Skip to contentSkip to conversation

nvme-ioctl: move capable() admin check to the end

PATCH In review Keith Busch 5 participants 8 messages Nov 30, 2023 – Dec 4, 2023 NVMe

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 nvme-ioctl: move capable() admin check to the end

From: Keith Busch <[email protected]>

This can be an expensive call on some distros. Move it to the end after checking the cheaper ways to determine if the command is allowed.

---
 drivers/nvme/host/ioctl.c | 21 +++++++++++----------
 1 file changed, 11 insertions(+), 10 deletions(-)
drivers/nvme/host/ioctl.c +11 -10
index 32c9bcf491a33..4d4453743d62c 100644
@@ -18,15 +18,12 @@ static bool nvme_cmd_allowed(struct nvme_ns *ns, struct nvme_command *c,
 {
 	u32 effects;
 
-	if (capable(CAP_SYS_ADMIN))
-		return true;
-
 	/*
 	 * Do not allow unprivileged passthrough on partitions, as that allows an
 	 * escape from the containment of the partition.
 	 */
 	if (flags & NVME_IOCTL_PARTITION)
-		return false;
+		goto admin;
 
 	/*
 	 * Do not allow unprivileged processes to send vendor specific or fabrics
@@ -34,7 +31,7 @@ static bool nvme_cmd_allowed(struct nvme_ns *ns, struct nvme_command *c,
 	 */
 	if (c->common.opcode >= nvme_cmd_vendor_start ||
 	    c->common.opcode == nvme_fabrics_command)
-		return false;
+		goto admin;
 
 	/*
 	 * Do not allow unprivileged passthrough of admin commands except
@@ -53,7 +50,7 @@ static bool nvme_cmd_allowed(struct nvme_ns *ns, struct nvme_command *c,
 				return true;
 			}
 		}
-		return false;
+		goto admin;
 	}
 
 	/*
@@ -63,7 +60,7 @@ static bool nvme_cmd_allowed(struct nvme_ns *ns, struct nvme_command *c,
 	 */
 	effects = nvme_command_effects(ns->ctrl, ns, c->common.opcode);
 	if (!(effects & NVME_CMD_EFFECTS_CSUPP))
-		return false;
+		goto admin;
 
 	/*
 	 * Don't allow passthrough for command that have intrusive (or unknown)
@@ -72,16 +69,20 @@ static bool nvme_cmd_allowed(struct nvme_ns *ns, struct nvme_command *c,
 	if (effects & ~(NVME_CMD_EFFECTS_CSUPP | NVME_CMD_EFFECTS_LBCC |
 			NVME_CMD_EFFECTS_UUID_SEL |
 			NVME_CMD_EFFECTS_SCOPE_MASK))
-		return false;
+		goto admin;
 
 	/*
 	 * Only allow I/O commands that transfer data to the controller or that
 	 * change the logical block contents if the file descriptor is open for
 	 * writing.
 	 */
-	if (nvme_is_write(c) || (effects & NVME_CMD_EFFECTS_LBCC))
-		return open_for_write;
+	if ((nvme_is_write(c) || (effects & NVME_CMD_EFFECTS_LBCC)) &&
+	    !open_for_write)
+		goto admin;
+
 	return true;
+admin:
+	return capable(CAP_SYS_ADMIN);
 }
 
 /*
Signature · 1 line
2.34.1
maintainer Reviewed-by on v1 1/1
#1 lore raw
Quoted 4 lines · Keith Busch

From: Keith Busch <[email protected]>

This can be an expensive call on some distros. Move it to the end after checking the cheaper ways to determine if the command is allowed.

Signature · 1 line
Jens Axboe
maintainer
#4 lore raw
Quoted 6 lines · Christoph Hellwig
Quoted 3 lines · Keith Busch

From: Keith Busch <[email protected]>

This can be an expensive call on some distros.

Huh? How would a distro matter for kernel code?

Security monitoring: audit, LSM, and bpf hooks all make this function undesirable for a fast path check.

#5 lore raw
Quoted 10 lines · Keith Busch
Quoted 6 lines · Christoph Hellwig
Quoted 3 lines · Keith Busch

From: Keith Busch <[email protected]>

This can be an expensive call on some distros.

Huh? How would a distro matter for kernel code?

Security monitoring: audit, LSM, and bpf hooks all make this function undesirable for a fast path check.

So what matters is kernel configuration, not "distro".

maintainer
#6 lore raw
Quoted 5 lines · Christoph Hellwig
Quoted 4 lines · Keith Busch
Quoted 3 lines · Christoph Hellwig
Quoted 2 lines · Keith Busch

From: Keith Busch <[email protected]>

Quoted 8 lines
Quoted 6 lines
Quoted 3 lines
Quoted 1 line

This can be an expensive call on some distros.

Huh? How would a distro matter for kernel code?

Security monitoring: audit, LSM, and bpf hooks all make this function undesirable for a fast path check.

So what matters is kernel configuration, not "distro".

Yeah, fair enough: the overhead comes from the kernel config. So just update the change log and good to go?