From: Eric Biggers <ebiggers@kernel.org> Date: 2021-01-15 18:20:50
[This patchset applies to v5.11-rc3]
Add an ioctl FS_IOC_READ_VERITY_METADATA which allows reading verity
metadata from a file that has fs-verity enabled, including:
- The Merkle tree
- The fsverity_descriptor (not including the signature if present)
- The built-in signature, if present
This ioctl has similar semantics to pread(). It is passed the type of
metadata to read (one of the above three), and a buffer, offset, and
size. It returns the number of bytes read or an error.
This ioctl doesn't make any assumption about where the metadata is
stored on-disk. It does assume the metadata is in a stable format, but
that's basically already the case:
- The Merkle tree and fsverity_descriptor are defined by how fs-verity
file digests are computed; see the "File digest computation" section
of Documentation/filesystems/fsverity.rst. Technically, the way in
which the levels of the tree are ordered relative to each other wasn't
previously specified, but it's logical to put the root level first.
- The built-in signature is the value passed to FS_IOC_ENABLE_VERITY.
This ioctl is useful because it allows writing a server program that
takes a verity file and serves it to a client program, such that the
client can do its own fs-verity compatible verification of the file.
This only makes sense if the client doesn't trust the server and if the
server needs to provide the storage for the client.
More concretely, there is interest in using this ability in Android to
export APK files (which are protected by fs-verity) to "protected VMs".
This would use Protected KVM (https://lwn.net/Articles/836693), which
provides an isolated execution environment without having to trust the
traditional "host". A "guest" VM can boot from a signed image and
perform specific tasks in a minimum trusted environment using files that
have fs-verity enabled on the host, without trusting the host or
requiring that the guest has its own trusted storage.
Technically, it would be possible to duplicate the metadata and store it
in separate files for serving. However, that would be less efficient
and would require extra care in userspace to maintain file consistency.
In addition to the above, the ability to read the built-in signatures is
useful because it allows a system that is using the in-kernel signature
verification to migrate to userspace signature verification.
This patchset has been tested by new xfstests which call this new ioctl
via a new subcommand for the 'fsverity' program from fsverity-utils.
Eric Biggers (6):
fs-verity: factor out fsverity_get_descriptor()
fs-verity: don't pass whole descriptor to fsverity_verify_signature()
fs-verity: add FS_IOC_READ_VERITY_METADATA ioctl
fs-verity: support reading Merkle tree with ioctl
fs-verity: support reading descriptor with ioctl
fs-verity: support reading signature with ioctl
Documentation/filesystems/fsverity.rst | 76 ++++++++++
fs/ext4/ioctl.c | 7 +
fs/f2fs/file.c | 11 ++
fs/verity/Makefile | 1 +
fs/verity/fsverity_private.h | 13 +-
fs/verity/open.c | 133 +++++++++++------
fs/verity/read_metadata.c | 195 +++++++++++++++++++++++++
fs/verity/signature.c | 20 +--
include/linux/fsverity.h | 12 ++
include/uapi/linux/fsverity.h | 14 ++
10 files changed, 417 insertions(+), 65 deletions(-)
create mode 100644 fs/verity/read_metadata.c
base-commit: 7c53f6b671f4aba70ff15e1b05148b10d58c2837
--
2.30.0
From: Eric Biggers <ebiggers@kernel.org> Date: 2021-01-15 18:20:24
From: Eric Biggers <redacted>
The FS_IOC_READ_VERITY_METADATA ioctl will need to return the fs-verity
descriptor (and signature) to userspace.
There are a few ways we could implement this:
- Save a copy of the descriptor (and signature) in the fsverity_info
struct that hangs off of the in-memory inode. However, this would
waste memory since most of the time it wouldn't be needed.
- Regenerate the descriptor from the merkle_tree_params in the
fsverity_info. However, this wouldn't work for the signature, nor for
the salt which the merkle_tree_params only contains indirectly as part
of the 'hashstate'. It would also be error-prone.
- Just get them from the filesystem again. The disadvantage is that in
general we can't trust that they haven't been maliciously changed
since the file has opened. However, the use cases for
FS_IOC_READ_VERITY_METADATA don't require that it verifies the chain
of trust. So this is okay as long as we do some basic validation.
In preparation for implementing the third option, factor out a helper
function fsverity_get_descriptor() which gets the descriptor (and
appended signature) from the filesystem and does some basic validation.
As part of this, start checking the sig_size field for overflow.
Currently fsverity_verify_signature() does this. But the new ioctl will
need this too, so do it earlier.
Signed-off-by: Eric Biggers <redacted>
---
fs/verity/fsverity_private.h | 7 +-
fs/verity/open.c | 130 +++++++++++++++++++++++------------
2 files changed, 91 insertions(+), 46 deletions(-)
@@ -142,45 +142,17 @@ static int compute_file_digest(struct fsverity_hash_alg *hash_alg,}/*-*Validatethegivenfsverity_descriptorandcreateanewfsverity_infofrom-*it.Thesignature(ifpresent)isalsochecked.+*Createanewfsverity_infofromthegivenfsverity_descriptor(withoptional+*appendedsignature),andcheckthesignatureifpresent.The+*fsverity_descriptormusthavealreadyundergonebasicvalidation.*/structfsverity_info*fsverity_create_info(conststructinode*inode,-void*_desc,size_tdesc_size)+structfsverity_descriptor*desc,+size_tdesc_size){-structfsverity_descriptor*desc=_desc;structfsverity_info*vi;interr;-if(desc_size<sizeof(*desc)){-fsverity_err(inode,"Unrecognized descriptor size: %zu bytes",-desc_size);-returnERR_PTR(-EINVAL);-}--if(desc->version!=1){-fsverity_err(inode,"Unrecognized descriptor version: %u",-desc->version);-returnERR_PTR(-EINVAL);-}--if(memchr_inv(desc->__reserved,0,sizeof(desc->__reserved))){-fsverity_err(inode,"Reserved bits set in descriptor");-returnERR_PTR(-EINVAL);-}--if(desc->salt_size>sizeof(desc->salt)){-fsverity_err(inode,"Invalid salt_size: %u",desc->salt_size);-returnERR_PTR(-EINVAL);-}--if(le64_to_cpu(desc->data_size)!=inode->i_size){-fsverity_err(inode,-"Wrong data_size: %llu (desc) != %lld (inode)",-le64_to_cpu(desc->data_size),inode->i_size);-returnERR_PTR(-EINVAL);-}-vi=kmem_cache_zalloc(fsverity_info_cachep,GFP_KERNEL);if(!vi)returnERR_PTR(-ENOMEM);
