xen.git
19 years ago[MINIOS] Fix the build after start_info interface changes.
kfraser@localhost.localdomain [Wed, 16 Aug 2006 15:48:45 +0000 (16:48 +0100)]
[MINIOS] Fix the build after start_info interface changes.
Signed-off-by: Anil Madhavapeddy <anil@xensource.com>
19 years ago[LINUX] Define is_initial_xendomain() to zero for HVM drivers.
kfraser@localhost.localdomain [Wed, 16 Aug 2006 15:16:32 +0000 (16:16 +0100)]
[LINUX] Define is_initial_xendomain() to zero for HVM drivers.
Signed-off-by: Steven Smith <ssmith@xensource.com>
19 years ago[LINUX] Update the hvm_op wrapper now that the return value of get_param is in
kfraser@localhost.localdomain [Wed, 16 Aug 2006 15:15:46 +0000 (16:15 +0100)]
[LINUX] Update the hvm_op wrapper now that the return value of get_param is in
the parameter structure rather than the hypercall return value.
Signed-off-by: Steven Smith <ssmith@xensource.com>
19 years ago[NET] front: Fix indentation of a comment.
kfraser@localhost.localdomain [Wed, 16 Aug 2006 15:14:03 +0000 (16:14 +0100)]
[NET] front: Fix indentation of a comment.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[NET] front: Allow packets to be copied on the receive path instead of flipped.
kfraser@localhost.localdomain [Wed, 16 Aug 2006 15:11:53 +0000 (16:11 +0100)]
[NET] front: Allow packets to be copied on the receive path instead of flipped.
Signed-off-by: Steven Smith <ssmith@xensource.com>
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[NET] back: Change 'copyall' xenstore node to 'request-rx-copy', which
kfraser@localhost.localdomain [Wed, 16 Aug 2006 15:11:12 +0000 (16:11 +0100)]
[NET] back: Change 'copyall' xenstore node to 'request-rx-copy', which
is slightly more informative and mirrors the 'feature-rx-copy' name
which communicates the feature in the other direction (back->front).
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years agoFix log rotation which was inadvertently broken by changeset 11129:939f75570a15.
emellor@leeni.uk.xensource.com [Wed, 16 Aug 2006 15:06:32 +0000 (16:06 +0100)]
Fix log rotation which was inadvertently broken by changeset 11129:939f75570a15.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
19 years ago[NET] back: Support copying packets to the frontend on receive path.
kfraser@localhost.localdomain [Wed, 16 Aug 2006 15:01:00 +0000 (16:01 +0100)]
[NET] back: Support copying packets to the frontend on receive path.
Signed-off-by: Steven Smith <ssmith@xensource.com>
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[HVM] Provide common support function for HLT emulation: hvm_hlt().
kfraser@localhost.localdomain [Wed, 16 Aug 2006 13:27:30 +0000 (14:27 +0100)]
[HVM] Provide common support function for HLT emulation: hvm_hlt().
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[NET] front: Fix features on resume when csum is off
kfraser@localhost.localdomain [Wed, 16 Aug 2006 13:26:59 +0000 (14:26 +0100)]
[NET] front: Fix features on resume when csum is off

When the netfront driver is resumed the features are renegotiated with
the backend.  However, I forgot take into account the status of the TX
checksum setting.  When TX checksum is disabled by the user, we cannot
enable SG or TSO since both require checksum offload.  This patch
makes
xennet check the checksum setting before renegotiating SG or TSO.

This bug was fixed thanks to a report from Anton Burtsev.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
19 years agoHeuristic to detect windows shutdown + power down domain.
shand@kneesaa.uk.xensource.com [Wed, 16 Aug 2006 13:01:03 +0000 (14:01 +0100)]
Heuristic to detect windows shutdown + power down domain.

Signed-off-by: Steven Hand <steven@xensource.com>
Signed-off-by: George Dunlap <george.dunlap@xensource.com>
19 years agoRename two xentrace tools to make it obvious that setmask and setsize are xen related.
kfraser@localhost.localdomain [Wed, 16 Aug 2006 12:53:10 +0000 (13:53 +0100)]
Rename two xentrace tools to make it obvious that setmask and setsize are xen related.

Based on a patch from Rob Gardner <rob.gardner@hp.com>

Signed-off-by: Keir Fraser <keir@xensource.com>
19 years agoImplement xm block-configure for CD-ROMs.
emellor@leeni.uk.xensource.com [Wed, 16 Aug 2006 11:59:30 +0000 (12:59 +0100)]
Implement xm block-configure for CD-ROMs.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
19 years ago[HVM] Fix the issue that the HVM 64bit guest cannot boot with 32G memory.
kfraser@localhost.localdomain [Wed, 16 Aug 2006 10:58:48 +0000 (11:58 +0100)]
[HVM] Fix the issue that the HVM 64bit guest cannot boot with 32G memory.
It enlarges the PGT_mfn_mask to support 27bit, and also remove the
score stuff on x86-64.

