[PATCH 0/3] x86/insn: Fix not using prefixes.nbytes for loop over prefixes.bytes

[PATCH 0/3] x86/insn: Fix not using prefixes.nbytes for loop over prefixes.bytes

[Masami Hiramatsu]

Hi,

Here are the patches to fix the wrong loop boundary check
on insn.prefixes.bytes[] array.

Kees Cook reported that this issue that there are similar
wrong boundary check patterns in the x86 code.

Since the insn.prefixes.nbytes can be bigger than the size of
insn.prefixes.bytes[] when a same prefix is repeated, we have to
check whether the insn.prefixes.bytes[i] != 0 (*) and i < 4 instead
of insn.prefixes.nbytes.

(*) Note that insn.prefixes.bytes[] should be zeroed in insn_init()
before decoding, and 0x00 is not a legacy prefix. So if you see 0
on insn.prefix.bytes[], it indicates the end of the array. Or,
if the prefixes.bytes[] is filled with prefix bytes, we can check
the index is less than 4.

Thank you,

---

Masami Hiramatsu (3):
      x86/sev-es: Fix not using prefixes.nbytes for loop over prefixes.bytes
      x86/uprobes: Fix not using prefixes.nbytes for loop over prefixes.bytes
      x86/insn-eval: Fix not using prefixes.nbytes for loop over prefixes.bytes


 arch/x86/boot/compressed/sev-es.c |    2 +-
 arch/x86/kernel/uprobes.c         |    4 ++--
 arch/x86/lib/insn-eval.c          |    4 ++--
 3 files changed, 5 insertions(+), 5 deletions(-)

-- 
Masami Hiramatsu (Linaro) <mhiramat@kernel.org>

[PATCH 1/3] x86/sev-es: Fix not using prefixes.nbytes for loop over prefixes.bytes

[Masami Hiramatsu]

Since the insn.prefixes.nbytes can be bigger than the size of
insn.prefixes.bytes[] when a same prefix is repeated, we have to
check whether the insn.prefixes.bytes[i] != 0 and i < 4 instead
of insn.prefixes.nbytes.

Fixes: 25189d08e516 ("x86/sev-es: Add support for handling IOIO exceptions")
Reported-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org>
---
 arch/x86/boot/compressed/sev-es.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/x86/boot/compressed/sev-es.c b/arch/x86/boot/compressed/sev-es.c
index 954cb2702e23..6a7a3027c9ac 100644
--- a/arch/x86/boot/compressed/sev-es.c
+++ b/arch/x86/boot/compressed/sev-es.c
@@ -36,7 +36,7 @@ static bool insn_has_rep_prefix(struct insn *insn)
 
 	insn_get_prefixes(insn);
 
-	for (i = 0; i < insn->prefixes.nbytes; i++) {
+	for (i = 0; insn->prefixes.bytes[i] && i < 4; i++) {
 		insn_byte_t p = insn->prefixes.bytes[i];
 
 		if (p == 0xf2 || p == 0xf3)

Re: [PATCH 1/3] x86/sev-es: Fix not using prefixes.nbytes for loop over prefixes.bytes

[Tom Lendacky]

On 12/2/20 2:51 AM, Masami Hiramatsu wrote:
> Since the insn.prefixes.nbytes can be bigger than the size of
> insn.prefixes.bytes[] when a same prefix is repeated, we have to
> check whether the insn.prefixes.bytes[i] != 0 and i < 4 instead
> of insn.prefixes.nbytes.
> 
> Fixes: 25189d08e516 ("x86/sev-es: Add support for handling IOIO exceptions")
> Reported-by: Kees Cook <keescook@chromium.org>
> Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org>
> ---
>   arch/x86/boot/compressed/sev-es.c |    2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/arch/x86/boot/compressed/sev-es.c b/arch/x86/boot/compressed/sev-es.c
> index 954cb2702e23..6a7a3027c9ac 100644
> --- a/arch/x86/boot/compressed/sev-es.c
> +++ b/arch/x86/boot/compressed/sev-es.c
> @@ -36,7 +36,7 @@ static bool insn_has_rep_prefix(struct insn *insn)
>   
>   	insn_get_prefixes(insn);
>   
> -	for (i = 0; i < insn->prefixes.nbytes; i++) {
> +	for (i = 0; insn->prefixes.bytes[i] && i < 4; i++) {

