- the name of the config file was computed incorrectly because of the
linux version number 4th digit being present in both LINUX_VER and
EXTRAVERSION extracted from the linux Makefile.
- DESTDIR was not set for the prep and config targets.
Signed-off-by: Christian Limpach <Christian.Limpach@xensource.com>
The number "31" means the idle domain ID.
In detail:
- display the idle domain ID with "Idle" instead of "31"
- write to the file "log-idle.log" instead of "log-dom31.log".
Signed-off-by: KUWAMURA Shin'ya <kuwa@jp.fujitsu.com>
are using 2M PSE pages. Now we can do kernel build on:
1) RHEL4 IA32 SMP guest on PAE host.
2) RHEL4 IA32 SMP guest on IA32e host.
3) IA32e SMP guest on IA32e host.
Signed-off-by: Xiaohui Xin <xiaohui.xin@intel.com>
Signed-off-by: Xin Li <xin.b.li@intel.com>
config file, with this change, RHEL4 IA32 SMP guest can boot.
But because we don't support 4Mbytes PSE pages, so x86_32 HVM guest
can NOT use PSE feature. However user may have "pae=1" in config file,
which causes PSE feature seen by x86_32 HVM guest, and it will then use
4Mbytes PSE pages. In this case, we will have to crash x86_32 HVM.
So when running x86_32 HVM guest, user need pay attention not to have
"pae=1" in config file.
Signed-off-by: Xin Li <xin.b.li@intel.com>
Signed-off-by: Xiaohui Xin <xiaohui.xin@intel.com>
When we reconnect to the backend we need to first zero all negotiated
bits as the functions xen_set_sg and xen_set_tso do not (and are not
supposed to) zero bits when they fail to set them.
This patch also permanently enables the NETIF_F_GSO_ROBUST bit as we
never parse any GSO fields ourselves (even if we did the backend could
not trust us so it's wasted effort).
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
This patch adds the as-yet unused GSO features which will contain
protocol-independent bits such as the ECN marker.
It also makes the backend check gso_size to ensure that it is non-zero.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The generalised extra request info patch introduced a bug with the use
of netbk_tx_err since it advanced the req_cons pointer by one. This
patch fixes thing by delaying the increment in netbk_tx_err.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
* Don't delete non-existent %.tar.bz2 in buildconfigs/Rules.mk's %-mrproper
* Add a mrpropper target to buildconfigs/mk.linux-2.6-xen, and
have this delete the local tarball
- This should also delete the local patch, if there is one, but I am not
sure how to do this with the current incarntation of the code, and I
will post a subsequent patch which makes this easier and effects this
chang. As there currently is no local patch this is not a big deal at
this moment. (Horms)
* Stop kclean from erroring out if the kernel directory doesn't exist
* Remove patches/*/.makedep in top level Makefile rather than in
buildconfigs/Rules.mk, as calling rules in the latter causes
patches/*/.make to be created (Magnus)
Signed-Off-By: Magnus Damm <magnus@valinux.co.jp>
Signed-Off-By: Horms <horms@verge.net.au>
Attached patch makes it always create valid devices on tty1-tty63, but
all but the one specified by xencons are dummy devices.
From: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Keir Fraser <keir@xensource.com>
- use 64-bit integral types for addresses in struct domain_start_info
- use stroull() to parse 64-bit values
- remove redundant _p(a) definition and add a comment
- printf format changes for the new types
Signed-off-by: Jimi Xenidis <jimix@watson.ibm.com>
Signed-off-by: Hollis Blanchard <hollisb@us.ibm.com>
The new method for printing unconnected device status is arguably
cleaner and also does not screw up reference counts (previous code
called bus_find_device() with no put_device() on the result.
Signed-off-by: Keir Fraser <keir@xensource.com>
New changeset merged upstream:
[TCP]: Reset gso_segs if packet is dodgy
I wasn't paranoid enough in verifying GSO information. A bogus gso_segs
could upset drivers as much as a bogus header would. Let's reset it in
the per-protocol gso_segment functions.
I didn't verify gso_size because that can be verified by the source of
the dodgy packets.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
call smp_processor_id() after any softirq, as rescheduling
may cause us to move to another processor on ia64
(spotted by Isaku Yamahata). Also get rid of many direct
callers of do_softirq() by creating new function
process_pending_timers().
Signed-off-by: Keir Fraser <keir@xensource.com>
If there's any consistency problems in the VMCB, the VMRUN will return
an exitcode of -1 to indicate that the guest-state is "bad". Prior to
this change, it would cause the hypervisor to reboot. This is not at
all necessary - the VMCB may be bad, but the system is most likely
still intact. [In particular, a guest could cause a crash of the
system by setting up PAE with an invalid set of PDPE's, which is
definitely not a desired behaviour]
Now also print out the VMCB, as that MAY help identify the cause of
the exitcode of -1.
Signed-off-by: Mats Petersson <mats.petersson@amd.com>
However xchg of IA64 has acquire semantics so that event
channel notification is lost sometimes. This patch fixes it.
Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>
