kvm-46 release

ANNOUNCE Applied Avi Kivity 9 participants 45 messages Oct 10, 2007 – Oct 18, 2007 38 messages not in the archive KVM

Share & tools

Avi Kivity

author

2007-10-10 08:28 UTC #0 lore raw

We've now switched to allocating guest memory in userspace rather than in the kernel. This is important if you have a mainframe, but also if you want to share memory between guests and implement nice features like swapping.

Changes since kvm-45:
- fix host oops on bad guest ioapic accesses
- handle NMIs before enabling host interrupts
- add general guest memory accessors (Izik Eidus)
- allow user control over shadow cache size (Izik Eidus)
- auto tune shadow cache size with guest memory (Izik Eidus)
- allocate guest memory in userspace instead of in kernel (Izik Eidus)
- code style fixes (Mike Day, Anthony Liguori)
- lapic cleanups (Rusty Russell)
- fix acpi interrupt reporting for FreeBSD
- fix color problems on certain displays
- Red Hat 7.1 support
- vmmouse support (Dan Kenigsberg)
- fix sdl window caption when keyboard is captured (Dan Kenigsberg)
- improve libkvm configure script (Jerone Young)
- improve bios compilation support on x86_64 hosts (Laurent Vivier)
- fix acpi processor reporting for Windows 2000 with the ACPI HAL (Sheng Yang)

Notes:
If you use the modules bundled with kvm-46, you can use any version
of Linux from 2.6.9 upwards.
If you use the modules bundled with Linux 2.6.20, you need to use
kvm-12.
If you use the modules bundled with Linux 2.6.21, you need to use
kvm-17.
Modules from Linux 2.6.22 and up will work with any kvm version from
kvm-22. Some features may only be available in newer releases.
For best performance, use Linux 2.6.23-rc2 or later as the host.

http://kvm.qumranet.com

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now >> http://get.splunk.com/

A message that is not in the archive (sent off-list or not yet received): 470C8D28.2060408-atKUWr5tajBWk0Htik3J/[email protected]

Jun Koi

2007-10-10 08:39 UTC #2 lore raw

Hi,

Quoted 4 lines · Avi Kivity

We've now switched to allocating guest memory in userspace rather than in the kernel. This is important if you have a mainframe, but also if you want to share memory between guests and implement nice features like swapping.

This is interesting! But how can we do that now? (share memory between guests)

Thanks,
Jun

A message that is not in the archive (sent off-list or not yet received): fdaac4d50710100139v204e0e1du21efc51df95f77e7-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org

Avi Kivity

author

2007-10-10 08:41 UTC #4 lore raw

Quoted 4 lines · Jun Koi

Hi,
On 10/10/07, Avi Kivity <avi-atKUWr5tajBWk0Htik3J/[email protected]> wrote:

Quoted 9 lines

Quoted 5 lines
We've now switched to allocating guest memory in userspace rather than in the kernel. This is important if you have a mainframe, but also if you want to share memory between guests and implement nice features like swapping.
This is interesting! But how can we do that now? (share memory between guests)

It's not exposed by qemu, but you can now mmap() some file (or use SysV shared memory) and use that as guest memory.

Signature · 8 lines

Do not meddle in the internals of kernels, for they are subtle and quick to panic.


-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/

A message that is not in the archive (sent off-list or not yet received): 470C9034.2000206-atKUWr5tajBWk0Htik3J/[email protected]

Jun Koi

2007-10-10 09:04 UTC #6 lore raw

Quoted 11 lines · Avi Kivity

Quoted 10 lines · Jun Koi
Hi,
Quoted 5 lines · Avi Kivity
We've now switched to allocating guest memory in userspace rather than in the kernel. This is important if you have a mainframe, but also if you want to share memory between guests and implement nice features like swapping.

Quoted 7 lines

Quoted 3 lines
This is interesting! But how can we do that now? (share memory between guests)
It's not exposed by qemu, but you can now mmap() some file (or use SysV shared memory) and use that as guest memory.

OK, lets say we have 2 guest VMs share a memory, like mmap() a tmpfs file (which is actually in memory). Now one writes to the memory (shared file). Can we guarantee that the memory immediately updated, so other will see the change immediately? Or the data might be cached for a while, befere being flushed to the shared memory?

Thanks,
Jun

A message that is not in the archive (sent off-list or not yet received): fdaac4d50710100204n23a3486bk3012fa995b14455d-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org

Avi Kivity

author

2007-10-10 09:07 UTC #8 lore raw

Quoted 24 lines · Jun Koi

Quoted 13 lines · Avi Kivity
Quoted 11 lines · Jun Koi
Hi,
Quoted 6 lines · Avi Kivity
We've now switched to allocating guest memory in userspace rather than in the kernel. This is important if you have a mainframe, but also if you want to share memory between guests and implement nice features like swapping.
Quoted 8 lines
Quoted 4 lines
This is interesting! But how can we do that now? (share memory between guests)
It's not exposed by qemu, but you can now mmap() some file (or use SysV shared memory) and use that as guest memory.

Quoted 6 lines

OK, lets say we have 2 guest VMs share a memory, like mmap() a tmpfs file (which is actually in memory). Now one writes to the memory (shared file). Can we guarantee that the memory immediately updated, so other will see the change immediately? Or the data might be cached for a while, befere being flushed to the shared memory?

Changes are visible immediately.

Signature · 8 lines

Do not meddle in the internals of kernels, for they are subtle and quick to panic.


-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/

A message that is not in the archive (sent off-list or not yet received): 470C9660.8090700-atKUWr5tajBWk0Htik3J/[email protected]

Jun Koi

2007-10-11 01:02 UTC #10 lore raw

Quoted 25 lines · Avi Kivity

Quoted 24 lines · Jun Koi
Quoted 21 lines · Avi Kivity
Quoted 5 lines · Jun Koi
Hi,
On 10/10/07, Avi Kivity <avi-atKUWr5tajBWk0Htik3J/[email protected]> wrote:
Quoted 10 lines
Quoted 6 lines
We've now switched to allocating guest memory in userspace rather than in the kernel. This is important if you have a mainframe, but also if you want to share memory between guests and implement nice features like swapping.
This is interesting! But how can we do that now? (share memory between guests)
It's not exposed by qemu, but you can now mmap() some file (or use SysV shared memory) and use that as guest memory.

Quoted 9 lines

Quoted 6 lines
OK, lets say we have 2 guest VMs share a memory, like mmap() a tmpfs file (which is actually in memory). Now one writes to the memory (shared file). Can we guarantee that the memory immediately updated, so other will see the change immediately? Or the data might be cached for a while, befere being flushed to the shared memory?
Changes are visible immediately.

OK, that is about sharing memory file. How about a shared file in which a file is a real physical file on the host?

As far as I remember, there is a caution about sharing a file using mmap() between 2 processes, that is: if the first process modifies the shared file, it might take a while for the changes to be flushed from cache to the file. That means the other process doesnt see the change immediately.

I guess this problem affects us in this case, when we use a real (physical) file to share between 2 VMs, correct?

Thanks,
Jun

A message that is not in the archive (sent off-list or not yet received): fdaac4d50710101802u27013e48l8c88641a904f6ef8-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org

Anthony Liguori

2007-10-11 01:13 UTC #12 lore raw

Quoted 44 lines · Jun Koi

Quoted 31 lines · Avi Kivity
Quoted 29 lines · Jun Koi
Quoted 26 lines · Avi Kivity
Quoted 6 lines · Jun Koi
Hi,
On 10/10/07, Avi Kivity <avi-atKUWr5tajBWk0Htik3J/[email protected]> wrote:
Quoted 12 lines
Quoted 7 lines
We've now switched to allocating guest memory in userspace rather than in the kernel. This is important if you have a mainframe, but also if you want to share memory between guests and implement nice features like swapping.
This is interesting! But how can we do that now? (share memory between guests)
It's not exposed by qemu, but you can now mmap() some file (or use SysV shared memory) and use that as guest memory.
Quoted 10 lines
Quoted 7 lines
OK, lets say we have 2 guest VMs share a memory, like mmap() a tmpfs file (which is actually in memory). Now one writes to the memory (shared file). Can we guarantee that the memory immediately updated, so other will see the change immediately? Or the data might be cached for a while, befere being flushed to the shared memory?
Changes are visible immediately.

Quoted 9 lines

OK, that is about sharing memory file. How about a shared file in which a file is a real physical file on the host?
As far as I remember, there is a caution about sharing a file using mmap() between 2 processes, that is: if the first process modifies the shared file, it might take a while for the changes to be flushed from cache to the file. That means the other process doesnt see the change immediately.

Changes to a mmap() file may not be flushed to the disk until an msync or munmap but if you've got it opened with MAP_SHARED, then everyone mapping will have access to the same physical memory IIUC. You still need to use barriers to take into account SMP. There may be some weird interactions if the area isn't mlocked but I'm not totally sure on that.

Regards,

Anthony Liguori

Quoted 6 lines

I guess this problem affects us in this case, when we use a real (physical) file to share between 2 VMs, correct?
Thanks,
Jun

Quoted 11 lines

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
kvm-devel mailing list [email protected]
https://lists.sourceforge.net/lists/listinfo/kvm-devel

Farkas Levente

2007-10-10 09:52 UTC #13 lore raw

In reply to a message not in the archive

Quoted 15 lines · Avi Kivity

We've now switched to allocating guest memory in userspace rather than in the kernel. This is important if you have a mainframe, but also if you want to share memory between guests and implement nice features like swapping.
Changes since kvm-45:
- fix host oops on bad guest ioapic accesses
- handle NMIs before enabling host interrupts
- add general guest memory accessors (Izik Eidus)
- allow user control over shadow cache size (Izik Eidus)
- auto tune shadow cache size with guest memory (Izik Eidus)
- allocate guest memory in userspace instead of in kernel (Izik Eidus)
- code style fixes (Mike Day, Anthony Liguori)
- lapic cleanups (Rusty Russell)
- fix acpi interrupt reporting for FreeBSD

Quoted 8 lines

- fix color problems on certain displays
- Red Hat 7.1 support
- vmmouse support (Dan Kenigsberg)
- fix sdl window caption when keyboard is captured (Dan Kenigsberg)
- improve libkvm configure script (Jerone Young)
- improve bios compilation support on x86_64 hosts (Laurent Vivier)
- fix acpi processor reporting for Windows 2000 with the ACPI HAL (Sheng Yang)

when i try to compile it on centos i've got this error (which is not happend in kvm-45) :
----------------------------------
iasl -tc -p acpi-dsdt.hex acpi-dsdt.dsl
make[1]: iasl: Command not found
----------------------------------
where this iasl comes from? and where can i found it? i can't found in any package even in external repos.

