## direct-io.hg

Move the booting description out of the VM class description into the VM
Lifecycle section of the doc.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
author Ewan Mellor Fri Mar 09 01:06:11 2007 +0000 (2007-03-09) 038e9138acae be686775bdb0 docs/xen-api/vm-lifecycle.tex docs/xen-api/xenapi-datamodel.tex tools/libxen/include/xen_vm.h
line diff
     1.1 --- a/docs/xen-api/vm-lifecycle.tex	Fri Mar 09 00:46:50 2007 +0000
1.2 +++ b/docs/xen-api/vm-lifecycle.tex	Fri Mar 09 01:06:11 2007 +0000
1.3 @@ -22,3 +22,19 @@
1.4
1.5  Figure~\ref{fig-vm-lifecycle} shows the states that a VM can be in
1.6  and the API calls that can be used to move the VM between these states.
1.7 +
1.8 +\section{VM boot parameters}
1.9 +
1.10 +The VM class contains a number of fields that control the way in which the VM is booted.
1.11 +With reference to the fields defined in the VM class (see later in this document),
1.12 +this section outlines the boot options available and the mechanisms provided for controlling them.
1.13 +
1.14 +VM booting is controlled by setting one of the two mutually exclusive groups: PV'', and HVM''.  If HVM.boot\_policy is the empty string, then paravirtual domain building and booting will be used; otherwise the VM will be loaded as an HVM domain, and booted using an emulated BIOS.
1.15 +
1.16 +When paravirtual booting is in use, the PV/bootloader field indicates the bootloader to use.  It may be pygrub'', in which case the platform's default installation of pygrub will be used, or a full path within the control domain to some other bootloader.  The other fields, PV/kernel, PV/ramdisk, PV/args and PV/bootloader\_args will be passed to the bootloader unmodified, and interpretation of those fields is then specific to the bootloader itself, including the possibility that the bootloader will ignore some or all of those given values. Finally the paths of all bootable disks are added to the bootloader commandline (a disk is bootable if its VBD has the bootable flag set). There may be zero, one or many bootable disks; the bootloader decides which disk (if any) to boot from.
1.17 +
1.18 +If the bootloader is pygrub, then the menu.lst is parsed if present in the guest's filesystem, otherwise the specified kernel and ramdisk are used, or an autodetected kernel is used if nothing is specified and autodetection is possible.  PV/args is appended to the kernel command line, no matter which mechanism is used for finding the kernel.
1.19 +
1.20 +If PV/bootloader is empty but PV/kernel is specified, then the kernel and ramdisk values will be treated as paths within the control domain.  If both PV/bootloader and PV/kernel are empty, then the behaviour is as if PV/bootloader was specified as pygrub''.
1.21 +
1.22 +When using HVM booting, HVM/boot\_policy and HVM/boot\_params specify the boot handling.  Only one policy is currently defined: BIOS order''.  In this case, HVM/boot\_params should contain one key-value pair order'' = N'' where N is the string that will be passed to QEMU.
1.23 \ No newline at end of file