@@ -245,15 +217,57 @@ void fsverity_free_info(struct fsverity_info *vi)kmem_cache_free(fsverity_info_cachep,vi);}-/* Ensure the inode has an ->i_verity_info */-staticintensure_verity_info(structinode*inode)+staticboolvalidate_fsverity_descriptor(structinode*inode,+conststructfsverity_descriptor*desc,+size_tdesc_size){-structfsverity_info*vi=fsverity_get_info(inode);-structfsverity_descriptor*desc;-intres;+if(desc_size<sizeof(*desc)){+fsverity_err(inode,"Unrecognized descriptor size: %zu bytes",+desc_size);+returnfalse;+}-if(vi)-return0;+if(desc->version!=1){+fsverity_err(inode,"Unrecognized descriptor version: %u",+desc->version);+returnfalse;+}++if(memchr_inv(desc->__reserved,0,sizeof(desc->__reserved))){+fsverity_err(inode,"Reserved bits set in descriptor");+returnfalse;+}++if(desc->salt_size>sizeof(desc->salt)){+fsverity_err(inode,"Invalid salt_size: %u",desc->salt_size);+returnfalse;+}++if(le64_to_cpu(desc->data_size)!=inode->i_size){+fsverity_err(inode,+"Wrong data_size: %llu (desc) != %lld (inode)",+le64_to_cpu(desc->data_size),inode->i_size);+returnfalse;+}++if(le32_to_cpu(desc->sig_size)>desc_size-sizeof(*desc)){+fsverity_err(inode,"Signature overflows verity descriptor");+returnfalse;+}++returntrue;+}++/*+*Readtheinode'sfsverity_descriptor(withoptionalappendedsignature)from+*thefilesystem,anddobasicvalidationofit.+*/+intfsverity_get_descriptor(structinode*inode,+structfsverity_descriptor**desc_ret,+size_t*desc_size_ret)+{+intres;+structfsverity_descriptor*desc;res=inode->i_sb->s_vop->get_verity_descriptor(inode,NULL,0);if(res<0){
@@ -272,20 +286,46 @@ static int ensure_verity_info(struct inode *inode)res=inode->i_sb->s_vop->get_verity_descriptor(inode,desc,res);if(res<0){fsverity_err(inode,"Error %d reading verity descriptor",res);-gotoout_free_desc;+kfree(desc);+returnres;+}++if(!validate_fsverity_descriptor(inode,desc,res)){+kfree(desc);+return-EINVAL;}-vi=fsverity_create_info(inode,desc,res);+*desc_ret=desc;+*desc_size_ret=res;+return0;+}++/* Ensure the inode has an ->i_verity_info */+staticintensure_verity_info(structinode*inode)+{+structfsverity_info*vi=fsverity_get_info(inode);+structfsverity_descriptor*desc;+size_tdesc_size;+interr;++if(vi)+return0;++err=fsverity_get_descriptor(inode,&desc,&desc_size);+if(err)+returnerr;++vi=fsverity_create_info(inode,desc,desc_size);if(IS_ERR(vi)){-res=PTR_ERR(vi);+err=PTR_ERR(vi);gotoout_free_desc;}fsverity_set_info(inode,vi);-res=0;+err=0;out_free_desc:kfree(desc);-returnres;+returnerr;}/**
From: Eric Biggers <ebiggers@kernel.org> Date: 2021-01-15 18:20:24
From: Eric Biggers <redacted>
Add an ioctl FS_IOC_READ_VERITY_METADATA which will allow reading verity
metadata from a file that has fs-verity enabled, including:
- The Merkle tree
- The fsverity_descriptor (not including the signature if present)
- The built-in signature, if present
This ioctl has similar semantics to pread(). It is passed the type of
metadata to read (one of the above three), and a buffer, offset, and
size. It returns the number of bytes read or an error.
Separate patches will add support for each of the above metadata types.
This patch just adds the ioctl itself.
This ioctl doesn't make any assumption about where the metadata is
stored on-disk. It does assume the metadata is in a stable format, but
that's basically already the case:
- The Merkle tree and fsverity_descriptor are defined by how fs-verity
file digests are computed; see the "File digest computation" section
of Documentation/filesystems/fsverity.rst. Technically, the way in
which the levels of the tree are ordered relative to each other wasn't
previously specified, but it's logical to put the root level first.
- The built-in signature is the value passed to FS_IOC_ENABLE_VERITY.
This ioctl is useful because it allows writing a server program that
takes a verity file and serves it to a client program, such that the
client can do its own fs-verity compatible verification of the file.
This only makes sense if the client doesn't trust the server and if the
server needs to provide the storage for the client.
More concretely, there is interest in using this ability in Android to
export APK files (which are protected by fs-verity) to "protected VMs".
This would use Protected KVM (https://lwn.net/Articles/836693), which
provides an isolated execution environment without having to trust the
traditional "host". A "guest" VM can boot from a signed image and
perform specific tasks in a minimum trusted environment using files that
have fs-verity enabled on the host, without trusting the host or
requiring that the guest has its own trusted storage.
Technically, it would be possible to duplicate the metadata and store it
in separate files for serving. However, that would be less efficient
and would require extra care in userspace to maintain file consistency.
In addition to the above, the ability to read the built-in signatures is
useful because it allows a system that is using the in-kernel signature
verification to migrate to userspace signature verification.
Signed-off-by: Eric Biggers <redacted>
---
Documentation/filesystems/fsverity.rst | 57 ++++++++++++++++++++++++++
fs/ext4/ioctl.c | 7 ++++
fs/f2fs/file.c | 11 +++++
fs/verity/Makefile | 1 +
fs/verity/read_metadata.c | 55 +++++++++++++++++++++++++
include/linux/fsverity.h | 12 ++++++
include/uapi/linux/fsverity.h | 10 +++++
7 files changed, 153 insertions(+)
create mode 100644 fs/verity/read_metadata.c
@@ -217,6 +217,63 @@ FS_IOC_MEASURE_VERITY can fail with the following errors:-``EOVERFLOW``: the digest is longer than the specified``digest_size`` bytes. Try providing a larger buffer.+FS_IOC_READ_VERITY_METADATA+---------------------------++The FS_IOC_READ_VERITY_METADATA ioctl reads verity metadata from a+verity file. This ioctl is available since Linux v5.12.++This ioctl allows writing a server program that takes a verity file+and serves it to a client program, such that the client can do its own+fs-verity compatible verification of the file. This only makes sense+if the client doesn't trust the server and if the server needs to+provide the storage for the client.++This is a fairly specialized use case, and most fs-verity users won't+need this ioctl.++This ioctl takes in a pointer to the following structure::++ struct fsverity_read_metadata_arg {+ __u64 metadata_type;+ __u64 offset;+ __u64 length;+ __u64 buf_ptr;+ __u64 __reserved;+ };++``metadata_type`` specifies the type of metadata to read.++The semantics are similar to those of ``pread()``. ``offset``+specifies the offset in bytes into the metadata item to read from, and+``length`` specifies the maximum number of bytes to read from the+metadata item. ``buf_ptr`` is the pointer to the buffer to read into,+cast to a 64-bit integer. ``__reserved`` must be 0. On success, the+number of bytes read is returned. 0 is returned at the end of the+metadata item. The returned length may be less than ``length``, for+example if the ioctl is interrupted.++The metadata returned by FS_IOC_READ_VERITY_METADATA isn't guaranteed+to be authenticated against the file digest that would be returned by+`FS_IOC_MEASURE_VERITY`_, as the metadata is expected to be used to+implement fs-verity compatible verification anyway (though absent a+malicious disk, the metadata will indeed match). E.g. to implement+this ioctl, the filesystem is allowed to just read the Merkle tree+blocks from disk without actually verifying the path to the root node.++FS_IOC_READ_VERITY_METADATA can fail with the following errors:++-``EFAULT``: the caller provided inaccessible memory+-``EINTR``: the ioctl was interrupted before any data was read+-``EINVAL``: reserved fields were set, or ``offset + length``+ overflowed+-``ENODATA``: the file is not a verity file+-``ENOTTY``: this type of filesystem does not implement fs-verity, or+ this ioctl is not yet implemented on it+-``EOPNOTSUPP``: the kernel was not configured with fs-verity+ support, or the filesystem superblock has not had the 'verity'+ feature enabled on it. (See `Filesystem support`_.)+ FS_IOC_GETFLAGS ---------------
