diff mbox series

[kvm-unit-tests,v3,3/6] s390x: sie: switch to home space mode before entering SIE

Message ID 20230601070202.152094-4-nrb@linux.ibm.com (mailing list archive)
State New, archived
Headers show
Series s390x: Add support for running guests without MSO/MSL | expand

Commit Message

Nico Boehr June 1, 2023, 7:01 a.m. UTC 
This is to prepare for running guests without MSO/MSL, which is
currently not possible.

We already have code in sie64a to setup a guest primary ASCE before
entering SIE, so we can in theory switch to the page tables which
translate gpa to hpa.

But the host is running in primary space mode already, so changing the
primary ASCE before entering SIE will also affect the host's code and
data.

To make this switch useful, the host should run in a different address
space mode. Hence, set up and change to home address space mode before
installing the guest ASCE.

The home space ASCE is just copied over from the primary space ASCE, so
no functional change is intended, also for tests that want to use
MSO/MSL. If a test intends to use a different primary space ASCE, it can
now just set the guest.asce in the save_area.

Signed-off-by: Nico Boehr <nrb@linux.ibm.com>
---
 lib/s390x/asm/arch_def.h |  1 +
 lib/s390x/sie.c          | 18 ++++++++++++++++++
 2 files changed, 19 insertions(+)

Comments

Janosch Frank June 5, 2023, 9:03 a.m. UTC | #1 
On 6/1/23 09:01, Nico Boehr wrote:
> This is to prepare for running guests without MSO/MSL, which is
> currently not possible.

This is a preparation patch for non-MSO/MSL guest support.

> 
> We already have code in sie64a to setup a guest primary ASCE before
> entering SIE, so we can in theory switch to the page tables which
> translate gpa to hpa.

The important information that's missing here is that SIE always uses 
the prim space for guest to host translations.

> 
> But the host is running in primary space mode already, so changing the
> primary ASCE before entering SIE will also affect the host's code and
> data.

And that's why this is an issue.

For the time being we'll copy the primary ASCE into the home space ASCE. 
No functional change is intended. But if a test intends to....

> 
> To make this switch useful, the host should run in a different address
> space mode. Hence, set up and change to home address space mode before
> installing the guest ASCE.
> 
> The home space ASCE is just copied over from the primary space ASCE, so
> no functional change is intended, also for tests that want to use
> MSO/MSL. If a test intends to use a different primary space ASCE, it can
> now just set the guest.asce in the save_area.
> 
> Signed-off-by: Nico Boehr <nrb@linux.ibm.com>
> ---
>   lib/s390x/asm/arch_def.h |  1 +
>   lib/s390x/sie.c          | 18 ++++++++++++++++++
>   2 files changed, 19 insertions(+)
> 
> diff --git a/lib/s390x/asm/arch_def.h b/lib/s390x/asm/arch_def.h
> index 84f6996c4d8c..099289e7550e 100644
> --- a/lib/s390x/asm/arch_def.h
> +++ b/lib/s390x/asm/arch_def.h
> @@ -90,6 +90,7 @@ struct cpu {
>   #define AS_HOME				3
>   
>   #define PSW_MASK_DAT			0x0400000000000000UL
> +#define PSW_MASK_HOME			0x0000C00000000000UL
>   #define PSW_MASK_IO			0x0200000000000000UL
>   #define PSW_MASK_EXT			0x0100000000000000UL
>   #define PSW_MASK_KEY			0x00F0000000000000UL
> diff --git a/lib/s390x/sie.c b/lib/s390x/sie.c
> index 9241b4b4a512..ffa8ec91a423 100644
> --- a/lib/s390x/sie.c
> +++ b/lib/s390x/sie.c
> @@ -46,6 +46,8 @@ void sie_handle_validity(struct vm *vm)
>   
>   void sie(struct vm *vm)
>   {
> +	uint64_t old_cr13;
> +
>   	if (vm->sblk->sdf == 2)
>   		memcpy(vm->sblk->pv_grregs, vm->save_area.guest.grs,
>   		       sizeof(vm->save_area.guest.grs));
> @@ -53,6 +55,16 @@ void sie(struct vm *vm)
>   	/* Reset icptcode so we don't trip over it below */
>   	vm->sblk->icptcode = 0;
>   
> +	/* set up home address space to match primary space */
> +	old_cr13 = stctg(13);
> +	lctlg(13, stctg(1));
> +
> +	/* switch to home space so guest tables can be different from host */
> +	psw_mask_set_bits(PSW_MASK_HOME);
> +
> +	/* also handle all interruptions in home space while in SIE */
> +	irq_set_dat_mode(IRQ_DAT_ON, AS_HOME);

