2023-01-22 03:03:25

by Tianyu Lan

Subject: [RFC PATCH V3 07/16] drivers: hv: Decrypt percpu hvcall input arg page in sev-snp enlightened guest

From: Tianyu Lan <[email protected]>

Hypervisor needs to access iput arg page and guest should decrypt
the page.

Signed-off-by: Tianyu Lan <[email protected]>
---
Change since RFC V2:
* Set inputarg to be zero after kfree()
* Not free mem when fail to encrypt mem in the hv_common_cpu_die().
---
drivers/hv/hv_common.c | 20 +++++++++++++++++++-
1 file changed, 19 insertions(+), 1 deletion(-)

diff --git a/drivers/hv/hv_common.c b/drivers/hv/hv_common.c
index f788c64de0bd..205b6380d794 100644
--- a/drivers/hv/hv_common.c
+++ b/drivers/hv/hv_common.c
@@ -21,6 +21,7 @@
#include <linux/ptrace.h>
#include <linux/slab.h>
#include <linux/dma-map-ops.h>
+#include <linux/set_memory.h>
#include <asm/hyperv-tlfs.h>
#include <asm/mshyperv.h>

@@ -125,6 +126,7 @@ int hv_common_cpu_init(unsigned int cpu)
u64 msr_vp_index;
gfp_t flags;
int pgcount = hv_root_partition ? 2 : 1;
+ int ret;

/* hv_cpu_init() can be called with IRQs disabled from hv_resume() */
flags = irqs_disabled() ? GFP_ATOMIC : GFP_KERNEL;
@@ -134,6 +136,17 @@ int hv_common_cpu_init(unsigned int cpu)
if (!(*inputarg))
return -ENOMEM;

+ if (hv_isolation_type_en_snp()) {
+ ret = set_memory_decrypted((unsigned long)*inputarg, pgcount);
+ if (ret) {
+ kfree(*inputarg);
+ *inputarg = NULL;
+ return ret;
+ }
+
+ memset(*inputarg, 0x00, PAGE_SIZE);
+ }
+
if (hv_root_partition) {
outputarg = (void **)this_cpu_ptr(hyperv_pcpu_output_arg);
*outputarg = (char *)(*inputarg) + HV_HYP_PAGE_SIZE;
@@ -168,7 +181,12 @@ int hv_common_cpu_die(unsigned int cpu)

local_irq_restore(flags);

- kfree(mem);
+ if (hv_isolation_type_en_snp()) {
+ if (!set_memory_encrypted((unsigned long)mem, 1))
+ kfree(mem);
+ } else {
+ kfree(mem);
+ }

return 0;
}
--
2.25.1

2023-01-31 18:02:50

by Michael Kelley (LINUX)

Subject: RE: [RFC PATCH V3 07/16] drivers: hv: Decrypt percpu hvcall input arg page in sev-snp enlightened guest

From: Tianyu Lan <[email protected]> Sent: Saturday, January 21, 2023 6:46 PM
>
> Hypervisor needs to access iput arg page and guest should decrypt
> the page.
>
> Signed-off-by: Tianyu Lan <[email protected]>
> ---
> Change since RFC V2:
> * Set inputarg to be zero after kfree()
> * Not free mem when fail to encrypt mem in the hv_common_cpu_die().
> ---
> drivers/hv/hv_common.c | 20 +++++++++++++++++++-
> 1 file changed, 19 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/hv/hv_common.c b/drivers/hv/hv_common.c
> index f788c64de0bd..205b6380d794 100644
> --- a/drivers/hv/hv_common.c
> +++ b/drivers/hv/hv_common.c
> @@ -21,6 +21,7 @@
> #include <linux/ptrace.h>
> #include <linux/slab.h>
> #include <linux/dma-map-ops.h>
> +#include <linux/set_memory.h>
> #include <asm/hyperv-tlfs.h>
> #include <asm/mshyperv.h>
>
> @@ -125,6 +126,7 @@ int hv_common_cpu_init(unsigned int cpu)
> u64 msr_vp_index;
> gfp_t flags;
> int pgcount = hv_root_partition ? 2 : 1;
> + int ret;
>
> /* hv_cpu_init() can be called with IRQs disabled from hv_resume() */
> flags = irqs_disabled() ? GFP_ATOMIC : GFP_KERNEL;
> @@ -134,6 +136,17 @@ int hv_common_cpu_init(unsigned int cpu)
> if (!(*inputarg))
> return -ENOMEM;
>
> + if (hv_isolation_type_en_snp()) {
> + ret = set_memory_decrypted((unsigned long)*inputarg, pgcount);

You used "pgcount" here in response to a comment on v2 of the
patch. But the corresponding re-encryption in hv_common_cpu_die()
uses a fixed value of "1". The two cases should be consistent. Either
assert that hv_root_partition will never be true in an SNP VM, in which
case hard coding "1" is OK. Or properly calculate the number of pages
in both cases so they are consistent.

> + if (ret) {
> + kfree(*inputarg);
> + *inputarg = NULL;
> + return ret;
> + }
> +
> + memset(*inputarg, 0x00, PAGE_SIZE);
> + }
> +
> if (hv_root_partition) {
> outputarg = (void **)this_cpu_ptr(hyperv_pcpu_output_arg);
> *outputarg = (char *)(*inputarg) + HV_HYP_PAGE_SIZE;
> @@ -168,7 +181,12 @@ int hv_common_cpu_die(unsigned int cpu)
>
> local_irq_restore(flags);
>
> - kfree(mem);
> + if (hv_isolation_type_en_snp()) {
> + if (!set_memory_encrypted((unsigned long)mem, 1))
> + kfree(mem);
> + } else {
> + kfree(mem);
> + }
>
> return 0;
> }
> --
> 2.25.1

2023-02-03 05:23:23

by Tianyu Lan

Subject: Re: [RFC PATCH V3 07/16] drivers: hv: Decrypt percpu hvcall input arg page in sev-snp enlightened guest

On 2/1/2023 2:02 AM, Michael Kelley (LINUX) wrote:
>> @@ -134,6 +136,17 @@ int hv_common_cpu_init(unsigned int cpu)
>> if (!(*inputarg))
>> return -ENOMEM;
>>
>> + if (hv_isolation_type_en_snp()) {
>> + ret = set_memory_decrypted((unsigned long)*inputarg, pgcount);
> You used "pgcount" here in response to a comment on v2 of the
> patch. But the corresponding re-encryption in hv_common_cpu_die()
> uses a fixed value of "1". The two cases should be consistent. Either
> assert that hv_root_partition will never be true in an SNP VM, in which
> case hard coding "1" is OK. Or properly calculate the number of pages
> in both cases so they are consistent.
>

Agree. We should keep the logic in both hv_common_cpu_init() and
hv_common_cpu_die(). Will fix it.