@@ -1306,6 +1306,12 @@ static long __ext4_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)return-EOPNOTSUPP;returnfsverity_ioctl_measure(filp,(void__user*)arg);+caseFS_IOC_READ_VERITY_METADATA:+if(!ext4_has_feature_verity(sb))+return-EOPNOTSUPP;+returnfsverity_ioctl_read_metadata(filp,+(constvoid__user*)arg);+default:return-ENOTTY;}
@@ -1388,6 +1394,7 @@ long ext4_compat_ioctl(struct file *file, unsigned int cmd, unsigned long arg)caseFS_IOC_GETFSMAP:caseFS_IOC_ENABLE_VERITY:caseFS_IOC_MEASURE_VERITY:+caseFS_IOC_READ_VERITY_METADATA:caseEXT4_IOC_CLEAR_ES_CACHE:caseEXT4_IOC_GETSTATE:caseEXT4_IOC_GET_ES_CACHE:
@@ -3357,6 +3357,14 @@ static int f2fs_ioc_measure_verity(struct file *filp, unsigned long arg)returnfsverity_ioctl_measure(filp,(void__user*)arg);}+staticintf2fs_ioc_read_verity_metadata(structfile*filp,unsignedlongarg)+{+if(!f2fs_sb_has_verity(F2FS_I_SB(file_inode(filp))))+return-EOPNOTSUPP;++returnfsverity_ioctl_read_metadata(filp,(constvoid__user*)arg);+}+staticintf2fs_ioc_getfslabel(structfile*filp,unsignedlongarg){structinode*inode=file_inode(filp);
@@ -4272,6 +4280,8 @@ static long __f2fs_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)returnf2fs_ioc_enable_verity(filp,arg);caseFS_IOC_MEASURE_VERITY:returnf2fs_ioc_measure_verity(filp,arg);+caseFS_IOC_READ_VERITY_METADATA:+returnf2fs_ioc_read_verity_metadata(filp,arg);caseFS_IOC_GETFSLABEL:returnf2fs_ioc_getfslabel(filp,arg);caseFS_IOC_SETFSLABEL:
@@ -4523,6 +4533,7 @@ long f2fs_compat_ioctl(struct file *file, unsigned int cmd, unsigned long arg)caseF2FS_IOC_RESIZE_FS:caseFS_IOC_ENABLE_VERITY:caseFS_IOC_MEASURE_VERITY:+caseFS_IOC_READ_VERITY_METADATA:caseFS_IOC_GETFSLABEL:caseFS_IOC_SETFSLABEL:caseF2FS_IOC_GET_COMPRESS_BLOCKS:
@@ -0,0 +1,55 @@+// SPDX-License-Identifier: GPL-2.0-only+/*+*Ioctltoreadveritymetadata+*+*Copyright2021GoogleLLC+*/++#include"fsverity_private.h"++#include<linux/uaccess.h>++/**+*fsverity_ioctl_read_metadata()-readveritymetadatafromafile+*@filp:filetoreadthemetadatafrom+*@uarg:userpointertofsverity_read_metadata_arg+*+*Return:lengthreadonsuccess,0onEOF,-errnoonfailure+*/+intfsverity_ioctl_read_metadata(structfile*filp,constvoid__user*uarg)+{+structinode*inode=file_inode(filp);+conststructfsverity_info*vi;+structfsverity_read_metadata_argarg;+intlength;+void__user*buf;++vi=fsverity_get_info(inode);+if(!vi)+return-ENODATA;/* not a verity file */+/*+*Notethatwedon'thavetoexplicitlycheckthatthefileisopenfor+*reading,sinceverityfilescanonlybeopenedforreading.+*/++if(copy_from_user(&arg,uarg,sizeof(arg)))+return-EFAULT;++if(arg.__reserved)+return-EINVAL;++/* offset + length must not overflow. */+if(arg.offset+arg.length<arg.offset)+return-EINVAL;++/* Ensure that the return value will fit in INT_MAX. */+length=min_t(u64,arg.length,INT_MAX);++buf=u64_to_user_ptr(arg.buf_ptr);++switch(arg.metadata_type){+default:+return-EINVAL;+}+}+EXPORT_SYMBOL_GPL(fsverity_ioctl_read_metadata);
From: Eric Biggers <ebiggers@kernel.org> Date: 2021-01-15 18:20:25
From: Eric Biggers <redacted>
Now that fsverity_get_descriptor() validates the sig_size field,
fsverity_verify_signature() doesn't need to do it.
Just change the prototype of fsverity_verify_signature() to take the
signature directly rather than take a fsverity_descriptor.
Signed-off-by: Eric Biggers <redacted>
---
fs/verity/fsverity_private.h | 6 ++----
fs/verity/open.c | 3 ++-
fs/verity/signature.c | 20 ++++++--------------
3 files changed, 10 insertions(+), 19 deletions(-)
From: Eric Biggers <ebiggers@kernel.org> Date: 2021-01-15 18:21:33
From: Eric Biggers <redacted>
Add support for FS_VERITY_METADATA_TYPE_DESCRIPTOR to
FS_IOC_READ_VERITY_METADATA. This allows a userspace server program to
retrieve the fs-verity descriptor of a file for serving to a client
which implements fs-verity compatible verification. See the patch which
introduced FS_IOC_READ_VERITY_METADATA for more details.
"fs-verity descriptor" here means only the part that userspace cares
about because it is hashed to produce the file digest. It doesn't
include the signature which ext4 and f2fs append to the
fsverity_descriptor struct when storing it on-disk, since that way of
storing the signature is an implementation detail. The next patch adds
a separate metadata_type value for retrieving the signature separately.
This has been tested using a new xfstest which calls this ioctl via a
new subcommand for the 'fsverity' program from fsverity-utils.
Signed-off-by: Eric Biggers <redacted>
---
Documentation/filesystems/fsverity.rst | 4 +++
fs/verity/read_metadata.c | 40 ++++++++++++++++++++++++++
include/uapi/linux/fsverity.h | 1 +
3 files changed, 45 insertions(+)
@@ -235,6 +235,7 @@ need this ioctl. This ioctl takes in a pointer to the following structure:: #define FS_VERITY_METADATA_TYPE_MERKLE_TREE 1+ #define FS_VERITY_METADATA_TYPE_DESCRIPTOR 2 struct fsverity_read_metadata_arg { __u64 metadata_type;
@@ -252,6 +253,9 @@ This ioctl takes in a pointer to the following structure:: the same order that their hashes are themselves hashed. See `Merkle tree`_ for more information.+-``FS_VERITY_METADATA_TYPE_DESCRIPTOR`` reads the fs-verity+ descriptor. See `fs-verity descriptor`_.+ The semantics are similar to those of ``pread()``. ``offset`` specifies the offset in bytes into the metadata item to read from, and``length`` specifies the maximum number of bytes to read from the
@@ -76,6 +76,44 @@ static int fsverity_read_merkle_tree(struct inode *inode,}returnretval?retval:err;}++/* Copy the requested portion of the buffer to userspace. */+staticintfsverity_read_buffer(void__user*dst,u64offset,intlength,+constvoid*src,size_tsrc_length)+{+if(offset>=src_length)+return0;+src+=offset;+src_length-=offset;++length=min_t(size_t,length,src_length);++if(copy_to_user(dst,src,length))+return-EFAULT;++returnlength;+}++staticintfsverity_read_descriptor(structinode*inode,+void__user*buf,u64offset,intlength)+{+structfsverity_descriptor*desc;+size_tdesc_size;+intres;++res=fsverity_get_descriptor(inode,&desc,&desc_size);+if(res)+returnres;++/* don't include the signature */+desc_size=offsetof(structfsverity_descriptor,signature);+desc->sig_size=0;++res=fsverity_read_buffer(buf,offset,length,desc,desc_size);++kfree(desc);+returnres;+}/***fsverity_ioctl_read_metadata()-readveritymetadatafromafile*@filp:filetoreadthemetadatafrom
From: Eric Biggers <ebiggers@kernel.org> Date: 2021-01-15 18:21:33
From: Eric Biggers <redacted>
Add support for FS_VERITY_METADATA_TYPE_SIGNATURE to
FS_IOC_READ_VERITY_METADATA. This allows a userspace server program to
retrieve the built-in signature (if present) of a verity file for
serving to a client which implements fs-verity compatible verification.