I'm wondering why this needs to be two calls when you clearly want to 
have a convenience function that does a full space change.

We could introduce:

#define AS_REAL 4

static void mmu_set_addr_space(uint8_t space)
{
	if (space < 4)
		psw_mask_set_bits(space);
	irq_set_dat_mode(space);
}

The "addr" in the function name is optional in my opinion.
@Claudio: What's your opinion?

> +
>   	while (vm->sblk->icptcode == 0) {
>   		sie64a(vm->sblk, &vm->save_area);
>   		sie_handle_validity(vm);
> @@ -60,6 +72,12 @@ void sie(struct vm *vm)
>   	vm->save_area.guest.grs[14] = vm->sblk->gg14;
>   	vm->save_area.guest.grs[15] = vm->sblk->gg15;
>   
> +	irq_set_dat_mode(IRQ_DAT_ON, AS_PRIM);
> +	psw_mask_clear_bits(PSW_MASK_HOME);
> +
> +	/* restore the old CR 13 */
> +	lctlg(13, old_cr13);
> +
>   	if (vm->sblk->sdf == 2)
>   		memcpy(vm->save_area.guest.grs, vm->sblk->pv_grregs,
>   		       sizeof(vm->save_area.guest.grs));
Nico Boehr June 15, 2023, 1:09 p.m. UTC | #2 
Quoting Janosch Frank (2023-06-05 11:03:19)
> On 6/1/23 09:01, Nico Boehr wrote:
> > This is to prepare for running guests without MSO/MSL, which is
> > currently not possible.
> 
> This is a preparation patch for non-MSO/MSL guest support.
> 
> > 
> > We already have code in sie64a to setup a guest primary ASCE before
> > entering SIE, so we can in theory switch to the page tables which
> > translate gpa to hpa.
> 
> The important information that's missing here is that SIE always uses 
> the prim space for guest to host translations.
> 
> > 
> > But the host is running in primary space mode already, so changing the
> > primary ASCE before entering SIE will also affect the host's code and
> > data.
> 
> And that's why this is an issue.
> 
> For the time being we'll copy the primary ASCE into the home space ASCE. 
> No functional change is intended. But if a test intends to....

Thanks, I have to following text now:

 s390x: sie: switch to home space mode before entering SIE

 This is a preparation patch for running guests without MSO/MSL.

 We already have code in sie64a to setup a guest primary ASCE before
 entering SIE. Since SIE always uses the primary ASCE to translate gpa to
 hpa we can in theory switch to seperate guest page tables.

 But the host is running in primary space mode already, so changing the
 primary ASCE before entering SIE will also affect the host's code and
 data. That's why running the host in primary space is an issue.

 To make this switch useful, the host should run in a different address
 space mode. Hence, set up and change to home address space mode before
 installing the guest ASCE.

 For the time being the primary ASCE is copied to the home space ASCE.
 No functional change is intended, but if a test intends to use a
 different primary space ASCE, it can now just set the guest.asce in the
 save_area.