Signature · 7 lines

  Levente                               "Si vis pacem para bellum!"

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/

A message that is not in the archive (sent off-list or not yet received): [email protected]

Avi Kivity

author

2007-10-10 10:08 UTC #15 lore raw

Quoted 27 lines · Farkas Levente

Quoted 15 lines · Avi Kivity
We've now switched to allocating guest memory in userspace rather than in the kernel. This is important if you have a mainframe, but also if you want to share memory between guests and implement nice features like swapping.
Changes since kvm-45:
- fix host oops on bad guest ioapic accesses
- handle NMIs before enabling host interrupts
- add general guest memory accessors (Izik Eidus)
- allow user control over shadow cache size (Izik Eidus)
- auto tune shadow cache size with guest memory (Izik Eidus)
- allocate guest memory in userspace instead of in kernel (Izik Eidus)
- code style fixes (Mike Day, Anthony Liguori)
- lapic cleanups (Rusty Russell)
- fix acpi interrupt reporting for FreeBSD
Quoted 9 lines
- fix color problems on certain displays
- Red Hat 7.1 support
- vmmouse support (Dan Kenigsberg)
- fix sdl window caption when keyboard is captured (Dan Kenigsberg)
- improve libkvm configure script (Jerone Young)
- improve bios compilation support on x86_64 hosts (Laurent Vivier)
- fix acpi processor reporting for Windows 2000 with the ACPI HAL (Sheng Yang)

Quoted 10 lines

when i try to compile it on centos i've got this error (which is not happend in kvm-45) :
----------------------------------
iasl -tc -p acpi-dsdt.hex acpi-dsdt.dsl
make[1]: iasl: Command not found
----------------------------------
where this iasl comes from? and where can i found it? i can't found in any package even in external repos.

How did you get this? Normal 'make' doesn't try to compile the bios.

You can download the iasl compiler from
http://www.intel.com/technology/iapc/acpi/downloads.htm.

Signature · 8 lines

error compiling committee.c: too many arguments to function


-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/

A message that is not in the archive (sent off-list or not yet received): 470CA4AA.6020909-atKUWr5tajBWk0Htik3J/[email protected]

Farkas Levente

2007-10-10 10:18 UTC #17 lore raw

Quoted 39 lines · Avi Kivity

Quoted 27 lines · Farkas Levente
Quoted 15 lines · Avi Kivity
We've now switched to allocating guest memory in userspace rather than in the kernel. This is important if you have a mainframe, but also if you want to share memory between guests and implement nice features like swapping.
Changes since kvm-45:
- fix host oops on bad guest ioapic accesses
- handle NMIs before enabling host interrupts
- add general guest memory accessors (Izik Eidus)
- allow user control over shadow cache size (Izik Eidus)
- auto tune shadow cache size with guest memory (Izik Eidus)
- allocate guest memory in userspace instead of in kernel (Izik Eidus)
- code style fixes (Mike Day, Anthony Liguori)
- lapic cleanups (Rusty Russell)
- fix acpi interrupt reporting for FreeBSD
Quoted 9 lines
- fix color problems on certain displays
- Red Hat 7.1 support
- vmmouse support (Dan Kenigsberg)
- fix sdl window caption when keyboard is captured (Dan Kenigsberg)
- improve libkvm configure script (Jerone Young)
- improve bios compilation support on x86_64 hosts (Laurent Vivier)
- fix acpi processor reporting for Windows 2000 with the ACPI HAL (Sheng Yang)
Quoted 10 lines
when i try to compile it on centos i've got this error (which is not happend in kvm-45) :
----------------------------------
iasl -tc -p acpi-dsdt.hex acpi-dsdt.dsl
make[1]: iasl: Command not found
----------------------------------
where this iasl comes from? and where can i found it? i can't found in any package even in external repos.

Quoted 1 line

How did you get this? Normal 'make' doesn't try to compile the bios.

yes, but my spec based on fedora spec which use Jeremy Katz's patch to bios boot menu, that's way i need it.

Signature · 7 lines

  Levente                               "Si vis pacem para bellum!"

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/

Farkas Levente

2007-10-10 10:16 UTC #18 lore raw

In reply to a message not in the archive

Quoted 25 lines · Farkas Levente

Quoted 15 lines · Avi Kivity
We've now switched to allocating guest memory in userspace rather than in the kernel. This is important if you have a mainframe, but also if you want to share memory between guests and implement nice features like swapping.
Changes since kvm-45:
- fix host oops on bad guest ioapic accesses
- handle NMIs before enabling host interrupts
- add general guest memory accessors (Izik Eidus)
- allow user control over shadow cache size (Izik Eidus)
- auto tune shadow cache size with guest memory (Izik Eidus)
- allocate guest memory in userspace instead of in kernel (Izik Eidus)
- code style fixes (Mike Day, Anthony Liguori)
- lapic cleanups (Rusty Russell)
- fix acpi interrupt reporting for FreeBSD
Quoted 8 lines
- fix color problems on certain displays
- Red Hat 7.1 support
- vmmouse support (Dan Kenigsberg)
- fix sdl window caption when keyboard is captured (Dan Kenigsberg)
- improve libkvm configure script (Jerone Young)
- improve bios compilation support on x86_64 hosts (Laurent Vivier)
- fix acpi processor reporting for Windows 2000 with the ACPI HAL (Sheng Yang)

Quoted 9 lines

when i try to compile it on centos i've got this error (which is not happend in kvm-45) :
----------------------------------
iasl -tc -p acpi-dsdt.hex acpi-dsdt.dsl
make[1]: iasl: Command not found
----------------------------------
where this iasl comes from? and where can i found it? i can't found in any package even in external repos.

ok i find it for fedora and recompile it from there.

Signature · 7 lines

  Levente                               "Si vis pacem para bellum!"

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/

Farkas Levente

2007-10-10 11:24 UTC #19 lore raw

In reply to a message not in the archive

Quoted 4 lines · Avi Kivity

We've now switched to allocating guest memory in userspace rather than in the kernel. This is important if you have a mainframe, but also if you want to share memory between guests and implement nice features like swapping.

we did a quick test against this version and it's turn out the smp guests are still not working. even if we start only one linux guest with 4cpu on 4cpu host the guest kernel hangs at random stages of the kernel loading process (if we run more smp guests then they hang earlier, but different random stage), but it's never reach the end of the kernel so hang somewhere inside the kernel. and the cpu usage of these process goes up to until the qemu-kvm processes eat all cpus (ie. 100%). the strange thing is that the host is running no error on console just these
and similar are in dmesg:
-------------------------------
Ignoring de-assert INIT to vcpu 1
SIPI to vcpu 1 vector 0x03
SIPI to vcpu 1 vector 0x03
Ignoring de-assert INIT to vcpu 2
SIPI to vcpu 2 vector 0x03
-------------------------------

but the good news that if we run more single cpu guests (both linux and windows) than it seems to working (at least in the last half an hour:-). which means that kvm-46 is as usable as kvm-36 for us (which was not the case with all previous version), but of course this's a very subjective 'fact':-)

Signature · 7 lines

  Levente                               "Si vis pacem para bellum!"

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/

A message that is not in the archive (sent off-list or not yet received): [email protected]

Avi Kivity

author

2007-10-10 11:28 UTC #21 lore raw

Quoted 8 lines · Farkas Levente

Quoted 5 lines · Avi Kivity
We've now switched to allocating guest memory in userspace rather than in the kernel. This is important if you have a mainframe, but also if you want to share memory between guests and implement nice features like swapping.

Quoted 7 lines

we did a quick test against this version and it's turn out the smp guests are still not working. even if we start only one linux guest with 4cpu on 4cpu host the guest kernel hangs at random stages of the kernel loading process (if we run more smp guests then they hang earlier, but different random stage), but it's never reach the end of the kernel so hang somewhere inside the kernel. and the cpu usage of these process goes up to until the qemu-kvm processes eat all cpus (ie. 100%). t

What host cpu are you using? What guest kernel version? 32-bit or 64-bit?

I have no problems running a 4-way Linux guest here.

Signature · 8 lines

error compiling committee.c: too many arguments to function


-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/

A message that is not in the archive (sent off-list or not yet received): 470CB75E.4000404-atKUWr5tajBWk0Htik3J/[email protected]

Farkas Levente

2007-10-10 11:35 UTC #23 lore raw

Quoted 18 lines · Avi Kivity

Quoted 8 lines · Farkas Levente
Quoted 5 lines · Avi Kivity
We've now switched to allocating guest memory in userspace rather than in the kernel. This is important if you have a mainframe, but also if you want to share memory between guests and implement nice features like swapping.
Quoted 7 lines
we did a quick test against this version and it's turn out the smp guests are still not working. even if we start only one linux guest with 4cpu on 4cpu host the guest kernel hangs at random stages of the kernel loading process (if we run more smp guests then they hang earlier, but different random stage), but it's never reach the end of the kernel so hang somewhere inside the kernel. and the cpu usage of these process goes up to until the qemu-kvm processes eat all cpus (ie. 100%). t

Quoted 3 lines

What host cpu are you using? What guest kernel version? 32-bit or 64-bit?
I have no problems running a 4-way Linux guest here.

- host:
- Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
- Intel S3000AHV
- 8GB RAM
- CentOS-5
- kernel-2.6.18-8.1.14.el5 x86_64 64bit
- guest-1:
- CentOS-5
- kernel-2.6.18-8.1.14.el5 i386 32bit
- guest-2:
- CentOS-5
- kernel-2.6.18-8.1.14.el5 x86_64 64bit
- guest-3:
- Mandrake-9
- kernel-2.4.19.16mdk-1-1mdk 32bit
- guest-4:
- Windows XP Professional 32bit

start the first guest-1 with 4cpu produce the above hang. (and of course using kvm-46:-)

Signature · 7 lines

  Levente                               "Si vis pacem para bellum!"

