Alastair Tse [Thu, 18 Jan 2007 15:52:23 +0000 (15:52 +0000)]
[XEND] Change SrvServer to not require sockets to be non-blocking.
Set all spawned server threads to be daemonic and make sure their
sockets are closed properly using shutdown() and close(), otherwise
reload will complain about the "Address still in use."
[XEN] Simplify CR3 switching code (remove slow path that allows old
page tables to be writable in new page tables -- shadow mode cannot
handle the failure case and it is easy to work around this in the
guest anyway). Also remove broken shadow-refcount compat-mode case. Signed-off-by: Keir Fraser <keir@xensource.com>
The function pit_load_count incorrectly references current for
determining whether the channel number is zero. This breaks
when starting a new guest because current points to dom0.
The fix is to explicitly pass the address for verification.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Microcode does not need to have the default size of 2000+48 bytes.
A corresponding patch has gone into Linux 2.6.19; this is the
port to Xen.
Signed-off-by: Kurt Garloff <kurt@garloff.de>
Reference: Patch to linux kernel from Sep 27 2006
> # User Shaohua Li <shaohua.li@intel.com>
> # Node ID 45898b908138b5d93c2cc7353f061ce54af145dc
> # Parent f962eab7b82c9bf1a6da69571046e764f5128395
> [PATCH] x86 microcode: don't check the size
>
> IA32 manual says if micorcode update's size is 0, then the size is
> default size (2048 bytes). But this doesn't suggest all microcode
> update's size should be above 2048 bytes to me. We actually had a
> microcode update whose size is 1024 bytes. The patch just removed
> the check.
>
> Signed-off-by: Shaohua Li <shaohua.li@intel.com>
> Cc: Tigran Aivazian <tigran@veritas.com>
> Signed-off-by: Andrew Morton <akpm@osdl.org>
> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
>
> committer: Linus Torvalds <torvalds@g5.osdl.org> 1159370778 -0700
linux/i386: enhance dump_fault_path() in the highpte case
As long as the pte page isn't really located in highmem, there is no
reason to not access it, in order to print the complete page table
hierarchy. A functionally similar patch will go to lkml for native
Linux.
George Dunlap [Tue, 16 Jan 2007 19:04:12 +0000 (14:04 -0500)]
[XEN] Fix early-unshadow detection for 3- or 4-level guest pagetables.
Early-unshadow will unshadow whenever 2 zero values are written
to the same page; for PAE, one PTE takes 2 writes. Only check
for the early unshadow when writing the low half of the PTE.
Steven Hand [Tue, 16 Jan 2007 10:02:50 +0000 (10:02 +0000)]
Enable lazy (on-demand) allocation of memory to a guest being restored; this
means that ballooned down domains only require as much memory as is currently
being used (rather than their max) when being restored from save, or when
being migrated.
Allow extra application objects to be linked against a partial link of
minios with non-global names localised. For now we assume that global
names ill be assigned a xenos_ prefix in due course. We may decide to
choose a different prefix.
Those who wish to finish the link from an external Makefile can link
their own objects against mini-os.o (which must be first in link
order!) and specify the appropriate architecture linker script.
[XEN] Fix some e820 start-of-day issues by clipping all E820_RAM
regions to be page-sized and -aligned. Signed-off-by: Keir Fraser <keir@xensource.com>
Ian Campbell [Fri, 12 Jan 2007 16:03:33 +0000 (16:03 +0000)]
[PATCH] kexec/kdump: allow zero start for crashkernel
Some architectures, notably ia64, can automatically place
the crash kernel into an appropriate place if the start
address for crashkernel is specified as zero or omitted.
E.g. crashkernel=256M or crashkernel=256M@0
While xen does not actually have support for ia64 kexec,
I am working on it. And in any case this change should be harmless.
Just a small bit of infastructure that will make things easier in the
future.
This should also be possible on x86, now relocatable kernels
are reloacatble on x86. So once xen moves up to using linux ~2.6.19
it would make sense to look into implementing this.
The patch does 3 things.
* Firstly, parse_crashkernel() is modified to allows the size and start elements of kexec_crash_area
to be set to any value, including zero. Previously if either size or
start were specified as zero, they would both end up as zero.
* Secondly, when kexec_get() is called, if either the size or start
elements of kexec_crash_area are zero, then zero is passed
back to the hypercall caller for both values.
This gives the same behaviour as having parse_crashkernel() set
size and start to zero if either of them are zero, but it allows
architecture sepcific code called between the invocation of
parse_crashkernel() and kexec_get() to modify start (and size if
there was a reason). In particular, this allows the
architecture specific code to find a good start point if 0 is
specified.
