Re: [PATCH] abstract out bits of ldt.c

2005-08-09 Thread Christoph Hellwig
On Tue, Aug 09, 2005 at 11:23:18AM +0200, Ingo Molnar wrote: > mine are mostly technical arguments. I just also wanted to vent away > this slowly gathering false notion of building 'interoperability', while > the only apparent goal seems to be to maximize benefits to the closed > hypervisors,

Re: [PATCH] abstract out bits of ldt.c

2005-08-09 Thread Ingo Molnar
* Andrew Morton <[EMAIL PROTECTED]> wrote: > > Furthermore, why should we hamper Xen by going to _any_ sort of > > "formalized" hypervisor API, when we dont even know what we want, as Xen > > is pretty much work in progress? And whatever Xen support is exported > > from the kernel, it should

Re: [PATCH] abstract out bits of ldt.c

2005-08-09 Thread Christoph Hellwig
On Tue, Aug 09, 2005 at 11:23:18AM +0200, Ingo Molnar wrote: mine are mostly technical arguments. I just also wanted to vent away this slowly gathering false notion of building 'interoperability', while the only apparent goal seems to be to maximize benefits to the closed hypervisors, while

Re: [PATCH] abstract out bits of ldt.c

2005-08-07 Thread Chris Wright
* Zachary Amsden ([EMAIL PROTECTED]) wrote: > Doesn't that require 16 pages per CPU? That seems excessive to impose > on a native build. Perhaps we could get away with 1 page per CPU for > the GDT on native boots and bump that up to 16 if compiling for a > virtualized sub-architecture - i.e.

Re: [PATCH] abstract out bits of ldt.c

2005-08-07 Thread Martin J. Bligh
> I like these patches. They greatly simplify a lot of the work I > had anticipated was necessary for Xen. I.e. - LDT / GDT accessors > are not needed for most updates, only updates to live descriptor > table entries (for GDT this is TLS, LDT, TSS?, entries and there > is 1 LDT update case).

Re: [PATCH] abstract out bits of ldt.c

2005-08-07 Thread Zachary Amsden
Chris Wright wrote: * Zachary Amsden ([EMAIL PROTECTED]) wrote: Does Xen assume page aligned descriptor tables? I assume from this Yes. patch and snippets I have gathered from others, that is a yes, and other things here imply that DT pages are not shadowed. If so, Xen itself

Re: [PATCH] abstract out bits of ldt.c

2005-08-07 Thread Chris Wright
* Zachary Amsden ([EMAIL PROTECTED]) wrote: > Does Xen assume page aligned descriptor tables? I assume from this Yes. > patch and snippets I have gathered from others, that is a yes, and other > things here imply that DT pages are not shadowed. If so, Xen itself > must have live segments in

Re: [PATCH] abstract out bits of ldt.c

2005-08-07 Thread Zachary Amsden
Chris Wright wrote: * Andrew Morton ([EMAIL PROTECTED]) wrote: "Martin J. Bligh" <[EMAIL PROTECTED]> wrote: xen_make_pages_readonly / xen_make_pages_writable ? Well we don't want to assume "xen" at this stage. We're faced with a choice at present: to make the

Re: [PATCH] abstract out bits of ldt.c

2005-08-07 Thread Martin J. Bligh
--Andrew Morton <[EMAIL PROTECTED]> wrote (on Sunday, August 07, 2005 17:41:29 -0700): > "Martin J. Bligh" <[EMAIL PROTECTED]> wrote: >> >> >> >> --Chris Wright <[EMAIL PROTECTED]> wrote (on Sunday, August 07, 2005 >> 16:44:11 -0700): >> >> > * Martin J. Bligh ([EMAIL PROTECTED]) wrote: >>

Re: [PATCH] abstract out bits of ldt.c

2005-08-07 Thread Chris Wright
* Andrew Morton ([EMAIL PROTECTED]) wrote: > "Martin J. Bligh" <[EMAIL PROTECTED]> wrote: > > xen_make_pages_readonly / xen_make_pages_writable ? > > Well we don't want to assume "xen" at this stage. We're faced with a > choice at present: to make the linux->hypervisor interface be some >

