On Tuesday, 5 June 2018 8:34:28 PM AEST George Dunlap wrote:
> On Mon, Jun 4, 2018 at 3:55 PM, Lars Kurth wrote:
> > 2.2.3 B. Git baseline of patches
> > This created quite a bit of discussion and we did learn a few things:
> > * From the thread, having to cherry pick a small (around 5-6) patches
flight 124751 freebsd-master real [real]
http://logs.test-lab.xenproject.org/osstest/logs/124751/
Perfect :-)
All tests in this flight passed as required
version targeted for testing:
freebsd 0246611c60edfa274ca0504bc3bbe6a884bfb25e
baseline version:
freebsd 635457b6a00
flight 124750 freebsd-master real [real]
http://logs.test-lab.xenproject.org/osstest/logs/124750/
Perfect :-)
All tests in this flight passed as required
version targeted for testing:
freebsd 635457b6a00a5f5ed87674e7362064f66bf58a9e
baseline version:
freebsd 120832ac318
One reoccuring code pattern is to read an MSR, modify one or more bits,
and write the result back. Introduce helpers for this purpose.
First, introduce rdmsr_split() and wrmsr_split() which are tiny static inline
wrappers which deal with the MSR value in two 32bit halves.
Next, construct msr_{se
On 26/06/18 14:18, Andrew Cooper wrote:
> xen/arch/x86/cpu/mwait-idle.c | 4 ++--
I forgot to say that this patch as shown may impact the ease of taking
new code from Linux.
While we don't want to proliferate the broken APIs of the current
rdmsr() infrastructure, one option we could do is
Am Thu, 7 Jun 2018 15:08:29 +0200
schrieb Olaf Hering :
> Add an option to control when vTSC emulation will be activated for a
> domU with tsc_mode=default. Without such option each TSC access from
> domU will be emulated, which causes a significant perfomance drop for
> workloads that make use o
On Tue, Jun 26, 2018 at 02:18:15PM +0100, Andrew Cooper wrote:
> These MSRs, while being Intel specific, are used to offer virtualised
> CPUID faulting support on AMD hardware, so remove the INTEL infix.
>
> The bit position constants are used by guest_rdmsr(), but the logic can
> be expressed usi
flight 124719 xen-unstable-smoke running [real]
http://logs.test-lab.xenproject.org/osstest/logs/124719/
Failures and problems with tests :-(
Tests which did not succeed and are blocking,
including tests which could not be run:
build-armhf broken
build-armhf
flight 124715 xen-unstable running [real]
http://logs.test-lab.xenproject.org/osstest/logs/124715/
Failures and problems with tests :-(
Tests which did not succeed and are blocking,
including tests which could not be run:
build-i386-xsm broken
build-amd64-prev
flight 124731 seabios running [real]
http://logs.test-lab.xenproject.org/osstest/logs/124731/
Failures and problems with tests :-(
Tests which did not succeed and are blocking,
including tests which could not be run:
build-amd64 broken
build-i386-pvops
flight 124697 linux-4.14 running [real]
http://logs.test-lab.xenproject.org/osstest/logs/124697/
Failures and problems with tests :-(
Tests which did not succeed and are blocking,
including tests which could not be run:
build-amd64-xsm broken
build-amd64-pvops
flight 124690 libvirt running [real]
http://logs.test-lab.xenproject.org/osstest/logs/124690/
Regressions :-(
Tests which did not succeed and are blocking,
including tests which could not be run:
build-amd64-pvopsbroken
build-amd64
flight 124717 ovmf running [real]
http://logs.test-lab.xenproject.org/osstest/logs/124717/
Failures and problems with tests :-(
Tests which did not succeed and are blocking,
including tests which could not be run:
build-amd64-xsm broken
build-amd64
flight 124723 linux-linus running [real]
http://logs.test-lab.xenproject.org/osstest/logs/124723/
Failures and problems with tests :-(
Tests which did not succeed and are blocking,
including tests which could not be run:
build-amd64 broken
build-i386-pvops
On Tue, Jun 26, 2018 at 02:18:14PM +0100, Andrew Cooper wrote:
> Begin the process of cleaning up msr-index.h. Order the MSRs at the
> head of the file by index, use spaces for indentation, _AC() for bit
> positions, and add a comment describing the expected style. Abbreviate
> the ARCH_CAPS_* co
flight 124691 qemu-mainline running [real]
http://logs.test-lab.xenproject.org/osstest/logs/124691/
Failures and problems with tests :-(
Tests which did not succeed and are blocking,
including tests which could not be run:
build-amd64-pvopsbroken
build-amd64
flight 124687 linux-next running [real]
http://logs.test-lab.xenproject.org/osstest/logs/124687/
Failures and problems with tests :-(
Tests which did not succeed and are blocking,
including tests which could not be run:
build-i386 broken
build-amd64
On Tue, Jun 26, 2018 at 02:18:13PM +0100, Andrew Cooper wrote:
> The bit position constants are only used by the trampoline asm, but the
> code is shorter and clearer when using the mask constants. This halves
> the number of constants used.