Wouldn't it be better to create a #define for the size rather than hard 
coding 4 in the various files? That would protect everything should the 
bytes array size ever change in the future.

Thanks,
Tom

>   		insn_byte_t p = insn->prefixes.bytes[i];
>   
>   		if (p == 0xf2 || p == 0xf3)
> 

Re: [PATCH 1/3] x86/sev-es: Fix not using prefixes.nbytes for loop over prefixes.bytes

[Kees Cook]

On Wed, Dec 02, 2020 at 09:31:57AM -0600, Tom Lendacky wrote:
> On 12/2/20 2:51 AM, Masami Hiramatsu wrote:
> > Since the insn.prefixes.nbytes can be bigger than the size of
> > insn.prefixes.bytes[] when a same prefix is repeated, we have to
> > check whether the insn.prefixes.bytes[i] != 0 and i < 4 instead
> > of insn.prefixes.nbytes.
> > 
> > Fixes: 25189d08e516 ("x86/sev-es: Add support for handling IOIO exceptions")
> > Reported-by: Kees Cook <keescook@chromium.org>
> > Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org>
> > ---
> >   arch/x86/boot/compressed/sev-es.c |    2 +-
> >   1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/arch/x86/boot/compressed/sev-es.c b/arch/x86/boot/compressed/sev-es.c
> > index 954cb2702e23..6a7a3027c9ac 100644
> > --- a/arch/x86/boot/compressed/sev-es.c
> > +++ b/arch/x86/boot/compressed/sev-es.c
> > @@ -36,7 +36,7 @@ static bool insn_has_rep_prefix(struct insn *insn)
> >   	insn_get_prefixes(insn);
> > -	for (i = 0; i < insn->prefixes.nbytes; i++) {
> > +	for (i = 0; insn->prefixes.bytes[i] && i < 4; i++) {

You must test "i" before bytes[i] or you still do the out-of-bounds-read.

> 
> Wouldn't it be better to create a #define for the size rather than hard
> coding 4 in the various files? That would protect everything should the
> bytes array size ever change in the future.

Agreed, and perhaps instead of repeating the idiom in the for loop, add
a helper like:

#define insn_prefix_valid(prefixes, i) (i >=0 && i < 4 && prefixes->bytes[i])

to be used like:

	for (i = 0; insn_prefix_valid(&insn->prefixes, i); i++) {

> 
> Thanks,
> Tom
> 
> >   		insn_byte_t p = insn->prefixes.bytes[i];
> >   		if (p == 0xf2 || p == 0xf3)
> > 

-- 
Kees Cook

Re: [PATCH 1/3] x86/sev-es: Fix not using prefixes.nbytes for loop over prefixes.bytes

[Masami Hiramatsu]

On Wed, 2 Dec 2020 11:07:26 -0800
Kees Cook <keescook@chromium.org> wrote:

