Skip to contentSkip to conversation

x86/sgx: Simplify struct sgx_enclave_restrict_permissions

RFC PATCH In review Jarkko Sakkinen 2 participants 4 messages Apr 5, 2022

Subsystems (2)
lore
Share & tools
lore

Jarkko Sakkinen

author maintainer
#0 lore raw

RFC PATCH 1/1 x86/sgx: Simplify struct sgx_enclave_restrict_permissions

The reasoning to change SECINFO to simply flags is stated in this inline comment:

/*
* Return valid permission fields from a secinfo structure provided by
* user space. The secinfo structure is required to only have bits in
* the permission fields set.
*/

It is better to simply change the parameter type than require to use a malformed version of a data structure.

---
Only compile-tested.

 arch/x86/include/uapi/asm/sgx.h |  5 ++-
 arch/x86/kernel/cpu/sgx/ioctl.c | 57 ++++++---------------------------
 2 files changed, 12 insertions(+), 50 deletions(-)
arch/x86/include/uapi/asm/sgx.h +2 -3
index feda7f85b2ce..627136798f2a 100644
@@ -88,15 +88,14 @@ struct sgx_enclave_provision {
  * @offset:	starting page offset (page aligned relative to enclave base
  *		address defined in SECS)
  * @length:	length of memory (multiple of the page size)
- * @secinfo:	address for the SECINFO data containing the new permission bits
- *		for pages in range described by @offset and @length
+ * @flags:	flags field of the SECINFO data
  * @result:	(output) SGX result code of ENCLS[EMODPR] function
  * @count:	(output) bytes successfully changed (multiple of page size)
  */
 struct sgx_enclave_restrict_permissions {
 	__u64 offset;
 	__u64 length;
-	__u64 secinfo;
+	__u64 flags;
 	__u64 result;
 	__u64 count;
 };
arch/x86/kernel/cpu/sgx/ioctl.c +10 -47

95 lines not on this page · raw message

Signature · 1 line
2.35.1
#1 lore raw

Hi Jarkko,

Quoted 11 lines · Jarkko Sakkinen

The reasoning to change SECINFO to simply flags is stated in this inline comment:

/*
* Return valid permission fields from a secinfo structure provided by
* user space. The secinfo structure is required to only have bits in
* the permission fields set.
*/

It is better to simply change the parameter type than require to use a malformed version of a data structure.

Could you please elaborate what is malformed?

Quoted 21 lines

---
Only compile-tested.

 arch/x86/include/uapi/asm/sgx.h |  5 ++-
 arch/x86/kernel/cpu/sgx/ioctl.c | 57 ++++++---------------------------
 2 files changed, 12 insertions(+), 50 deletions(-)
arch/x86/include/uapi/asm/sgx.h +1 -2
index feda7f85b2ce..627136798f2a 100644
@@ -88,15 +88,14 @@ struct sgx_enclave_provision {
  * @offset:	starting page offset (page aligned relative to enclave base
  *		address defined in SECS)
  * @length:	length of memory (multiple of the page size)
- * @secinfo:	address for the SECINFO data containing the new permission bits
- *		for pages in range described by @offset and @length
+ * @flags:	flags field of the SECINFO data
  * @result:	(output) SGX result code of ENCLS[EMODPR] function

arch/x86/include/uapi/asm/sgx.h struct sgx_enclave_provision { · line 101

Quoted 10 lines
arch/x86/include/uapi/asm/sgx.h +1 -1
  * @count:	(output) bytes successfully changed (multiple of page size)
  */
 struct sgx_enclave_restrict_permissions {
 	__u64 offset;
 	__u64 length;
-	__u64 secinfo;
+	__u64 flags;
 	__u64 result;
 	__u64 count;
 };

What is the motivation for using the flags field of the SECINFO data? If the goal is to only provide necessary information, why not just provide the permission bits since none of the other bits are used? If the goal is to make room for future SECINFO changes/demands, why not include the reserved field of SECINFO where future changes are most likely to reside?

Reinette

author maintainer
#2 lore raw
Quoted 12 lines · Reinette Chatre

Hi Jarkko,

Quoted 9 lines · Jarkko Sakkinen

The reasoning to change SECINFO to simply flags is stated in this inline comment:

/*
 * Return valid permission fields from a secinfo structure provided by
 * user space. The secinfo structure is required to only have bits in
 * the permission fields set.
 */

Quoted 4 lines
Quoted 2 lines

It is better to simply change the parameter type than require to use a malformed version of a data structure.

Could you please elaborate what is malformed?

The structure that is accepted by the API. According to SDM permission changes are done with a structure where PT_REG is set, which gives -EINVAL. I categorize it as a bug.

BR, Jarkko

#3 lore raw

Hi Jarkko,

Quoted 13 lines · Jarkko Sakkinen
Quoted 12 lines · Reinette Chatre

Hi Jarkko,

Quoted 9 lines · Jarkko Sakkinen

The reasoning to change SECINFO to simply flags is stated in this inline comment:

/*
 * Return valid permission fields from a secinfo structure provided by
 * user space. The secinfo structure is required to only have bits in
 * the permission fields set.
 */

Quoted 8 lines
Quoted 4 lines
Quoted 2 lines

It is better to simply change the parameter type than require to use a malformed version of a data structure.

Could you please elaborate what is malformed?

The structure that is accepted by the API. According to SDM permission changes are done with a structure where PT_REG is set, which gives -EINVAL. I categorize it as a bug.

I assume that you are referring to this line from the SDM:

IF (EPCM(DS:RCX).PT is not PT_REG)

    THEN #PF(DS:RCX); FI;

Please note that the above tests the PT bit of the EPCM entry, not the PT field in the provided SECINFO.

Reinette