Signed-off-by: Jun Nakajima <jun.nakajima@intel.com>
Signed-off-by: Xiaohui Xin <xiaohui.xin@intel.com>
19 years ago[HVM] Fix the HVM hypercall issue when paging is not enabled yet on 64bit host.
kfraser@localhost.localdomain [Wed, 16 Aug 2006 10:55:26 +0000 (11:55 +0100)]
[HVM] Fix the HVM hypercall issue when paging is not enabled yet on 64bit host.
Signed-off-by: Xin Li <xin.b.li@intel.com>
19 years agoThis is a small patch that makes a cross compiler successfully compile
kfraser@localhost.localdomain [Wed, 16 Aug 2006 10:53:37 +0000 (11:53 +0100)]
This is a small patch that makes a cross compiler successfully compile
the sources under the tools directory.  This patch became necessary
after qemu-dm is updated to a newer version.

Signed-off-by: Hiroya INAKOSHI <inakoshi.hiroya@jp.fujitsu.com>
19 years agoWhitespace clean-ups.
shand@kneesaa.uk.xensource.com [Wed, 16 Aug 2006 10:36:13 +0000 (11:36 +0100)]
Whitespace clean-ups.

Signed-off-by: Steven Hand <steven@xensource.com>
19 years agoFix VT after hvm changes.
shand@kneesaa.uk.xensource.com [Tue, 15 Aug 2006 18:53:55 +0000 (19:53 +0100)]
Fix VT after hvm changes.

Signed-off-by: Steven Hand <steven@xensource.com>
19 years ago[XEND] xc_save/xc_restore open the libxc interface independently
kfraser@localhost.localdomain [Tue, 15 Aug 2006 18:04:43 +0000 (19:04 +0100)]
[XEND] xc_save/xc_restore open the libxc interface independently
of their parent. This is required now that the interface fd is
marked for close-on-exec.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years agoVarious HVM clean-ups.
shand@kneesaa.uk.xensource.com [Tue, 15 Aug 2006 17:20:03 +0000 (18:20 +0100)]
Various HVM clean-ups.

Signed-off-by: Steven Hand <steven@xensource.com>
19 years agoAdded missing decl.
emellor@leeni.uk.xensource.com [Tue, 15 Aug 2006 16:03:06 +0000 (17:03 +0100)]
Added missing decl.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
19 years ago[XEN] Add tracing to slow path in setting new guest %cr3 value.
kfraser@localhost.localdomain [Tue, 15 Aug 2006 15:31:14 +0000 (16:31 +0100)]
[XEN] Add tracing to slow path in setting new guest %cr3 value.
Also, allow x86/64 user cr3 value to be reset to zero.
Signed-off-by: Gerd Hoffmann <kraxel@suse.de>
19 years agoSet close-on-exec flag on /var/log/xend.log logging file.
kfraser@localhost.localdomain [Tue, 15 Aug 2006 15:27:26 +0000 (16:27 +0100)]
Set close-on-exec flag on /var/log/xend.log logging file.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
19 years agoSet the close-on-exec flag on the RPC socket in tool stack.
kfraser@localhost.localdomain [Tue, 15 Aug 2006 15:26:39 +0000 (16:26 +0100)]
Set the close-on-exec flag on the RPC socket in tool stack.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
19 years agoSet the close-on-exec flag on the xenstore fd in libxenstore.
kfraser@localhost.localdomain [Tue, 15 Aug 2006 15:25:39 +0000 (16:25 +0100)]
Set the close-on-exec flag on the xenstore fd in libxenstore.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
19 years ago[LIBXC] Set close-on-exec on the privcmd fd in libxc.
kfraser@localhost.localdomain [Tue, 15 Aug 2006 15:25:04 +0000 (16:25 +0100)]
[LIBXC] Set close-on-exec on the privcmd fd in libxc.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
19 years ago[HVM] Ensure the read-only M2P table is mapped without _PAGE_GLOBAL
kfraser@localhost.localdomain [Tue, 15 Aug 2006 15:21:12 +0000 (16:21 +0100)]
[HVM] Ensure the read-only M2P table is mapped without _PAGE_GLOBAL
otherwise context switch to shadow-mode-translate guests does not
work properly (they reuse that area of virtual address space).
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years agoMove vga options to their own 'vga=' boot parameter.
kfraser@localhost.localdomain [Tue, 15 Aug 2006 15:17:24 +0000 (16:17 +0100)]
Move vga options to their own 'vga=' boot parameter.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years agomerge
ack@localhost.localdomain [Tue, 15 Aug 2006 15:04:12 +0000 (16:04 +0100)]
merge

19 years agomerge
kfraser@localhost.localdomain [Tue, 15 Aug 2006 14:51:47 +0000 (15:51 +0100)]
merge

19 years agoClean up console fields in start_info structure. Add compat code for
kfraser@localhost.localdomain [Tue, 15 Aug 2006 14:50:36 +0000 (15:50 +0100)]
Clean up console fields in start_info structure. Add compat code for
older kernels so they can use the old start_info API.