> On Wed, Dec 02, 2020 at 09:31:57AM -0600, Tom Lendacky wrote:
> > On 12/2/20 2:51 AM, Masami Hiramatsu wrote:
> > > Since the insn.prefixes.nbytes can be bigger than the size of
> > > insn.prefixes.bytes[] when a same prefix is repeated, we have to
> > > check whether the insn.prefixes.bytes[i] != 0 and i < 4 instead
> > > of insn.prefixes.nbytes.
> > > 
> > > Fixes: 25189d08e516 ("x86/sev-es: Add support for handling IOIO exceptions")
> > > Reported-by: Kees Cook <keescook@chromium.org>
> > > Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org>
> > > ---
> > >   arch/x86/boot/compressed/sev-es.c |    2 +-
> > >   1 file changed, 1 insertion(+), 1 deletion(-)
> > > 
> > > diff --git a/arch/x86/boot/compressed/sev-es.c b/arch/x86/boot/compressed/sev-es.c
> > > index 954cb2702e23..6a7a3027c9ac 100644
> > > --- a/arch/x86/boot/compressed/sev-es.c
> > > +++ b/arch/x86/boot/compressed/sev-es.c
> > > @@ -36,7 +36,7 @@ static bool insn_has_rep_prefix(struct insn *insn)
> > >   	insn_get_prefixes(insn);
> > > -	for (i = 0; i < insn->prefixes.nbytes; i++) {
> > > +	for (i = 0; insn->prefixes.bytes[i] && i < 4; i++) {
> 
> You must test "i" before bytes[i] or you still do the out-of-bounds-read.

Oops, thanks.

> 
> > 
> > Wouldn't it be better to create a #define for the size rather than hard
> > coding 4 in the various files? That would protect everything should the
> > bytes array size ever change in the future.
> 
> Agreed, and perhaps instead of repeating the idiom in the for loop, add
> a helper like:
> 
> #define insn_prefix_valid(prefixes, i) (i >=0 && i < 4 && prefixes->bytes[i])
> 
> to be used like:
> 
> 	for (i = 0; insn_prefix_valid(&insn->prefixes, i); i++) {

Hm, for all of these usage, they are looping on the prefixes, so

for_each_insn_prefix(insn, idx, prefix) {
...
}

will be simpler.

Thank you,

> 
> > 
> > Thanks,
> > Tom
> > 
> > >   		insn_byte_t p = insn->prefixes.bytes[i];
> > >   		if (p == 0xf2 || p == 0xf3)
> > > 
> 
> -- 
> Kees Cook


-- 
Masami Hiramatsu <mhiramat@kernel.org>

[PATCH 2/3] x86/uprobes: Fix not using prefixes.nbytes for loop over prefixes.bytes

[Masami Hiramatsu]

Since the insn.prefixes.nbytes can be bigger than the size of
insn.prefixes.bytes[] when a same prefix is repeated, we have to
check whether the insn.prefixes.bytes[i] != 0 and i < 4 instead
of insn.prefixes.nbytes.

Fixes: 2b1444983508 ("uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints")
Cc: stable@vger.kernel.org
Reported-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org>
---
 arch/x86/kernel/uprobes.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/arch/x86/kernel/uprobes.c b/arch/x86/kernel/uprobes.c
index 3fdaa042823d..bb3ea3705b99 100644
--- a/arch/x86/kernel/uprobes.c
+++ b/arch/x86/kernel/uprobes.c
@@ -257,7 +257,7 @@ static bool is_prefix_bad(struct insn *insn)
 {
 	int i;
 
-	for (i = 0; i < insn->prefixes.nbytes; i++) {
+	for (i = 0; insn->prefixes.bytes[i] && i < 4; i++) {
 		insn_attr_t attr;
 
 		attr = inat_get_opcode_attribute(insn->prefixes.bytes[i]);
@@ -746,7 +746,7 @@ static int branch_setup_xol_ops(struct arch_uprobe *auprobe, struct insn *insn)
 	 * Intel and AMD behavior differ in 64-bit mode: Intel ignores 66 prefix.
 	 * No one uses these insns, reject any branch insns with such prefix.
 	 */
-	for (i = 0; i < insn->prefixes.nbytes; i++) {
+	for (i = 0; insn->prefixes.bytes[i] && i < 4; i++) {
 		if (insn->prefixes.bytes[i] == 0x66)
 			return -ENOTSUPP;
 	}

