Many files:
- watch now takes a token, returned when reading watch
- More tests
- Fix domain shared page communication (flush output)
- Add "home" path for domains
- More permissions checks in various functions
- Simplify watch acknowledgement code and fix occasional bug
xs_watch_stress.c, 12readonly.sh, 11domain-watch.sh, 10domain-homedir.sh:
new file
xs_stress.c, xs_lib.h, xs_lib.c:
Cleanup whitespace.
ignore:
Add tools/xenstore/xs_watch_stress
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
Makefile:
Add xs_dom0_test program to test IOCTL_PRIVCMD_INITDOMAIN_STORE.
xs_dom0_test.c:
new file
From: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
Makefile:
Build libxenstore-pic.a.
setup.py:
Link with libxenstore-pic.a.
Makefile:
Build libxc-pic.a.
Build xenstore once is enough.
Signed-off-by: Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
privcmd.h, privcmd.c:
Add IOCTL_PRIVCMD_INITDOMAIN_STORE to setup dom0 xenstore page
and event channel.
From: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
Slightly disgusting hack to avoid using lots of lock instructions on a uniprocessor
machine just because we happened to compile with CONFIG_SMP. Essentially, we
make a big table of all of the instruction sequences which differ in ``easy''
ways between UP and SMP kernels, and then select which one to use at run time.
Signed-off-by: Steven Smith <sos22@cam.ac.uk>
[PATCH] x86_64 entry.S cleanup - take 3
This patch cleans up x86_64 entry.S. Namely, it updates the Xen relevant
macros to be the simpler version that's found in i386. This means that:
- XEN_[UN]BLOCK_EVENTS interface now takes care of dealing with
SMP issues and is no longer conditionally defined
- XEN_LOCKED_[UN]BLOCK_EVENTS is identical in both cases (SMP and UP)
and no longer needs to be conditionally defined
- XEN_[UN]LOCK_VPCU_INFO_SMP is dropped in favor of XEN_GET/PUT_VCPU_INFO
This cleans up the code, minimizes the differences with i386 code, and
lays the groundwork for SMP support (the real reason I did this ;-).
It's booting, executing syscalls, taking interrupts, etc (it's what I'm
using to send this e-mail).
Signed-off-by: Chris Wright <chrisw@osdl.org>
===== linux-2.6.11-xen-sparse/arch/xen/x86_64/kernel/entry.S 1.10 vs edited =====
[PATCH] 64-session-id.patch
session_id is an opaque pointer, so should be wider on 64 bit platforms.
Signed-off-by: Kevin Tian <kevin.tian@intel.com>
Signed-off-by: Arun Sharma <arun.sharma@intel.com>
[PATCH] x64-vmxassist-fix2.patch
Fix 64 bit build. Tested the binary built on a 64 bit platform on a 32 bit platform.
Signed-off-by: Arun Sharma <arun.sharma@intel.com>
fix compilation with newer binutils that don't accept the l suffix for mov when using segment register
Signed-off-by: Flávio Bruno Leitner <fbl@conectiva.com.br>
Signed-off-by: Vincent Hanquez <vincent@xensource.com>
user -> kernel syscall path must enable interrupts after all, since
it calls create_bounce_frame which can potentially fault. Thus we must
also test_all_events.
Signed-off-by: Keir Fraser <keir@xensource.com>
x86/64 linux needs rtc driver so that it can set correct kernel time.
todo: emulate (or pass-through) read accesses to rtc ports within xen,
so that domu can access rtc too.
Signed-off-by: Li B Xin <li.b.xin@intel.com>
Signed-off-by: Jun Nakajima <jun.nakajima@intel.com>
Adding the VMCS encodings as per "Intel Virtualization Technology
Specification for the IA-32 Intel Architecture", and removing hard
coded values. Code cleanup & preparation for 64bit VMX guest support.
Signed-Off-By: Chengyuan Li <chengyuan.li@intel.com>
Signed-Off-By: Yunhong Jiang <hunhong.jiang@intel.com>
Signed-Off-By: Jun Nakajima <jun.nakajima@intel.com>
Signed-Off-By: Nitin A Kamble <nitin.a.kamble@intel.com>