Re: [PATCH] abstract out bits of ldt.c

2005-08-07 Thread Andrew Morton
"Martin J. Bligh" <[EMAIL PROTECTED]> wrote: > > > > --Chris Wright <[EMAIL PROTECTED]> wrote (on Sunday, August 07, 2005 16:44:11 > -0700): > > > * Martin J. Bligh ([EMAIL PROTECTED]) wrote: > >> Starting on the work to merge xen cleanly as a subarch. > >> Introduce make_pages_readonly and

Re: [PATCH] abstract out bits of ldt.c

2005-08-07 Thread Dave Hansen
On Sun, 2005-08-07 at 16:20 -0700, Martin J. Bligh wrote: > Starting on the work to merge xen cleanly as a subarch. > Introduce make_pages_readonly and make_pages_writable where appropriate > for Xen, defined as a no-op on other subarches. Same for > add_context_to_unpinned and

Re: [PATCH] abstract out bits of ldt.c

2005-08-07 Thread Chris Wright
* Martin J. Bligh ([EMAIL PROTECTED]) wrote: > You're welcome to suggest something else if you want, though it would > have been easier if you'd done it the first time you saw this patch, > not now. Going through this stuff multiple times is going to get very > boring very fast. Sorry, that's my

Re: [PATCH] abstract out bits of ldt.c

2005-08-07 Thread Martin J. Bligh
--Chris Wright <[EMAIL PROTECTED]> wrote (on Sunday, August 07, 2005 16:44:11 -0700): > * Martin J. Bligh ([EMAIL PROTECTED]) wrote: >> Starting on the work to merge xen cleanly as a subarch. >> Introduce make_pages_readonly and make_pages_writable where appropriate >> for Xen, defined as a

Re: [PATCH] abstract out bits of ldt.c

2005-08-07 Thread Chris Wright
* Martin J. Bligh ([EMAIL PROTECTED]) wrote: > Starting on the work to merge xen cleanly as a subarch. > Introduce make_pages_readonly and make_pages_writable where appropriate > for Xen, defined as a no-op on other subarches. Same for Maybe this is a bad name, since

[PATCH] abstract out bits of ldt.c

2005-08-07 Thread Martin J. Bligh
Starting on the work to merge xen cleanly as a subarch. Introduce make_pages_readonly and make_pages_writable where appropriate for Xen, defined as a no-op on other subarches. Same for add_context_to_unpinned and del_context_from_unpinned. Abstract out install_ldt_entry(). This will do have no

[PATCH] abstract out bits of ldt.c

2005-08-07 Thread Martin J. Bligh
Starting on the work to merge xen cleanly as a subarch. Introduce make_pages_readonly and make_pages_writable where appropriate for Xen, defined as a no-op on other subarches. Same for add_context_to_unpinned and del_context_from_unpinned. Abstract out install_ldt_entry(). This will do have no

Re: [PATCH] abstract out bits of ldt.c

2005-08-07 Thread Chris Wright
* Martin J. Bligh ([EMAIL PROTECTED]) wrote: Starting on the work to merge xen cleanly as a subarch. Introduce make_pages_readonly and make_pages_writable where appropriate for Xen, defined as a no-op on other subarches. Same for Maybe this is a bad name, since make_pages_readonly/writable

Re: [PATCH] abstract out bits of ldt.c

2005-08-07 Thread Martin J. Bligh
--Chris Wright [EMAIL PROTECTED] wrote (on Sunday, August 07, 2005 16:44:11 -0700): * Martin J. Bligh ([EMAIL PROTECTED]) wrote: Starting on the work to merge xen cleanly as a subarch. Introduce make_pages_readonly and make_pages_writable where appropriate for Xen, defined as a no-op on