Re: [PATCH 2/3] x86/uprobes: Fix not using prefixes.nbytes for loop over prefixes.bytes

[Srikar Dronamraju]

* Masami Hiramatsu <mhiramat@kernel.org> [2020-12-02 17:51:16]:

> Since the insn.prefixes.nbytes can be bigger than the size of
> insn.prefixes.bytes[] when a same prefix is repeated, we have to
> check whether the insn.prefixes.bytes[i] != 0 and i < 4 instead
> of insn.prefixes.nbytes.
> 
> Fixes: 2b1444983508 ("uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints")
> Cc: stable@vger.kernel.org
> Reported-by: Kees Cook <keescook@chromium.org>
> Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org>

Looks good to me.

Reviewed-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>

> ---
>  arch/x86/kernel/uprobes.c |    4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/arch/x86/kernel/uprobes.c b/arch/x86/kernel/uprobes.c
> index 3fdaa042823d..bb3ea3705b99 100644
> --- a/arch/x86/kernel/uprobes.c
> +++ b/arch/x86/kernel/uprobes.c
> @@ -257,7 +257,7 @@ static bool is_prefix_bad(struct insn *insn)
>  {
>  	int i;
> 
> -	for (i = 0; i < insn->prefixes.nbytes; i++) {
> +	for (i = 0; insn->prefixes.bytes[i] && i < 4; i++) {
>  		insn_attr_t attr;
> 
>  		attr = inat_get_opcode_attribute(insn->prefixes.bytes[i]);
> @@ -746,7 +746,7 @@ static int branch_setup_xol_ops(struct arch_uprobe *auprobe, struct insn *insn)
>  	 * Intel and AMD behavior differ in 64-bit mode: Intel ignores 66 prefix.
>  	 * No one uses these insns, reject any branch insns with such prefix.
>  	 */
> -	for (i = 0; i < insn->prefixes.nbytes; i++) {
> +	for (i = 0; insn->prefixes.bytes[i] && i < 4; i++) {
>  		if (insn->prefixes.bytes[i] == 0x66)
>  			return -ENOTSUPP;
>  	}
> 

-- 
Thanks and Regards
Srikar Dronamraju

Re: [PATCH 2/3] x86/uprobes: Fix not using prefixes.nbytes for loop over prefixes.bytes

[Masami Hiramatsu]

On Wed, 2 Dec 2020 20:21:35 +0530
Srikar Dronamraju <srikar@linux.vnet.ibm.com> wrote:

> * Masami Hiramatsu <mhiramat@kernel.org> [2020-12-02 17:51:16]:
> 
> > Since the insn.prefixes.nbytes can be bigger than the size of
> > insn.prefixes.bytes[] when a same prefix is repeated, we have to
> > check whether the insn.prefixes.bytes[i] != 0 and i < 4 instead
> > of insn.prefixes.nbytes.
> > 
> > Fixes: 2b1444983508 ("uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints")
> > Cc: stable@vger.kernel.org
> > Reported-by: Kees Cook <keescook@chromium.org>
> > Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org>
> 
> Looks good to me.
> 
> Reviewed-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>

Thanks Srikar!