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/

A message that is not in the archive (sent off-list or not yet received): [email protected]

Avi Kivity

author

2007-10-10 11:58 UTC #25 lore raw

Quoted 31 lines · Farkas Levente

Quoted 8 lines
Quoted 8 lines
we did a quick test against this version and it's turn out the smp guests are still not working. even if we start only one linux guest with 4cpu on 4cpu host the guest kernel hangs at random stages of the kernel loading process (if we run more smp guests then they hang earlier, but different random stage), but it's never reach the end of the kernel so hang somewhere inside the kernel. and the cpu usage of these process goes up to until the qemu-kvm processes eat all cpus (ie. 100%). t
Quoted 4 lines
What host cpu are you using? What guest kernel version? 32-bit or 64-bit?
I have no problems running a 4-way Linux guest here.
- host:
- Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
- Intel S3000AHV
- 8GB RAM
- CentOS-5
- kernel-2.6.18-8.1.14.el5 x86_64 64bit
- guest-1:
- CentOS-5
- kernel-2.6.18-8.1.14.el5 i386 32bit
- guest-2:
- CentOS-5
- kernel-2.6.18-8.1.14.el5 x86_64 64bit
- guest-3:
- Mandrake-9
- kernel-2.4.19.16mdk-1-1mdk 32bit
- guest-4:
- Windows XP Professional 32bit

Quoted 4 lines

start the first guest-1 with 4cpu produce the above hang. (and of course using kvm-46:-)

Is it the only guest that hangs with -smp 4? Or all of them?

Signature · 8 lines

error compiling committee.c: too many arguments to function


-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/

A message that is not in the archive (sent off-list or not yet received): 470CBE58.3040807-atKUWr5tajBWk0Htik3J/[email protected]

Farkas Levente

2007-10-10 12:49 UTC #27 lore raw

Quoted 34 lines · Avi Kivity

Quoted 33 lines · Farkas Levente
Quoted 9 lines
Quoted 9 lines
we did a quick test against this version and it's turn out the smp guests are still not working. even if we start only one linux guest with
4cpu on 4cpu host the guest kernel hangs at random stages of the kernel loading process (if we run more smp guests then they hang earlier, but different random stage), but it's never reach the end of the kernel so hang somewhere inside the kernel. and the cpu usage of these process goes up to until the qemu-kvm processes eat all cpus (ie. 100%). t
Quoted 5 lines
What host cpu are you using? What guest kernel version? 32-bit or 64-bit?
I have no problems running a 4-way Linux guest here.
- host:
- Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
- Intel S3000AHV
- 8GB RAM
- CentOS-5
- kernel-2.6.18-8.1.14.el5 x86_64 64bit
- guest-1:
- CentOS-5
- kernel-2.6.18-8.1.14.el5 i386 32bit
- guest-2:
- CentOS-5
- kernel-2.6.18-8.1.14.el5 x86_64 64bit
- guest-3:
- Mandrake-9
- kernel-2.4.19.16mdk-1-1mdk 32bit
- guest-4:
- Windows XP Professional 32bit

Quoted 6 lines

Quoted 4 lines
start the first guest-1 with 4cpu produce the above hang. (and of course using kvm-46:-)
Is it the only guest that hangs with -smp 4? Or all of them?

if i start this guest alone it's hang. if i start all guest-{1,2,3,4} than all linux guest hangs expect windows which runs but see only one cpu. ok now i try it for you:-) so:

- host + guest-1 -> hang

- host + guest-2 -> hang
in this case i've got a stack trace too on the host, and the host working 2 more minutes (!?) so i see the top part of the stack (shift-pageup) but when i try to take a picture the host crash and i can only see the end:-( but next time i bale to catch it:
-----------------------------------------------

BUG: soft lockup detected on CPU#0!

Call Trace:
 <IRQ>  [<ffffffff800b2c30>] softlockup_tick+0xdb/0xed
 [<ffffffff800933ec>] update_process_times+0x42/0x68
 [<ffffffff80073d61>] smp_local_timer_interrupt+0x23/0x47
 [<ffffffff80074423>] smp_apic_timer_interrupt+0x41/0x47
 [<ffffffff8005bcc2>] apic_timer_interrupt+0x66/0x6c
 <EOI>  [<ffffffff88208aa6>] :kvm:kvm_flush_remote_tlbs+0xfb/0x109
 [<ffffffff88208a95>] :kvm:kvm_flush_remote_tlbs+0xea/0x109
 [<ffffffff8820e0a7>] :kvm:kvm_mmu_pte_write+0x1fc/0x330
 [<ffffffff88207657>] :kvm:kvm_write_guest_page+0x98/0xa3
 [<ffffffff882093fd>] :kvm:emulator_write_emulated_onepage+0x6e/0xce
 [<ffffffff88211a91>] :kvm:x86_emulate_insn+0x2b99/0x3e58
 [<ffffffff8005fc15>] __memset+0x39/0xc0
 [<ffffffff88229404>] :kvm_intel:vmcs_readl+0x17/0x1c
 [<ffffffff88209251>] :kvm:emulate_instruction+0x152/0x290
 [<ffffffff8822bb85>] :kvm_intel:handle_exception+0x170/0x250
 [<ffffffff88209945>] :kvm:kvm_vcpu_ioctl+0x34b/0x10ba
 [<ffffffff88213082>] :kvm:kvm_vcpu_kick+0x34/0x8e
 [<ffffffff8820ae33>] :kvm:kvm_vm_ioctl+0x48a/0x6af
 [<ffffffff800bea1e>] zone_statistics+0x3e/0x6d
 [<ffffffff8000a778>] get_page_from_freelist+0x35d/0x3cf
 [<ffffffff800bea1e>] zone_statistics+0x3e/0x6d
 [<ffffffff8000a778>] get_page_from_freelist+0x35d/0x3cf
 [<ffffffff8000ee4e>] __alloc_pages+0x65/0x2b2
 [<ffffffff8013b1b2>] __next_cpu+0x19/0x28
 [<ffffffff800857cf>] find_busiest_group+0x20d/0x621
 [<ffffffff8013b1b2>] __next_cpu+0x19/0x28
 [<ffffffff80072a51>] smp_send_reschedule+0x4e/0x53
 [<ffffffff80086458>] enqueue_task+0x41/0x56
 [<ffffffff80086494>] __activate_task+0x27/0x39
 [<ffffffff80044d32>] try_to_wake_up+0x407/0x418
 [<ffffffff800850b5>] __wake_up_common+0x3e/0x68
 [<ffffffff8002dd9c>] __wake_up+0x38/0x4f
 [<ffffffff80021b8a>] __up_read+0x19/0x7f
 [<ffffffff8011735a>] avc_has_perm+0x43/0x55
 [<ffffffff80117e91>] inode_has_perm+0x56/0x63
 [<ffffffff80086c37>] default_wake_function+0x0/0xe
 [<ffffffff80117f32>] file_has_perm+0x94/0xa3
 [<ffffffff8003fc5b>] do_ioctl+0x21/0x6b
 [<ffffffff8002fa61>] vfs_ioctl+0x248/0x261
 [<ffffffff8004a267>] sys_ioctl+0x59/0x78
 [<ffffffff8005b2c1>] tracesys+0xd1/0xdc
-----------------------------------------------

i hope shouldn't have to test all other guests:-(
but if i can help you in something more specific let me know:-)

Signature · 7 lines

  Levente                               "Si vis pacem para bellum!"

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/

A message that is not in the archive (sent off-list or not yet received): [email protected]

Avi Kivity

author

2007-10-13 07:18 UTC #29 lore raw

Quoted 52 lines · Farkas Levente

Quoted 36 lines · Avi Kivity
Quoted 34 lines · Farkas Levente
Quoted 10 lines
Quoted 10 lines
we did a quick test against this version and it's turn out the smp guests are still not working. even if we start only one linux guest with
4cpu on 4cpu host the guest kernel hangs at random stages of the kernel loading process (if we run more smp guests then they hang earlier, but different random stage), but it's never reach the end of the kernel so hang somewhere inside the kernel. and the cpu usage of these process goes up to until the qemu-kvm processes eat all cpus (ie. 100%). t
Quoted 6 lines
What host cpu are you using? What guest kernel version? 32-bit or 64-bit?
I have no problems running a 4-way Linux guest here.
- host:
- Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
- Intel S3000AHV
- 8GB RAM
- CentOS-5
- kernel-2.6.18-8.1.14.el5 x86_64 64bit
- guest-1:
- CentOS-5
- kernel-2.6.18-8.1.14.el5 i386 32bit
- guest-2:
- CentOS-5
- kernel-2.6.18-8.1.14.el5 x86_64 64bit
- guest-3:
- Mandrake-9
- kernel-2.4.19.16mdk-1-1mdk 32bit
- guest-4:
- Windows XP Professional 32bit
Quoted 7 lines
Quoted 5 lines
start the first guest-1 with 4cpu produce the above hang. (and of course using kvm-46:-)
Is it the only guest that hangs with -smp 4? Or all of them?
if i start this guest alone it's hang. if i start all guest-{1,2,3,4} than all linux guest hangs expect windows which runs but see only one cpu. ok now i try it for you:-) so:
- host + guest-1 -> hang

Quoted 8 lines

- host + guest-2 -> hang
in this case i've got a stack trace too on the host, and the host working 2 more minutes (!?) so i see the top part of the stack (shift-pageup) but when i try to take a picture the host crash and i can only see the end:-( but next time i bale to catch it:
-----------------------------------------------
BUG: soft lockup detected on CPU#0!

Yunfeng reports that this has been fixed, can you try the latest snapshot from http://people.qumranet.com/avi/snapshots/?

Signature · 8 lines

Do not meddle in the internals of kernels, for they are subtle and quick to panic.