See the patch which introduced FS_IOC_READ_VERITY_METADATA for more
details.
The ability for userspace to read the built-in signatures is also useful
because it allows a system that is using the in-kernel signature
verification to migrate to userspace signature verification.
This has been tested using a new xfstest which calls this ioctl via a
new subcommand for the 'fsverity' program from fsverity-utils.
Signed-off-by: Eric Biggers <redacted>
---
Documentation/filesystems/fsverity.rst | 9 +++++++-
fs/verity/read_metadata.c | 30 ++++++++++++++++++++++++++
include/uapi/linux/fsverity.h | 1 +
3 files changed, 39 insertions(+), 1 deletion(-)
@@ -236,6 +236,7 @@ This ioctl takes in a pointer to the following structure:: #define FS_VERITY_METADATA_TYPE_MERKLE_TREE 1 #define FS_VERITY_METADATA_TYPE_DESCRIPTOR 2+ #define FS_VERITY_METADATA_TYPE_SIGNATURE 3 struct fsverity_read_metadata_arg { __u64 metadata_type;
@@ -256,6 +257,10 @@ This ioctl takes in a pointer to the following structure::-``FS_VERITY_METADATA_TYPE_DESCRIPTOR`` reads the fs-verity descriptor. See `fs-verity descriptor`_.+-``FS_VERITY_METADATA_TYPE_SIGNATURE`` reads the signature which was+ passed to FS_IOC_ENABLE_VERITY, if any. See `Built-in signature+ verification`_.+ The semantics are similar to those of ``pread()``. ``offset`` specifies the offset in bytes into the metadata item to read from, and``length`` specifies the maximum number of bytes to read from the
@@ -279,7 +284,9 @@ FS_IOC_READ_VERITY_METADATA can fail with the following errors:-``EINTR``: the ioctl was interrupted before any data was read-``EINVAL``: reserved fields were set, or ``offset + length`` overflowed--``ENODATA``: the file is not a verity file+-``ENODATA``: the file is not a verity file, or+ FS_VERITY_METADATA_TYPE_SIGNATURE was requested but the file doesn't+ have a built-in signature-``ENOTTY``: this type of filesystem does not implement fs-verity, or this ioctl is not yet implemented on it-``EOPNOTSUPP``: the kernel was not configured with fs-verity
From: Eric Biggers <ebiggers@kernel.org> Date: 2021-01-15 18:21:34
From: Eric Biggers <redacted>
Add support for FS_VERITY_METADATA_TYPE_MERKLE_TREE to
FS_IOC_READ_VERITY_METADATA. This allows a userspace server program to
retrieve the Merkle tree of a verity file for serving to a client which
implements fs-verity compatible verification. See the patch which
introduced FS_IOC_READ_VERITY_METADATA for more details.
This has been tested using a new xfstest which calls this ioctl via a
new subcommand for the 'fsverity' program from fsverity-utils.
Signed-off-by: Eric Biggers <redacted>
---
Documentation/filesystems/fsverity.rst | 10 +++-
fs/verity/read_metadata.c | 70 ++++++++++++++++++++++++++
include/uapi/linux/fsverity.h | 2 +
3 files changed, 81 insertions(+), 1 deletion(-)
@@ -234,6 +234,8 @@ need this ioctl. This ioctl takes in a pointer to the following structure::+ #define FS_VERITY_METADATA_TYPE_MERKLE_TREE 1+ struct fsverity_read_metadata_arg { __u64 metadata_type; __u64 offset;
@@ -242,7 +244,13 @@ This ioctl takes in a pointer to the following structure:: __u64 __reserved; };-``metadata_type`` specifies the type of metadata to read.+``metadata_type`` specifies the type of metadata to read:++-``FS_VERITY_METADATA_TYPE_MERKLE_TREE`` reads the blocks of the+ Merkle tree. The blocks are returned in order from the root level+ to the leaf level. Within each level, the blocks are returned in+ the same order that their hashes are themselves hashed.+ See `Merkle tree`_ for more information. The semantics are similar to those of ``pread()``. ``offset`` specifies the offset in bytes into the metadata item to read from, and
From: Victor Hsieh <hidden> Date: 2021-01-22 23:27:52
LGTM. Thanks!
Reviewed-by: Victor Hsieh <redacted>
On Fri, Jan 15, 2021 at 10:19 AM Eric Biggers [off-list ref] wrote:
[This patchset applies to v5.11-rc3]
Add an ioctl FS_IOC_READ_VERITY_METADATA which allows reading verity
metadata from a file that has fs-verity enabled, including:
- The Merkle tree
- The fsverity_descriptor (not including the signature if present)
- The built-in signature, if present
This ioctl has similar semantics to pread(). It is passed the type of
metadata to read (one of the above three), and a buffer, offset, and
size. It returns the number of bytes read or an error.
This ioctl doesn't make any assumption about where the metadata is
stored on-disk. It does assume the metadata is in a stable format, but
that's basically already the case:
- The Merkle tree and fsverity_descriptor are defined by how fs-verity
file digests are computed; see the "File digest computation" section
of Documentation/filesystems/fsverity.rst. Technically, the way in
which the levels of the tree are ordered relative to each other wasn't
previously specified, but it's logical to put the root level first.
- The built-in signature is the value passed to FS_IOC_ENABLE_VERITY.
This ioctl is useful because it allows writing a server program that
takes a verity file and serves it to a client program, such that the
client can do its own fs-verity compatible verification of the file.
This only makes sense if the client doesn't trust the server and if the
server needs to provide the storage for the client.
More concretely, there is interest in using this ability in Android to
export APK files (which are protected by fs-verity) to "protected VMs".
This would use Protected KVM (https://lwn.net/Articles/836693), which
provides an isolated execution environment without having to trust the
traditional "host". A "guest" VM can boot from a signed image and
perform specific tasks in a minimum trusted environment using files that
have fs-verity enabled on the host, without trusting the host or
requiring that the guest has its own trusted storage.
Technically, it would be possible to duplicate the metadata and store it
in separate files for serving. However, that would be less efficient
and would require extra care in userspace to maintain file consistency.
In addition to the above, the ability to read the built-in signatures is
useful because it allows a system that is using the in-kernel signature
verification to migrate to userspace signature verification.
This patchset has been tested by new xfstests which call this new ioctl
via a new subcommand for the 'fsverity' program from fsverity-utils.
Eric Biggers (6):
fs-verity: factor out fsverity_get_descriptor()
fs-verity: don't pass whole descriptor to fsverity_verify_signature()
fs-verity: add FS_IOC_READ_VERITY_METADATA ioctl
fs-verity: support reading Merkle tree with ioctl
fs-verity: support reading descriptor with ioctl
fs-verity: support reading signature with ioctl
Documentation/filesystems/fsverity.rst | 76 ++++++++++
fs/ext4/ioctl.c | 7 +
fs/f2fs/file.c | 11 ++
fs/verity/Makefile | 1 +
fs/verity/fsverity_private.h | 13 +-
fs/verity/open.c | 133 +++++++++++------
fs/verity/read_metadata.c | 195 +++++++++++++++++++++++++
fs/verity/signature.c | 20 +--
include/linux/fsverity.h | 12 ++
include/uapi/linux/fsverity.h | 14 ++
10 files changed, 417 insertions(+), 65 deletions(-)
create mode 100644 fs/verity/read_metadata.c
base-commit: 7c53f6b671f4aba70ff15e1b05148b10d58c2837
--
2.30.0
From: Eric Biggers <ebiggers@kernel.org> Date: 2021-01-26 20:10:15
On Fri, Jan 22, 2021 at 03:26:48PM -0800, Victor Hsieh wrote:
LGTM. Thanks!
