similar to: [RFC, PATCH 14/24] i386 Vmi reboot fixes

Attached and below are a patch that unifies the shutdown code paths in Xen, including those from EARLY_FAIL. This will facilitate the use of Intel(R) TXT for measured launch by ensuring that all shutdowns will call the necessary hook to tear down the measured environment. It also centralizes any future shutdown-related changes. The patch also postpones clearing the online status APs in

Fairly straightforward code motion in system.h into the sub-arch layer. Affected functionality include control register accessors, which are virtualizable but with great overhead due to the #GP cost; wbinvd, and most importantly, halt and interrupt control, which is non-virtualizable. Since read_cr4_safe can never fault on a VMI kernel (P5+ processor is required for VMI), we can omit the fault

Fairly straightforward code motion in system.h into the sub-arch layer. Affected functionality include control register accessors, which are virtualizable but with great overhead due to the #GP cost; wbinvd, and most importantly, halt and interrupt control, which is non-virtualizable. Since read_cr4_safe can never fault on a VMI kernel (P5+ processor is required for VMI), we can omit the fault

--text follows this line-- This patch uses the standard generic header for the definition of machine_emergency_restart() and reverses the call path for XenoLinux. Signed-off-by: Jimi Xenidis <jimix@watson.ibm.com> --- diff -r 7680b29c03a6 drivers/xen/core/reboot.c --- a/drivers/xen/core/reboot.c Sat Aug 05 11:33:24 2006 -0400 +++ b/drivers/xen/core/reboot.c Mon Aug 07 09:50:33 2006 -0400

In a virtualized environment, virtual machines will time share the system with each other and with other processes running on the host system. Therefore, a VM's virtual CPUs (VCPUs) will be executing on the host's physical CPUs (pcpus) for only some portion of time. The VMI exposes a paravirtual view of time to the guest operating systems so that they may operate more effectively in a

In a virtualized environment, virtual machines will time share the system with each other and with other processes running on the host system. Therefore, a VM's virtual CPUs (VCPUs) will be executing on the host's physical CPUs (pcpus) for only some portion of time. The VMI exposes a paravirtual view of time to the guest operating systems so that they may operate more effectively in a

Hi Andrew, This series of patches lays the basic ground work for the paravirtualized kernel patches coming later on. I think this lot is ready for the rough-and-tumble world of the -mm tree. The main change from the last posting is that all the page-table related patches have been moved out, and will be posted separately. Also, the off-by-one in reserving the top of address space has been

Hi Andrew, This series of patches lays the basic ground work for the paravirtualized kernel patches coming later on. I think this lot is ready for the rough-and-tumble world of the -mm tree. The main change from the last posting is that all the page-table related patches have been moved out, and will be posted separately. Also, the off-by-one in reserving the top of address space has been

This patch uses the standard generic header for the definition of machine_emergency_restart() and reverses the call path for XenoLinux. Signed-off-by: Jimi Xenidis <jimix@watson.ibm.com> --- diff -r 7680b29c03a6 drivers/xen/core/reboot.c --- a/drivers/xen/core/reboot.c Sat Aug 05 11:33:24 2006 -0400 +++ b/drivers/xen/core/reboot.c Mon Aug 07 09:50:33 2006 -0400 @@ -9,6 +9,7 @@ #include

I'm using VMWare Server 2 RC1 to on top of CentOS 5.2 x86_64 running a CentOS 5.2 i386 guest. I have enabled VMI in VMware, so I guess it won't let me install if VMI wasn't available in the kernel? How do I know whether VMI is supported/enabled and what performance benefits can I expect from it? I'm still not getting full hard drive speeds (only getting about 1/3 when using

So there is a logic error that would prevent us from ever using NOP = relocations, thus stopping any unnecessary callouts into a VMI ROM. It = doesn't affect us so much, but could conceivably affect open source = implementations of a VMI ROM in the future. It is not absolutely = critical, but it would be extremely nice to get it fixed in 2.6.21 so = there are no broken older kernels

So there is a logic error that would prevent us from ever using NOP = relocations, thus stopping any unnecessary callouts into a VMI ROM. It = doesn't affect us so much, but could conceivably affect open source = implementations of a VMI ROM in the future. It is not absolutely = critical, but it would be extremely nice to get it fixed in 2.6.21 so = there are no broken older kernels

VMI is broken under COMPAT_VDSO, as Xen and other non hardware assisted hypervisors will be. I have been working on a fix for this which works for older glibcs that panic when the new relocatable VDSO is used. However, I believe at this time that the fix is going to be too radical to consider at this stage in the release of 2.6.21. We don't expect this config option to be turned on by

VMI is broken under COMPAT_VDSO, as Xen and other non hardware assisted hypervisors will be. I have been working on a fix for this which works for older glibcs that panic when the new relocatable VDSO is used. However, I believe at this time that the fix is going to be too radical to consider at this stage in the release of 2.6.21. We don't expect this config option to be turned on by

Linker changes to add the VMI MACH_TEXT area into the kernel text section. This area layout is heavily influenced by magic to overlap exactly with the 32-byte VMI ROM entry points, allowing the kernel to copy the VMI ROM over the native section. As use of magic is on the decline, this approach is being phased out in favor of the more modern approach of the inline implementation. The translation

Linker changes to add the VMI MACH_TEXT area into the kernel text section. This area layout is heavily influenced by magic to overlap exactly with the 32-byte VMI ROM entry points, allowing the kernel to copy the VMI ROM over the native section. As use of magic is on the decline, this approach is being phased out in favor of the more modern approach of the inline implementation. The translation

In paravirt builds with VMI compiled in, vmi_bringup is called unconditionally, not via a paravirt-ops table (as no other hypervisor uses the APIC startup technique). Make the calls to setup VMI state conditional on the presence of the VMI ROM. Signed-off-by: Zachary Amsden <zach@vmware.com> diff -r 1915e2685a3c arch/i386/kernel/vmi.c --- a/arch/i386/kernel/vmi.c Thu Jan 04 15:56:40 2007

In paravirt builds with VMI compiled in, vmi_bringup is called unconditionally, not via a paravirt-ops table (as no other hypervisor uses the APIC startup technique). Make the calls to setup VMI state conditional on the presence of the VMI ROM. Signed-off-by: Zachary Amsden <zach@vmware.com> diff -r 1915e2685a3c arch/i386/kernel/vmi.c --- a/arch/i386/kernel/vmi.c Thu Jan 04 15:56:40 2007

Add a sub-arch specifier to the module identification string to avoid cross loading modules from different subarch compiles. XXX FIXME. Module loading is broken in paravirtualized VMI kernels, since there is no annotation fixup applied to modules. Signed-off-by: Zachary Amsden <zach@vmware.com> Index: linux-2.6.16-rc3/include/asm-i386/module.h

Add a sub-arch specifier to the module identification string to avoid cross loading modules from different subarch compiles. XXX FIXME. Module loading is broken in paravirtualized VMI kernels, since there is no annotation fixup applied to modules. Signed-off-by: Zachary Amsden <zach@vmware.com> Index: linux-2.6.16-rc3/include/asm-i386/module.h