-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/

A message that is not in the archive (sent off-list or not yet received): 47107150.3000106-atKUWr5tajBWk0Htik3J/[email protected]

Farkas Levente

2007-10-15 14:22 UTC #31 lore raw

Quoted 61 lines · Avi Kivity

Quoted 51 lines · Farkas Levente
Quoted 36 lines · Avi Kivity
Quoted 34 lines · Farkas Levente
Quoted 10 lines
Quoted 10 lines
we did a quick test against this version and it's turn out the smp guests are still not working. even if we start only one linux guest with
4cpu on 4cpu host the guest kernel hangs at random stages of the kernel loading process (if we run more smp guests then they hang earlier, but different random stage), but it's never reach the end of the kernel so hang somewhere inside the kernel. and the cpu usage of these process goes up to until the qemu-kvm processes eat all cpus (ie. 100%). t
Quoted 6 lines
What host cpu are you using? What guest kernel version? 32-bit or 64-bit?
I have no problems running a 4-way Linux guest here.
- host:
- Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
- Intel S3000AHV
- 8GB RAM
- CentOS-5
- kernel-2.6.18-8.1.14.el5 x86_64 64bit
- guest-1:
- CentOS-5
- kernel-2.6.18-8.1.14.el5 i386 32bit
- guest-2:
- CentOS-5
- kernel-2.6.18-8.1.14.el5 x86_64 64bit
- guest-3:
- Mandrake-9
- kernel-2.4.19.16mdk-1-1mdk 32bit
- guest-4:
- Windows XP Professional 32bit
Quoted 7 lines
Quoted 5 lines
start the first guest-1 with 4cpu produce the above hang. (and of course using kvm-46:-)
Is it the only guest that hangs with -smp 4? Or all of them?
if i start this guest alone it's hang. if i start all guest-{1,2,3,4} than all linux guest hangs expect windows which runs but see only one cpu. ok now i try it for you:-) so:
- host + guest-1 -> hang
Quoted 8 lines
- host + guest-2 -> hang
in this case i've got a stack trace too on the host, and the host working 2 more minutes (!?) so i see the top part of the stack (shift-pageup) but when i try to take a picture the host crash and i can only see the end:-( but next time i bale to catch it:
-----------------------------------------------
BUG: soft lockup detected on CPU#0!

Quoted 2 lines

Yunfeng reports that this has been fixed, can you try the latest snapshot from http://people.qumranet.com/avi/snapshots/?

and the rpms in my repo, but i'd rather wait for kvm-47 since this's a production server and i'd like to play with it only once a week:-)

Signature · 7 lines

  Levente                               "Si vis pacem para bellum!"

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/

Daniel P. Berrange

2007-10-10 13:39 UTC #32 lore raw

In reply to a message not in the archive

Quoted 4 lines · Avi Kivity

We've now switched to allocating guest memory in userspace rather than in the kernel. This is important if you have a mainframe, but also if you want to share memory between guests and implement nice features like swapping.

Is the memory allocation swappable by default, or still pinned in RAM by default ?

Dan.

Signature · 10 lines

|=- Red Hat, Engineering, Emerging Technologies, Boston.  +1 978 392 2496 -=|
|=-           Perl modules: http://search.cpan.org/~danberr/              -=|
|=-               Projects: http://freshmeat.net/~danielpb/               -=|
|=-  GnuPG: 7D3B9505   F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505  -=| 

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/

A message that is not in the archive (sent off-list or not yet received): [email protected]

Avi Kivity

author

2007-10-10 13:39 UTC #34 lore raw

Quoted 12 lines · Daniel P. Berrange

Quoted 5 lines · Avi Kivity
We've now switched to allocating guest memory in userspace rather than in the kernel. This is important if you have a mainframe, but also if you want to share memory between guests and implement nice features like swapping.
Is the memory allocation swappable by default, or still pinned in RAM by default ?

It's still not swappable. That will come in a future release.

Signature · 8 lines

error compiling committee.c: too many arguments to function


-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/

A message that is not in the archive (sent off-list or not yet received): 470CD61A.4090103-atKUWr5tajBWk0Htik3J/[email protected]

Haydn Solomon

2007-10-10 13:49 UTC #36 lore raw

Should we be seeing a mmap file when running guests now or do we have to use a special flag when running kvm?

Quoted 23 lines · Avi Kivity

Quoted 5 lines · Daniel P. Berrange
Quoted 3 lines · Avi Kivity
We've now switched to allocating guest memory in userspace rather than in the kernel. This is important if you have a mainframe, but also if you want to share memory between guests and implement nice features
like
Quoted 7 lines
Quoted 2 lines
swapping.
Is the memory allocation swappable by default, or still pinned in RAM by default ?
It's still not swappable. That will come in a future release.
--
error compiling committee.c: too many arguments to function

Quoted 10 lines

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
kvm-devel mailing list [email protected]
https://lists.sourceforge.net/lists/listinfo/kvm-devel

_______________________________________________
kvm-devel mailing list [email protected]
https://lists.sourceforge.net/lists/listinfo/kvm-devel

A message that is not in the archive (sent off-list or not yet received): b75785ba0710100649y3903f0ccma8c687d033fffec7-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org

Dor Laor

2007-10-10 14:59 UTC #38 lore raw

Quoted 3 lines · Haydn Solomon

Should we be seeing a mmap file when running guests now or do we have to use a special flag when running kvm?

Nope, the default is user space allocation were a regular malloc is used for allocating the guest ram from userspace.
The mmap is for future shared memory implementations for parts of the guest memory space.

Quoted 51 lines

On 10/10/07, *Avi Kivity* < avi-atKUWr5tajBWk0Htik3J/[email protected] <mailto:avi-atKUWr5tajBWk0Htik3J/[email protected]>> wrote:
Daniel P. Berrange wrote:
> On Wed, Oct 10, 2007 at 10:28:24AM +0200, Avi Kivity wrote:
>
>> We've now switched to allocating guest memory in userspace rather than
>> in the kernel. This is important if you have a mainframe, but also if
>> you want to share memory between guests and implement nice features like
>> swapping.
>>
>
> Is the memory allocation swappable by default, or still pinned in RAM by
> default ?
>
>
It's still not swappable. That will come in a future release.
--
error compiling committee.c: too many arguments to function
-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
kvm-devel mailing list [email protected] <mailto:[email protected]>
https://lists.sourceforge.net/lists/listinfo/kvm-devel
<https://lists.sourceforge.net/lists/listinfo/kvm-devel>
------------------------------------------------------------------------
-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now >> http://get.splunk.com/
------------------------------------------------------------------------

Quoted 5 lines

_______________________________________________
kvm-devel mailing list [email protected]
https://lists.sourceforge.net/lists/listinfo/kvm-devel

A message that is not in the archive (sent off-list or not yet received): 470CE8E7.7010001-atKUWr5tajBWk0Htik3J/[email protected]

Haydn Solomon

2007-10-10 15:22 UTC #40 lore raw

Thanks for clearing that up.

Quoted 52 lines · Dor Laor

Quoted 3 lines · Haydn Solomon
Should we be seeing a mmap file when running guests now or do we have to use a special flag when running kvm?
Nope, the default is user space allocation were a regular malloc is used for allocating the guest ram from userspace.
The mmap is for future shared memory implementations for parts of the guest memory space.
Quoted 21 lines
On 10/10/07, *Avi Kivity* < avi-atKUWr5tajBWk0Htik3J/[email protected] <mailto:avi-atKUWr5tajBWk0Htik3J/[email protected]>> wrote:
Daniel P. Berrange wrote:
> On Wed, Oct 10, 2007 at 10:28:24AM +0200, Avi Kivity wrote:
>
>> We've now switched to allocating guest memory in userspace rather than
>> in the kernel. This is important if you have a mainframe, but also if
>> you want to share memory between guests and implement nice features like
>> swapping.
>>
>
> Is the memory allocation swappable by default, or still pinned in RAM by
> default ?
>
>
Quoted 8 lines
It's still not swappable. That will come in a future release.
--
error compiling committee.c: too many arguments to function
-------------------------------------------------------------------------
Quoted 13 lines
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
kvm-devel mailing list [email protected] <mailto:[email protected]>
https://lists.sourceforge.net/lists/listinfo/kvm-devel
<https://lists.sourceforge.net/lists/listinfo/kvm-devel>

Quoted 17 lines

Quoted 3 lines
------------------------------------------------------------------------
-------------------------------------------------------------------------
Quoted 11 lines
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now >> http://get.splunk.com/
------------------------------------------------------------------------
_______________________________________________
kvm-devel mailing list [email protected]
https://lists.sourceforge.net/lists/listinfo/kvm-devel

_______________________________________________
kvm-devel mailing list [email protected]
https://lists.sourceforge.net/lists/listinfo/kvm-devel

Gerd Hoffmann

2007-10-12 07:53 UTC #41 lore raw

In reply to a message not in the archive

Quoted 2 lines · Avi Kivity

We've now switched to allocating guest memory in userspace rather than in the kernel.

Hmm, a quick glimpse over kvmctl.h doesn't show an obvious way how to use that. If I want to back vm memory with a file mapping, how can I do that?

cheers,
Gerd

A message that is not in the archive (sent off-list or not yet received): [email protected]

Avi Kivity

author

2007-10-13 07:13 UTC #43 lore raw

Quoted 10 lines · Gerd Hoffmann

Quoted 3 lines · Avi Kivity
We've now switched to allocating guest memory in userspace rather than in the kernel.
Hmm, a quick glimpse over kvmctl.h doesn't show an obvious way how to use that. If I want to back vm memory with a file mapping, how can I do that?

kvmctl.h doesn't expose an API for that currently, though is should be fairly trivial to do so. The kernel API is straightforward: take this bunch of memory and present it to a guest at some address. The kernel doesn't care if the memory came from a file mmap() or some other source.

Signature · 8 lines

Do not meddle in the internals of kernels, for they are subtle and quick to panic.


-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/

A message that is not in the archive (sent off-list or not yet received): 4710701C.2040200-atKUWr5tajBWk0Htik3J/[email protected]

Gerd Hoffmann

2007-10-15 10:44 UTC #45 lore raw

Quoted 6 lines · Avi Kivity

Quoted 5 lines · Gerd Hoffmann
Quoted 3 lines · Avi Kivity
We've now switched to allocating guest memory in userspace rather than in the kernel.