>
> Consistently use _AC() for the bit constants, and s
flight 124729 linux-3.18 running [real]
http://logs.test-lab.xenproject.org/osstest/logs/124729/
Failures and problems with tests :-(
Tests which did not succeed and are blocking,
including tests which could not be run:
build-amd64-pvopsbroken
build-i386-pvops
flight 124712 linux-4.9 running [real]
http://logs.test-lab.xenproject.org/osstest/logs/124712/
Failures and problems with tests :-(
Tests which did not succeed and are blocking,
including tests which could not be run:
build-i386 broken
build-amd64-xsm
flight 124732 freebsd-master running [real]
http://logs.test-lab.xenproject.org/osstest/logs/124732/
Failures and problems with tests :-(
Tests which did not succeed and are blocking,
including tests which could not be run:
build-amd64-freebsd-again 4 syslog-serverrunning
bu
>>> On 25.06.18 at 10:33, wrote:
> On Thu, Jun 14, 2018 at 08:33:17AM -0600, Jan Beulich wrote:
>> > So far I've not been able to make Xen panic with the new xen kernel.
>> > Attached is a log of the serial console, in case you notice something.
>>
>> None of the printk()s replacing ASSERT()s hav
When SECONDARY_EXEC_ENABLE_VIRT_EXCEPTIONS is set,
vmx_vcpu_update_eptp() __vmwrites() EPTP_INDEX in
altp2m_vcpu_destroy(). This means that when disabling altp2m on a
domain after xc_altp2m_set_vcpu_enable_notify() has been
successfully called, EPTP_INDEX ends up being stored as
INVALID_ALTP2M. Thi
On Mon, Jun 25, 2018 at 03:53:09PM +0100, Ian Jackson wrote:
> From: Ian Jackson
>
> Today's run through the branching checklist generated these
> doc improvement suggestions.
I don't think an ack is needed by in any case:
Acked-by: Wei Liu
___
Xen-
On Tue, Jun 26, 2018 at 03:11:39PM +0100, Wei Liu wrote:
> On Mon, Jun 25, 2018 at 05:50:37PM +0300, Alexandru Isaila wrote:
> > This is used to save data from a single instance.
> >
> > Signed-off-by: Alexandru Isaila
>
> Reviewed-by: Wei Liu
Oh, actually Paul asked you to define the return v
On Mon, Jun 25, 2018 at 05:50:37PM +0300, Alexandru Isaila wrote:
> This is used to save data from a single instance.
>
> Signed-off-by: Alexandru Isaila
Reviewed-by: Wei Liu
___
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xen
>>> On 25.06.18 at 16:50, wrote:
> --- a/xen/arch/x86/hvm/hvm.c
> +++ b/xen/arch/x86/hvm/hvm.c
> @@ -786,117 +786,127 @@ static int hvm_load_tsc_adjust(struct domain *d,
> hvm_domain_context_t *h)
> HVM_REGISTER_SAVE_RESTORE(TSC_ADJUST, hvm_save_tsc_adjust,
>hvm_load_
>>> On 26.06.18 at 14:49, wrote:
> So assuming that the HVMOPs are properly handled atomically with respect
> to their corresponding VCPU (which I now believe to be the case), the
> only possible issue that remains (that I can think of) is the case where
> EPTP_INDEX has been saved in a previous r
The existing bit names are excessively long (45 chars!), and can be trimmed
down substantially. Drop the IA32 prefix and abbreviate FEATURE_CONTROL to
FEAT_CTL. Furthermore, all of these are feature enablement bits, so drop
ENABLE/ON parts of the constants.