Reviewed-by: Victor Hsieh <redacted>
On Fri, Jan 15, 2021 at 10:19 AM Eric Biggers [off-list ref] wrote:
quoted
[This patchset applies to v5.11-rc3]
Add an ioctl FS_IOC_READ_VERITY_METADATA which allows reading verity
metadata from a file that has fs-verity enabled, including:
Thanks Victor. Does anyone else have comments on this patchset?
- Eric
From: Jaegeuk Kim <jaegeuk@kernel.org> Date: 2021-01-28 01:12:47
On 01/15, Eric Biggers wrote:
From: Eric Biggers <redacted>
Now that fsverity_get_descriptor() validates the sig_size field,
fsverity_verify_signature() doesn't need to do it.
Just change the prototype of fsverity_verify_signature() to take the
signature directly rather than take a fsverity_descriptor.
Signed-off-by: Eric Biggers <redacted>
From: Jaegeuk Kim <jaegeuk@kernel.org> Date: 2021-01-28 01:12:48
On 01/15, Eric Biggers wrote:
From: Eric Biggers <redacted>
The FS_IOC_READ_VERITY_METADATA ioctl will need to return the fs-verity
descriptor (and signature) to userspace.
There are a few ways we could implement this:
- Save a copy of the descriptor (and signature) in the fsverity_info
struct that hangs off of the in-memory inode. However, this would
waste memory since most of the time it wouldn't be needed.
- Regenerate the descriptor from the merkle_tree_params in the
fsverity_info. However, this wouldn't work for the signature, nor for
the salt which the merkle_tree_params only contains indirectly as part
of the 'hashstate'. It would also be error-prone.
- Just get them from the filesystem again. The disadvantage is that in
general we can't trust that they haven't been maliciously changed
since the file has opened. However, the use cases for
FS_IOC_READ_VERITY_METADATA don't require that it verifies the chain
of trust. So this is okay as long as we do some basic validation.
In preparation for implementing the third option, factor out a helper
function fsverity_get_descriptor() which gets the descriptor (and
appended signature) from the filesystem and does some basic validation.
As part of this, start checking the sig_size field for overflow.
Currently fsverity_verify_signature() does this. But the new ioctl will
need this too, so do it earlier.
Signed-off-by: Eric Biggers <redacted>
@@ -142,45 +142,17 @@ static int compute_file_digest(struct fsverity_hash_alg *hash_alg,}/*-*Validatethegivenfsverity_descriptorandcreateanewfsverity_infofrom-*it.Thesignature(ifpresent)isalsochecked.+*Createanewfsverity_infofromthegivenfsverity_descriptor(withoptional+*appendedsignature),andcheckthesignatureifpresent.The+*fsverity_descriptormusthavealreadyundergonebasicvalidation.*/structfsverity_info*fsverity_create_info(conststructinode*inode,-void*_desc,size_tdesc_size)+structfsverity_descriptor*desc,+size_tdesc_size){-structfsverity_descriptor*desc=_desc;structfsverity_info*vi;interr;-if(desc_size<sizeof(*desc)){-fsverity_err(inode,"Unrecognized descriptor size: %zu bytes",-desc_size);-returnERR_PTR(-EINVAL);-}--if(desc->version!=1){-fsverity_err(inode,"Unrecognized descriptor version: %u",-desc->version);-returnERR_PTR(-EINVAL);-}--if(memchr_inv(desc->__reserved,0,sizeof(desc->__reserved))){-fsverity_err(inode,"Reserved bits set in descriptor");-returnERR_PTR(-EINVAL);-}--if(desc->salt_size>sizeof(desc->salt)){-fsverity_err(inode,"Invalid salt_size: %u",desc->salt_size);-returnERR_PTR(-EINVAL);-}--if(le64_to_cpu(desc->data_size)!=inode->i_size){-fsverity_err(inode,-"Wrong data_size: %llu (desc) != %lld (inode)",-le64_to_cpu(desc->data_size),inode->i_size);-returnERR_PTR(-EINVAL);-}-vi=kmem_cache_zalloc(fsverity_info_cachep,GFP_KERNEL);if(!vi)returnERR_PTR(-ENOMEM);
@@ -245,15 +217,57 @@ void fsverity_free_info(struct fsverity_info *vi)kmem_cache_free(fsverity_info_cachep,vi);}-/* Ensure the inode has an ->i_verity_info */-staticintensure_verity_info(structinode*inode)+staticboolvalidate_fsverity_descriptor(structinode*inode,+conststructfsverity_descriptor*desc,+size_tdesc_size){-structfsverity_info*vi=fsverity_get_info(inode);-structfsverity_descriptor*desc;-intres;+if(desc_size<sizeof(*desc)){+fsverity_err(inode,"Unrecognized descriptor size: %zu bytes",+desc_size);+returnfalse;+}-if(vi)-return0;+if(desc->version!=1){+fsverity_err(inode,"Unrecognized descriptor version: %u",+desc->version);+returnfalse;+}++if(memchr_inv(desc->__reserved,0,sizeof(desc->__reserved))){+fsverity_err(inode,"Reserved bits set in descriptor");+returnfalse;+}++if(desc->salt_size>sizeof(desc->salt)){+fsverity_err(inode,"Invalid salt_size: %u",desc->salt_size);+returnfalse;+}++if(le64_to_cpu(desc->data_size)!=inode->i_size){+fsverity_err(inode,+"Wrong data_size: %llu (desc) != %lld (inode)",+le64_to_cpu(desc->data_size),inode->i_size);+returnfalse;+}++if(le32_to_cpu(desc->sig_size)>desc_size-sizeof(*desc)){+fsverity_err(inode,"Signature overflows verity descriptor");+returnfalse;+}++returntrue;+}++/*+*Readtheinode'sfsverity_descriptor(withoptionalappendedsignature)from+*thefilesystem,anddobasicvalidationofit.+*/+intfsverity_get_descriptor(structinode*inode,+structfsverity_descriptor**desc_ret,+size_t*desc_size_ret)+{+intres;+structfsverity_descriptor*desc;res=inode->i_sb->s_vop->get_verity_descriptor(inode,NULL,0);if(res<0){
@@ -272,20 +286,46 @@ static int ensure_verity_info(struct inode *inode)res=inode->i_sb->s_vop->get_verity_descriptor(inode,desc,res);if(res<0){fsverity_err(inode,"Error %d reading verity descriptor",res);-gotoout_free_desc;+kfree(desc);+returnres;+}++if(!validate_fsverity_descriptor(inode,desc,res)){+kfree(desc);+return-EINVAL;}-vi=fsverity_create_info(inode,desc,res);+*desc_ret=desc;+*desc_size_ret=res;+return0;+}++/* Ensure the inode has an ->i_verity_info */+staticintensure_verity_info(structinode*inode)+{+structfsverity_info*vi=fsverity_get_info(inode);+structfsverity_descriptor*desc;+size_tdesc_size;+interr;++if(vi)+return0;++err=fsverity_get_descriptor(inode,&desc,&desc_size);+if(err)+returnerr;++vi=fsverity_create_info(inode,desc,desc_size);if(IS_ERR(vi)){-res=PTR_ERR(vi);+err=PTR_ERR(vi);gotoout_free_desc;}fsverity_set_info(inode,vi);-res=0;+err=0;out_free_desc:kfree(desc);-returnres;+returnerr;}/**
From: Jaegeuk Kim <jaegeuk@kernel.org> Date: 2021-01-28 01:13:28
On 01/15, Eric Biggers wrote:
quoted hunk
From: Eric Biggers <redacted>
Add support for FS_VERITY_METADATA_TYPE_MERKLE_TREE to
FS_IOC_READ_VERITY_METADATA. This allows a userspace server program to
retrieve the Merkle tree of a verity file for serving to a client which
implements fs-verity compatible verification. See the patch which
introduced FS_IOC_READ_VERITY_METADATA for more details.
