Message ID | 20180511231236.5501-3-mjg59@google.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
On Fri, 2018-05-11 at 16:12 -0700, Matthew Garrett wrote: > Sites may wish to provide additional metadata alongside files in order > to make more fine-grained security decisions[1]. The security of this is > enhanced if this metadata is protected, something that EVM makes > possible. However, the kernel cannot know about the set of extended > attributes that local admins may wish to protect, and hardcoding this > policy in the kernel makes it difficult to change over time and less > convenient for distributions to enable. > > This patch adds a new /sys/kernel/security/evm_xattrs node, which can be > read to obtain the current set of EVM-protected extended attributes or > written to in order to add new entries. Extending this list will not > change the validity of any existing signatures provided that the file > in question does not have any of the additional extended attributes - > missing xattrs are skipped when calculating the EVM hash. The new pathname introduced in the new patch needs to be propagated to the patch description, Documentation and Kconfig in this patch. > > [1] For instance, a package manager could install information about the > package uploader in an additional extended attribute. Local LSM policy > could then be associated with that extended attribute in order to > restrict the privileges available to packages from less trusted > uploaders. > > Signed-off-by: Matthew Garrett <mjg59@google.com> > --- > > I think this covers the review comments. Yes, it looks good. The pathname changes introduced in the new patch, need to be propagated to the patch description, above, Documentation, and Kconfig. The only other change is that with the introduction of requiring the xattr names to be prefixed with "security", locking the xattr names list fails. After making these tw changes, there's no need to re-post the entire patch set. Please re-post just this one. > Documentation/ABI/testing/evm | 13 +++ > include/uapi/linux/audit.h | 1 + > security/integrity/evm/Kconfig | 11 ++ > security/integrity/evm/evm_crypto.c | 2 +- > security/integrity/evm/evm_main.c | 6 +- > security/integrity/evm/evm_secfs.c | 173 ++++++++++++++++++++++++++++ > 6 files changed, 202 insertions(+), 4 deletions(-) > > diff --git a/Documentation/ABI/testing/evm b/Documentation/ABI/testing/evm > index d12cb2eae9ee..fa31df7fd30b 100644 > --- a/Documentation/ABI/testing/evm > +++ b/Documentation/ABI/testing/evm > @@ -57,3 +57,16 @@ Description: > dracut (via 97masterkey and 98integrity) and systemd (via > core/ima-setup) have support for loading keys at boot > time. > + > +What: security/evm_xattrs Update > +Date: April 2018 > +Contact: Matthew Garrett <mjg59@google.com> > +Description: > + Shows the set of extended attributes used to calculate or > + validate the EVM signature, and allows additional attributes > + to be added at runtime. Any signatures generated after > + additional attributes are added (and on files posessing those > + additional attributes) will only be valid if the same > + additional attributes are configured on system boot. Writing > + a single period (.) will lock the xattr list from any further > + modification. Writing '.', doesn't match the code. > +config EVM_ADD_XATTRS > + bool "Add additional EVM extended attributes at runtime" > + depends on EVM > + default n > + help > + Allow userland to provide additional xattrs for HMAC calculation. > + > + When this option is enabled, root can add additional xattrs to the > + list used by EVM by writing them into > + /sys/kernel/security/evm_xattrs. Update > + > + if (strcmp(xattr->name, ".") == 0) { > + evm_xattrs_locked = 1; > + inode = evm_xattrs->d_inode; > + inode_lock(inode); > + newattrs.ia_mode = S_IFREG | 0440; > + newattrs.ia_valid = ATTR_MODE; > + err = notify_change(evm_xattrs, &newattrs, NULL); > + inode_unlock(inode); > + audit_log_format(ab, "locked"); > + if (!err) > + err = count; > + goto out; > + } > + > + audit_log_format(ab, "xattr="); > + audit_log_untrustedstring(ab, xattr->name); > + > + if (strncmp(xattr->name, XATTR_SECURITY_PREFIX, > + XATTR_SECURITY_PREFIX_LEN) != 0) { > + err = -EINVAL; > + goto out; > + } This test now prevents locking the xattr names list. Making this an else clause will fix it. thanks, Mimi