* Lastly, it ads an additional check to the x86 setup code.
As this code currently does not know how to deal with a
0 start address, it doesn't reserve memory if start is 0.
This is neccessary as previously if start was specified as zero,
parse_crashkernel() would set size to zero, but that is
no longer the case, so a stronger check is needed.
I would really appreciate it if this patch was merged,
as I don't believe it harms anything, and it does allow
a nice path for moving forwards.
linux: Add trivial forwarding of dma_{,un}map_page when not using highmem
Use non-trivial dma_{un,}map_page only when CONFIG_HIGHMEM, allowing
to not compile swiotlb_{un,}map_page in that case, which in neither
case need exporting.
[LINUX] ipv6: Disable addrconf on Xen bridge device
The ipv6-no-autoconf patch didn't disable IPv6 addrconf completely.
This means that the Xen bridge device still interfered with normal
IPv6 operation by engaging the IPv6 network with a bogus MAC address.
For details please refer to
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=200360
The following patch completely disables IPv6 on the Xen bridge device
by temporarily setting the MTU to a value less than the minimum
allowed for IPv6.
Upstream will provide a cleaner way to disable IPv6 addrconf in
future, possibly in the form of a proc sysctl. Of course if the Xen
loopback device is removed it would render this change unnecessary.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
[IA64]: noreturn cannot be used if function may return
The functions die_if_kernel and vmx_die_if_kernel can certainly
return. This disqualifies them from using the noreturn keyword
which is reserved for functions that never return.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
x86: properly handle LOCK prefix in privileged PV opcode emulation
- include LOCK prefix presence in calculation of which CR/DR is being
targeted by moves to/from these registers
- fail any other opcodes when beingused with lock prefix
While, as discussed, the performance impact of this option is
certainly higher than on native Linux, the option should not be
entirely disallowed if people want to sacrifice performance for less
lowmem pressure.
[XEN] Consistent assumption that PTEs contain MFNs in both ptwr
page-fault handler and ptwr emulation handler. Signed-off-by: Keir Fraser <keir@xensource.com>
x64 SMP Vista HVM guest uses HPET as the main system timer, and it
uses physical destination mode with broadcast to deliver the interrupts
generated by HPET. In current code, timer interrupts are injected only
to VCPU0 in vioapic.c, but this doesn't satisfy x64 SMP Vista -- when
it boots, it complains "a clock interrupt was not received on a
secondary processor within the allocated time interval" with Bug Check
0x101.
Ewan Mellor [Thu, 11 Jan 2007 19:00:35 +0000 (19:00 +0000)]
This patch does the following:
- renames the XenManagedDomain.py file to XenAPIDomain.py, since this
name better reflects its functionality/purpose
- adds domain tracking to the XenAPIDomain class so that xend-managed
domains can be deleted in an 'atexit' handler upon test case termination
- adds one basic xapi-related test which is part of the grouptests
'xapi'
- refactors the vtpm-related test using xen-api and adds it to the
grouptest 'xapi'
- adds documentation to the README for how to configure xm and xend to
use XML-RPC or Xen-API for communication
Ewan Mellor [Thu, 11 Jan 2007 18:56:59 +0000 (18:56 +0000)]
I added a parameter '-md' to 'runtest.sh' for running the xm test suite
in a mode where all created domains a created as xend-managed domains.
This patch also fixes a problem related to calling 'xm domid' on a
currently suspended domain if that domain is a managed domain. In that
case a 'None' is returned by Xend, which I default to '-1'.
Use strstr() to look for "bimodal" string in ELF notes, to allow guests to use
"yes,bimodal", so they are correctly identified as PAE on older hypervisors.
[XEN] More emulator fixes:
1. Emulate LAHF/SAHF instructions
2. #GP if instruction is longer than 15 bytes
3. Accept any number of prefix bytes (up to
15-byte total instruction length)
4. Repeated addr/data-size overrides are sticky
rather than toggling.
Ian Campbell [Wed, 10 Jan 2007 09:39:24 +0000 (09:39 +0000)]
[PATCH] kexec/kdump: remove unnecessary incusion of asm/fixmap.h
I think this a legacy of an older revision of this code,
but asm/fixmap.h does not seem to be needed in
asm-x86/x86_32/kexec.h or asm-x86/x86_64/kexec.h
That is, neither of these incarntations of kexec.h seem
to do anything related to fixmap directly or indirectly.
projects / xen.git / log