* [PATCH] Fix SEV-ES INS/OUTS instructions for word, dword, and qword.
@ 2021-11-18 2:13 Michael Sterritt
2021-11-18 2:41 ` Peter Gonda
` (2 more replies)
0 siblings, 3 replies; 4+ messages in thread
From: Michael Sterritt @ 2021-11-18 2:13 UTC (permalink / raw)
To: Thomas Gleixner, Ingo Molnar, Borislav Petkov, Dave Hansen, x86,
H. Peter Anvin, Joerg Roedel, Peter Zijlstra, Tom Lendacky,
linux-kernel, kvm, virtualization, linux-coco
Cc: marcorr, pgonda, Michael Sterritt
Properly type the operands being passed to __put_user()/__get_user().
Otherwise, these routines truncate data for dependent instructions
(e.g., INSW) and only read/write one byte.
Tested: Tested by sending a string with `REP OUTSW` to a port and then
reading it back in with `REP INSW` on the same port. Previous behavior
was to only send and receive the first char of the size. For example,
word operations for "abcd" would only read/write "ac". With change, the
full string is now written and read back.
Signed-off-by: Michael Sterritt <sterritt@google.com>
Reviewed-by: Marc Orr <marcorr@google.com>
Reviewed-by: Peter Gonda <pgonda@google.com>
---
arch/x86/kernel/sev.c | 57 +++++++++++++++++++++++++++++--------------
1 file changed, 39 insertions(+), 18 deletions(-)
diff --git a/arch/x86/kernel/sev.c b/arch/x86/kernel/sev.c
index 74f0ec955384..a9fc2ac7a8bd 100644
--- a/arch/x86/kernel/sev.c
+++ b/arch/x86/kernel/sev.c
@@ -294,11 +294,6 @@ static enum es_result vc_write_mem(struct es_em_ctxt *ctxt,
char *dst, char *buf, size_t size)
{
unsigned long error_code = X86_PF_PROT | X86_PF_WRITE;
- char __user *target = (char __user *)dst;
- u64 d8;
- u32 d4;
- u16 d2;
- u8 d1;
/*
* This function uses __put_user() independent of whether kernel or user
@@ -320,26 +315,42 @@ static enum es_result vc_write_mem(struct es_em_ctxt *ctxt,
* instructions here would cause infinite nesting.
*/
switch (size) {
- case 1:
+ case 1: {
+ u8 d1;
+ u8 __user *target = (u8 __user *)dst;
+
memcpy(&d1, buf, 1);
if (__put_user(d1, target))
goto fault;
break;
- case 2:
+ }
+ case 2: {
+ u16 d2;
+ u16 __user *target = (u16 __user *)dst;
+
memcpy(&d2, buf, 2);
if (__put_user(d2, target))
goto fault;
break;
- case 4:
+ }
+ case 4: {
+ u32 d4;
+ u32 __user *target = (u32 __user *)dst;
+
memcpy(&d4, buf, 4);
if (__put_user(d4, target))
goto fault;
break;
- case 8:
+ }
+ case 8: {
+ u64 d8;
+ u64 __user *target = (u64 __user *)dst;
+
memcpy(&d8, buf, 8);
if (__put_user(d8, target))
goto fault;
break;
+ }
default:
WARN_ONCE(1, "%s: Invalid size: %zu\n", __func__, size);
return ES_UNSUPPORTED;
@@ -362,11 +373,6 @@ static enum es_result vc_read_mem(struct es_em_ctxt *ctxt,
char *src, char *buf, size_t size)
{
unsigned long error_code = X86_PF_PROT;
- char __user *s = (char __user *)src;
- u64 d8;
- u32 d4;
- u16 d2;
- u8 d1;
/*
* This function uses __get_user() independent of whether kernel or user
@@ -388,26 +394,41 @@ static enum es_result vc_read_mem(struct es_em_ctxt *ctxt,
* instructions here would cause infinite nesting.
*/
switch (size) {
- case 1:
+ case 1: {
+ u8 d1;
+ u8 __user *s = (u8 __user *)src;
+
if (__get_user(d1, s))
goto fault;
memcpy(buf, &d1, 1);
break;
- case 2:
+ }
+ case 2: {
+ u16 d2;
+ u16 __user *s = (u16 __user *)src;
+
if (__get_user(d2, s))
goto fault;
memcpy(buf, &d2, 2);
break;
- case 4:
+ }
+ case 4: {
+ u32 d4;
+ u32 __user *s = (u32 __user *)src;
+
if (__get_user(d4, s))
goto fault;
memcpy(buf, &d4, 4);
break;
- case 8:
+ }
+ case 8: {
+ u64 d8;
+ u64 __user *s = (u64 __user *)src;
if (__get_user(d8, s))
goto fault;
memcpy(buf, &d8, 8);
break;
+ }
default:
WARN_ONCE(1, "%s: Invalid size: %zu\n", __func__, size);
return ES_UNSUPPORTED;
--
2.34.0.rc2.393.gf8c9666880-goog
^ permalink raw reply related [flat|nested] 4+ messages in thread
* Re: [PATCH] Fix SEV-ES INS/OUTS instructions for word, dword, and qword.