This has been tested using a new xfstest which calls this ioctl via a
new subcommand for the 'fsverity' program from fsverity-utils.
Signed-off-by: Eric Biggers <redacted>
---
Documentation/filesystems/fsverity.rst | 10 +++-
fs/verity/read_metadata.c | 70 ++++++++++++++++++++++++++
include/uapi/linux/fsverity.h | 2 +
3 files changed, 81 insertions(+), 1 deletion(-)
@@ -234,6 +234,8 @@ need this ioctl. This ioctl takes in a pointer to the following structure::+ #define FS_VERITY_METADATA_TYPE_MERKLE_TREE 1+ struct fsverity_read_metadata_arg { __u64 metadata_type; __u64 offset;
@@ -242,7 +244,13 @@ This ioctl takes in a pointer to the following structure:: __u64 __reserved; };-``metadata_type`` specifies the type of metadata to read.+``metadata_type`` specifies the type of metadata to read:++-``FS_VERITY_METADATA_TYPE_MERKLE_TREE`` reads the blocks of the+ Merkle tree. The blocks are returned in order from the root level+ to the leaf level. Within each level, the blocks are returned in+ the same order that their hashes are themselves hashed.+ See `Merkle tree`_ for more information. The semantics are similar to those of ``pread()``. ``offset`` specifies the offset in bytes into the metadata item to read from, and
From: Jaegeuk Kim <jaegeuk@kernel.org> Date: 2021-01-28 01:14:38
On 01/15, Eric Biggers wrote:
From: Eric Biggers <redacted>
Add support for FS_VERITY_METADATA_TYPE_SIGNATURE to
FS_IOC_READ_VERITY_METADATA. This allows a userspace server program to
retrieve the built-in signature (if present) of a verity file for
serving to a client which implements fs-verity compatible verification.
See the patch which introduced FS_IOC_READ_VERITY_METADATA for more
details.
The ability for userspace to read the built-in signatures is also useful
because it allows a system that is using the in-kernel signature
verification to migrate to userspace signature verification.
This has been tested using a new xfstest which calls this ioctl via a
new subcommand for the 'fsverity' program from fsverity-utils.
Signed-off-by: Eric Biggers <redacted>
@@ -236,6 +236,7 @@ This ioctl takes in a pointer to the following structure:: #define FS_VERITY_METADATA_TYPE_MERKLE_TREE 1 #define FS_VERITY_METADATA_TYPE_DESCRIPTOR 2+ #define FS_VERITY_METADATA_TYPE_SIGNATURE 3 struct fsverity_read_metadata_arg { __u64 metadata_type;
@@ -256,6 +257,10 @@ This ioctl takes in a pointer to the following structure::-``FS_VERITY_METADATA_TYPE_DESCRIPTOR`` reads the fs-verity descriptor. See `fs-verity descriptor`_.+-``FS_VERITY_METADATA_TYPE_SIGNATURE`` reads the signature which was+ passed to FS_IOC_ENABLE_VERITY, if any. See `Built-in signature+ verification`_.+ The semantics are similar to those of ``pread()``. ``offset`` specifies the offset in bytes into the metadata item to read from, and``length`` specifies the maximum number of bytes to read from the
@@ -279,7 +284,9 @@ FS_IOC_READ_VERITY_METADATA can fail with the following errors:-``EINTR``: the ioctl was interrupted before any data was read-``EINVAL``: reserved fields were set, or ``offset + length`` overflowed--``ENODATA``: the file is not a verity file+-``ENODATA``: the file is not a verity file, or+ FS_VERITY_METADATA_TYPE_SIGNATURE was requested but the file doesn't+ have a built-in signature-``ENOTTY``: this type of filesystem does not implement fs-verity, or this ioctl is not yet implemented on it-``EOPNOTSUPP``: the kernel was not configured with fs-verity
From: Jaegeuk Kim <jaegeuk@kernel.org> Date: 2021-01-28 01:14:42
On 01/15, Eric Biggers wrote:
From: Eric Biggers <redacted>
Add support for FS_VERITY_METADATA_TYPE_DESCRIPTOR to
FS_IOC_READ_VERITY_METADATA. This allows a userspace server program to
retrieve the fs-verity descriptor of a file for serving to a client
which implements fs-verity compatible verification. See the patch which
introduced FS_IOC_READ_VERITY_METADATA for more details.
"fs-verity descriptor" here means only the part that userspace cares
about because it is hashed to produce the file digest. It doesn't
include the signature which ext4 and f2fs append to the
fsverity_descriptor struct when storing it on-disk, since that way of
storing the signature is an implementation detail. The next patch adds
a separate metadata_type value for retrieving the signature separately.
This has been tested using a new xfstest which calls this ioctl via a
new subcommand for the 'fsverity' program from fsverity-utils.
Signed-off-by: Eric Biggers <redacted>
@@ -235,6 +235,7 @@ need this ioctl. This ioctl takes in a pointer to the following structure:: #define FS_VERITY_METADATA_TYPE_MERKLE_TREE 1+ #define FS_VERITY_METADATA_TYPE_DESCRIPTOR 2 struct fsverity_read_metadata_arg { __u64 metadata_type;
@@ -252,6 +253,9 @@ This ioctl takes in a pointer to the following structure:: the same order that their hashes are themselves hashed. See `Merkle tree`_ for more information.+-``FS_VERITY_METADATA_TYPE_DESCRIPTOR`` reads the fs-verity+ descriptor. See `fs-verity descriptor`_.+ The semantics are similar to those of ``pread()``. ``offset`` specifies the offset in bytes into the metadata item to read from, and``length`` specifies the maximum number of bytes to read from the
@@ -76,6 +76,44 @@ static int fsverity_read_merkle_tree(struct inode *inode,}returnretval?retval:err;}++/* Copy the requested portion of the buffer to userspace. */+staticintfsverity_read_buffer(void__user*dst,u64offset,intlength,+constvoid*src,size_tsrc_length)+{+if(offset>=src_length)+return0;+src+=offset;+src_length-=offset;++length=min_t(size_t,length,src_length);++if(copy_to_user(dst,src,length))+return-EFAULT;++returnlength;+}++staticintfsverity_read_descriptor(structinode*inode,+void__user*buf,u64offset,intlength)+{+structfsverity_descriptor*desc;+size_tdesc_size;+intres;++res=fsverity_get_descriptor(inode,&desc,&desc_size);+if(res)+returnres;++/* don't include the signature */+desc_size=offsetof(structfsverity_descriptor,signature);+desc->sig_size=0;++res=fsverity_read_buffer(buf,offset,length,desc,desc_size);++kfree(desc);+returnres;+}/***fsverity_ioctl_read_metadata()-readveritymetadatafromafile*@filp:filetoreadthemetadatafrom
From: Jaegeuk Kim <jaegeuk@kernel.org> Date: 2021-01-28 01:17:28
On 01/15, Eric Biggers wrote:
From: Eric Biggers <redacted>
Add an ioctl FS_IOC_READ_VERITY_METADATA which will allow reading verity
metadata from a file that has fs-verity enabled, including:
- The Merkle tree
- The fsverity_descriptor (not including the signature if present)
- The built-in signature, if present
This ioctl has similar semantics to pread(). It is passed the type of
metadata to read (one of the above three), and a buffer, offset, and
size. It returns the number of bytes read or an error.