Quoted 7 lines

Quoted 4 lines
Hmm, a quick glimpse over kvmctl.h doesn't show an obvious way how to use that. If I want to back vm memory with a file mapping, how can I do that?
kvmctl.h doesn't expose an API for that currently, though is should be fairly trivial to do so.

Something like this? (compiles, not tested yet).

cheers,
Gerd

* make kvm_create() skip all memory setup in case phys_mem_bytes == 0
* add kvm_register_userspace_phys_mem() to register any userspace memory as guest physical memory.

Signed-off-by Gerd Hoffmann <[email protected]>

---
 user/kvmctl.c |   53 +++++++++++++++++++++++++++++++++++++++++------------
 user/kvmctl.h |    3 +++
 2 files changed, 44 insertions(+), 12 deletions(-)

user/kvmctl.c +41 ~~-12~~

Index: kvm-46/user/kvmctl.c
===================================================================
@@ -443,20 +443,22 @@ int kvm_create(kvm_context_t kvm, unsign
 	}
 	kvm->vm_fd = fd;
 
-	r = ioctl(kvm->fd, KVM_CHECK_EXTENSION, KVM_CAP_USER_MEMORY);
-	if (r > 0)
-		r = kvm_alloc_userspace_memory(kvm, memory, vm_mem);
-	else
-		r = kvm_alloc_kernel_memory(kvm, memory, vm_mem);
-	if (r < 0)
-		return r;
+	if (phys_mem_bytes) {
+		r = ioctl(kvm->fd, KVM_CHECK_EXTENSION, KVM_CAP_USER_MEMORY);
+		if (r > 0)
+			r = kvm_alloc_userspace_memory(kvm, memory, vm_mem);
+		else
+			r = kvm_alloc_kernel_memory(kvm, memory, vm_mem);
+		if (r < 0)
+			return r;
 
-        zfd = open("/dev/zero", O_RDONLY);
-        mmap(*vm_mem + 0xa8000, 0x8000, PROT_READ|PROT_WRITE,
-             MAP_PRIVATE|MAP_FIXED, zfd, 0);
-        close(zfd);
+	        zfd = open("/dev/zero", O_RDONLY);
+	        mmap(*vm_mem + 0xa8000, 0x8000, PROT_READ|PROT_WRITE,
+	             MAP_PRIVATE|MAP_FIXED, zfd, 0);
+	        close(zfd);
 
-	kvm->physical_memory = *vm_mem;
+		kvm->physical_memory = *vm_mem;
+	}
 
 	kvm->irqchip_in_kernel = 0;
 	if (!kvm->no_irqchip_creation) {

@@ -558,6 +560,33 @@ void *kvm_create_phys_mem(kvm_context_t 
 								log, writable);
 }
 
+int kvm_register_userspace_phys_mem(kvm_context_t kvm,
+			unsigned long phys_start, void *userspace_addr,
+			unsigned long len, int slot, int log)
+{
+	int r;
+	struct kvm_userspace_memory_region memory = {
+		.slot = slot,
+		.memory_size = len,
+		.guest_phys_addr = phys_start,
+		.userspace_addr = (intptr_t)userspace_addr,
+		.flags = log ? KVM_MEM_LOG_DIRTY_PAGES : 0,
+	};
+
+	if (!kvm->physical_memory)
+		kvm->physical_memory = userspace_addr - phys_start;
+
+	r = ioctl(kvm->vm_fd, KVM_SET_USER_MEMORY_REGION, &memory);
+	if (r == -1) {
+		fprintf(stderr, "create_userspace_phys_mem: %s", strerror(errno));
+		return -1;
+	}
+
+	kvm_userspace_memory_region_save_params(kvm, &memory);
+
+	return 0;
+}
+
 /* destroy/free a whole slot.
  * phys_start, len and slot are the params passed to kvm_create_phys_mem()
  */

user/kvmctl.h +3 -0

Index: kvm-46/user/kvmctl.h
===================================================================
@@ -404,6 +404,9 @@ void *kvm_create_phys_mem(kvm_context_t,
 			  unsigned long len, int slot, int log, int writable);
 void kvm_destroy_phys_mem(kvm_context_t, unsigned long phys_start, 
 			  unsigned long len, int slot);
+int kvm_register_userspace_phys_mem(kvm_context_t kvm,
+			unsigned long phys_start, void *userspace_addr,
+			unsigned long len, int slot, int log);
 int kvm_get_dirty_pages(kvm_context_t, int slot, void *buf);

_______________________________________________
kvm-devel mailing list [email protected]
https://lists.sourceforge.net/lists/listinfo/kvm-devel

A message that is not in the archive (sent off-list or not yet received): [email protected]

Gerd Hoffmann

2007-10-15 11:00 UTC #47 lore raw

Quoted 1 line · Gerd Hoffmann

Something like this? (compiles, not tested yet).

Hmm, no-go, results in "kvm_create_vcpu: Cannot allocate memory".

cheers,
Gerd

A message that is not in the archive (sent off-list or not yet received): [email protected]

Izik Eidus

2007-10-15 11:06 UTC #49 lore raw

Quoted 5 lines · Gerd Hoffmann

Quoted 1 line · Gerd Hoffmann
Something like this? (compiles, not tested yet).
Hmm, no-go, results in "kvm_create_vcpu: Cannot allocate memory".

try use slot bigger bigger than 5 and smaller than 8

Quoted 12 lines

cheers,
Gerd
-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
kvm-devel mailing list [email protected]
https://lists.sourceforge.net/lists/listinfo/kvm-devel

A message that is not in the archive (sent off-list or not yet received): 1192446374.6578.0.camel-siXIhNkUrCXckEVJwWePHtCfPAL7FxvL@public.gmane.org

Gerd Hoffmann

2007-10-15 11:15 UTC #51 lore raw

Quoted 7 lines · Izik Eidus

Quoted 4 lines · Gerd Hoffmann
Quoted 1 line · Gerd Hoffmann
Something like this? (compiles, not tested yet).
Hmm, no-go, results in "kvm_create_vcpu: Cannot allocate memory".
try use slot bigger bigger than 5 and smaller than 8

It doesn't come that far, kvm_create(&foo, 0, NULL) fails, seems doing vcpu_create() before memory setup doesn't work.

cheers,
Gerd

A message that is not in the archive (sent off-list or not yet received): [email protected]

Izik Eidus

2007-10-15 11:24 UTC #53 lore raw

Quoted 7 lines · Gerd Hoffmann

Quoted 6 lines · Izik Eidus
Quoted 4 lines · Gerd Hoffmann
Quoted 1 line · Gerd Hoffmann
Something like this? (compiles, not tested yet).
Hmm, no-go, results in "kvm_create_vcpu: Cannot allocate memory".

Quoted 9 lines

Quoted 1 line
try use slot bigger bigger than 5 and smaller than 8
It doesn't come that far, kvm_create(&foo, 0, NULL) fails, seems doing vcpu_create() before memory setup doesn't work.
cheers,
Gerd

can you send me your kvmctl.c file so i will look at it? it is somewhat hard to me understand how you did the things that way

A message that is not in the archive (sent off-list or not yet received): 1192447442.6911.5.camel-siXIhNkUrCXckEVJwWePHtCfPAL7FxvL@public.gmane.org

Gerd Hoffmann

2007-10-15 11:44 UTC #55 lore raw

Quoted 2 lines · Izik Eidus

can you send me your kvmctl.c file so i will look at it? it is somewhat hard to me understand how you did the things that way

I'm trying to back a vm with memory coming from a file map, see patch attached against kvm-46 (was also attached earlier in this thread).

I've made kvm_create() optionally skip the memory setup, so I can create my own later on. That doesn't work though because creating the vcpu fails then.

cheers,
Gerd

* make kvm_create() skip all memory setup in case phys_mem_bytes == 0
* add kvm_register_userspace_phys_mem() to register any userspace memory as guest physical memory.

Signed-off-by Gerd Hoffmann <[email protected]>

---
 user/kvmctl.c |   53 +++++++++++++++++++++++++++++++++++++++++------------
 user/kvmctl.h |    3 +++
 2 files changed, 44 insertions(+), 12 deletions(-)

user/kvmctl.c +41 ~~-12~~

Index: kvm-46/user/kvmctl.c
===================================================================
@@ -443,20 +443,22 @@ int kvm_create(kvm_context_t kvm, unsign
 	}
 	kvm->vm_fd = fd;
 
-	r = ioctl(kvm->fd, KVM_CHECK_EXTENSION, KVM_CAP_USER_MEMORY);
-	if (r > 0)
-		r = kvm_alloc_userspace_memory(kvm, memory, vm_mem);
-	else
-		r = kvm_alloc_kernel_memory(kvm, memory, vm_mem);
-	if (r < 0)
-		return r;
+	if (phys_mem_bytes) {
+		r = ioctl(kvm->fd, KVM_CHECK_EXTENSION, KVM_CAP_USER_MEMORY);
+		if (r > 0)
+			r = kvm_alloc_userspace_memory(kvm, memory, vm_mem);
+		else
+			r = kvm_alloc_kernel_memory(kvm, memory, vm_mem);
+		if (r < 0)
+			return r;
 
-        zfd = open("/dev/zero", O_RDONLY);
-        mmap(*vm_mem + 0xa8000, 0x8000, PROT_READ|PROT_WRITE,
-             MAP_PRIVATE|MAP_FIXED, zfd, 0);
-        close(zfd);
+	        zfd = open("/dev/zero", O_RDONLY);
+	        mmap(*vm_mem + 0xa8000, 0x8000, PROT_READ|PROT_WRITE,
+	             MAP_PRIVATE|MAP_FIXED, zfd, 0);
+	        close(zfd);
 
-	kvm->physical_memory = *vm_mem;
+		kvm->physical_memory = *vm_mem;
+	}
 
 	kvm->irqchip_in_kernel = 0;
 	if (!kvm->no_irqchip_creation) {

@@ -558,6 +560,33 @@ void *kvm_create_phys_mem(kvm_context_t 
 								log, writable);
 }
 