     2.1 --- a/docs/xen-api/xenapi-datamodel.tex	Fri Mar 09 00:46:50 2007 +0000
2.2 +++ b/docs/xen-api/xenapi-datamodel.tex	Fri Mar 09 01:06:11 2007 +0000
2.3 @@ -1022,40 +1022,7 @@ references to objects with match names
2.4  \hline
2.5  \multicolumn{1}{|l}{Name} & \multicolumn{3}{l|}{\bf VM} \\
2.6  \multicolumn{1}{|l}{Description} & \multicolumn{3}{l|}{\parbox{11cm}{\em A
2.7 -virtual machine (or 'guest').
2.8 -
2.9 -VM booting is controlled by setting one of the two mutually exclusive
2.10 -groups: "PV", and "HVM".  If HVM.boot\_policy is the empty string, then
2.11 -paravirtual domain building and booting will be used; otherwise the VM will
2.12 -be loaded as an HVM domain, and booted using an emulated BIOS.
2.13 -
2.14 -When paravirtual booting is in use, the PV/bootloader field indicates the
2.15 -bootloader to use.  It may be "pygrub", in which case the platform's
2.16 -default installation of pygrub will be used, or a full path within the
2.17 -control domain to some other bootloader.  The other fields, PV/kernel,
2.18 -PV/ramdisk, PV/args and PV/bootloader\_args will be passed to the
2.19 -bootloader unmodified, and interpretation of those fields is then specific
2.21 -will ignore some or all of those given values. Finally the paths of all
2.22 -bootable disks are added to the bootloader commandline (a disk is bootable
2.23 -if its VBD has the bootable flag set). There may be zero, one or many
2.24 -bootable disks; the bootloader decides which disk (if any) to boot from.
2.25 -
2.26 -If the bootloader is pygrub, then the menu.lst is parsed if present in the
2.27 -guest's filesystem, otherwise the specified kernel and ramdisk are used, or
2.28 -an autodetected kernel is used if nothing is specified and autodetection is
2.29 -possible.  PV/args is appended to the kernel command line, no matter which
2.30 -mechanism is used for finding the kernel.
2.31 -
2.32 -If PV/bootloader is empty but PV/kernel is specified, then the kernel and
2.33 -ramdisk values will be treated as paths within the control domain.  If both
2.34 -PV/bootloader and PV/kernel are empty, then the behaviour is as if
2.35 -PV/bootloader was specified as "pygrub".
2.36 -
2.37 -When using HVM booting, HVM/boot\_policy and HVM/boot\_params specify the
2.38 -boot handling.  Only one policy is currently defined: "BIOS order".  In
2.39 -this case, HVM/boot\_params should contain one key-value pair "order" = "N"
2.40 -where N is the string that will be passed to QEMU.}} \\
2.41 +virtual machine (or 'guest').}} \\
2.42  \hline
2.43  Quals & Field & Type & Description \\
2.44  \hline

     3.1 --- a/tools/libxen/include/xen_vm.h	Fri Mar 09 00:46:50 2007 +0000
3.2 +++ b/tools/libxen/include/xen_vm.h	Fri Mar 09 01:06:11 2007 +0000
3.3 @@ -40,39 +40,6 @@
3.4   * The VM class.
3.5   *
3.6   * A virtual machine (or 'guest').
3.7 - *
3.8 - * VM booting is controlled by setting one of the two mutually exclusive
3.9 - * groups: "PV", and "HVM".  If HVM.boot_policy is the empty string, then
3.10 - * paravirtual domain building and booting will be used; otherwise the VM will
3.11 - * be loaded as an HVM domain, and booted using an emulated BIOS.
3.12 - *
3.13 - * When paravirtual booting is in use, the PV/bootloader field indicates the
3.14 - * bootloader to use.  It may be "pygrub", in which case the platform's
3.15 - * default installation of pygrub will be used, or a full path within the
3.16 - * control domain to some other bootloader.  The other fields, PV/kernel,
3.17 - * PV/ramdisk, PV/args and PV/bootloader_args will be passed to the bootloader
3.18 - * unmodified, and interpretation of those fields is then specific to the
3.19 - * bootloader itself, including the possibility that the bootloader will
3.20 - * ignore some or all of those given values. Finally the paths of all bootable
3.21 - * disks are added to the bootloader commandline (a disk is bootable if its
3.22 - * VBD has the bootable flag set). There may be zero, one or many bootable
3.23 - * disks; the bootloader decides which disk (if any) to boot from.
3.24 - *
3.25 - * If the bootloader is pygrub, then the menu.lst is parsed if present in the
3.26 - * guest's filesystem, otherwise the specified kernel and ramdisk are used, or
3.27 - * an autodetected kernel is used if nothing is specified and autodetection is
3.28 - * possible.  PV/args is appended to the kernel command line, no matter which
3.29 - * mechanism is used for finding the kernel.
3.30 - *
3.31 - * If PV/bootloader is empty but PV/kernel is specified, then the kernel and
3.32 - * ramdisk values will be treated as paths within the control domain.  If both
3.33 - * PV/bootloader and PV/kernel are empty, then the behaviour is as if
3.34 - * PV/bootloader was specified as "pygrub".
3.35 - *
3.36 - * When using HVM booting, HVM/boot_policy and HVM/boot_params specify the
3.37 - * boot handling.  Only one policy is currently defined: "BIOS order".  In
3.38 - * this case, HVM/boot_params should contain one key-value pair "order" = "N"
3.39 - * where N is the string that will be passed to QEMU..
3.40   */
3.41
3.42