Signed-off-by: Keir Fraser <keir@xensource.com>
19 years agomerge
ack@localhost.localdomain [Tue, 15 Aug 2006 14:49:16 +0000 (15:49 +0100)]
merge

19 years agoDon't copy all incoming fragmented packets.
ack@localhost.localdomain [Tue, 15 Aug 2006 14:48:31 +0000 (15:48 +0100)]
Don't copy all incoming fragmented packets.
Some drivers (ie the e1000) hand up packets with the header in
the main area and the data in a fragment. Unless there are
multiple references on any of the data, we don't need to
make a full copy of those packets.
Signed-off-by: Emmanuel Ackaouy <ack@xensource.com>
19 years ago[qemu patches] Update patches for changeset 11117:60eacb336e35.
chris@kneesaa.uk.xensource.com [Tue, 15 Aug 2006 12:54:09 +0000 (13:54 +0100)]
[qemu patches] Update patches for changeset 11117:60eacb336e35.

Signed-off-by: Christian Limpach <Christian.Limpach@xensource.com>
19 years ago[qemu/ia64] Sync ID cache only guest do DMA read.
chris@kneesaa.uk.xensource.com [Tue, 15 Aug 2006 12:53:18 +0000 (13:53 +0100)]
[qemu/ia64] Sync ID cache only guest do DMA read.
Fix ID cache sync time. Only when guest does DMA
read, ID sync cache is needed for VTi.

From: Zhang, Xiantao <xiantao.zhang@intel.com>
Signed-off-by: Christian Limpach <Christian.Limpach@xensource.com>
19 years agoClean up vga.c and change default font slot to zero, as other
kfraser@localhost.localdomain [Tue, 15 Aug 2006 10:36:10 +0000 (11:36 +0100)]
Clean up vga.c and change default font slot to zero, as other
slots are not supported by ATI video cards.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years agomerge
ack@localhost.localdomain [Tue, 15 Aug 2006 10:17:36 +0000 (11:17 +0100)]
merge

19 years agomerge
ack@localhost.localdomain [Tue, 15 Aug 2006 10:16:46 +0000 (11:16 +0100)]
merge

19 years agoSupport for vga text modes bigger than 80x25.
kfraser@localhost.localdomain [Tue, 15 Aug 2006 10:16:44 +0000 (11:16 +0100)]
Support for vga text modes bigger than 80x25.
Signed-off-by: Jan Beulich <jbeulich@novell.com>
19 years agoAdjust truesize in netfront to more closely reflect real data payload
ack@localhost.localdomain [Tue, 15 Aug 2006 10:13:04 +0000 (11:13 +0100)]
Adjust truesize in netfront to more closely reflect real data payload
plus supervisor overheads but keeping out gross hypervisor overheads.
Signed-off-by: Emmanuel Ackaouy <ack@xensource.com>
19 years ago[XEND] Apply the domain cpumask fully to every vcpu in the domain.
kfraser@localhost.localdomain [Tue, 15 Aug 2006 09:56:59 +0000 (10:56 +0100)]
[XEND] Apply the domain cpumask fully to every vcpu in the domain.

Apply the domain cpumask to each vcpu rather than mapping vcpus to
cpus in the list.  This is more inline with the comments for the
cpus parameter and also allows the credit scheduler to balance
vcpus within the domain cpumask.

Signed-off-by: Ryan Harper <ryanh@us.ibm.com>
19 years agoRemove another external-to-libxc user of xc_dom0_op().
kfraser@localhost.localdomain [Tue, 15 Aug 2006 09:47:26 +0000 (10:47 +0100)]
Remove another external-to-libxc user of xc_dom0_op().
Signed-off-by: John Levon <john.levon@sun.com>
19 years ago[VMX] Fix vmxassist for booting Vista.
kfraser@localhost.localdomain [Tue, 15 Aug 2006 09:45:05 +0000 (10:45 +0100)]
[VMX] Fix vmxassist for booting Vista.

This patch fixes 2 vmxassist bugs to boot Vista:
1) When switching from VM86_REAL_TO_PROTECTED to VM86_PROTECTED,
protected_mode() sets return eip to switch_to_protected_mode and
segments to vmxassist's. But in later set_mode(), it still consides
these segments are guest's (Bios at this time) and use its gdt table
to calculate address(), which results in segment limitation fault. 2) In
current vmxassist, when entering VM86_PROTECTED_TO_REAL, we need to
decode every instruction manually until we catch instructions like
jmpl to reset cs to really return to real mode. Vista boot manager uses
more instructions.