> 
> > ---
> >  arch/x86/kernel/uprobes.c |    4 ++--
> >  1 file changed, 2 insertions(+), 2 deletions(-)
> > 
> > diff --git a/arch/x86/kernel/uprobes.c b/arch/x86/kernel/uprobes.c
> > index 3fdaa042823d..bb3ea3705b99 100644
> > --- a/arch/x86/kernel/uprobes.c
> > +++ b/arch/x86/kernel/uprobes.c
> > @@ -257,7 +257,7 @@ static bool is_prefix_bad(struct insn *insn)
> >  {
> >  	int i;
> > 
> > -	for (i = 0; i < insn->prefixes.nbytes; i++) {
> > +	for (i = 0; insn->prefixes.bytes[i] && i < 4; i++) {
> >  		insn_attr_t attr;
> > 
> >  		attr = inat_get_opcode_attribute(insn->prefixes.bytes[i]);
> > @@ -746,7 +746,7 @@ static int branch_setup_xol_ops(struct arch_uprobe *auprobe, struct insn *insn)
> >  	 * Intel and AMD behavior differ in 64-bit mode: Intel ignores 66 prefix.
> >  	 * No one uses these insns, reject any branch insns with such prefix.
> >  	 */
> > -	for (i = 0; i < insn->prefixes.nbytes; i++) {
> > +	for (i = 0; insn->prefixes.bytes[i] && i < 4; i++) {
> >  		if (insn->prefixes.bytes[i] == 0x66)
> >  			return -ENOTSUPP;
> >  	}
> > 
> 
> -- 
> Thanks and Regards
> Srikar Dronamraju


-- 
Masami Hiramatsu <mhiramat@kernel.org>

Re: [PATCH 2/3] x86/uprobes: Fix not using prefixes.nbytes for loop over prefixes.bytes

[Kees Cook]

On Wed, Dec 02, 2020 at 05:51:16PM +0900, Masami Hiramatsu wrote:
> Since the insn.prefixes.nbytes can be bigger than the size of
> insn.prefixes.bytes[] when a same prefix is repeated, we have to
> check whether the insn.prefixes.bytes[i] != 0 and i < 4 instead
> of insn.prefixes.nbytes.
> 
> Fixes: 2b1444983508 ("uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints")
> Cc: stable@vger.kernel.org
> Reported-by: Kees Cook <keescook@chromium.org>

This should probably be:

Reported-by: syzbot+9b64b619f10f19d19a7c@syzkaller.appspotmail.com
Debugged-by: Kees Cook <keescook@chromium.org>

> Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org>
> ---
>  arch/x86/kernel/uprobes.c |    4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/arch/x86/kernel/uprobes.c b/arch/x86/kernel/uprobes.c
> index 3fdaa042823d..bb3ea3705b99 100644
> --- a/arch/x86/kernel/uprobes.c
> +++ b/arch/x86/kernel/uprobes.c
> @@ -257,7 +257,7 @@ static bool is_prefix_bad(struct insn *insn)
>  {
>  	int i;
>  
> -	for (i = 0; i < insn->prefixes.nbytes; i++) {
> +	for (i = 0; insn->prefixes.bytes[i] && i < 4; i++) {
>  		insn_attr_t attr;
>  
>  		attr = inat_get_opcode_attribute(insn->prefixes.bytes[i]);
> @@ -746,7 +746,7 @@ static int branch_setup_xol_ops(struct arch_uprobe *auprobe, struct insn *insn)
>  	 * Intel and AMD behavior differ in 64-bit mode: Intel ignores 66 prefix.
>  	 * No one uses these insns, reject any branch insns with such prefix.
>  	 */
> -	for (i = 0; i < insn->prefixes.nbytes; i++) {
> +	for (i = 0; insn->prefixes.bytes[i] && i < 4; i++) {
>  		if (insn->prefixes.bytes[i] == 0x66)
>  			return -ENOTSUPP;
>  	}
> 


-- 
Kees Cook

Re: [PATCH 2/3] x86/uprobes: Fix not using prefixes.nbytes for loop over prefixes.bytes

[Masami Hiramatsu]

On Wed, 2 Dec 2020 11:04:41 -0800
Kees Cook <keescook@chromium.org> wrote:

> On Wed, Dec 02, 2020 at 05:51:16PM +0900, Masami Hiramatsu wrote:
> > Since the insn.prefixes.nbytes can be bigger than the size of
> > insn.prefixes.bytes[] when a same prefix is repeated, we have to
> > check whether the insn.prefixes.bytes[i] != 0 and i < 4 instead
> > of insn.prefixes.nbytes.
> > 
> > Fixes: 2b1444983508 ("uprobes, mm, x86: Add the ability to install and remove uprobes breakpoints")
> > Cc: stable@vger.kernel.org
> > Reported-by: Kees Cook <keescook@chromium.org>
> 
> This should probably be:
> 
> Reported-by: syzbot+9b64b619f10f19d19a7c@syzkaller.appspotmail.com
> Debugged-by: Kees Cook <keescook@chromium.org>

OK, let me fix it.