2021-11-18 2:13 [PATCH] Fix SEV-ES INS/OUTS instructions for word, dword, and qword Michael Sterritt
@ 2021-11-18 2:41 ` Peter Gonda
2021-11-18 11:48 ` Paolo Bonzini
2021-11-19 8:36 ` Joerg Roedel
2 siblings, 0 replies; 4+ messages in thread
From: Peter Gonda @ 2021-11-18 2:41 UTC (permalink / raw)
To: Michael Sterritt
Cc: Thomas Gleixner, Ingo Molnar, Borislav Petkov, Dave Hansen, x86,
H. Peter Anvin, Joerg Roedel, Peter Zijlstra, Tom Lendacky,
linux-kernel, kvm, virtualization, linux-coco, marcorr
On Wed, Nov 17, 2021 at 7:13 PM Michael Sterritt <sterritt@google.com> wrote:
>
> Properly type the operands being passed to __put_user()/__get_user().
> Otherwise, these routines truncate data for dependent instructions
> (e.g., INSW) and only read/write one byte.
>
> Tested: Tested by sending a string with `REP OUTSW` to a port and then
> reading it back in with `REP INSW` on the same port. Previous behavior
> was to only send and receive the first char of the size. For example,
> word operations for "abcd" would only read/write "ac". With change, the
> full string is now written and read back.
I think this should include:
Fixes: f980f9c31a923 (x86/sev-es: Compile early handler code into kernel image)
>
> Signed-off-by: Michael Sterritt <sterritt@google.com>
> Reviewed-by: Marc Orr <marcorr@google.com>
> Reviewed-by: Peter Gonda <pgonda@google.com>
> ---
> arch/x86/kernel/sev.c | 57 +++++++++++++++++++++++++++++--------------
> 1 file changed, 39 insertions(+), 18 deletions(-)
>
> diff --git a/arch/x86/kernel/sev.c b/arch/x86/kernel/sev.c
> index 74f0ec955384..a9fc2ac7a8bd 100644
> --- a/arch/x86/kernel/sev.c
> +++ b/arch/x86/kernel/sev.c
> @@ -294,11 +294,6 @@ static enum es_result vc_write_mem(struct es_em_ctxt *ctxt,
> char *dst, char *buf, size_t size)
> {
> unsigned long error_code = X86_PF_PROT | X86_PF_WRITE;
> - char __user *target = (char __user *)dst;
> - u64 d8;
> - u32 d4;
> - u16 d2;
> - u8 d1;
>
> /*
> * This function uses __put_user() independent of whether kernel or user
> @@ -320,26 +315,42 @@ static enum es_result vc_write_mem(struct es_em_ctxt *ctxt,
> * instructions here would cause infinite nesting.
> */
> switch (size) {
> - case 1:
> + case 1: {
> + u8 d1;
> + u8 __user *target = (u8 __user *)dst;
> +
> memcpy(&d1, buf, 1);
> if (__put_user(d1, target))
> goto fault;
> break;
> - case 2:
> + }
> + case 2: {
> + u16 d2;
> + u16 __user *target = (u16 __user *)dst;
> +
> memcpy(&d2, buf, 2);
> if (__put_user(d2, target))
> goto fault;
> break;
> - case 4:
> + }
> + case 4: {
> + u32 d4;
> + u32 __user *target = (u32 __user *)dst;
> +
> memcpy(&d4, buf, 4);
> if (__put_user(d4, target))
> goto fault;
> break;
> - case 8:
> + }
> + case 8: {
> + u64 d8;
> + u64 __user *target = (u64 __user *)dst;
> +
> memcpy(&d8, buf, 8);
> if (__put_user(d8, target))
> goto fault;
> break;
> + }
> default:
> WARN_ONCE(1, "%s: Invalid size: %zu\n", __func__, size);
> return ES_UNSUPPORTED;
> @@ -362,11 +373,6 @@ static enum es_result vc_read_mem(struct es_em_ctxt *ctxt,
> char *src, char *buf, size_t size)
> {
> unsigned long error_code = X86_PF_PROT;
> - char __user *s = (char __user *)src;
> - u64 d8;
> - u32 d4;
> - u16 d2;
> - u8 d1;
>
> /*
> * This function uses __get_user() independent of whether kernel or user
> @@ -388,26 +394,41 @@ static enum es_result vc_read_mem(struct es_em_ctxt *ctxt,
> * instructions here would cause infinite nesting.