Signed-off-by: Xiaowei Yang <xiaowei.yang@intel.com>
19 years ago[VMX] Fix VMX ap boot.
kfraser@localhost.localdomain [Tue, 15 Aug 2006 09:43:43 +0000 (10:43 +0100)]
[VMX] Fix VMX ap boot.
Signed-off-by: Xin Li <xin.b.li@intel.com>
19 years ago[POWERPC] Update .hgignore
kfraser@localhost.localdomain [Tue, 15 Aug 2006 09:38:59 +0000 (10:38 +0100)]
[POWERPC] Update .hgignore
Signed-off-by: Hollis Blanchard <hollisb@us.ibm.com>
19 years agoMerge with xenppc-unstable-merge.hg
kfraser@localhost.localdomain [Tue, 15 Aug 2006 09:30:19 +0000 (10:30 +0100)]
Merge with xenppc-unstable-merge.hg

19 years agoAdd a series file to order the patches in patches/linux-<version>
Ian.Campbell@xensource.com [Tue, 15 Aug 2006 09:06:13 +0000 (10:06 +0100)]
Add a series file to order the patches in patches/linux-<version>

The initial ordering is the same as produced by the previous shell
globbing method.

The series file is a simple list of patch files relative to the
patches/linux-<version>/ directory. We do not support comments,
trailing directives (e.g. "-pN") or blank lines. Apart from these
restrictions the series file should be compatible with patch queue
management tools such as quilt.

Signed-off-by: Ian Campbell <ian.campbell@xensource.com>
19 years ago[POWERPC] Support console=comX option for Xen console.
Hollis Blanchard [Mon, 14 Aug 2006 20:22:22 +0000 (15:22 -0500)]
[POWERPC] Support console=comX option for Xen console.
Let Xen/PPC support the console=comX option that Xen/x86 uses, so one
need simply supply CMDLINE="console=com2" on the make invocation to get
proper use of the physical serial port on a JS21.

We do this by pruning all serial devices from the device tree, and then
initializing both com1 and com2 (like arch/x86/setup.c does).

Note that `setenv direct-serial? true; nvupdate' should in theory help
with this, but it does not work at present, and will only be supported
in some future version of SLOF.  Regardless, it is useful to support the
console=comX option, since there will always be environments where it is
difficult to get access to the firmware console.

Tested on JS20 and JS21 blades.

Signed-off-by: Amos Waterland <apw@us.ibm.com>
Signed-off-by: Hollis Blanchard <hollisb@us.ibm.com>
19 years ago[IA64] Fix IA64 build.
kfraser@localhost.localdomain [Mon, 14 Aug 2006 17:38:28 +0000 (18:38 +0100)]
[IA64] Fix IA64 build.

This patch gets ia64 building again on xen-unstable.hg due to the
addition of the uint64_t type in public/arch-ia64.h.  I also cleaned
up directly including this file versus public/xen.h as ia64 seems to be
the only arch using it directly.

Signed-off-by: Alex Williamson <alex.williamson@hp.com>
19 years ago[HVM] Fix roms.h to depend on acpi data.
kfraser@localhost.localdomain [Mon, 14 Aug 2006 16:57:37 +0000 (17:57 +0100)]
[HVM] Fix roms.h to depend on acpi data.
Signed-off-by: Xiaowei Yang <xiaowei.yang@intel.com>
19 years ago[VMXASSIST] Fix linker script to define start/end section labels
kfraser@localhost.localdomain [Mon, 14 Aug 2006 16:56:59 +0000 (17:56 +0100)]
[VMXASSIST] Fix linker script to define start/end section labels
in the usual idiomatic way. This avoids some BSS symbols being
defined oustide of range _sbss/_ebss.

Issue raised by Xiaowei Yang from Intel.

Signed-off-by: Keir Fraser <keir@xensource.com>
19 years agoPick up SMBIOS entry point in the ROMBIOS.
kfraser@localhost.localdomain [Mon, 14 Aug 2006 16:38:23 +0000 (17:38 +0100)]
Pick up SMBIOS entry point in the ROMBIOS.
Signed-off-by: Andrew D. Ball <aball@us.ibm.com>
19 years agoAdd code to generate SMBIOS tables to hvmloader.
kfraser@localhost.localdomain [Mon, 14 Aug 2006 16:37:07 +0000 (17:37 +0100)]
Add code to generate SMBIOS tables to hvmloader.
Signed-off-by: Andrew D. Ball <aball@us.ibm.com>
19 years agoMake room in e820map for SMBIOS tables.
kfraser@localhost.localdomain [Mon, 14 Aug 2006 16:29:11 +0000 (17:29 +0100)]
Make room in e820map for SMBIOS tables.
Signed-off-by: Andrew D. Ball <aball@us.ibm.com>
19 years agoAdd utilities needed for SMBIOS generation to hvmloader.
kfraser@localhost.localdomain [Mon, 14 Aug 2006 16:28:25 +0000 (17:28 +0100)]
Add utilities needed for SMBIOS generation to hvmloader.
Signed-off-by: Andrew D. Ball <aball@us.ibm.com>
19 years ago[HVMLOADER] Include xen.h for __HYPERVISOR* constants.
kfraser@localhost.localdomain [Mon, 14 Aug 2006 15:41:20 +0000 (16:41 +0100)]
[HVMLOADER] Include xen.h for __HYPERVISOR* constants.