+int kvm_register_userspace_phys_mem(kvm_context_t kvm,
+			unsigned long phys_start, void *userspace_addr,
+			unsigned long len, int slot, int log)
+{
+	int r;
+	struct kvm_userspace_memory_region memory = {
+		.slot = slot,
+		.memory_size = len,
+		.guest_phys_addr = phys_start,
+		.userspace_addr = (intptr_t)userspace_addr,
+		.flags = log ? KVM_MEM_LOG_DIRTY_PAGES : 0,
+	};
+
+	if (!kvm->physical_memory)
+		kvm->physical_memory = userspace_addr - phys_start;
+
+	r = ioctl(kvm->vm_fd, KVM_SET_USER_MEMORY_REGION, &memory);
+	if (r == -1) {
+		fprintf(stderr, "create_userspace_phys_mem: %s", strerror(errno));
+		return -1;
+	}
+
+	kvm_userspace_memory_region_save_params(kvm, &memory);
+
+	return 0;
+}
+
 /* destroy/free a whole slot.
  * phys_start, len and slot are the params passed to kvm_create_phys_mem()
  */

user/kvmctl.h +3 -0

Index: kvm-46/user/kvmctl.h
===================================================================
@@ -404,6 +404,9 @@ void *kvm_create_phys_mem(kvm_context_t,
 			  unsigned long len, int slot, int log, int writable);
 void kvm_destroy_phys_mem(kvm_context_t, unsigned long phys_start, 
 			  unsigned long len, int slot);
+int kvm_register_userspace_phys_mem(kvm_context_t kvm,
+			unsigned long phys_start, void *userspace_addr,
+			unsigned long len, int slot, int log);
 int kvm_get_dirty_pages(kvm_context_t, int slot, void *buf);

_______________________________________________
kvm-devel mailing list [email protected]
https://lists.sourceforge.net/lists/listinfo/kvm-devel

A message that is not in the archive (sent off-list or not yet received): [email protected]

Gerd Hoffmann

2007-10-17 10:07 UTC #57 lore raw

PATCH 1/1 kvmctl.c: allow custom memory setup.

Quoted 3 lines · Gerd Hoffmann

I've made kvm_create() optionally skip the memory setup, so I can create my own later on. That doesn't work though because creating the vcpu fails then.

Ugh, vmx grabs last 4 pages from slot 0 (looks like for real mode emulation). Thus memory must exist before creating the vcpu. Which makes fitting filemap-backed memory into the current libkvm init procedure a bit difficuilt.

I've decided to split the kvm_create() into a bunch of pieces which can be called as needed, so I can first create the vm, then do my custom memory setup, then create the vcpu without making vmx unhappy ...

cheers,
Gerd

* split kvm_create() into smaller pieces which can be individually called if needed.
* add kvm_register_userspace_phys_mem() function.

---
 user/kvmctl.c |   61 +++++++++++++++++++++++++++++++++++++++++++++++++++++-----
 user/kvmctl.h |    8 +++++++
 2 files changed, 64 insertions(+), 5 deletions(-)

user/kvmctl.c +56 -5

Index: kvm-46/user/kvmctl.c
===================================================================
@@ -427,12 +427,9 @@ int kvm_alloc_userspace_memory(kvm_conte
 	return 0;
 }
 
-int kvm_create(kvm_context_t kvm, unsigned long phys_mem_bytes, void **vm_mem)
+int kvm_create_vm(kvm_context_t kvm)
 {
-	unsigned long memory = (phys_mem_bytes + PAGE_SIZE - 1) & PAGE_MASK;
 	int fd = kvm->fd;
-	int zfd;
-	int r;
 
 	kvm->vcpu_fd[0] = -1;

@@ -442,6 +439,15 @@ int kvm_create(kvm_context_t kvm, unsign
 		return -1;
 	}
 	kvm->vm_fd = fd;
+	return 0;
+}
+
+int kvm_create_default_phys_mem(kvm_context_t kvm, unsigned long phys_mem_bytes,
+				void **vm_mem)
+{
+	unsigned long memory = (phys_mem_bytes + PAGE_SIZE - 1) & PAGE_MASK;
+	int zfd;
+	int r;
 
 	r = ioctl(kvm->fd, KVM_CHECK_EXTENSION, KVM_CAP_USER_MEMORY);
 	if (r > 0)

@@ -457,18 +463,37 @@ int kvm_create(kvm_context_t kvm, unsign
         close(zfd);
 
 	kvm->physical_memory = *vm_mem;
+	return 0;
+}
+
+void kvm_create_irqchip(kvm_context_t kvm)
+{
+	int r;
 
 	kvm->irqchip_in_kernel = 0;
 	if (!kvm->no_irqchip_creation) {
 		r = ioctl(kvm->fd, KVM_CHECK_EXTENSION, KVM_CAP_IRQCHIP);
 		if (r > 0) {	/* kernel irqchip supported */
-			r = ioctl(fd, KVM_CREATE_IRQCHIP);
+			r = ioctl(kvm->vm_fd, KVM_CREATE_IRQCHIP);
 			if (r >= 0)
 				kvm->irqchip_in_kernel = 1;
 			else
 				printf("Create kernel PIC irqchip failed\n");
 		}
 	}
+}
+
+int kvm_create(kvm_context_t kvm, unsigned long phys_mem_bytes, void **vm_mem)
+{
+	int r;
+
+	r = kvm_create_vm(kvm);
+	if (r < 0)
+		return r;
+	r = kvm_create_default_phys_mem(kvm, phys_mem_bytes, vm_mem);
+	if (r < 0)
+		return r;
+	kvm_create_irqchip(kvm);
 	r = kvm_create_vcpu(kvm, 0);
 	if (r < 0)
 		return r;

@@ -558,6 +583,32 @@ void *kvm_create_phys_mem(kvm_context_t 
 								log, writable);
 }
 
+int kvm_register_userspace_phys_mem(kvm_context_t kvm,
+			unsigned long phys_start, void *userspace_addr,
+			unsigned long len, int slot, int log)
+{
+	struct kvm_userspace_memory_region memory = {
+		.slot = slot,
+		.memory_size = len,
+		.guest_phys_addr = phys_start,
+		.userspace_addr = (intptr_t)userspace_addr,
+		.flags = log ? KVM_MEM_LOG_DIRTY_PAGES : 0,
+	};
+	int r;
+
+	if (!kvm->physical_memory)
+		kvm->physical_memory = userspace_addr - phys_start;
+
+	r = ioctl(kvm->vm_fd, KVM_SET_USER_MEMORY_REGION, &memory);
+	if (r == -1) {
+		fprintf(stderr, "create_userspace_phys_mem: %s", strerror(errno));
+		return -1;
+	}
+
+	kvm_userspace_memory_region_save_params(kvm, &memory);
+        return 0;
+}
+
 /* destroy/free a whole slot.
  * phys_start, len and slot are the params passed to kvm_create_phys_mem()
  */

user/kvmctl.h +8 -0

Index: kvm-46/user/kvmctl.h
===================================================================
@@ -137,6 +137,11 @@ int kvm_get_shadow_pages(kvm_context_t k
 int kvm_create(kvm_context_t kvm,
 	       unsigned long phys_mem_bytes,
 	       void **phys_mem);
+int kvm_create_vm(kvm_context_t kvm);
+int kvm_create_default_phys_mem(kvm_context_t kvm, unsigned long phys_mem_bytes,
+				void **vm_mem);
+void kvm_create_irqchip(kvm_context_t kvm);
+
 /*!
  * \brief Create a new virtual cpu
  *

@@ -404,6 +409,9 @@ void *kvm_create_phys_mem(kvm_context_t,
 			  unsigned long len, int slot, int log, int writable);
 void kvm_destroy_phys_mem(kvm_context_t, unsigned long phys_start, 
 			  unsigned long len, int slot);
+int kvm_register_userspace_phys_mem(kvm_context_t kvm,
+			unsigned long phys_start, void *userspace_addr,
+			unsigned long len, int slot, int log);
 int kvm_get_dirty_pages(kvm_context_t, int slot, void *buf);

_______________________________________________
kvm-devel mailing list [email protected]
https://lists.sourceforge.net/lists/listinfo/kvm-devel

A message that is not in the archive (sent off-list or not yet received): [email protected]

Izik Eidus

2007-10-17 22:46 UTC #59 lore raw

Re: [patch] kvmctl.c: allow custom memory setup.

Quoted 12 lines · Gerd Hoffmann

Quoted 4 lines · Gerd Hoffmann
I've made kvm_create() optionally skip the memory setup, so I can create my own later on. That doesn't work though because creating the vcpu fails then.
Ugh, vmx grabs last 4 pages from slot 0 (looks like for real mode emulation). Thus memory must exist before creating the vcpu. Which makes fitting filemap-backed memory into the current libkvm init procedure a bit difficuilt.

Quoted 8 lines

I've decided to split the kvm_create() into a bunch of pieces which can be called as needed, so I can first create the vm, then do my custom memory setup, then create the vcpu without making vmx unhappy ...
cheers,
Gerd

hi,
why not making kvm_create_userspace_memory() recive a pointer to a userspace allocated memory (that was allocated from file or from normal malloc)
and make all the changes before kvm_create_userspace_memory() get called? this way you just need to remove the memory allocation out of kvm_create_userspace_memory() and alloc as you want.

what you think about that?

A message that is not in the archive (sent off-list or not yet received): 471690D2.2060204-atKUWr5tajBWk0Htik3J/[email protected]

Izik Eidus

2007-10-17 22:56 UTC #61 lore raw

Re: [patch] kvmctl.c: allow custom memory setup.

