Skip to contentSkip to conversation

x86/sgx: Remove unnecessary kmap() from sgx_ioc_enclave_init()

PATCH Applied [email protected] 3 participants 3 messages Mar 24, 2021 – Mar 26, 2021

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 x86/sgx: Remove unnecessary kmap() from sgx_ioc_enclave_init()

From: Ira Weiny <[email protected]>

kmap is inefficient and is being replaced by kmap_local_page(), if possible.

That said, there is no readily apparent reason why initp_page needs to be allocated and kmap'ed() except that 'sigstruct' needs to be page aligned and 'token' 512 byte aligned.

Rather than change this kmap() to kmap_local_page() use kmalloc() instead because kmalloc() can gives this alignment when allocating PAGE_SIZE bytes.

Remove the alloc_page()/kmap() and replace with kmalloc(PAGE_SIZE, ...) to get a page aligned kernel address.

In addition add a comment to document the alignment requirements so that others don't attempt to 'fix' this again.

Cc Sean Christopherson <[email protected]>
Cc Jethro Beekman <[email protected]>
Cc Jarkko Sakkinen <[email protected]>
Cc Dave Hansen <[email protected]>

---
Changes from v5[5]:
From Boris
Clean up commit msg and comment

Changes from v4[4]:
Add Ack and Reviews
Send to the correct maintainers

Changes from v3[3]:
Remove BUILD_BUG_ONs

Changes from v2[2]:
When allocating a power of 2 size kmalloc() now guarantees the alignment of the respective size. So go back to using kmalloc() but with a PAGE_SIZE allocation to get the alignment. This also follows the pattern in sgx_ioc_enclave_create()

Changes from v1[1]:
Use page_address() instead of kcmalloc() to ensure sigstruct is page aligned
Use BUILD_BUG_ON to ensure token and sigstruct don't collide.

[1] https://lore.kernel.org/lkml/[email protected]/
[2] https://lore.kernel.org/lkml/[email protected]/
[3] https://lore.kernel.org/lkml/[email protected]/#t
[4] https://lore.kernel.org/lkml/[email protected]/
[5] https://lore.kernel.org/lkml/[email protected]/

---
 arch/x86/kernel/cpu/sgx/ioctl.c | 14 ++++++++------
 1 file changed, 8 insertions(+), 6 deletions(-)
arch/x86/kernel/cpu/sgx/ioctl.c +8 -6
index 90a5caf76939..2e10367ea66c 100644
@@ -604,7 +604,6 @@ static long sgx_ioc_enclave_init(struct sgx_encl *encl, void __user *arg)
 {
 	struct sgx_sigstruct *sigstruct;
 	struct sgx_enclave_init init_arg;
-	struct page *initp_page;
 	void *token;
 	int ret;
 
@@ -615,11 +614,15 @@ static long sgx_ioc_enclave_init(struct sgx_encl *encl, void __user *arg)
 	if (copy_from_user(&init_arg, arg, sizeof(init_arg)))
 		return -EFAULT;
 
-	initp_page = alloc_page(GFP_KERNEL);
-	if (!initp_page)
+	/*
+	 * 'sigstruct' must be on a page boundary and 'token' on a 512 byte
+	 * boundary.  kmalloc() will give this alignment when allocating
+	 * PAGE_SIZE bytes.
+	 */
+	sigstruct = kmalloc(PAGE_SIZE, GFP_KERNEL);
+	if (!sigstruct)
 		return -ENOMEM;
 
-	sigstruct = kmap(initp_page);
 	token = (void *)((unsigned long)sigstruct + PAGE_SIZE / 2);
 	memset(token, 0, SGX_LAUNCH_TOKEN_SIZE);
 
@@ -645,8 +648,7 @@ static long sgx_ioc_enclave_init(struct sgx_encl *encl, void __user *arg)
 	ret = sgx_encl_init(encl, sigstruct, token);
 
 out:
-	kunmap(initp_page);
-	__free_page(initp_page);
+	kfree(sigstruct);
 	return ret;
 }
 
Signature · 1 line
2.28.0.rc0.12.gb6a658bd00c9
merge bot Applied
#1 lore raw
The following commit has been merged into the x86/sgx branch of tip:

The following commit has been merged into the x86/sgx branch of tip:

Commit-ID:     633b0616cfe085679471a4c0fae02e8c3a1a9866
Gitweb:        https://git.kernel.org/tip/633b0616cfe085679471a4c0fae02e8c3a1a9866
Author:        Ira Weiny <[email protected]>
AuthorDate:    Wed, 24 Mar 2021 11:22:46 -07:00
Committer:     Borislav Petkov <[email protected]>
CommitterDate: Thu, 25 Mar 2021 09:50:32 +01:00

x86/sgx: Remove unnecessary kmap() from sgx_ioc_enclave_init()

kmap() is inefficient and is being replaced by kmap_local_page(), if possible. There is no readily apparent reason why initp_page needs to be allocated and kmap'ed() except that 'sigstruct' needs to be page-aligned and 'token' 512 byte-aligned.

Rather than change it to kmap_local_page(), use kmalloc() instead because kmalloc() can give this alignment when allocating PAGE_SIZE bytes.

Remove the alloc_page()/kmap() and replace with kmalloc(PAGE_SIZE, ...) to get a page aligned kernel address.

In addition, add a comment to document the alignment requirements so that others don't attempt to 'fix' this again.

[ bp: Massage commit message. ]

---
 arch/x86/kernel/cpu/sgx/ioctl.c | 14 ++++++++------
 1 file changed, 8 insertions(+), 6 deletions(-)
arch/x86/kernel/cpu/sgx/ioctl.c +8 -6
index 90a5caf..2e10367 100644
@@ -604,7 +604,6 @@ static long sgx_ioc_enclave_init(struct sgx_encl *encl, void __user *arg)
 {
 	struct sgx_sigstruct *sigstruct;
 	struct sgx_enclave_init init_arg;
-	struct page *initp_page;
 	void *token;
 	int ret;
 
@@ -615,11 +614,15 @@ static long sgx_ioc_enclave_init(struct sgx_encl *encl, void __user *arg)
 	if (copy_from_user(&init_arg, arg, sizeof(init_arg)))
 		return -EFAULT;
 
-	initp_page = alloc_page(GFP_KERNEL);
-	if (!initp_page)
+	/*
+	 * 'sigstruct' must be on a page boundary and 'token' on a 512 byte
+	 * boundary.  kmalloc() will give this alignment when allocating
+	 * PAGE_SIZE bytes.
+	 */
+	sigstruct = kmalloc(PAGE_SIZE, GFP_KERNEL);
+	if (!sigstruct)
 		return -ENOMEM;
 
-	sigstruct = kmap(initp_page);
 	token = (void *)((unsigned long)sigstruct + PAGE_SIZE / 2);
 	memset(token, 0, SGX_LAUNCH_TOKEN_SIZE);
 
@@ -645,8 +648,7 @@ static long sgx_ioc_enclave_init(struct sgx_encl *encl, void __user *arg)
 	ret = sgx_encl_init(encl, sigstruct, token);
 
 out:
-	kunmap(initp_page);
-	__free_page(initp_page);
+	kfree(sigstruct);
 	return ret;
 }
 
maintainer
#2 lore raw
In reply to [email protected]
Quoted 4 lines · On Wed, Mar 24, 2021 at 11:22:46AM -0700, [email protected] wrote:

From: Ira Weiny <[email protected]>

kmap is inefficient and is being replaced by kmap_local_page(), if possible.

Nit: just saying that something is inefficient is not enough. Always should have something to back it up.

Quoted 8 lines

That said, there is no readily apparent reason why initp_page needs to be allocated and kmap'ed() except that 'sigstruct' needs to be page aligned and 'token' 512 byte aligned.

Rather than change this kmap() to kmap_local_page() use kmalloc() instead because kmalloc() can gives this alignment when allocating PAGE_SIZE bytes.

Quoted 5 lines

Remove the alloc_page()/kmap() and replace with kmalloc(PAGE_SIZE, ...) to get a page aligned kernel address.

In addition add a comment to document the alignment requirements so that others don't attempt to 'fix' this again.

I'm a bit confused based on this commit message. Why is kmap_local_page() better, and why kmalloc() is the ultimate choice of all three options?

Quoted 6 lines
Cc Sean Christopherson <[email protected]>
Cc Jethro Beekman <[email protected]>
Cc Jarkko Sakkinen <[email protected]>
Cc Dave Hansen <[email protected]>

/Jarkko