diff options
| author | Paulo Alcantara <pc@manguebit.org> | 2026-09-26 19:20:30 -0300 |
|---|---|---|
| committer | Paulo Alcantara <pc@manguebit.org> | 2026-09-30 14:24:12 -0300 |
| commit | dd05add7b2b6004c5fa3a1f276f56d8668f86b65 (patch) | |
| tree | 23efdcaa0b8031f6ac6f57b6734d18d6c1adc6f2 /Documentation/security/secrets/coco.rst | |
| download | linux-stable-dd05add7b2b6004c5fa3a1f276f56d8668f86b65.tar.gz linux-stable-dd05add7b2b6004c5fa3a1f276f56d8668f86b65.zip | |
netfs: zero the tail of a short DIO/unbuffered readgrafted
The buffered read collector zero-fills the tail of a short read that
stops below the inode's i_size (netfs_clear_unread()), so a read that
races an extending write still returns the expected number of bytes.
The non-buffered collector path does no such thing: it just records
how much was transferred.
Add the same zero-fill for the non-buffered case, gated on
NETFS_SREQ_CLEAR_TAIL: a subreq's source sets that flag when a short
result from it is known to be safe to treat as a hole, as opposed to
NETFS_SREQ_HIT_EOF, which means the read genuinely ran off the end of
the file and should be reported short as-is. Only CLEAR_TAIL should
zero-fill here; a real EOF must stay a real short read.
No source currently sets CLEAR_TAIL on an unbuffered/DIO subrequest,
so this is inert on its own -- a following change teaches cifs to set
it in the one case that needs it.
Fixes: e2d46f2ec332 ("netfs: Change the read result collector to only use one work item")
Reviewed-by: David Howells <dhowells@redhat.com>
Reviewed-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Paulo Alcantara <pc@manguebit.org>
Cc: Christian Brauner <brauner@kernel.org>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Ronnie Sahlberg <ronniesahlberg@gmail.com>
Cc: Shyam Prasad N <sprasad@microsoft.com>
Cc: Tom Talpey <tom@talpey.com>
Cc: Bharath SM <bharathsm@microsoft.com>
Cc: stable@vger.kernel.org
Diffstat (limited to 'Documentation/security/secrets/coco.rst')
| -rw-r--r-- | Documentation/security/secrets/coco.rst | 103 |
1 files changed, 103 insertions, 0 deletions
diff --git a/Documentation/security/secrets/coco.rst b/Documentation/security/secrets/coco.rst new file mode 100644 index 000000000..a1210db8e --- /dev/null +++ b/Documentation/security/secrets/coco.rst @@ -0,0 +1,103 @@ +.. SPDX-License-Identifier: GPL-2.0 + +============================== +Confidential Computing secrets +============================== + +This document describes how Confidential Computing secret injection is handled +from the firmware to the operating system, in the EFI driver and the efi_secret +kernel module. + + +Introduction +============ + +Confidential Computing (coco) hardware such as AMD SEV (Secure Encrypted +Virtualization) allows guest owners to inject secrets into the VMs +memory without the host/hypervisor being able to read them. In SEV, +secret injection is performed early in the VM launch process, before the +guest starts running. + +The efi_secret kernel module allows userspace applications to access these +secrets via securityfs. + + +Secret data flow +================ + +The guest firmware may reserve a designated memory area for secret injection, +and publish its location (base GPA and length) in the EFI configuration table +under a ``LINUX_EFI_COCO_SECRET_AREA_GUID`` entry +(``adf956ad-e98c-484c-ae11-b51c7d336447``). This memory area should be marked +by the firmware as ``EFI_RESERVED_TYPE``, and therefore the kernel should not +be use it for its own purposes. + +During the VM's launch, the virtual machine manager may inject a secret to that +area. In AMD SEV and SEV-ES this is performed using the +``KVM_SEV_LAUNCH_SECRET`` command (see [sev]_). The structure of the injected +Guest Owner secret data should be a GUIDed table of secret values; the binary +format is described in ``drivers/virt/coco/efi_secret/efi_secret.c`` under +"Structure of the EFI secret area". + +On kernel start, the kernel's EFI driver saves the location of the secret area +(taken from the EFI configuration table) in the ``efi.coco_secret`` field. +Later it checks if the secret area is populated: it maps the area and checks +whether its content begins with ``EFI_SECRET_TABLE_HEADER_GUID`` +(``1e74f542-71dd-4d66-963e-ef4287ff173b``). If the secret area is populated, +the EFI driver will autoload the efi_secret kernel module, which exposes the +secrets to userspace applications via securityfs. The details of the +efi_secret filesystem interface are in [secrets-coco-abi]_. + + +Application usage example +========================= + +Consider a guest performing computations on encrypted files. The Guest Owner +provides the decryption key (= secret) using the secret injection mechanism. +The guest application reads the secret from the efi_secret filesystem and +proceeds to decrypt the files into memory and then performs the needed +computations on the content. + +In this example, the host can't read the files from the disk image +because they are encrypted. Host can't read the decryption key because +it is passed using the secret injection mechanism (= secure channel). +Host can't read the decrypted content from memory because it's a +confidential (memory-encrypted) guest. + +Here is a simple example for usage of the efi_secret module in a guest +to which an EFI secret area with 4 secrets was injected during launch:: + + # ls -la /sys/kernel/security/secrets/coco + total 0 + drwxr-xr-x 2 root root 0 Jun 28 11:54 . + drwxr-xr-x 3 root root 0 Jun 28 11:54 .. + -r--r----- 1 root root 0 Jun 28 11:54 736870e5-84f0-4973-92ec-06879ce3da0b + -r--r----- 1 root root 0 Jun 28 11:54 83c83f7f-1356-4975-8b7e-d3a0b54312c6 + -r--r----- 1 root root 0 Jun 28 11:54 9553f55d-3da2-43ee-ab5d-ff17f78864d2 + -r--r----- 1 root root 0 Jun 28 11:54 e6f5a162-d67f-4750-a67c-5d065f2a9910 + + # hd /sys/kernel/security/secrets/coco/e6f5a162-d67f-4750-a67c-5d065f2a9910 + 00000000 74 68 65 73 65 2d 61 72 65 2d 74 68 65 2d 6b 61 |these-are-the-ka| + 00000010 74 61 2d 73 65 63 72 65 74 73 00 01 02 03 04 05 |ta-secrets......| + 00000020 06 07 |..| + 00000022 + + # rm /sys/kernel/security/secrets/coco/e6f5a162-d67f-4750-a67c-5d065f2a9910 + + # ls -la /sys/kernel/security/secrets/coco + total 0 + drwxr-xr-x 2 root root 0 Jun 28 11:55 . + drwxr-xr-x 3 root root 0 Jun 28 11:54 .. + -r--r----- 1 root root 0 Jun 28 11:54 736870e5-84f0-4973-92ec-06879ce3da0b + -r--r----- 1 root root 0 Jun 28 11:54 83c83f7f-1356-4975-8b7e-d3a0b54312c6 + -r--r----- 1 root root 0 Jun 28 11:54 9553f55d-3da2-43ee-ab5d-ff17f78864d2 + + +References +========== + +See [sev-api-spec]_ for more info regarding SEV ``LAUNCH_SECRET`` operation. + +.. [sev] Documentation/virt/kvm/x86/amd-memory-encryption.rst +.. [secrets-coco-abi] Documentation/ABI/testing/securityfs-secrets-coco +.. [sev-api-spec] https://www.amd.com/system/files/TechDocs/55766_SEV-KM_API_Specification.pdf |