Signed-off-by: Andrew D. BAll <aball@us.ibm.com>
19 years ago[LIBXC] Remove unused xc_evtchn_status().
kfraser@localhost.localdomain [Mon, 14 Aug 2006 15:39:35 +0000 (16:39 +0100)]
[LIBXC] Remove unused xc_evtchn_status().

Signed-off-by: John Levon <john.levon@sun.com>
19 years ago[LINUX] Mark /dev/mem regions as VM_PFNMAP.
kfraser@localhost.localdomain [Mon, 14 Aug 2006 15:35:49 +0000 (16:35 +0100)]
[LINUX] Mark /dev/mem regions as VM_PFNMAP.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[LINUX] Oney 'nosmp' and 'max_cpus=' command line options.
kfraser@localhost.localdomain [Mon, 14 Aug 2006 15:12:43 +0000 (16:12 +0100)]
[LINUX] Oney 'nosmp' and 'max_cpus=' command line options.
Based on a patch from Jan Beulich.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[NET] back: Fix an __init that should be an __exit in loopback driver.
kfraser@localhost.localdomain [Mon, 14 Aug 2006 14:26:37 +0000 (15:26 +0100)]
[NET] back: Fix an __init that should be an __exit in loopback driver.

This was pointed out by the newer build system used in 2.6.18-rc4.

Signed-off-by: Jan Beulich <jbeulich@novell.com>
19 years ago[LINUX][ Fix x86/64 entry.S to use paravirtualized sti/cli properly.
kfraser@localhost.localdomain [Mon, 14 Aug 2006 14:24:59 +0000 (15:24 +0100)]
[LINUX][ Fix x86/64 entry.S to use paravirtualized sti/cli properly.

This fixes
- one instance of using XEN_BLOCK_EVENTS where XEN_UNBLOCK_EVENTS was
  meant
- four instances of left-over cli-s that should have been replaced
  with XEN_BLOCK_EVENTS
- one instance of slightly wrong placement (without resulting in ill
  behavior, but needlessly making the diff to native more complex)

Signed-off-by: Jan Beulich <jbeulich@novell.com>
19 years ago[XEN] grant tables: update pair of entries using a union to ensure endian neutrality.
kfraser@localhost.localdomain [Mon, 14 Aug 2006 14:17:42 +0000 (15:17 +0100)]
[XEN] grant tables: update pair of entries using a union to ensure endian neutrality.

The first two members of a grant entry are updated as a combined pair.
The following patch uses a union so updates can happen in an endian neutral fashion.

Signed-off-by: Jimi Xenidis <jimix@watson.ibm.com>
19 years ago[POWERPC] memory cleanup (phase 2), destroy (et. al) now work
Jimi Xenidis [Mon, 14 Aug 2006 13:53:46 +0000 (09:53 -0400)]
[POWERPC] memory cleanup (phase 2), destroy (et. al) now work

The patch fixes the following problems:
  - Fix initializing the heaps so we can Link anywhere < RMA.  We not
    link xen at 4MiB, which is way saner.
  - We track and allocate pages (especially RMA) based on order.
  - free domain memory as a resource not as domain
  - clarify "order" arithmetic
  - update stale def of fix IS_XEN_HEAP_FRAME()

Signed-off-by: Jimi Xenidis <jimix@watson.ibm.com>
Signed-off-by: Hollis Blanchard <hollisb@us.ibm.com>
19 years ago[LINUX] Provide a handy macro for detrmining if we are domain0.
kfraser@localhost.localdomain [Mon, 14 Aug 2006 13:44:31 +0000 (14:44 +0100)]
[LINUX] Provide a handy macro for detrmining if we are domain0.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[XENBUS] Avoid direct use of xen_start_info. It's unevailable when
kfraser@localhost.localdomain [Mon, 14 Aug 2006 13:17:26 +0000 (14:17 +0100)]
[XENBUS] Avoid direct use of xen_start_info. It's unevailable when
building as a separate driver for an HVM guest.
Signed-off-by: Steven Smith <ssmith@xensource.com>
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[NET] linux: Kill the WARN_ON() calls for checksum fixups.
kfraser@localhost.localdomain [Mon, 14 Aug 2006 12:55:17 +0000 (13:55 +0100)]
[NET] linux: Kill the WARN_ON() calls for checksum fixups.