> */
> switch (size) {
> - case 1:
> + case 1: {
> + u8 d1;
> + u8 __user *s = (u8 __user *)src;
> +
> if (__get_user(d1, s))
> goto fault;
> memcpy(buf, &d1, 1);
> break;
> - case 2:
> + }
> + case 2: {
> + u16 d2;
> + u16 __user *s = (u16 __user *)src;
> +
> if (__get_user(d2, s))
> goto fault;
> memcpy(buf, &d2, 2);
> break;
> - case 4:
> + }
> + case 4: {
> + u32 d4;
> + u32 __user *s = (u32 __user *)src;
> +
> if (__get_user(d4, s))
> goto fault;
> memcpy(buf, &d4, 4);
> break;
> - case 8:
> + }
> + case 8: {
> + u64 d8;
> + u64 __user *s = (u64 __user *)src;
> if (__get_user(d8, s))
> goto fault;
> memcpy(buf, &d8, 8);
> break;
> + }
> default:
> WARN_ONCE(1, "%s: Invalid size: %zu\n", __func__, size);
> return ES_UNSUPPORTED;
> --
> 2.34.0.rc2.393.gf8c9666880-goog
>
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH] Fix SEV-ES INS/OUTS instructions for word, dword, and qword.
2021-11-18 2:13 [PATCH] Fix SEV-ES INS/OUTS instructions for word, dword, and qword Michael Sterritt
2021-11-18 2:41 ` Peter Gonda
@ 2021-11-18 11:48 ` Paolo Bonzini
2021-11-19 8:36 ` Joerg Roedel
2 siblings, 0 replies; 4+ messages in thread
From: Paolo Bonzini @ 2021-11-18 11:48 UTC (permalink / raw)
To: Michael Sterritt, Thomas Gleixner, Ingo Molnar, Borislav Petkov,
Dave Hansen, x86, H. Peter Anvin, Joerg Roedel, Peter Zijlstra,
Tom Lendacky, linux-kernel, kvm, virtualization, linux-coco
Cc: marcorr, pgonda
On 11/18/21 03:13, Michael Sterritt wrote:
> Properly type the operands being passed to __put_user()/__get_user().
> Otherwise, these routines truncate data for dependent instructions
> (e.g., INSW) and only read/write one byte.
>
> Tested: Tested by sending a string with `REP OUTSW` to a port and then
> reading it back in with `REP INSW` on the same port. Previous behavior
> was to only send and receive the first char of the size. For example,
> word operations for "abcd" would only read/write "ac". With change, the
> full string is now written and read back.
>
> Signed-off-by: Michael Sterritt <sterritt@google.com>
> Reviewed-by: Marc Orr <marcorr@google.com>
> Reviewed-by: Peter Gonda <pgonda@google.com>
> ---
> arch/x86/kernel/sev.c | 57 +++++++++++++++++++++++++++++--------------
> 1 file changed, 39 insertions(+), 18 deletions(-)
>
> diff --git a/arch/x86/kernel/sev.c b/arch/x86/kernel/sev.c
> index 74f0ec955384..a9fc2ac7a8bd 100644
> --- a/arch/x86/kernel/sev.c
> +++ b/arch/x86/kernel/sev.c
> @@ -294,11 +294,6 @@ static enum es_result vc_write_mem(struct es_em_ctxt *ctxt,
> char *dst, char *buf, size_t size)
> {
> unsigned long error_code = X86_PF_PROT | X86_PF_WRITE;
> - char __user *target = (char __user *)dst;
> - u64 d8;
> - u32 d4;
> - u16 d2;
> - u8 d1;
>
> /*
> * This function uses __put_user() independent of whether kernel or user
> @@ -320,26 +315,42 @@ static enum es_result vc_write_mem(struct es_em_ctxt *ctxt,
> * instructions here would cause infinite nesting.
