCH] XEN: Avoid double reclaim dom0 image and dom0 initrd

2006-03-01 Thread Xu, Anthony
VMM doesn't need to reclaim dom0 image and dom0 initrd by using 
init_domheap_pages, because they have been reclaimed by below code:
in function start_kernel of xensetup.c
efi_memmap_walk(filter_rsvd_memory, init_boot_pages);
Reclaiming again may cause xen hang at very beginning.

BTW, dom0 initrd is for dom0 not for xen, it will be copied into dom0 
memory space, so VMM doesn't need to reserve memory for dom0 initrd.
Commenting below definition makes dom0 initrd reclaimable.
-#define CONFIG_BLK_DEV_INITRD // needed to reserve memory for domain0
+//#define CONFIG_BLK_DEV_INITRD // needed to reserve memory for domain0

Signed-off-by: Anthony Xu [EMAIL PROTECTED]

Thanks,
-Anthony 



avoid_double_mem_reclaim.patch
Description: avoid_double_mem_reclaim.patch
___
Xen-ia64-devel mailing list
Xen-ia64-devel@lists.xensource.com
http://lists.xensource.com/xen-ia64-devel