Thank you,

> 
> > Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org>
> > ---
> >  arch/x86/kernel/uprobes.c |    4 ++--
> >  1 file changed, 2 insertions(+), 2 deletions(-)
> > 
> > diff --git a/arch/x86/kernel/uprobes.c b/arch/x86/kernel/uprobes.c
> > index 3fdaa042823d..bb3ea3705b99 100644
> > --- a/arch/x86/kernel/uprobes.c
> > +++ b/arch/x86/kernel/uprobes.c
> > @@ -257,7 +257,7 @@ static bool is_prefix_bad(struct insn *insn)
> >  {
> >  	int i;
> >  
> > -	for (i = 0; i < insn->prefixes.nbytes; i++) {
> > +	for (i = 0; insn->prefixes.bytes[i] && i < 4; i++) {
> >  		insn_attr_t attr;
> >  
> >  		attr = inat_get_opcode_attribute(insn->prefixes.bytes[i]);
> > @@ -746,7 +746,7 @@ static int branch_setup_xol_ops(struct arch_uprobe *auprobe, struct insn *insn)
> >  	 * Intel and AMD behavior differ in 64-bit mode: Intel ignores 66 prefix.
> >  	 * No one uses these insns, reject any branch insns with such prefix.
> >  	 */
> > -	for (i = 0; i < insn->prefixes.nbytes; i++) {
> > +	for (i = 0; insn->prefixes.bytes[i] && i < 4; i++) {
> >  		if (insn->prefixes.bytes[i] == 0x66)
> >  			return -ENOTSUPP;
> >  	}
> > 
> 
> 
> -- 
> Kees Cook


-- 
Masami Hiramatsu <mhiramat@kernel.org>

[PATCH 3/3] x86/insn-eval: Fix not using prefixes.nbytes for loop over prefixes.bytes

[Masami Hiramatsu]

Since the insn.prefixes.nbytes can be bigger than the size of
insn.prefixes.bytes[] when a same prefix is repeated, we have to
check whether the insn.prefixes.bytes[i] != 0 and i < 4 instead
of insn.prefixes.nbytes.

Fixes: 32d0b95300db ("x86/insn-eval: Add utility functions to get segment selector")
Cc: stable@vger.kernel.org
Reported-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org>
---
 arch/x86/lib/insn-eval.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/arch/x86/lib/insn-eval.c b/arch/x86/lib/insn-eval.c
index 58f7fb95c7f4..c52c91461f52 100644
--- a/arch/x86/lib/insn-eval.c
+++ b/arch/x86/lib/insn-eval.c
@@ -67,7 +67,7 @@ bool insn_has_rep_prefix(struct insn *insn)
 
 	insn_get_prefixes(insn);
 
-	for (i = 0; i < insn->prefixes.nbytes; i++) {
+	for (i = 0; insn->prefixes.bytes[i] && i < 4; i++) {
 		insn_byte_t p = insn->prefixes.bytes[i];
 
 		if (p == 0xf2 || p == 0xf3)
@@ -99,7 +99,7 @@ static int get_seg_reg_override_idx(struct insn *insn)
 	insn_get_prefixes(insn);
 
 	/* Look for any segment override prefixes. */
-	for (i = 0; i < insn->prefixes.nbytes; i++) {
+	for (i = 0; insn->prefixes.bytes[i] && i < 4; i++) {
 		insn_attr_t attr;
 
 		attr = inat_get_opcode_attribute(insn->prefixes.bytes[i]);