> */
> switch (size) {
> - case 1:
> + case 1: {
> + u8 d1;
> + u8 __user *target = (u8 __user *)dst;
> +
> memcpy(&d1, buf, 1);
> if (__put_user(d1, target))
> goto fault;
> break;
> - case 2:
> + }
> + case 2: {
> + u16 d2;
> + u16 __user *target = (u16 __user *)dst;
> +
> memcpy(&d2, buf, 2);
> if (__put_user(d2, target))
> goto fault;
> break;
> - case 4:
> + }
> + case 4: {
> + u32 d4;
> + u32 __user *target = (u32 __user *)dst;
> +
> memcpy(&d4, buf, 4);
> if (__put_user(d4, target))
> goto fault;
> break;
> - case 8:
> + }
> + case 8: {
> + u64 d8;
> + u64 __user *target = (u64 __user *)dst;
> +
> memcpy(&d8, buf, 8);
> if (__put_user(d8, target))
> goto fault;
> break;
> + }
> default:
> WARN_ONCE(1, "%s: Invalid size: %zu\n", __func__, size);
> return ES_UNSUPPORTED;
> @@ -362,11 +373,6 @@ static enum es_result vc_read_mem(struct es_em_ctxt *ctxt,
> char *src, char *buf, size_t size)
> {
> unsigned long error_code = X86_PF_PROT;
> - char __user *s = (char __user *)src;
> - u64 d8;
> - u32 d4;
> - u16 d2;
> - u8 d1;
>
> /*
> * This function uses __get_user() independent of whether kernel or user
> @@ -388,26 +394,41 @@ static enum es_result vc_read_mem(struct es_em_ctxt *ctxt,
> * instructions here would cause infinite nesting.
> */
> switch (size) {
> - case 1:
> + case 1: {
> + u8 d1;
> + u8 __user *s = (u8 __user *)src;
> +
> if (__get_user(d1, s))
> goto fault;
> memcpy(buf, &d1, 1);
> break;
> - case 2:
> + }
> + case 2: {
> + u16 d2;
> + u16 __user *s = (u16 __user *)src;
> +
> if (__get_user(d2, s))
> goto fault;
> memcpy(buf, &d2, 2);
> break;
> - case 4:
> + }
> + case 4: {
> + u32 d4;
> + u32 __user *s = (u32 __user *)src;
> +
> if (__get_user(d4, s))
> goto fault;
> memcpy(buf, &d4, 4);
> break;
> - case 8:
> + }
> + case 8: {
> + u64 d8;
> + u64 __user *s = (u64 __user *)src;
> if (__get_user(d8, s))
> goto fault;
> memcpy(buf, &d8, 8);
> break;
> + }
> default:
> WARN_ONCE(1, "%s: Invalid size: %zu\n", __func__, size);
> return ES_UNSUPPORTED;
>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH] Fix SEV-ES INS/OUTS instructions for word, dword, and qword.
2021-11-18 2:13 [PATCH] Fix SEV-ES INS/OUTS instructions for word, dword, and qword Michael Sterritt
2021-11-18 2:41 ` Peter Gonda
2021-11-18 11:48 ` Paolo Bonzini
@ 2021-11-19 8:36 ` Joerg Roedel
2 siblings, 0 replies; 4+ messages in thread
From: Joerg Roedel @ 2021-11-19 8:36 UTC (permalink / raw)
To: Michael Sterritt
Cc: Thomas Gleixner, Ingo Molnar, Borislav Petkov, Dave Hansen, x86,
H. Peter Anvin, Peter Zijlstra, Tom Lendacky, linux-kernel, kvm,
virtualization, linux-coco, marcorr, pgonda
Hi Michael,
On Wed, Nov 17, 2021 at 06:13:26PM -0800, Michael Sterritt wrote:
> Properly type the operands being passed to __put_user()/__get_user().
> Otherwise, these routines truncate data for dependent instructions
> (e.g., INSW) and only read/write one byte.
>
> Tested: Tested by sending a string with `REP OUTSW` to a port and then
> reading it back in with `REP INSW` on the same port. Previous behavior
> was to only send and receive the first char of the size. For example,
> word operations for "abcd" would only read/write "ac". With change, the
> full string is now written and read back.
Thanks for fixing this! When you re-send, please change the subject to
x86/sev-es: Fix SEV-ES INS/OUTS instructions for word, dword, and qword
Regards,
--
Jörg Rödel
jroedel@suse.de
SUSE Software Solutions Germany GmbH
Maxfeldstr. 5
90409 Nürnberg
Germany
(HRB 36809, AG Nürnberg)
Geschäftsführer: Ivo Totev
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2021-11-19 8:36 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-11-18 2:13 [PATCH] Fix SEV-ES INS/OUTS instructions for word, dword, and qword Michael Sterritt
2021-11-18 2:41 ` Peter Gonda
2021-11-18 11:48 ` Paolo Bonzini
2021-11-19 8:36 ` Joerg Roedel
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).