This changeset is from upstream Linux.  We need it so Xen users
using NAT are not unnecessarily alarmed by these WARN_ON errors
especially since they're not in a position to do anything about
it.  Original changelog:

        [NET]: Kill the WARN_ON() calls for checksum fixups.

        We have a more complete solution in the works, involving
        the seperation of CHECKSUM_HW on input vs. output, and
        having netfilter properly do incremental checksums.

        But that is a very involved patch and is thus 2.6.19
        material.

        What we have now is infinitely better than the past,
        wherein all TSO packets were dropped due to corrupt
        checksums as soon at the NAT module was loaded.  At
        least now, the checksums do get fixed up, it just
        isn't the cleanest nor most optimal solution.

Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
19 years ago[LINUX] Split useful bits of page.h into new header maddr.h.
kfraser@localhost.localdomain [Mon, 14 Aug 2006 12:54:10 +0000 (13:54 +0100)]
[LINUX] Split useful bits of page.h into new header maddr.h.

Split the bits of page.h which are needed by PV-on-HVM drivers out
into a separate header file, so that it's a bit easier to get the xenolinux
machine address functions when compiled against a native kernel.

Signed-off-by: Steven Smith <ssmith@xensource.com>
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[HVM][LINUX] Add a header file with wrappers for hvm_op hypercall.
kfraser@localhost.localdomain [Mon, 14 Aug 2006 10:35:57 +0000 (11:35 +0100)]
[HVM][LINUX] Add a header file with wrappers for hvm_op hypercall.
Signed-off-by: Steven Smith <ssmith@xensource.com>
19 years ago[HVM] Add stubs to Linux for the new hvm_op hypercall.
kfraser@localhost.localdomain [Mon, 14 Aug 2006 10:33:50 +0000 (11:33 +0100)]
[HVM] Add stubs to Linux for the new hvm_op hypercall.
Signed-off-by: Steven Smith <ssmith@xensource.com>
19 years ago[NET] back: Tidy up and remove communications via references to global structures.
kfraser@localhost.localdomain [Mon, 14 Aug 2006 10:32:24 +0000 (11:32 +0100)]
[NET] back: Tidy up and remove communications via references to global structures.
Signed-off-by: Steven Smith <ssmith@xensource.com>
19 years ago[HVM] More info when failing to emulate instruction group 1.
kfraser@localhost.localdomain [Mon, 14 Aug 2006 10:18:37 +0000 (11:18 +0100)]
[HVM] More info when failing to emulate instruction group 1.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[XEN] Remove sync_pagetable_state(). No longer needed.
kfraser@localhost.localdomain [Mon, 14 Aug 2006 09:58:02 +0000 (10:58 +0100)]
[XEN] Remove sync_pagetable_state(). No longer needed.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[XEN] Add some missing put_page()s to __gnttab_copy.
ssmith@weybridge.uk.xensource.com [Mon, 14 Aug 2006 09:47:59 +0000 (10:47 +0100)]
[XEN] Add some missing put_page()s to __gnttab_copy.

Signed-off-by: Steven Smith <sos22@cam.ac.uk>
19 years ago[XEN] Add a missing access check to __gnttab_copy.
ssmith@weybridge.uk.xensource.com [Mon, 14 Aug 2006 09:45:45 +0000 (10:45 +0100)]
[XEN] Add a missing access check to __gnttab_copy.

Signed-off-by: Steven Smith <sos22@cam.ac.uk>
19 years ago[POWERPC] memory cleanup
Jimi Xenidis [Sun, 13 Aug 2006 23:19:37 +0000 (19:19 -0400)]
[POWERPC] memory cleanup

The following changes have been made:
  - htab is allocated by "order" using RMA size, soon it will be
    assigned memory.
  - keep track of RMA in terms of its allocated page struct and order
  - free htab and RMA memory in arch_domain_destroy()
  - get rid of PTEG FULL message, everyone thinks its a bug

Signed-off-by: Jimi Xenidis <jimix@watson.ibm.com>
Signed-off-by: Hollis Blanchard <hollisb@us.ibm.com>
19 years ago[POWERPC] need a real definition for gnttab_log_dirty()
Jimi Xenidis [Sun, 13 Aug 2006 18:43:50 +0000 (14:43 -0400)]
[POWERPC] need a real definition for gnttab_log_dirty()