While altering all the users, take the
The name MSR_IA32_APICBASE_MSR doesn't logically relate to its purpose.
Rename it to MSR_X2APIC_FIRST and introduce a corresponding
MSR_X2APIC_LAST to avoid opencoding the length of the x2APIC MSR range.
For the specific registers, drop the IA32 infix, break the APIC part
away from the register na
These MSRs, while being Intel specific, are used to offer virtualised
CPUID faulting support on AMD hardware, so remove the INTEL infix.
The bit position constants are used by guest_rdmsr(), but the logic can
be expressed using MASK_INSR() which allows the removal of the bit
position constants.
S
We currently have MSR_IA32_APICBASE and MSR_IA32_APICBASE_MSR which are
synonymous from a naming point of view, but refer to very different
things.
Cleave out the handling of MSR_APIC_BASE (0x1b), and rename
MSR_IA32_APICBASE_BASE to APIC_BASE_ADDR_MASK to better describe its
purpose.
Signed-off-
The bit position constants are only used by the trampoline asm, but the
code is shorter and clearer when using the mask constants. This halves
the number of constants used.
Consistently use _AC() for the bit constants, and start to use spaces
for indentation. Furthermore, EFER contains the NX-En
This series is a combination of fragments already posted, and cleanup work I
did while travelling in/around XenSummit.
There is no real functional change, but there is a lot of code volume
reduction and consistency improvements. This is the start of a very large
amount of MSR improvement work.
A
Begin the process of cleaning up msr-index.h. Order the MSRs at the
head of the file by index, use spaces for indentation, _AC() for bit
positions, and add a comment describing the expected style. Abbreviate
the ARCH_CAPS_* constants to reduce code volume.
Leave a trailing comment to logically s
>>> On 26.06.18 at 14:42, wrote:
> On 26/06/18 13:04, Jan Beulich wrote:
> On 26.06.18 at 13:09, wrote:
>>> Future changes will introduce a new SILO mode, which is intended to be
>>> useful
>>> for cloud and enterprise setups where all domUs are unprivileged and have no
>>> buisness communic
On 06/26/2018 03:06 PM, Razvan Cojocaru wrote:
> On 06/26/2018 02:56 PM, Jan Beulich wrote:
> On 26.06.18 at 12:55, wrote:
>>> On 06/26/2018 01:26 PM, Jan Beulich wrote:
>>> On 25.06.18 at 16:10, wrote:
> --- a/xen/arch/x86/mm/altp2m.c
> +++ b/xen/arch/x86/mm/altp2m.c
> @@ -58
On 26/06/18 13:04, Jan Beulich wrote:
On 26.06.18 at 13:09, wrote:
>> Future changes will introduce a new SILO mode, which is intended to be useful
>> for cloud and enterprise setups where all domUs are unprivileged and have no
>> buisness communicating directly.
>>
>> This was discussed at X
On Ma, 2018-06-26 at 11:07 +, Paul Durrant wrote:
> >
> > -Original Message-
> > From: Alexandru Isaila [mailto:aisa...@bitdefender.com]
> > Sent: 25 June 2018 15:51
> > To: xen-de...@lists.xen.org
> > Cc: Ian Jackson ; Wei Liu > com>;
> > jbeul...@suse.com; Andrew Cooper ; Paul
> > D
On 06/26/2018 02:56 PM, Jan Beulich wrote:
On 26.06.18 at 12:55, wrote:
>> On 06/26/2018 01:26 PM, Jan Beulich wrote:
>> On 25.06.18 at 16:10, wrote:
--- a/xen/arch/x86/mm/altp2m.c
+++ b/xen/arch/x86/mm/altp2m.c
@@ -58,8 +58,8 @@ altp2m_vcpu_destroy(struct vcpu *v)
>>> On 26.06.18 at 13:09, wrote:
> Future changes will introduce a new SILO mode, which is intended to be useful
> for cloud and enterprise setups where all domUs are unprivileged and have no
> buisness communicating directly.
