similar to: porting lguest to x86_64

plain text document attachment (lguest64.patch) This is the main core code for the lguest64. Have fun, and don't hurt the puppies! Signed-off-by: Steven Rostedt <srostedt@redhat.com> Signed-off-by: Glauber de Oliveira Costa <glommer@gmail.com> Cc: Chris Wright <chrisw@sous-sol.org> Index: work-pv/arch/x86_64/lguest/Makefile

plain text document attachment (lguest64.patch) This is the main core code for the lguest64. Have fun, and don't hurt the puppies! Signed-off-by: Steven Rostedt <srostedt@redhat.com> Signed-off-by: Glauber de Oliveira Costa <glommer@gmail.com> Cc: Chris Wright <chrisw@sous-sol.org> Index: work-pv/arch/x86_64/lguest/Makefile

Hi all! Lately, Glauber and I have been working on getting both paravirt_ops and lguest running on the x86_64. I already pushed the x86_64 patches and as promised, I'm pushing now the lguest64 patches. This patches are greatly influenced by Rusty Russell, and we tried to stay somewhat consistent with his work on the i386. But there are some major differences that we had to over come.

Hi all! Lately, Glauber and I have been working on getting both paravirt_ops and lguest running on the x86_64. I already pushed the x86_64 patches and as promised, I'm pushing now the lguest64 patches. This patches are greatly influenced by Rusty Russell, and we tried to stay somewhat consistent with his work on the i386. But there are some major differences that we had to over come.

plain text document attachment (hvvm.patch) OK, some explaination is needed here. The goal of lguest with paravirt ops, is to have one kernel that can be loaded both as a host and a guest. To do this, we need to map an area in virtual memory that both the host and guest share. But I don't want any conflicts with the guest. One solution is just to do a single area for boot up, and then use the

plain text document attachment (hvvm.patch) OK, some explaination is needed here. The goal of lguest with paravirt ops, is to have one kernel that can be loaded both as a host and a guest. To do this, we need to map an area in virtual memory that both the host and guest share. But I don't want any conflicts with the guest. One solution is just to do a single area for boot up, and then use the

Impact: potential bugfix In theory, the kernel could reuse the same page as pgdir for a new process while the hypervisor keeps it cached. This would have undesirable results. Signed-off-by: Rusty Russell <rusty at rustcorp.com.au> --- arch/x86/include/asm/lguest_hcall.h | 1 + arch/x86/lguest/boot.c | 8 ++++++++ drivers/lguest/hypercalls.c | 3 +++

Impact: potential bugfix In theory, the kernel could reuse the same page as pgdir for a new process while the hypervisor keeps it cached. This would have undesirable results. Signed-off-by: Rusty Russell <rusty at rustcorp.com.au> --- arch/x86/include/asm/lguest_hcall.h | 1 + arch/x86/lguest/boot.c | 8 ++++++++ drivers/lguest/hypercalls.c | 3 +++

this patch makes room for the vcpu structure in lguest, already used in this very same way at lguest64. It's the first part of our plan to have lguest and lguest64 unified too. When two dogs hang out, you don't have new puppies right in the other day. Some time has to be elapsed. They have to grow first. In this same spirit, having these patches _do not_ mean smp guests can be launched

this patch makes room for the vcpu structure in lguest, already used in this very same way at lguest64. It's the first part of our plan to have lguest and lguest64 unified too. When two dogs hang out, you don't have new puppies right in the other day. Some time has to be elapsed. They have to grow first. In this same spirit, having these patches _do not_ mean smp guests can be launched

Glauber noticed long delays between hitting a key, and seeing data come up on the virtual console. Looking into this, I found that the wake_parent routine that reads from all devices was actually starving out the parent after sending the parent a signal to wake up. The thing is, the child which takes the console input is recognized by the scheduler as an interactive process. The parent,

Glauber noticed long delays between hitting a key, and seeing data come up on the virtual console. Looking into this, I found that the wake_parent routine that reads from all devices was actually starving out the parent after sending the parent a signal to wake up. The thing is, the child which takes the console input is recognized by the scheduler as an interactive process. The parent,

Andrew Morton commented out some stuff to make lguest compile. This should actually make it work (in the latest -mm). Signed-off-by: Rusty Russell <rusty@rustcorp.com.au> --- drivers/lguest/lguest.c | 11 ++--------- 1 file changed, 2 insertions(+), 9 deletions(-) =================================================================== --- a/drivers/lguest/lguest.c +++

Andrew Morton commented out some stuff to make lguest compile. This should actually make it work (in the latest -mm). Signed-off-by: Rusty Russell <rusty@rustcorp.com.au> --- drivers/lguest/lguest.c | 11 ++--------- 1 file changed, 2 insertions(+), 9 deletions(-) =================================================================== --- a/drivers/lguest/lguest.c +++

plain text document attachment (lguest64-kconfig.patch) Put the kconfig options for lguest64 in. Signed-off-by: Steven Rostedt <srostedt@redhat.com> Signed-off-by: Glauber de Oliveira Costa <glommer@gmail.com> Cc: Chris Wright <chrisw@sous-sol.org> Index: work-pv/arch/x86_64/Kconfig =================================================================== ---

plain text document attachment (lguest64-kconfig.patch) Put the kconfig options for lguest64 in. Signed-off-by: Steven Rostedt <srostedt@redhat.com> Signed-off-by: Glauber de Oliveira Costa <glommer@gmail.com> Cc: Chris Wright <chrisw@sous-sol.org> Index: work-pv/arch/x86_64/Kconfig =================================================================== ---

plain text document attachment (lguest64-debug-utils.patch) This patch contains some nice features used to debug the lguest64 guest. It has a way to print page tables for either the host or the guest. It incorporates kallsyms, and can do a nice back trace of a guest when it crashes. The guest needs kallsyms obviously compiled in. Note: This code needs to be fixed to be more secure! Implements

plain text document attachment (lguest64-debug-utils.patch) This patch contains some nice features used to debug the lguest64 guest. It has a way to print page tables for either the host or the guest. It incorporates kallsyms, and can do a nice back trace of a guest when it crashes. The guest needs kallsyms obviously compiled in. Note: This code needs to be fixed to be more secure! Implements

The assumption that we have an overall irqs_pending flags, and a one-to-one lguest <-> task mapping fails to hold on x86_64, where we can have multiple puppies, aka vcpus. Although ifdefs could be used, it makes the code much more unreadable, and other ports are on the way, anyway. So some sort of acessor is preferred anyway. Signed-off-by: Glauber de Oliveira Costa

The assumption that we have an overall irqs_pending flags, and a one-to-one lguest <-> task mapping fails to hold on x86_64, where we can have multiple puppies, aka vcpus. Although ifdefs could be used, it makes the code much more unreadable, and other ports are on the way, anyway. So some sort of acessor is preferred anyway. Signed-off-by: Glauber de Oliveira Costa