Signed-off-by: Jimi Xenidis <jimix@watson.ibm.com>
Signed-off-by: Hollis Blanchard <hollisb@us.ibm.com>
19 years ago[XEN] Explicitly include domain_page.h for map_domain_page() and fix a
kaf24@localhost.localdomain [Sun, 13 Aug 2006 18:10:57 +0000 (19:10 +0100)]
[XEN] Explicitly include domain_page.h for map_domain_page() and fix a
warning about arithmetic on 'void *' pointers.
Based on a patch from Jimi Xenidis <jimix@watson.ibm.com>
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[XEN] Use gnttab_clear_flag() in new grant-table copy operation.
kaf24@localhost.localdomain [Sun, 13 Aug 2006 18:00:19 +0000 (19:00 +0100)]
[XEN] Use gnttab_clear_flag() in new grant-table copy operation.
From: Jimi Xenidis <jimix@watson.ibm.com>
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[XEN] Clean up e820 definitions. Define shared set in hvm public dir.
kaf24@localhost.localdomain [Sun, 13 Aug 2006 16:47:59 +0000 (17:47 +0100)]
[XEN] Clean up e820 definitions. Define shared set in hvm public dir.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[HVM] HVMOP_get_param return parameter value within the provided parameter struct,
kaf24@firebug.cl.cam.ac.uk [Sun, 13 Aug 2006 08:44:07 +0000 (09:44 +0100)]
[HVM] HVMOP_get_param return parameter value within the provided parameter struct,
not as a direct return value.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[XEN] Add a warning about changeable size of shared_info structure.
kaf24@localhost.localdomain [Sat, 12 Aug 2006 15:18:08 +0000 (16:18 +0100)]
[XEN] Add a warning about changeable size of shared_info structure.

sizeof(shared_info_t) can change even in compatible hypervisor
versions.
Add a clear warning about this, and fix the few places relying on it.
In addition, add some padding to arch_shared_info, to allow future
additions to the structure without breaking known offsets of members
following it in the shared info.

Signed-off-by: John Levon <john.levon@sun.com>
19 years ago[XM-TEST] Fix network tests for -route and -nat
kaf24@localhost.localdomain [Sat, 12 Aug 2006 14:58:52 +0000 (15:58 +0100)]
[XM-TEST] Fix network tests for -route and -nat

vif-nat and vif-route need to know domU ethX IP addresses in order to set
up the appropriate routes.The new network test infrastructure is not
providing the IP address on the vif line in /tmp/xm-test.conf; adding
it turns out to be a one-liner.

Tested with xm-test running with -bridge, -route, and -nat network
scripts.

Signed-off-by: Jim Dykman <dykman@us.ibm.com>
19 years ago[NET] front: Restore 16 bytes of reservation
kaf24@localhost.localdomain [Sat, 12 Aug 2006 14:56:00 +0000 (15:56 +0100)]
[NET] front: Restore 16 bytes of reservation

By replacing dev_alloc_skb with alloc_skb we need to do the 16-byte
reservation ourselves.  Also, it's unnecessary to align the size
when calling alloc_skb since the alignment is an implementation
detail of alloc_skb and it already takes care of it for us.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
19 years ago[LINUX][XENBUS] Wait for devices to connect when frontend driver is loaded as a module.
kaf24@localhost.localdomain [Sat, 12 Aug 2006 14:50:25 +0000 (15:50 +0100)]
[LINUX][XENBUS] Wait for devices to connect when frontend driver is loaded as a module.

Thsi fixes problems where a driver is loaded from a script, and the
following parts of the script depend on devices having connected
by the time the modprobe/insmod returns.

Based on a tested patch from Rik van Riel at Red Hat.

Signed-off-by: Keir Fraser <keir@xensource.com>
19 years agoNew xm uptime command.
emellor@leeni.uk.xensource.com [Sat, 12 Aug 2006 11:25:42 +0000 (12:25 +0100)]
New xm uptime command.

From Alex Brett <alex.brett@xensource.com>

Signed-off-by: Ewan Mellor <ewan@xensource.com>
19 years ago[POWERPC] start-cpu has no return code
Jimi Xenidis [Fri, 11 Aug 2006 17:30:48 +0000 (13:30 -0400)]
[POWERPC] start-cpu has no return code

According to this document:

 PowerPC Processor binding to: IEEE 1275-1994
 Standard for Boot (Initialization, Configuration) Firmware
 Revision: 2.1 (Approved Version)
 Date: November 6, 1996

The start-cpu client interface has the following signature:

 start-cpu
  IN: nodeid, pc, arg
  OUT: none

This patch reflects that, and makes us stop seeing a bogus OF_FAILURE if
we do actually check the return code of of_start_cpu.  Tested on JS20
and JS21 blades.

Signed-off-by: Amos Waterland <apw@us.ibm.com>
Signed-off-by: Hollis Blanchard <hollisb@us.ibm.com>
19 years ago[IA64] Add two extra empty files and add a comment to every empty
kfraser@localhost.localdomain [Fri, 11 Aug 2006 17:09:13 +0000 (18:09 +0100)]
[IA64] Add two extra empty files and add a comment to every empty
source file so that they can be handled by patch/diff.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[XEN] Remove batched writable pagetable logic.
kfraser@localhost.localdomain [Fri, 11 Aug 2006 15:07:22 +0000 (16:07 +0100)]
[XEN] Remove batched writable pagetable logic.