>
> This was discussed at XenSummit, but I'll leave further details to
>>> On 26.06.18 at 12:55, wrote:
> On 06/26/2018 01:26 PM, Jan Beulich wrote:
> On 25.06.18 at 16:10, wrote:
>>> --- a/xen/arch/x86/mm/altp2m.c
>>> +++ b/xen/arch/x86/mm/altp2m.c
>>> @@ -58,8 +58,8 @@ altp2m_vcpu_destroy(struct vcpu *v)
>>>
>>> altp2m_vcpu_reset(v);
>>>
>>> -altp
On 26/06/18 12:53, Andrew Cooper wrote:
> On 26/06/18 12:50, Jan Beulich wrote:
> On 26.06.18 at 12:52, wrote:
>>> On 26/06/18 10:52, Jan Beulich wrote:
>>> On 26.06.18 at 10:45, wrote:
> On 26/06/2018 08:32, Jan Beulich wrote:
>> Use EFLAGS.IF for all ordinary purposes; there's i
flight 124724 freebsd-master real [real]
http://logs.test-lab.xenproject.org/osstest/logs/124724/
Perfect :-)
All tests in this flight passed as required
version targeted for testing:
freebsd 120832ac318f335dedfba859043b74375bba5e51
baseline version:
freebsd a550f635ac0
On 26/06/18 12:50, Jan Beulich wrote:
On 26.06.18 at 12:52, wrote:
>> On 26/06/18 10:52, Jan Beulich wrote:
>> On 26.06.18 at 10:45, wrote:
On 26/06/2018 08:32, Jan Beulich wrote:
> Use EFLAGS.IF for all ordinary purposes; there's in particular no need
> to unduly defer NMI/
>>> On 26.06.18 at 12:52, wrote:
> On 26/06/18 10:52, Jan Beulich wrote:
> On 26.06.18 at 10:45, wrote:
>>> On 26/06/2018 08:32, Jan Beulich wrote:
Use EFLAGS.IF for all ordinary purposes; there's in particular no need
to unduly defer NMI/#MC. Clear/set GIF solely around VMRUN itsel
>>> On 26.06.18 at 13:14, wrote:
> On 26/06/18 08:22, Jan Beulich wrote:
>> This currently shows up for basically every domain, when originally it
>> was logged only when going beyond the default table size. Restore that
>> behavior.
>>
>> Signed-off-by: Jan Beulich
>>
>> --- a/xen/common/grant_t
This code is believed to be vestigial remnant of the PV Windows XP port. It
is not used by Linux, NetBSD, Solaris or MiniOS. Furthermore the
implementation is incomplete; it only functions for a present => not-present
transition, rather than a present => read/write transition.
The for_each_vcpu(
flight 124713 seabios real [real]
http://logs.test-lab.xenproject.org/osstest/logs/124713/
Failures and problems with tests :-(
Tests which did not succeed and are blocking,
including tests which could not be run:
build-amd64-pvopsbroken
build-i386-pvops
On Tue, Jun 26, 2018 at 02:11:32AM -0600, Jan Beulich wrote:
> >>> On 26.06.18 at 09:57, wrote:
> > On Tue, Jun 26, 2018 at 01:47:12AM -0600, Jan Beulich wrote:
> >> In this case, just like xcr0_accum, nonlazy_xstate_used should always be
> >> set to the intended new value, rather than possibly le
> -Original Message-
> From: Xen-devel [mailto:xen-devel-boun...@lists.xenproject.org] On Behalf
> Of Andrew Cooper
> Sent: 26 June 2018 12:15
> To: Jan Beulich ; xen-devel de...@lists.xenproject.org>
> Cc: Stefano Stabellini ; Wei Liu
> ; George Dunlap ; Tim
> (Xen.org) ; Ian Jackson ; Ju
On 25/06/18 14:17, Jan Beulich wrote:
> We must not leave a vCPU with CR0.TS clear when it is not in fully eager
> mode and has not touched non-lazy state. Instead of adding a 3rd
> invocation of stts() to vcpu_restore_fpu_eager(), consolidate all of
> them into a single one done at the end of the
On 26/06/18 08:22, Jan Beulich wrote:
> This currently shows up for basically every domain, when originally it
> was logged only when going beyond the default table size. Restore that
> behavior.