Separate patches will add support for each of the above metadata types.
This patch just adds the ioctl itself.
This ioctl doesn't make any assumption about where the metadata is
stored on-disk. It does assume the metadata is in a stable format, but
that's basically already the case:
- The Merkle tree and fsverity_descriptor are defined by how fs-verity
file digests are computed; see the "File digest computation" section
of Documentation/filesystems/fsverity.rst. Technically, the way in
which the levels of the tree are ordered relative to each other wasn't
previously specified, but it's logical to put the root level first.
- The built-in signature is the value passed to FS_IOC_ENABLE_VERITY.
This ioctl is useful because it allows writing a server program that
takes a verity file and serves it to a client program, such that the
client can do its own fs-verity compatible verification of the file.
This only makes sense if the client doesn't trust the server and if the
server needs to provide the storage for the client.
More concretely, there is interest in using this ability in Android to
export APK files (which are protected by fs-verity) to "protected VMs".
This would use Protected KVM (https://lwn.net/Articles/836693), which
provides an isolated execution environment without having to trust the
traditional "host". A "guest" VM can boot from a signed image and
perform specific tasks in a minimum trusted environment using files that
have fs-verity enabled on the host, without trusting the host or
requiring that the guest has its own trusted storage.
Technically, it would be possible to duplicate the metadata and store it
in separate files for serving. However, that would be less efficient
and would require extra care in userspace to maintain file consistency.
In addition to the above, the ability to read the built-in signatures is
useful because it allows a system that is using the in-kernel signature
verification to migrate to userspace signature verification.
Signed-off-by: Eric Biggers <redacted>
@@ -217,6 +217,63 @@ FS_IOC_MEASURE_VERITY can fail with the following errors:-``EOVERFLOW``: the digest is longer than the specified``digest_size`` bytes. Try providing a larger buffer.+FS_IOC_READ_VERITY_METADATA+---------------------------++The FS_IOC_READ_VERITY_METADATA ioctl reads verity metadata from a+verity file. This ioctl is available since Linux v5.12.++This ioctl allows writing a server program that takes a verity file+and serves it to a client program, such that the client can do its own+fs-verity compatible verification of the file. This only makes sense+if the client doesn't trust the server and if the server needs to+provide the storage for the client.++This is a fairly specialized use case, and most fs-verity users won't+need this ioctl.++This ioctl takes in a pointer to the following structure::++ struct fsverity_read_metadata_arg {+ __u64 metadata_type;+ __u64 offset;+ __u64 length;+ __u64 buf_ptr;+ __u64 __reserved;+ };++``metadata_type`` specifies the type of metadata to read.++The semantics are similar to those of ``pread()``. ``offset``+specifies the offset in bytes into the metadata item to read from, and+``length`` specifies the maximum number of bytes to read from the+metadata item. ``buf_ptr`` is the pointer to the buffer to read into,+cast to a 64-bit integer. ``__reserved`` must be 0. On success, the+number of bytes read is returned. 0 is returned at the end of the+metadata item. The returned length may be less than ``length``, for+example if the ioctl is interrupted.++The metadata returned by FS_IOC_READ_VERITY_METADATA isn't guaranteed+to be authenticated against the file digest that would be returned by+`FS_IOC_MEASURE_VERITY`_, as the metadata is expected to be used to+implement fs-verity compatible verification anyway (though absent a+malicious disk, the metadata will indeed match). E.g. to implement+this ioctl, the filesystem is allowed to just read the Merkle tree+blocks from disk without actually verifying the path to the root node.++FS_IOC_READ_VERITY_METADATA can fail with the following errors:++-``EFAULT``: the caller provided inaccessible memory+-``EINTR``: the ioctl was interrupted before any data was read+-``EINVAL``: reserved fields were set, or ``offset + length``+ overflowed+-``ENODATA``: the file is not a verity file+-``ENOTTY``: this type of filesystem does not implement fs-verity, or+ this ioctl is not yet implemented on it+-``EOPNOTSUPP``: the kernel was not configured with fs-verity+ support, or the filesystem superblock has not had the 'verity'+ feature enabled on it. (See `Filesystem support`_.)+ FS_IOC_GETFLAGS ---------------
@@ -1306,6 +1306,12 @@ static long __ext4_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)return-EOPNOTSUPP;returnfsverity_ioctl_measure(filp,(void__user*)arg);+caseFS_IOC_READ_VERITY_METADATA:+if(!ext4_has_feature_verity(sb))+return-EOPNOTSUPP;+returnfsverity_ioctl_read_metadata(filp,+(constvoid__user*)arg);+default:return-ENOTTY;}
@@ -1388,6 +1394,7 @@ long ext4_compat_ioctl(struct file *file, unsigned int cmd, unsigned long arg)caseFS_IOC_GETFSMAP:caseFS_IOC_ENABLE_VERITY:caseFS_IOC_MEASURE_VERITY:+caseFS_IOC_READ_VERITY_METADATA:caseEXT4_IOC_CLEAR_ES_CACHE:caseEXT4_IOC_GETSTATE:caseEXT4_IOC_GET_ES_CACHE:
@@ -3357,6 +3357,14 @@ static int f2fs_ioc_measure_verity(struct file *filp, unsigned long arg)returnfsverity_ioctl_measure(filp,(void__user*)arg);}+staticintf2fs_ioc_read_verity_metadata(structfile*filp,unsignedlongarg)+{+if(!f2fs_sb_has_verity(F2FS_I_SB(file_inode(filp))))+return-EOPNOTSUPP;++returnfsverity_ioctl_read_metadata(filp,(constvoid__user*)arg);+}+staticintf2fs_ioc_getfslabel(structfile*filp,unsignedlongarg){structinode*inode=file_inode(filp);
@@ -4272,6 +4280,8 @@ static long __f2fs_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)returnf2fs_ioc_enable_verity(filp,arg);caseFS_IOC_MEASURE_VERITY:returnf2fs_ioc_measure_verity(filp,arg);+caseFS_IOC_READ_VERITY_METADATA:+returnf2fs_ioc_read_verity_metadata(filp,arg);caseFS_IOC_GETFSLABEL:returnf2fs_ioc_getfslabel(filp,arg);caseFS_IOC_SETFSLABEL:
@@ -4523,6 +4533,7 @@ long f2fs_compat_ioctl(struct file *file, unsigned int cmd, unsigned long arg)caseF2FS_IOC_RESIZE_FS:caseFS_IOC_ENABLE_VERITY:caseFS_IOC_MEASURE_VERITY:+caseFS_IOC_READ_VERITY_METADATA:caseFS_IOC_GETFSLABEL:caseFS_IOC_SETFSLABEL:caseF2FS_IOC_GET_COMPRESS_BLOCKS:
@@ -0,0 +1,55 @@+// SPDX-License-Identifier: GPL-2.0-only+/*+*Ioctltoreadveritymetadata+*+*Copyright2021GoogleLLC+*/++#include"fsverity_private.h"++#include<linux/uaccess.h>++/**+*fsverity_ioctl_read_metadata()-readveritymetadatafromafile+*@filp:filetoreadthemetadatafrom+*@uarg:userpointertofsverity_read_metadata_arg+*+*Return:lengthreadonsuccess,0onEOF,-errnoonfailure+*/+intfsverity_ioctl_read_metadata(structfile*filp,constvoid__user*uarg)+{+structinode*inode=file_inode(filp);+conststructfsverity_info*vi;+structfsverity_read_metadata_argarg;+intlength;+void__user*buf;++vi=fsverity_get_info(inode);+if(!vi)+return-ENODATA;/* not a verity file */+/*+*Notethatwedon'thavetoexplicitlycheckthatthefileisopenfor+*reading,sinceverityfilescanonlybeopenedforreading.+*/++if(copy_from_user(&arg,uarg,sizeof(arg)))+return-EFAULT;++if(arg.__reserved)+return-EINVAL;++/* offset + length must not overflow. */+if(arg.offset+arg.length<arg.offset)+return-EINVAL;++/* Ensure that the return value will fit in INT_MAX. */+length=min_t(u64,arg.length,INT_MAX);++buf=u64_to_user_ptr(arg.buf_ptr);++switch(arg.metadata_type){+default:+return-EINVAL;+}+}+EXPORT_SYMBOL_GPL(fsverity_ioctl_read_metadata);
From: Eric Biggers <ebiggers@kernel.org> Date: 2021-01-28 02:18:04
On Wed, Jan 27, 2021 at 05:10:39PM -0800, Jaegeuk Kim wrote:
Minor thought:
How about invalidating or truncating merkel tree pages?
