kfraser@localhost.localdomain
6b5e36a305
linux: improve x86 page table handling performance
...
Where possible,
- use hypercalls instead of writing to read-only pages
- fold TLB flushes into page table update hypercalls
- on PAE, use single-access updates instead of two-access ones
The single change to PAE pte_clear() yields a 25-30% boost for kernel
builds on a 4x2x2 CPUs, 8Gb box; the other changes together yield
improvements of 2-5%.
Also, adjust backward compatibility handling in a few more places.
Signed-off-by: Jan Beulich <jbeulich@novell.com >
2007-04-05 09:10:33 +01:00
kfraser@localhost.localdomain
c937f696d7
linux/x86-64: fix unwind information
...
Mostly adjusting mis-merges dating back to the 2.6.14 merge.
Signed-off-by: Jan Beulich <jbeulich@novell.com >
2007-04-05 08:59:12 +01:00
kfraser@localhost.localdomain
c93368e0fb
linux: make compatibility config options more scalable
...
Signed-off-by: Jan Beulich <jbeulich@novell.com >
2007-04-05 08:58:32 +01:00
Steven Hand
1b46e1faf3
Fix hvm balloon driver; missed this hunk in last checkin.
...
Signed-off-by: Steven Hand <steven@xensource.com >
2007-04-05 08:24:56 +01:00
Keir Fraser
ed34f48aa8
Various cleanups and fixes for PV-on-HVM drivers.
...
Signed-off-by: Keir Fraser <keir@xensource.com >
2007-04-04 21:49:04 +01:00
Steven Hand
1d3e27e802
Initial check-in to support PV balloon driver within HVM guests.
...
Still todo:
- fix mapcache invalidation (should happen in Xen)
- support 32-on-64 mode correctly
Signed-off-by: Steven Hand <steven@xensource.com >
2007-04-04 19:59:10 +01:00
Ian Campbell
76de76e0f3
RFLAGS is in a known good state at start of day so no need to explicitly initialize.
...
Signed-off-by: Ian Campbell <ian.campbell@xensource.com >
2007-04-04 15:57:49 +01:00
Keir Fraser
75ba97629b
linux: Remove a few redundant config options. Build util.c for ia64
...
and ifdef out a couple of functions that are not yet fully generic.
Signed-off-by: Keir Fraser <keir@xensource.com >
2007-03-31 19:02:09 +01:00
Keir Fraser
f24c3877e5
linux: Avoid use of ptep_get_and_clear_full name as it's a macro on
...
some architectures.
Signed-off-by: Keir Fraser <keir@xensource.com >
2007-03-31 17:22:46 +01:00
Keir Fraser
c15c303a40
linux gntdev: Mark x86-specific code as such with ifdef.
...
Signed-off-by: Keir Fraser <keir@xensource.com >
2007-03-31 16:39:14 +01:00
Keir Fraser
990f0419bd
linux: User-space grant table device.
...
A character device for accessing (in user-space) pages that have been
granted by other domains.
Signed-off-by: Derek Murray <Derek.Murray@cl.cam.ac.uk >
Fix ioctl interface to be 32/64-bit invariant. Move xen_class
mechanism to common util.c.
Signed-off-by: Keir Fraser <keir@xensource.com >
2007-03-31 13:53:24 +01:00
Keir Fraser
7fd0f937ba
linux: Add a hook before a page table entry is cleared, for use with
...
the grant-table device.
Signed-off-by: Derek Murray <Derek.Murray@cl.cam.ac.uk >
2007-03-31 12:42:02 +01:00
Ian Campbell
ea1303c633
[LINUX] Explcitly include linux/mutex.h in files which use mutexes, rather then
...
relying on an include via some other header.
Signed-off-by: Ian Campbell <ian.campbell@xensource.com >
2007-03-30 17:33:31 +01:00
Ian Campbell
ea62867372
[LINUX] Make sure to initialiase the xen features bitmap before first use on x86_64.
...
Signed-off-by: Ian Campbell <ian.campbell@xensource.com >
2007-03-30 14:13:53 +01:00
Tim Deegan
9dd742c907
[HVM] Save/restore: PV-on-HVM driver save/restore support
...
Signed-off-by: Zhai Edwin <edwin.zhai@intel.com >
2007-03-29 16:27:52 +00:00
Keir Fraser
abaff6993d
pciback: Devices can lose config space info on D3->D0 transition, yet
...
they may not (and even if they try, can not) restore the BARs. Since
Linux does not do this for us, pciback must do it for itself manually.
Possible todo: Restore all first 64 bytes of the config space?