>
> Signed-off-by: Jan Beulich
>
> --- a/xen/common/grant_table.c
> +++ b/xen/common/grant_table.c
> @
Signed-off-by: Wei Liu
---
Cc: Ian Jackson
---
config/Tools.mk.in | 2 --
1 file changed, 2 deletions(-)
diff --git a/config/Tools.mk.in b/config/Tools.mk.in
index 2d6c440324..4cc9f29090 100644
--- a/config/Tools.mk.in
+++ b/config/Tools.mk.in
@@ -20,8 +20,6 @@ BCC := @BCC@
IAS
The paths shouldn't be set to "yes".
Reported-by: Anthony Perard
Signed-off-by: Wei Liu
---
v3: really check for absolute paths.
Cc: Ian Jackson
new in v2
---
tools/configure| 9 ++---
tools/configure.ac | 9 ++---
2 files changed, 12 insertions(+), 6 deletions(-)
diff --git a/t
Do not embed IPXE into Rombios anymore. Instead, it is loaded by the
toolstack from a file as a separate module.
Ability to let user specify an IPXE blob will come later.
No user visible change.
Signed-off-by: Wei Liu
Acked-by: Jan Beulich
---
v3: adjust libxl code a bit, addressed Jan's comme
And switch hvmloader/Makefile to use that binary. This will help later
when we change hvmloader to pick a user provided binary.
No functional change.
Signed-off-by: Wei Liu
Acked-by: Jan Beulich
---
v2: use intermediary file
Cc: Ian Jackson
---
tools/firmware/etherboot/Makefile | 7 ++-
Lift the loading code out of PVH specific branch. Take the chance to
make the debug message more useful. Now the code needs to take into
account virt_base.
IPXE will be loaded as a module of Rombios.
Signed-off-by: Wei Liu
Reviewed-by: Roger Pau Monné
---
Cc: Ian Jackson
---
tools/libxc/xc_d
Addressed Jan and Roger's comments.
Wei Liu (6):
Tools.mk.in: drop unused variables
ipxe: produce a single binary from its build
libxc: allow HVM guest to have modules
tools: load IPXE from standalone file
tools: provide --with-system-ipxe
tools: --with-system-{ovmf,seabios,ipxe} shoul
Signed-off-by: Wei Liu
---
v3: ipxe should require rombios
Cc: Ian Jackson
---
config/Tools.mk.in| 1 +
tools/config.h.in | 3 +++
tools/configure | 58 +++
tools/configure.ac| 23 +++
tools/libxl/li
The embedded policy is specific flask, so update the infrastructure to reflect
this.
Signed-off-by: Andrew Cooper
---
CC: Daniel De Graaf
CC: George Dunlap
CC: Jan Beulich
CC: Konrad Rzeszutek Wilk
CC: Stefano Stabellini
CC: Tim Deegan
CC: Wei Liu
CC: Sergey Dyasli
CC: Xin Li
CC: Ming Lu
Flask is one single XSM module, and another is about to be introduced.
Properly namespace the symbols for clarity.
No functional change.
Signed-off-by: Andrew Cooper
---
CC: Daniel De Graaf
CC: George Dunlap
CC: Jan Beulich
CC: Konrad Rzeszutek Wilk
CC: Stefano Stabellini
CC: Tim Deegan
CC
Future changes will introduce a new SILO mode, which is intended to be useful
for cloud and enterprise setups where all domUs are unprivileged and have no
buisness communicating directly.