Quoted 30 lines · Izik Eidus

Quoted 13 lines · Gerd Hoffmann
Quoted 5 lines · Gerd Hoffmann
I've made kvm_create() optionally skip the memory setup, so I can create
my own later on. That doesn't work though because creating the vcpu fails then.
Ugh, vmx grabs last 4 pages from slot 0 (looks like for real mode emulation). Thus memory must exist before creating the vcpu. Which makes fitting filemap-backed memory into the current libkvm init procedure a bit difficuilt.
Quoted 8 lines
I've decided to split the kvm_create() into a bunch of pieces which can be called as needed, so I can first create the vm, then do my custom memory setup, then create the vcpu without making vmx unhappy ...
cheers,
Gerd
hi,
why not making kvm_create_userspace_memory() recive a pointer to a userspace allocated memory (that was allocated from file or from normal malloc)
and make all the changes before kvm_create_userspace_memory() get called? this way you just need to remove the memory allocation out of kvm_create_userspace_memory() and alloc as you want.

Quoted 2 lines

what you think about that?

that would bring the ability to qemu allocate the memory and send us the pointer to the allocated memory.

Gerd Hoffmann

2007-10-18 07:39 UTC #62 lore raw

Re: [patch] kvmctl.c: allow custom memory setup.

In reply to a message not in the archive

Quoted 5 lines · Izik Eidus

hi,
why not making kvm_create_userspace_memory() recive a pointer to a userspace allocated memory (that was allocated from file or from normal malloc)
and make all the changes before kvm_create_userspace_memory() get called?

I don't see how I can pass a pointer to kvm_create_userspace_memory() via kvm_create() without breaking the libkvm interface. There is no flags field or similar which could be used to signal "vm_mem is a valid pointer, please use that instead of mmap()'ing anonymous memory".

Oh, and I also don't want special the treatment of the 0xa0000 -> 0xfffff memory window.

cheers,
Gerd

A message that is not in the archive (sent off-list or not yet received): [email protected]

Dor Laor

2007-10-18 08:10 UTC #64 lore raw

Re: [patch] kvmctl.c: allow custom memory setup.

Quoted 7 lines · Gerd Hoffmann

Quoted 5 lines · Izik Eidus
hi,
why not making kvm_create_userspace_memory() recive a pointer to a userspace allocated memory (that was allocated from file or from normal malloc)
and make all the changes before kvm_create_userspace_memory() get

Quoted 7 lines

called?
I don't see how I can pass a pointer to kvm_create_userspace_memory() via kvm_create() without breaking the libkvm interface. There is no flags field or similar which could be used to signal "vm_mem is a valid pointer, please use that instead of mmap()'ing anonymous memory".

There is no need to keep binary compat to libkvm as long as it is linked statically.
We only need to keep binary compat with the kernel.

Quoted 7 lines

Oh, and I also don't want special the treatment of the 0xa0000 -> 0xfffff memory window.
cheers,
Gerd

A message that is not in the archive (sent off-list or not yet received): 471714FC.6050703-atKUWr5tajBWk0Htik3J/[email protected]

Gerd Hoffmann

2007-10-18 10:18 UTC #66 lore raw

Re: [patch] kvmctl.c: allow custom memory setup.

Quoted 8 lines · Dor Laor

Quoted 4 lines · Gerd Hoffmann
I don't see how I can pass a pointer to kvm_create_userspace_memory() via kvm_create() without breaking the libkvm interface. There is no flags field or similar which could be used to signal "vm_mem is a valid pointer, please use that instead of mmap()'ing anonymous memory".
There is no need to keep binary compat to libkvm as long as it is linked statically.

What about source compatibility?

cheers,
Gerd

A message that is not in the archive (sent off-list or not yet received): [email protected]

Izik Eidus

2007-10-18 10:30 UTC #68 lore raw

Re: [patch] kvmctl.c: allow custom memory setup.

Quoted 7 lines · Gerd Hoffmann

Quoted 5 lines · Dor Laor
Quoted 4 lines · Gerd Hoffmann
I don't see how I can pass a pointer to kvm_create_userspace_memory() via kvm_create() without breaking the libkvm interface. There is no flags field or similar which could be used to signal "vm_mem is a valid pointer, please use that instead of mmap()'ing anonymous memory".

Quoted 9 lines

Quoted 2 lines
There is no need to keep binary compat to libkvm as long as it is linked statically.
What about source compatibility?
cheers,
Gerd

anyone who use kvmctl, should not call kvm_create_userspace_memory directly,

instead should call kvm_create()...

A message that is not in the archive (sent off-list or not yet received): 1192703432.3175.18.camel-siXIhNkUrCXckEVJwWePHtCfPAL7FxvL@public.gmane.org

Avi Kivity

author

2007-10-18 10:39 UTC #70 lore raw

Re: [patch] kvmctl.c: allow custom memory setup.

Quoted 21 lines · Izik Eidus

Quoted 9 lines · Gerd Hoffmann
Quoted 7 lines · Dor Laor
Quoted 5 lines · Gerd Hoffmann
I don't see how I can pass a pointer to kvm_create_userspace_memory() via kvm_create() without breaking the libkvm interface. There is no flags field or similar which could be used to signal "vm_mem is a valid pointer, please use that instead of mmap()'ing anonymous memory".
Quoted 10 lines
Quoted 3 lines
There is no need to keep binary compat to libkvm as long as it is linked statically.
What about source compatibility?
cheers,
Gerd

Quoted 6 lines

anyone who use kvmctl, should not call kvm_create_userspace_memory directly,
instead should call kvm_create()...

Why not have an API for creating memory slots? It allows much more flexibility and is more in line with what qemu wants.

Signature · 8 lines

error compiling committee.c: too many arguments to function


-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/

A message that is not in the archive (sent off-list or not yet received): 471737E8.6000708-atKUWr5tajBWk0Htik3J/[email protected]

Gerd Hoffmann

2007-10-18 11:06 UTC #72 lore raw

Re: [patch] kvmctl.c: allow custom memory setup.

Quoted 10 lines · Avi Kivity

Quoted 9 lines · Izik Eidus
Quoted 7 lines · Gerd Hoffmann
Quoted 7 lines
Quoted 7 lines
I don't see how I can pass a pointer to kvm_create_userspace_memory() via kvm_create() without breaking the libkvm interface. There is no flags field or similar which could be used to signal "vm_mem is a valid pointer, please use that instead of mmap()'ing anonymous memory".

Quoted 5 lines

Quoted 5 lines
Quoted 2 lines
Quoted 2 lines
There is no need to keep binary compat to libkvm as long as it is linked statically.
anyone who use kvmctl, should not call kvm_create_userspace_memory directly, instead should call kvm_create()...

I'm talking about the kvm_create() interface, dammit. Sure I can modify kvm_create_userspace_memory() without breaking anyone as this isn't part of the public (kvmctl.h) interface. That doesn't buy us much though. I'd have to pass the pointer to kvm_create() somehow so it can be passed down to kvm_create_userspace_memory(). I can't without breaking the public library interface though.

Quoted 2 lines

Why not have an API for creating memory slots? It allows much more flexibility and is more in line with what qemu wants.

This *is* what the patch does. It adds a function to add a userspace-memory-backed memory slot: kvm_register_userspace_phys_mem().

That alone doesn't cut it though as there are some more constrains:

* kvm_create() builds a standard memory layout, which I want be able to skip.
* kvm_create() fails to create a vcpu in case no memory is available, which makes simple approach to just ask kvm_create() for 0 bytes physical memory for the guest unusable.

Thats why I went the route to additionally split the job kvm_create() does into multiple, individually callable pieces. So I can first create the vm, then create my custom memory slots (instead of the standard setup built by kvm_create_userspace_memory()), then create the vcpu.

cheers,
Gerd

ps: patch attached again for reference.

* split kvm_create() into smaller pieces which can be individually called if needed.
* add kvm_register_userspace_phys_mem() function.

---
 user/kvmctl.c |   61 +++++++++++++++++++++++++++++++++++++++++++++++++++++-----
 user/kvmctl.h |    8 +++++++
 2 files changed, 64 insertions(+), 5 deletions(-)

user/kvmctl.c +56 -5

Index: kvm-46/user/kvmctl.c
===================================================================
@@ -427,12 +427,9 @@ int kvm_alloc_userspace_memory(kvm_conte
 	return 0;
 }
 
-int kvm_create(kvm_context_t kvm, unsigned long phys_mem_bytes, void **vm_mem)
+int kvm_create_vm(kvm_context_t kvm)
 {
-	unsigned long memory = (phys_mem_bytes + PAGE_SIZE - 1) & PAGE_MASK;
 	int fd = kvm->fd;
-	int zfd;
-	int r;
 
 	kvm->vcpu_fd[0] = -1;

@@ -442,6 +439,15 @@ int kvm_create(kvm_context_t kvm, unsign
 		return -1;
 	}
 	kvm->vm_fd = fd;
+	return 0;
+}
+
+int kvm_create_default_phys_mem(kvm_context_t kvm, unsigned long phys_mem_bytes,
+				void **vm_mem)
+{
+	unsigned long memory = (phys_mem_bytes + PAGE_SIZE - 1) & PAGE_MASK;
+	int zfd;
+	int r;
 
 	r = ioctl(kvm->fd, KVM_CHECK_EXTENSION, KVM_CAP_USER_MEMORY);
 	if (r > 0)

@@ -457,18 +463,37 @@ int kvm_create(kvm_context_t kvm, unsign
         close(zfd);
 
 	kvm->physical_memory = *vm_mem;
+	return 0;
+}
+
+void kvm_create_irqchip(kvm_context_t kvm)
+{
+	int r;
 
 	kvm->irqchip_in_kernel = 0;
 	if (!kvm->no_irqchip_creation) {
 		r = ioctl(kvm->fd, KVM_CHECK_EXTENSION, KVM_CAP_IRQCHIP);
 		if (r > 0) {	/* kernel irqchip supported */
-			r = ioctl(fd, KVM_CREATE_IRQCHIP);
+			r = ioctl(kvm->vm_fd, KVM_CREATE_IRQCHIP);
 			if (r >= 0)
 				kvm->irqchip_in_kernel = 1;
 			else
 				printf("Create kernel PIC irqchip failed\n");
 		}
 	}
+}
+
+int kvm_create(kvm_context_t kvm, unsigned long phys_mem_bytes, void **vm_mem)
+{
+	int r;
+
+	r = kvm_create_vm(kvm);
+	if (r < 0)
+		return r;
+	r = kvm_create_default_phys_mem(kvm, phys_mem_bytes, vm_mem);
+	if (r < 0)
+		return r;
+	kvm_create_irqchip(kvm);
 	r = kvm_create_vcpu(kvm, 0);
 	if (r < 0)
 		return r;