Benchmarks show it provides little or no benefit (except
on synthetic benchmarks). Also it is complicated and
likely to hinder efforts to reduce lockign granularity.

Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[NET] front: Do not allocate unnecessary page-sized main data area
kfraser@localhost.localdomain [Fri, 11 Aug 2006 13:22:54 +0000 (14:22 +0100)]
[NET] front: Do not allocate unnecessary page-sized main data area
for receive skbuffs. The bulk data is stored in fragments.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[XEN] Fix bug in spurious pagefault detection which could
kfraser@localhost.localdomain [Thu, 10 Aug 2006 14:51:38 +0000 (15:51 +0100)]
[XEN] Fix bug in spurious pagefault detection which could
be exploited by unprivileged guests. Thanks to Matt Yourst
for finding this and providing the patch.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[HVM] Remove unused apic_enabled field from hvm_info_table.
kfraser@localhost.localdomain [Thu, 10 Aug 2006 14:45:47 +0000 (15:45 +0100)]
[HVM] Remove unused apic_enabled field from hvm_info_table.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years agoRevert inadvertent code change from "replace tabs with spaces" changeset.
emellor@leeni.uk.xensource.com [Thu, 10 Aug 2006 13:29:04 +0000 (14:29 +0100)]
Revert inadvertent code change from "replace tabs with spaces" changeset.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
19 years ago[XEN] Add a new type of grant copy operation.
kfraser@localhost.localdomain [Thu, 10 Aug 2006 10:54:15 +0000 (11:54 +0100)]
[XEN] Add a new type of grant copy operation.
Based on an original patch from Rolf Neugebauer at Intel.
Signed-off-by: Steven Smith <ssmith@xensource.com>
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years agoAdd support for sending event channel interrupts to HVM guests.
kfraser@localhost.localdomain [Thu, 10 Aug 2006 10:36:27 +0000 (11:36 +0100)]
Add support for sending event channel interrupts to HVM guests.
Signed-off-by: Steven Smith <ssmith@xensource.com>
19 years ago[LIBXC] clean specified files instead of wildcarding
kfraser@localhost.localdomain [Thu, 10 Aug 2006 09:55:54 +0000 (10:55 +0100)]
[LIBXC] clean specified files instead of wildcarding
Signed-off-by: Hollis Blanchard <hollisb@us.ibm.com>
19 years ago[LIBXC] replace tabs with spaces
kfraser@localhost.localdomain [Thu, 10 Aug 2006 09:52:46 +0000 (10:52 +0100)]
[LIBXC] replace tabs with spaces
Signed-off-by: Hollis Blanchard <hollisb@us.ibm.com>
19 years ago[HVM] Fix some IOAPIC and LAPIC device model bugs.
kfraser@localhost.localdomain [Thu, 10 Aug 2006 09:47:37 +0000 (10:47 +0100)]
[HVM] Fix some IOAPIC and LAPIC device model bugs.

Fix some boundary checking errors.
Fix the confusion using variables 'level' and 'trig_mode'.
Also fix some other misc mistakes and do some cleanup.

Signed-off-by: Dexuan Cui <dexuan.cui@intel.com>
19 years ago[HVM] Rmove unused field pae_enabled from hvm_info_table.
kfraser@localhost.localdomain [Thu, 10 Aug 2006 09:45:10 +0000 (10:45 +0100)]
[HVM] Rmove unused field pae_enabled from hvm_info_table.
Signed-off-by: Xin Li <xin.b.li@intel.com>
19 years ago[MINIOS] A first step to re-architecture mini-os for a port to ia64.
kfraser@localhost.localdomain [Thu, 10 Aug 2006 09:43:20 +0000 (10:43 +0100)]
[MINIOS] A first step to re-architecture mini-os for a port to ia64.
Create architecture specific directories below mini-os for sources
and below mini-os/include for headers.

Signed-off-by: Dietmar Hahn <dietmar.hahn@fujitsu-siemens.com>
19 years ago[CONSOLE] Always specify CON_ENABLED in Linux ocnsole driver.
kfraser@localhost.localdomain [Thu, 10 Aug 2006 09:39:58 +0000 (10:39 +0100)]
[CONSOLE] Always specify CON_ENABLED in Linux ocnsole driver.

I cannot think of any circumstance that the Xen console device would
not have CON_ENABLED.  With out this, DomUs that panic() before the
init process has started (and maybe even after) do not get output hi
priority log_levels (like panic()) from LOG_BUF().

Signed-off-by: Jimi Xenidis <jimix@watson.ibm.com>