This was discussed at XenSummit, but I'll leave further details to the series
which introduces it. However, t
> -Original Message-
> From: Alexandru Isaila [mailto:aisa...@bitdefender.com]
> Sent: 25 June 2018 15:51
> To: xen-de...@lists.xen.org
> Cc: Ian Jackson ; Wei Liu ;
> jbeul...@suse.com; Andrew Cooper ; Paul
> Durrant ; Alexandru Isaila
>
> Subject: [PATCH v8 07/11] x86/hvm: Introduce
> vi
> -Original Message-
> From: Alexandru Isaila [mailto:aisa...@bitdefender.com]
> Sent: 25 June 2018 15:51
> To: xen-de...@lists.xen.org
> Cc: Ian Jackson ; Wei Liu ;
> jbeul...@suse.com; Andrew Cooper ; Paul
> Durrant
> Subject: [Xen-devel] [PATCH v8 00/11] x86/domctl: Save info for one vc
> -Original Message-
> From: Alexandru Isaila [mailto:aisa...@bitdefender.com]
> Sent: 25 June 2018 15:51
> To: xen-de...@lists.xen.org
> Cc: Ian Jackson ; Wei Liu ;
> jbeul...@suse.com; Andrew Cooper ; Paul
> Durrant ; Alexandru Isaila
>
> Subject: [PATCH v8 03/11] x86/hvm: Introduce hvm_
On 06/26/2018 01:26 PM, Jan Beulich wrote:
On 25.06.18 at 16:10, wrote:
>> When SECONDARY_EXEC_ENABLE_VIRT_EXCEPTIONS is set,
>> vmx_vcpu_update_eptp() __vmwrites() EPTP_INDEX in
>> altp2m_vcpu_destroy(). This means that when disabling altp2m on a
>> domain after xc_altp2m_set_vcpu_enable_not
On 26/06/18 08:35, Jan Beulich wrote:
> Correct a log message and, to clarify code as well, rename the
> respective function parameter too.
>
> Signed-off-by: Jan Beulich
>
> --- a/xen/arch/x86/acpi/cpu_idle.c
> +++ b/xen/arch/x86/acpi/cpu_idle.c
> @@ -1172,7 +1172,7 @@ static void print_cx_pminfo
On 26/06/18 10:52, Jan Beulich wrote:
On 26.06.18 at 10:45, wrote:
>> On 26/06/2018 08:32, Jan Beulich wrote:
>>> Use EFLAGS.IF for all ordinary purposes; there's in particular no need
>>> to unduly defer NMI/#MC. Clear/set GIF solely around VMRUN itself. This
>>> has the additional advantage
flight 124678 linux-3.18 real [real]
http://logs.test-lab.xenproject.org/osstest/logs/124678/
Failures and problems with tests :-(
Tests which did not succeed and are blocking,
including tests which could not be run:
build-i386 broken
build-i386-xsm
flight 74910 distros-debian-snapshot real [real]
http://osstest.xs.citrite.net/~osstest/testlogs/logs/74910/
Failures and problems with tests :-(
Tests which did not succeed and are blocking,
including tests which could not be run:
build-armhf-pvopsbroken
build-i
On 26/06/18 07:38, Jan Beulich wrote:
> While we don't expect CR0 to change behind our backs, cope with this
> happening, but other than for CR4 also log a (debug) message.
>
> Signed-off-by: Jan Beulich
>
> --- a/xen/arch/x86/hvm/vmx/vmcs.c
> +++ b/xen/arch/x86/hvm/vmx/vmcs.c
> @@ -1676,7 +1676,7
>>> On 25.06.18 at 16:10, wrote:
> When SECONDARY_EXEC_ENABLE_VIRT_EXCEPTIONS is set,
> vmx_vcpu_update_eptp() __vmwrites() EPTP_INDEX in
> altp2m_vcpu_destroy(). This means that when disabling altp2m on a
> domain after xc_altp2m_set_vcpu_enable_notify() has been
> successfully called, EPTP_INDEX
On 26/06/18 10:57, Jan Beulich wrote:
>
>>> +_p(regs->rip), _p(regs->rip), _p(fixup));
>>> +/*
>>> + * We mustn't come here, but for release builds have the recovery
>> "We shouldn't be able to reach here..."
> Well, okay - I did consider this weaker wording, but had
>>> On 26.06.18 at 11:41, wrote:
> On 26/06/18 07:36, Jan Beulich wrote:
>> --- a/xen/arch/x86/traps.c
>> +++ b/xen/arch/x86/traps.c
>> @@ -1749,7 +1749,21 @@ void do_device_not_available(struct cpu_
>> {
>> struct vcpu *curr = current;
>>
>> -BUG_ON(!guest_mode(regs));
>> +if ( !g
>>> On 26.06.18 at 10:45, wrote:
> On 26/06/2018 08:32, Jan Beulich wrote:
>> Use EFLAGS.IF for all ordinary purposes; there's in particular no need
>> to unduly defer NMI/#MC. Clear/set GIF solely around VMRUN itself. This
>> has the additional advantage that svm_stgi_label now indeed marks the
>
On 26/06/18 07:36, Jan Beulich wrote:
> Just in case we still don't get CR0.TS handling right, prevent a host
> crash by honoring exception fixups in do_device_not_available(). This
> would in particular cover emulator stubs raising #NM.