Signed-off-by: Alex Williamson <alex.williamson@hp.com >
Signed-off-by: Keir Fraser <keir@xensource.com >
2007-03-28 20:19:11 +01:00
kfraser@localhost.localdomain
6a6ebb9210
linux pciback: Proper fix for build failure.
...
Signed-off-by: Jan Beulich <jbeulich@novell.com >
2007-03-28 15:01:24 +01:00
kfraser@localhost.localdomain
573fcbc6f4
linux: Fix the pciback build (temporary hack).
...
Signed-off-by: Keir Fraser <keir@xensource.com >
2007-03-28 14:42:54 +01:00
kfraser@localhost.localdomain
8588d662d4
xenbus: check function return values of functions declared with
...
__must_check post-2.6.18.
This implies completely suppressing frontend/backend operations if the
respective bus or device registration failed.
Signed-off-by: Jan Beulich <jbeulich@novell.com >
2007-03-28 13:56:22 +01:00
kfraser@localhost.localdomain
3352268c6c
pciback: check function return values of functions declared with
...
__must_check post-2.6.18.
Signed-off-by: Jan Beulich <jbeulich@novell.com >
2007-03-28 13:55:01 +01:00
kfraser@localhost.localdomain
2405a261cd
linux evtchn: Read function in evtchn driver requires locking.
...
Signed-off-by: Keir Fraser <keir@xensource.com >
2007-03-28 13:52:03 +01:00
Christian Limpach
a712e471e2
[linux] Fix xenbus device to allow replies larger than 4096 bytes.
...
Also fix reply_queue mutex across fragmented replies and for concurrent
readers.
Signed-off-by: Christian Limpach <Christian.Limpach@xensource.com >
2007-03-28 11:01:58 +01:00
kfraser@localhost.localdomain
8e90ca4a3d
Merge with xen-ia64-unstable.hg
2007-03-28 10:38:41 +01:00
kfraser@localhost.localdomain
f3bdb5082b
linux/x86-64: cleanup the Xen-specific macros in entry-xen.S
...
- SWITCH_TO_KERNEL is dead (and was broken)
- comments to HYPERVISOR_IRET were somewhat outdated
- updates of selecor RPLs shouldn't be done with byte size to avoid
store-to-load forwarding issues
Signed-off-by: Jan Beulich <jbeulich@novell.com >
2007-03-27 17:23:33 +01:00
kfraser@localhost.localdomain
d0e5cf5462
linux: Fix for missing response to XS_UNWATCH request.
...
From: Magnus Carlsson <magnus@galois.com >
Signed-off-by: Keir Fraser <keir@xensource.com >
2007-03-27 16:19:25 +01:00
kfraser@localhost.localdomain
2ebb7bdf0d
linux: Fix a compiler warning (unused local variable).
...
Signed-off-by: Jan Beulich <jbeulich@novell.com >
2007-03-27 11:54:16 +01:00
awilliam@xenbuild2.aw
5b73247b5a
[IA64] Remove P2M workaround
...
The ia64 p2m semantic is now fixed. remove unnecessary work around patch.
Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp >
2007-03-26 09:17:25 -06:00
awilliam@xenbuild2.aw
5ea306faca
merge with xen-unstable.hg
2007-03-22 09:30:54 -06:00
Ian Campbell
27c451323a
LINUX: x86_64: start_pfn should be initialised even when XENFEAT_auto_translated_physmap is set.
...
Signed-off-by: Ian Campbell <ian.campbell@xensource.com >
2007-03-21 08:51:51 +00:00
awilliam@xenbuild2.aw
ebdb5845be
merge with xen-unstable.hg
2007-03-20 15:19:38 -06:00
Ian Campbell
3b8a13159a
[LINUX] Update crash-kernel-32-on-64.patch with more sensible version from -mm.
...
The previous version had side effects outside of the crash dump code.
Signed-off-by: Ian Campbell <ian.campbell@xensource.com >
2007-03-19 14:00:07 +00:00
Keir Fraser
c95d97855e
Fix up nosegneg vsyscall elfnote.
...
Signed-off-by: Keir Fraser <keir@xensource.com >
2007-03-17 12:24:32 +00:00
kfraser@localhost.localdomain
fbf7473b3a
linux: Various cleanup and locking clarification (and fixing!)
...
of mm pinning/unpinning logic.
Signed-off-by: Keir Fraser <keir@xensource.com >
2007-03-16 15:19:08 +00:00
kfraser@localhost.localdomain
3a4b9dc822
Revert CONFIG_SPLIT_PTLOCK_CPUS change of 14209:1061c7bff38f3e6e66fee96c407977ad55b3f2f2.