> > diff --git a/lib/s390x/sie.c b/lib/s390x/sie.c
> > index 9241b4b4a512..ffa8ec91a423 100644
> > --- a/lib/s390x/sie.c
> > +++ b/lib/s390x/sie.c
> > @@ -46,6 +46,8 @@ void sie_handle_validity(struct vm *vm)
> >   
> >   void sie(struct vm *vm)
> >   {
[...]
> > +     /* switch to home space so guest tables can be different from host */
> > +     psw_mask_set_bits(PSW_MASK_HOME);
> > +
> > +     /* also handle all interruptions in home space while in SIE */
> > +     irq_set_dat_mode(IRQ_DAT_ON, AS_HOME);
> 
> I'm wondering why this needs to be two calls when you clearly want to 
> have a convenience function that does a full space change.
> 
> We could introduce:
> 
> #define AS_REAL 4
> 
> static void mmu_set_addr_space(uint8_t space)
> {
>         if (space < 4)
>                 psw_mask_set_bits(space);
>         irq_set_dat_mode(space);
> }
> 
> The "addr" in the function name is optional in my opinion.
> @Claudio: What's your opinion?

I don't agree; this convenience function would just hide your intention.

With the code right now it is perfectly clear that we're running in home and
handling all interruptions in home. It is perfectly valid to handle
interruptions in a space different from the one you're currently running in;
I guess that's what every OS does for userspace ;-)
diff mbox series

Patch

diff --git a/lib/s390x/asm/arch_def.h b/lib/s390x/asm/arch_def.h
index 84f6996c4d8c..099289e7550e 100644
--- a/lib/s390x/asm/arch_def.h
+++ b/lib/s390x/asm/arch_def.h
@@ -90,6 +90,7 @@  struct cpu {
 #define AS_HOME				3
 
 #define PSW_MASK_DAT			0x0400000000000000UL
+#define PSW_MASK_HOME			0x0000C00000000000UL
 #define PSW_MASK_IO			0x0200000000000000UL
 #define PSW_MASK_EXT			0x0100000000000000UL
 #define PSW_MASK_KEY			0x00F0000000000000UL
diff --git a/lib/s390x/sie.c b/lib/s390x/sie.c
index 9241b4b4a512..ffa8ec91a423 100644
--- a/lib/s390x/sie.c
+++ b/lib/s390x/sie.c
@@ -46,6 +46,8 @@  void sie_handle_validity(struct vm *vm)
 
 void sie(struct vm *vm)
 {
+	uint64_t old_cr13;
+
 	if (vm->sblk->sdf == 2)
 		memcpy(vm->sblk->pv_grregs, vm->save_area.guest.grs,
 		       sizeof(vm->save_area.guest.grs));
@@ -53,6 +55,16 @@  void sie(struct vm *vm)
 	/* Reset icptcode so we don't trip over it below */
 	vm->sblk->icptcode = 0;
 
+	/* set up home address space to match primary space */
+	old_cr13 = stctg(13);
+	lctlg(13, stctg(1));
+
+	/* switch to home space so guest tables can be different from host */
+	psw_mask_set_bits(PSW_MASK_HOME);
+
+	/* also handle all interruptions in home space while in SIE */
+	irq_set_dat_mode(IRQ_DAT_ON, AS_HOME);
+
 	while (vm->sblk->icptcode == 0) {
 		sie64a(vm->sblk, &vm->save_area);
 		sie_handle_validity(vm);
@@ -60,6 +72,12 @@  void sie(struct vm *vm)
 	vm->save_area.guest.grs[14] = vm->sblk->gg14;
 	vm->save_area.guest.grs[15] = vm->sblk->gg15;
 
+	irq_set_dat_mode(IRQ_DAT_ON, AS_PRIM);
+	psw_mask_clear_bits(PSW_MASK_HOME);
+
+	/* restore the old CR 13 */
+	lctlg(13, old_cr13);
+
 	if (vm->sblk->sdf == 2)
 		memcpy(vm->save_area.guest.grs, vm->sblk->pv_grregs,
 		       sizeof(vm->save_area.guest.grs));