Re: [PATCH] abstract out bits of ldt.c

2005-08-07 Thread Chris Wright
* Martin J. Bligh ([EMAIL PROTECTED]) wrote: You're welcome to suggest something else if you want, though it would have been easier if you'd done it the first time you saw this patch, not now. Going through this stuff multiple times is going to get very boring very fast. Sorry, that's my

Re: [PATCH] abstract out bits of ldt.c

2005-08-07 Thread Dave Hansen
On Sun, 2005-08-07 at 16:20 -0700, Martin J. Bligh wrote: Starting on the work to merge xen cleanly as a subarch. Introduce make_pages_readonly and make_pages_writable where appropriate for Xen, defined as a no-op on other subarches. Same for add_context_to_unpinned and

Re: [PATCH] abstract out bits of ldt.c

2005-08-07 Thread Andrew Morton
Martin J. Bligh [EMAIL PROTECTED] wrote: --Chris Wright [EMAIL PROTECTED] wrote (on Sunday, August 07, 2005 16:44:11 -0700): * Martin J. Bligh ([EMAIL PROTECTED]) wrote: Starting on the work to merge xen cleanly as a subarch. Introduce make_pages_readonly and make_pages_writable

Re: [PATCH] abstract out bits of ldt.c

2005-08-07 Thread Chris Wright
* Andrew Morton ([EMAIL PROTECTED]) wrote: Martin J. Bligh [EMAIL PROTECTED] wrote: xen_make_pages_readonly / xen_make_pages_writable ? Well we don't want to assume xen at this stage. We're faced with a choice at present: to make the linux-hypervisor interface be some xen-specific and

Re: [PATCH] abstract out bits of ldt.c

2005-08-07 Thread Martin J. Bligh
--Andrew Morton [EMAIL PROTECTED] wrote (on Sunday, August 07, 2005 17:41:29 -0700): Martin J. Bligh [EMAIL PROTECTED] wrote: --Chris Wright [EMAIL PROTECTED] wrote (on Sunday, August 07, 2005 16:44:11 -0700): * Martin J. Bligh ([EMAIL PROTECTED]) wrote: Starting on the work to

Re: [PATCH] abstract out bits of ldt.c

2005-08-07 Thread Zachary Amsden
Chris Wright wrote: * Andrew Morton ([EMAIL PROTECTED]) wrote: Martin J. Bligh [EMAIL PROTECTED] wrote: xen_make_pages_readonly / xen_make_pages_writable ? Well we don't want to assume xen at this stage. We're faced with a choice at present: to make the linux-hypervisor

Re: [PATCH] abstract out bits of ldt.c

2005-08-07 Thread Chris Wright
* Zachary Amsden ([EMAIL PROTECTED]) wrote: Does Xen assume page aligned descriptor tables? I assume from this Yes. patch and snippets I have gathered from others, that is a yes, and other things here imply that DT pages are not shadowed. If so, Xen itself must have live segments in the

Re: [PATCH] abstract out bits of ldt.c

2005-08-07 Thread Zachary Amsden
Chris Wright wrote: * Zachary Amsden ([EMAIL PROTECTED]) wrote: Does Xen assume page aligned descriptor tables? I assume from this Yes. patch and snippets I have gathered from others, that is a yes, and other things here imply that DT pages are not shadowed. If so, Xen itself

Re: [PATCH] abstract out bits of ldt.c

2005-08-07 Thread Martin J. Bligh
I like these patches. They greatly simplify a lot of the work I had anticipated was necessary for Xen. I.e. - LDT / GDT accessors are not needed for most updates, only updates to live descriptor table entries (for GDT this is TLS, LDT, TSS?, entries and there is 1 LDT update case). I'm

Re: [PATCH] abstract out bits of ldt.c

2005-08-07 Thread Chris Wright
* Zachary Amsden ([EMAIL PROTECTED]) wrote: Doesn't that require 16 pages per CPU? That seems excessive to impose on a native build. Perhaps we could get away with 1 page per CPU for the GDT on native boots and bump that up to 16 if compiling for a virtualized sub-architecture - i.e. move