Reviewed-by: Jaegeuk Kim <jaegeuk@kernel.org>
Removing them from the page cache after the read, you mean? I'm not sure we can
assume that users of this ioctl would want the pages to *not* be cached, any
more than we could assume that for any regular read(). I think we should just
leave the pages cached (like a regular read) and not do anything special. Like
other pagecache pages, the kernel will evict the Merkle tree pages eventually if
they aren't being accessed anymore and memory needs to be reclaimed.
- Eric
On Fri, Jan 15, 2021 at 10:22 AM Eric Biggers [off-list ref] wrote:
From: Eric Biggers <redacted>
Now that fsverity_get_descriptor() validates the sig_size field,
fsverity_verify_signature() doesn't need to do it.
Just change the prototype of fsverity_verify_signature() to take the
signature directly rather than take a fsverity_descriptor.
Signed-off-by: Eric Biggers <redacted>
From: Eric Biggers <ebiggers@kernel.org> Date: 2021-02-01 17:42:30
On Fri, Jan 15, 2021 at 10:18:13AM -0800, Eric Biggers wrote:
[This patchset applies to v5.11-rc3]
Add an ioctl FS_IOC_READ_VERITY_METADATA which allows reading verity
metadata from a file that has fs-verity enabled, including:
- The Merkle tree
- The fsverity_descriptor (not including the signature if present)
- The built-in signature, if present
This ioctl has similar semantics to pread(). It is passed the type of
metadata to read (one of the above three), and a buffer, offset, and
size. It returns the number of bytes read or an error.
This ioctl doesn't make any assumption about where the metadata is
stored on-disk. It does assume the metadata is in a stable format, but
that's basically already the case:
- The Merkle tree and fsverity_descriptor are defined by how fs-verity
file digests are computed; see the "File digest computation" section
of Documentation/filesystems/fsverity.rst. Technically, the way in
which the levels of the tree are ordered relative to each other wasn't
previously specified, but it's logical to put the root level first.
- The built-in signature is the value passed to FS_IOC_ENABLE_VERITY.
This ioctl is useful because it allows writing a server program that
takes a verity file and serves it to a client program, such that the
client can do its own fs-verity compatible verification of the file.
This only makes sense if the client doesn't trust the server and if the
server needs to provide the storage for the client.
More concretely, there is interest in using this ability in Android to
export APK files (which are protected by fs-verity) to "protected VMs".
This would use Protected KVM (https://lwn.net/Articles/836693), which
provides an isolated execution environment without having to trust the
traditional "host". A "guest" VM can boot from a signed image and
perform specific tasks in a minimum trusted environment using files that
have fs-verity enabled on the host, without trusting the host or
requiring that the guest has its own trusted storage.
Technically, it would be possible to duplicate the metadata and store it
in separate files for serving. However, that would be less efficient
and would require extra care in userspace to maintain file consistency.
In addition to the above, the ability to read the built-in signatures is
useful because it allows a system that is using the in-kernel signature
verification to migrate to userspace signature verification.
This patchset has been tested by new xfstests which call this new ioctl
via a new subcommand for the 'fsverity' program from fsverity-utils.
Eric Biggers (6):
fs-verity: factor out fsverity_get_descriptor()
fs-verity: don't pass whole descriptor to fsverity_verify_signature()
fs-verity: add FS_IOC_READ_VERITY_METADATA ioctl
fs-verity: support reading Merkle tree with ioctl
fs-verity: support reading descriptor with ioctl
fs-verity: support reading signature with ioctl
Documentation/filesystems/fsverity.rst | 76 ++++++++++
fs/ext4/ioctl.c | 7 +
fs/f2fs/file.c | 11 ++
fs/verity/Makefile | 1 +
fs/verity/fsverity_private.h | 13 +-
fs/verity/open.c | 133 +++++++++++------
fs/verity/read_metadata.c | 195 +++++++++++++++++++++++++
fs/verity/signature.c | 20 +--
include/linux/fsverity.h | 12 ++
include/uapi/linux/fsverity.h | 14 ++
10 files changed, 417 insertions(+), 65 deletions(-)
create mode 100644 fs/verity/read_metadata.c
All applied to fscrypt.git#fsverity for 5.12.
- Eric
+static int f2fs_ioc_read_verity_metadata(struct file *filp, unsigned long arg)
+{
+ if (!f2fs_sb_has_verity(F2FS_I_SB(file_inode(filp))))
+ return -EOPNOTSUPP;
One case is after we update kernel image, f2fs module may no longer support
compress algorithm which current file was compressed with, to avoid triggering
IO with empty compress engine (struct f2fs_compress_ops pointer):
It needs to add f2fs_is_compress_backend_ready() check condition here?
Thanks,
From: Eric Biggers <ebiggers@kernel.org> Date: 2021-02-07 08:02:35
On Sun, Feb 07, 2021 at 03:46:43PM +0800, Chao Yu wrote:
Hi Eric,
On 2021/1/16 2:18, Eric Biggers wrote:
quoted
+static int f2fs_ioc_read_verity_metadata(struct file *filp, unsigned long arg)
+{
+ if (!f2fs_sb_has_verity(F2FS_I_SB(file_inode(filp))))
+ return -EOPNOTSUPP;
One case is after we update kernel image, f2fs module may no longer support
compress algorithm which current file was compressed with, to avoid triggering
IO with empty compress engine (struct f2fs_compress_ops pointer):
It needs to add f2fs_is_compress_backend_ready() check condition here?
Thanks,
In that case it wouldn't have been possible to open the file, because
f2fs_file_open() checks for it. So it's not necessary to repeat the same check
in every operation on the file descriptor.
- Eric
On Sun, Feb 07, 2021 at 03:46:43PM +0800, Chao Yu wrote:
quoted
Hi Eric,
On 2021/1/16 2:18, Eric Biggers wrote:
quoted
+static int f2fs_ioc_read_verity_metadata(struct file *filp, unsigned long arg)
+{
+ if (!f2fs_sb_has_verity(F2FS_I_SB(file_inode(filp))))
+ return -EOPNOTSUPP;
One case is after we update kernel image, f2fs module may no longer support
compress algorithm which current file was compressed with, to avoid triggering
IO with empty compress engine (struct f2fs_compress_ops pointer):
It needs to add f2fs_is_compress_backend_ready() check condition here?
Thanks,
In that case it wouldn't have been possible to open the file, because
f2fs_file_open() checks for it. So it's not necessary to repeat the same check
in every operation on the file descriptor.
Oh, yes, it's safe now.
I'm thinking we need to remove the check in f2fs_file_open(), because the check
will fail metadata access/update (via f{g,s}etxattr/ioctl), however original
intention of that check is only to avoid syscalls to touch compressed data w/o
the engine, anyway this is another topic.
The whole patchset looks fine to me, feel free to add:
Reviewed-by: Chao Yu <redacted>
Thanks,