>
> Signed-off-by: Jan Beulich
> ---
> v2: Add ASSERT_UNREAC
On 26/06/18 07:36, Jan Beulich wrote:
> The changes for XSA-267 did not touch management of CR0.TS for HVM
> guests. In fully eager mode this bit should never be set when
> respective vCPU-s are active, or else hvmemul_get_fpu() might leave it
> wrongly set, leading to #NM in hypervisor context.
>
>>> On 26.06.18 at 10:43, wrote:
> On 26/06/18 08:24, Jan Beulich wrote:
>> During early boot timestamps aren't very useful, as they're all zero
>> (in "boot" mode) or absent altogether (in "date" and "datems" modes).
>> Log "boot" format timestamps when the date formats aren't available yet,
>> a
>>> On 26.06.18 at 09:39, wrote:
> On 26/06/18 08:24, Jan Beulich wrote:
>> The first patch alone should be sufficient to address an issue with the
>> XSA-267 changes, causing a BUG_ON() to be triggered. The other two
>> patches would, afaict, each individually also have avoided the issue,
>> henc
On 26/06/2018 08:32, Jan Beulich wrote:
> Use EFLAGS.IF for all ordinary purposes; there's in particular no need
> to unduly defer NMI/#MC. Clear/set GIF solely around VMRUN itself. This
> has the additional advantage that svm_stgi_label now indeed marks the
> only place where GIF is being set.
>
>
Hi Jan,
On 26/06/18 08:24, Jan Beulich wrote:
During early boot timestamps aren't very useful, as they're all zero
(in "boot" mode) or absent altogether (in "date" and "datems" modes).
Log "boot" format timestamps when the date formats aren't available yet,
and log raw timestamps when boot ones
On Mon, Jun 25, 2018 at 06:09:53PM +0200, Roger Pau Monné wrote:
> On Mon, Jun 25, 2018 at 12:35:32PM +0100, Wei Liu wrote:
> > diff --git a/tools/configure.ac b/tools/configure.ac
> > index 0826af8cbc..8e4b173d6f 100644
> > --- a/tools/configure.ac
> > +++ b/tools/configure.ac
> > @@ -84,6 +84,7 @
On Mon, Jun 25, 2018 at 06:21:06PM +0200, Roger Pau Monné wrote:
> On Mon, Jun 25, 2018 at 12:35:33PM +0100, Wei Liu wrote:
> > The paths shouldn't be set to "yes".
>
> If you want to check if an absolute path is provided I would rather
> do:
>
> case $withval in
> no) ovmf_path= ;;
> /*)
On Mon, Jun 25, 2018 at 06:05:01PM +0200, Roger Pau Monné wrote:
> On Mon, Jun 25, 2018 at 12:35:31PM +0100, Wei Liu wrote:
> > diff --git a/tools/libxl/libxl_dom.c b/tools/libxl/libxl_dom.c
> > index f0fd5fd3a3..1b9e45c492 100644
> > --- a/tools/libxl/libxl_dom.c
> > +++ b/tools/libxl/libxl_dom.c
On Mon, Jun 25, 2018 at 05:58:52PM +0200, Roger Pau Monné wrote:
> On Mon, Jun 25, 2018 at 12:35:30PM +0100, Wei Liu wrote:
> > Lift the loading code out of PVH specific branch. Take the chance to
> > make the debug message more useful.
>
> Would be nice to mention that you also account for virt_b
>>> On 26.06.18 at 09:57, wrote:
> On Tue, Jun 26, 2018 at 01:47:12AM -0600, Jan Beulich wrote:
>> In this case, just like xcr0_accum, nonlazy_xstate_used should always be
>> set to the intended new value, rather than possibly leaving the flag set
>> from a prior state load.