...
Jan Beulich points out that it's not safe against mm pinning and unpinning.
Signed-off-by: Keir Fraser <keir@xensource.com >
2007-03-16 15:18:33 +00:00
kfraser@localhost.localdomain
f326023b83
xenbus: Clarify and simplify barrier usage.
...
Signed-off-by: Keir Fraser <keir@xensource.com >
2007-03-16 10:42:25 +00:00
Ian Campbell
6b0c303dfe
[LINUX] Don't pin the initial page tables on x86_64 when XENFEAT_writable_page_tables.
...
Signed-off-by: Ian Campbell <ian.campbell@xensource.com >
2007-03-16 10:24:56 +00:00
awilliam@xenbuild2.aw
bd56728806
merge with xen-unstable.hg
2007-03-15 09:04:23 -06:00
kfraser@localhost.localdomain
4da77d5b3f
net back: Quieten loud error path.
...
Signed-off-by: Keir Fraser <keir@xensource.com >
2007-03-15 14:51:32 +00:00
Jake Wires
5fc295e4de
[BLKTAP] remove spurious error message
...
Signed-off-by: Jake Wires <jwires@xensource.com >
2007-03-14 14:18:51 -07:00
Ian Campbell
a84c4f75e5
[LINUX] Kexec: Ensure that pages allocated for kexec have MFNs which are within
...
the required limits.
Previously we were lucky since most dom0 pages would be low enough but now Xen
is more aggressive in giving out higher MFNs when it can.
Signed-off-by: Ian Campbell <ian.campbell@xensource.com >
2007-03-14 16:39:32 +00:00
kfraser@localhost.localdomain
26e24b46d1
linux: No need to force pcibios_scan_all_fns() on modern Xen.
...
Signed-off-by: Keir Fraser <keir@xensource.com >
2007-03-13 13:28:26 +00:00
Ian Campbell
f63d9525d5
[LINUX] Fix typo so 3.0.4 compatability works.
...
Signed-off-by: Ian Campbell <ian.campbell@xensource.com >
2007-03-13 09:07:55 +00:00
awilliam@xenbuild2.aw
04a0164bad
merge with xen-unstable.hg
2007-03-12 10:07:08 -06:00
awilliam@xenbuild2.aw
5171dd3120
[IA64] PV steal time accounting
...
Signed-off-by: Atsushi SAKAI <sakaia@jp.fujitsu.com >
2007-03-12 09:39:59 -06:00
kfraser@localhost.localdomain
79ae7e51e1
linux: Adjust memory allocation condition in netback to reduce memory
...
pressure on Xen and avoid unnecessarily stalling receives.
Signed-off-by: Jan Beulich <jbeulich@novell.com >
2007-03-12 14:55:02 +00:00
kfraser@localhost.localdomain
415ecbe777
linux: No need to manually bring VCPUs online across 'fast' save/restore.
...
Signed-off-by: Keir Fraser <keir@xensource.com >
2007-03-12 14:11:43 +00:00
kfraser@localhost.localdomain
e260cd0a07
The Xen console intercepts ^O and prevents it from propagating
...
through to the guest. However, ^O is a useful key in bash and in
Emacs. This patch changes the console to pass a single ^O
through to the guest when it is pressed twice.
Signed-off-by: Ben Pfaff <blp@cs.stanford.edu >
2007-03-12 11:04:34 +00:00
kfraser@localhost.localdomain
aca8807ff0
Fix valid_phys_addr_range() prototype of xen /dev/mem.
...
The second argument is size_t, not size_t *.
Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp >
2007-03-12 11:03:07 +00:00
kfraser@localhost.localdomain
ac4fb3c4ec
pci back: Various fixes.
...
- remove superfluous call to pciback_field_is_dup()
- fix a variable type mismatch in pciback_field_is_dup()
- make readability improvements by using the OFFSET macro
- revises quirk data output via the sysfs quirks node so that
- displayed offset information includes base_offset.
Thanks to Jambunathan K. for giving such specific bug diagnosis.
Signed-off-by: Chris Bookholt <hap10@tycho.ncsc.mil >
2007-03-10 17:01:21 +00:00
kfraser@localhost.localdomain
505ef3ea86
Add flags field to VCPUOP_set_singlsehot_timer.
...
Flag 'future' causes Xen to check if the timeout is in the past and
return -ETIME if so.
From: Jeremy Fitzhardinge <jeremy@goop.org >
Signed-off-by: Keir Fraser <keir@xensource.com >
2007-03-10 16:58:11 +00:00