@@ -558,6 +583,32 @@ void *kvm_create_phys_mem(kvm_context_t 
 								log, writable);
 }
 
+int kvm_register_userspace_phys_mem(kvm_context_t kvm,
+			unsigned long phys_start, void *userspace_addr,
+			unsigned long len, int slot, int log)
+{
+	struct kvm_userspace_memory_region memory = {
+		.slot = slot,
+		.memory_size = len,
+		.guest_phys_addr = phys_start,
+		.userspace_addr = (intptr_t)userspace_addr,
+		.flags = log ? KVM_MEM_LOG_DIRTY_PAGES : 0,
+	};
+	int r;
+
+	if (!kvm->physical_memory)
+		kvm->physical_memory = userspace_addr - phys_start;
+
+	r = ioctl(kvm->vm_fd, KVM_SET_USER_MEMORY_REGION, &memory);
+	if (r == -1) {
+		fprintf(stderr, "create_userspace_phys_mem: %s", strerror(errno));
+		return -1;
+	}
+
+	kvm_userspace_memory_region_save_params(kvm, &memory);
+        return 0;
+}
+
 /* destroy/free a whole slot.
  * phys_start, len and slot are the params passed to kvm_create_phys_mem()
  */

user/kvmctl.h +8 -0

Index: kvm-46/user/kvmctl.h
===================================================================
@@ -137,6 +137,11 @@ int kvm_get_shadow_pages(kvm_context_t k
 int kvm_create(kvm_context_t kvm,
 	       unsigned long phys_mem_bytes,
 	       void **phys_mem);
+int kvm_create_vm(kvm_context_t kvm);
+int kvm_create_default_phys_mem(kvm_context_t kvm, unsigned long phys_mem_bytes,
+				void **vm_mem);
+void kvm_create_irqchip(kvm_context_t kvm);
+
 /*!
  * \brief Create a new virtual cpu
  *

@@ -404,6 +409,9 @@ void *kvm_create_phys_mem(kvm_context_t,
 			  unsigned long len, int slot, int log, int writable);
 void kvm_destroy_phys_mem(kvm_context_t, unsigned long phys_start, 
 			  unsigned long len, int slot);
+int kvm_register_userspace_phys_mem(kvm_context_t kvm,
+			unsigned long phys_start, void *userspace_addr,
+			unsigned long len, int slot, int log);
 int kvm_get_dirty_pages(kvm_context_t, int slot, void *buf);

_______________________________________________
kvm-devel mailing list [email protected]
https://lists.sourceforge.net/lists/listinfo/kvm-devel

A message that is not in the archive (sent off-list or not yet received): [email protected]

Avi Kivity

author

2007-10-18 11:21 UTC #74 lore raw

Re: [patch] kvmctl.c: allow custom memory setup.

Quoted 5 lines · Gerd Hoffmann

Quoted 4 lines
Quoted 4 lines
anyone who use kvmctl, should not call kvm_create_userspace_memory directly, instead should call kvm_create()...

Quoted 8 lines

I'm talking about the kvm_create() interface, dammit. Sure I can modify kvm_create_userspace_memory() without breaking anyone as this isn't part of the public (kvmctl.h) interface. That doesn't buy us much though. I'd have to pass the pointer to kvm_create() somehow so it can be passed down to kvm_create_userspace_memory(). I can't without breaking the public library interface though.

Break it. It has just one user, our qemu, which is included in the same package.

We will freeze the lib kvm ABI once we have merged our changes into upstream qemu. But that won't happen unless we get a sane interface for that library (the current interface is not acceptable).

Quoted 4 lines

Quoted 3 lines
Why not have an API for creating memory slots? It allows much more flexibility and is more in line with what qemu wants.

Quoted 8 lines

This *is* what the patch does. It adds a function to add a userspace-memory-backed memory slot: kvm_register_userspace_phys_mem().
That alone doesn't cut it though as there are some more constrains:
* kvm_create() builds a standard memory layout, which I want be able to skip.

kvm_create() should not be creating memory.

Quoted 4 lines

* kvm_create() fails to create a vcpu in case no memory is available, which makes simple approach to just ask kvm_create() for 0 bytes physical memory for the guest unusable.

kvm_create() should not be creating vcpus.

Quoted 5 lines

Thats why I went the route to additionally split the job kvm_create() does into multiple, individually callable pieces. So I can first create the vm, then create my custom memory slots (instead of the standard setup built by kvm_create_userspace_memory()), then create the vcpu.

That's exactly what's needed.

The patch looks good, except that I wouldn't export kvm_create_default_phys_mem(). Let qemu do the hard work instead. Once qemu switches to the new interface, we can remove kvm_create().

Signature · 8 lines

error compiling committee.c: too many arguments to function


-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/

A message that is not in the archive (sent off-list or not yet received): 471741D1.7000500-atKUWr5tajBWk0Htik3J/[email protected]

Gerd Hoffmann

2007-10-18 12:24 UTC #76 lore raw

Re: [patch] kvmctl.c: allow custom memory setup.

Hi,

Quoted 2 lines

Break it. It has just one user, our qemu, which is included in the same package.

No, I'm hacking up one more user ;)

But maybe I'm better off shipping a private copy of kvmctl.c as long as the library interface isn't finalized yet and subject to change.

Quoted 10 lines

Quoted 5 lines
Thats why I went the route to additionally split the job kvm_create() does into multiple, individually callable pieces. So I can first create the vm, then create my custom memory slots (instead of the standard setup built by kvm_create_userspace_memory()), then create the vcpu.
That's exactly what's needed.
The patch looks good, except that I wouldn't export kvm_create_default_phys_mem().

Fine with me. If you one uses the splitted versions, then for creating a non-default memory layout, so there is no point in exporting that one.

Should I send an updated patch or do you just drop these lines when merging?

cheers,
Gerd

A message that is not in the archive (sent off-list or not yet received): [email protected]

Avi Kivity

author

2007-10-18 12:34 UTC #78 lore raw

Re: [patch] kvmctl.c: allow custom memory setup.

Quoted 10 lines · Gerd Hoffmann

Hi,
Quoted 3 lines
Break it. It has just one user, our qemu, which is included in the same package.
No, I'm hacking up one more user ;)

Nice. What will it do?

Quoted 4 lines

But maybe I'm better off shipping a private copy of kvmctl.c as long as the library interface isn't finalized yet and subject to change.

Yes. Patches to improve libkvm's interface so we can decouple it from qemu will be most welcome.

Quoted 8 lines

Quoted 8 lines
Quoted 6 lines
Thats why I went the route to additionally split the job kvm_create() does into multiple, individually callable pieces. So I can first create the vm, then create my custom memory slots (instead of the standard setup built by kvm_create_userspace_memory()), then create the vcpu.
That's exactly what's needed.

Quoted 9 lines

Quoted 3 lines
The patch looks good, except that I wouldn't export kvm_create_default_phys_mem().
Fine with me. If you one uses the splitted versions, then for creating a non-default memory layout, so there is no point in exporting that one.
Should I send an updated patch or do you just drop these lines when merging?

Please send a rebased and retested patch.

Signature · 8 lines

error compiling committee.c: too many arguments to function


-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/

A message that is not in the archive (sent off-list or not yet received): 471752BC.90600-atKUWr5tajBWk0Htik3J/[email protected]

Gerd Hoffmann

2007-10-18 15:29 UTC #80 lore raw

Re: [patch] kvmctl.c: allow custom memory setup.

Quoted 4 lines · Avi Kivity

Quoted 1 line · Gerd Hoffmann
No, I'm hacking up one more user ;)
Nice. What will it do?

Run xenified guest kernels without Xen.

Quoted 4 lines

Quoted 2 lines
Should I send an updated patch or do you just drop these lines when merging?
Please send a rebased and retested patch.

Oh, -47 is there. Updated patch attached.

cheers,
Gerd

kvmctl.h renamed from kvmctl.h.upstream +6 -0

diff -up ./kvmctl.h.upstream ./kvmctl.h
@@ -146,6 +146,9 @@ int kvm_get_shadow_pages(kvm_context_t k
 int kvm_create(kvm_context_t kvm,
 	       unsigned long phys_mem_bytes,
 	       void **phys_mem);
+int kvm_create_vm(kvm_context_t kvm);
+void kvm_create_irqchip(kvm_context_t kvm);
+
 /*!
  * \brief Create a new virtual cpu
  *

@@ -413,6 +416,9 @@ void *kvm_create_phys_mem(kvm_context_t,
 			  unsigned long len, int slot, int log, int writable);
 void kvm_destroy_phys_mem(kvm_context_t, unsigned long phys_start, 
 			  unsigned long len, int slot);
+int kvm_register_userspace_phys_mem(kvm_context_t kvm,
+			unsigned long phys_start, void *userspace_addr,
+			unsigned long len, int slot, int log);
 int kvm_get_dirty_pages(kvm_context_t, int slot, void *buf);

kvmctl.c renamed from kvmctl.c.upstream +56 -5

99 lines not on this page · raw message

_______________________________________________
kvm-devel mailing list [email protected]
https://lists.sourceforge.net/lists/listinfo/kvm-devel

A message that is not in the archive (sent off-list or not yet received): [email protected]

Avi Kivity

author Applied

2007-10-18 16:43 UTC #82 lore raw

Re: [patch] kvmctl.c: allow custom memory setup.

Quoted 10 lines · Gerd Hoffmann

Quoted 6 lines
Quoted 4 lines
Should I send an updated patch or do you just drop these lines when merging?
Please send a rebased and retested patch.
Oh, -47 is there. Updated patch attached.

Applied. But please:

- always supply the changelog comment so I don't have to hunt for it
- base diffs so they can be applied from the base directory with patch -p1 (this one was -p0 against the user/ directory). in general 'git diff' (or 'git format-patch' produces the best results with a minimum of fuss.

Signature · 8 lines

error compiling committee.c: too many arguments to function


-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/