>>
>> Signed-off-by:
On Tue, Jun 26, 2018 at 01:47:12AM -0600, Jan Beulich wrote:
> In this case, just like xcr0_accum, nonlazy_xstate_used should always be
> set to the intended new value, rather than possibly leaving the flag set
> from a prior state load.
>
> Signed-off-by: Jan Beulich
Why? Looking at the code it
On Tue, Jun 26, 2018 at 01:37:02AM -0600, Jan Beulich wrote:
> Including the header solely at the assembler level means the compiler
> can't record the header as a dependency in .*.o.d files. Arrange for the
> header to be include-able at the C level instead.
>
> Signed-off-by: Jan Beulich
Revie
First of all introduce a helper function instead of replicating almost
the same code for PV and HVM. The differences between the two pieces of
code actually points out an issue (which is also addressed here): In
the HVM case FCW would not have been set to FCW_RESET in certain cases
(note for exampl
> -Original Message-
> From: Roger Pau Monne
> Sent: 26 June 2018 07:57
> To: Andreas Kinzler
> Cc: xen-de...@lists.xen.org; Paul Durrant
> Subject: Re: [Xen-devel] PCI passthrough performance loss with Skylake-SP
>
> On Mon, Jun 25, 2018 at 05:32:58PM +0200, Andreas Kinzler wrote:
> > I
In this case, just like xcr0_accum, nonlazy_xstate_used should always be
set to the intended new value, rather than possibly leaving the flag set
from a prior state load.
Signed-off-by: Jan Beulich
--- a/xen/arch/x86/domctl.c
+++ b/xen/arch/x86/domctl.c
@@ -1187,8 +1187,7 @@ long arch_do_domctl(
On Tue, Jun 26, 2018 at 01:35:18AM -0600, Jan Beulich wrote:
> Correct a log message and, to clarify code as well, rename the
> respective function parameter too.
>
> Signed-off-by: Jan Beulich
Reviewed-by: Wei Liu
___
Xen-devel mailing list
Xen-deve
flight 124708 freebsd-master real [real]
http://logs.test-lab.xenproject.org/osstest/logs/124708/
Perfect :-)
All tests in this flight passed as required
version targeted for testing:
freebsd a550f635ac0ec8163d698717f0e31b9a7f998ef3
baseline version:
freebsd c8b8ae9609e
On 26/06/18 08:24, Jan Beulich wrote:
> The first patch alone should be sufficient to address an issue with the
> XSA-267 changes, causing a BUG_ON() to be triggered. The other two
> patches would, afaict, each individually also have avoided the issue,
> hence I'm also including them here. While th
Including the header solely at the assembler level means the compiler
can't record the header as a dependency in .*.o.d files. Arrange for the
header to be include-able at the C level instead.
Signed-off-by: Jan Beulich
--- a/xen/include/asm-x86/asm_defns.h
+++ b/xen/include/asm-x86/asm_defns.h
On Mon, Jun 25, 2018 at 09:34:16PM +, osstest service owner wrote:
> flight 124705 freebsd-master real [real]
> http://logs.test-lab.xenproject.org/osstest/logs/124705/
>
> Regressions :-(
>
> Tests which did not succeed and are blocking,
> including tests which could not be run:
> build-amd
Correct a log message and, to clarify code as well, rename the
respective function parameter too.
Signed-off-by: Jan Beulich
--- a/xen/arch/x86/acpi/cpu_idle.c
+++ b/xen/arch/x86/acpi/cpu_idle.c
@@ -1172,7 +1172,7 @@ static void print_cx_pminfo(uint32_t cpu
#define print_cx_pminfo(c, p)
#endif
Use EFLAGS.IF for all ordinary purposes; there's in particular no need
to unduly defer NMI/#MC. Clear/set GIF solely around VMRUN itself. This
has the additional advantage that svm_stgi_label now indeed marks the
only place where GIF is being set.
A note regarding the main STI placement: Orignally
On Tue, Jun 26, 2018 at 01:22:01AM -0600, Jan Beulich wrote:
> This currently shows up for basically every domain, when originally it
> was logged only when going beyond the default table size. Restore that
> behavior.
>
> Signed-off-by: Jan Beulich
Acked-by: Wei Liu
__
1 - 100 of 106 matches
Mail list logo