On Sun, May 13, 2018 at 9:41 AM Mimi Zohar <zohar@linux.vnet.ibm.com> wrote: > On Fri, 2018-05-11 at 16:12 -0700, Matthew Garrett wrote: > + > > + if (strcmp(xattr->name, ".") == 0) { > > + evm_xattrs_locked = 1; > > + inode = evm_xattrs->d_inode; > > + inode_lock(inode); > > + newattrs.ia_mode = S_IFREG | 0440; > > + newattrs.ia_valid = ATTR_MODE; > > + err = notify_change(evm_xattrs, &newattrs, NULL); > > + inode_unlock(inode); > > + audit_log_format(ab, "locked"); > > + if (!err) > > + err = count; > > + goto out; > > + } > > + > > + audit_log_format(ab, "xattr="); > > + audit_log_untrustedstring(ab, xattr->name); > > + > > + if (strncmp(xattr->name, XATTR_SECURITY_PREFIX, > > + XATTR_SECURITY_PREFIX_LEN) != 0) { > > + err = -EINVAL; > > + goto out; > > + } > This test now prevents locking the xattr names list. Making this an > else clause will fix it. Are you sure? The check for "." happens before this, and jumps over the rest of the function.
On Mon, 2018-05-14 at 10:01 -0700, Matthew Garrett wrote: > On Sun, May 13, 2018 at 9:41 AM Mimi Zohar <zohar@linux.vnet.ibm.com> wrote: > > > On Fri, 2018-05-11 at 16:12 -0700, Matthew Garrett wrote: > > + > > > + if (strcmp(xattr->name, ".") == 0) { > > > + evm_xattrs_locked = 1; > > > + inode = evm_xattrs->d_inode; > > > + inode_lock(inode); > > > + newattrs.ia_mode = S_IFREG | 0440; > > > + newattrs.ia_valid = ATTR_MODE; > > > + err = notify_change(evm_xattrs, &newattrs, NULL); > > > + inode_unlock(inode); > > > + audit_log_format(ab, "locked"); > > > + if (!err) > > > + err = count; > > > + goto out; > > > + } > > > + > > > + audit_log_format(ab, "xattr="); > > > + audit_log_untrustedstring(ab, xattr->name); > > > + > > > + if (strncmp(xattr->name, XATTR_SECURITY_PREFIX, > > > + XATTR_SECURITY_PREFIX_LEN) != 0) { > > > + err = -EINVAL; > > > + goto out; > > > + } > > > This test now prevents locking the xattr names list. Making this an > > else clause will fix it. > > Are you sure? The check for "." happens before this, and jumps over the > rest of the function. Oh! It did work, but the messages are confusing. # cat /sys/kernel/security/integrity/evm/evm_xattrs security.selinux security.ima security.capability security.foo # echo foo > /sys/kernel/security/integrity/evm/evm_xattrs bash: echo: write error: Operation not permitted # echo security.foo > /sys/kernel/security/integrity/evm/evm_xattrs bash: echo: write error: Operation not permitted # cat /sys/kernel/security/integrity/evm/evm_xattrs security.selinux security.ima security.capability security.foo # echo . > /sys/kernel/security/integrity/evm/evm_xattrs bash: echo: write error: Operation not permitted # echo security.boo > /sys/kernel/security/integrity/evm/evm_xattrs bash: echo: write error: Operation not permitted # ls -lat //sys/kernel/security/integrity/evm/evm_xattrs -rw-rw----. 1 root root 0 May 14 13:10 //sys/kernel/security/integrity/evm/evm_xattrs "security.boo" wasn't added. So writing '.' worked, but the chmod did not work. Maybe the error messages are coming from there. Mimi
On Mon, 2018-05-14 at 13:19 -0400, Mimi Zohar wrote: > On Mon, 2018-05-14 at 10:01 -0700, Matthew Garrett wrote: > > On Sun, May 13, 2018 at 9:41 AM Mimi Zohar <zohar@linux.vnet.ibm.com> wrote: > > > > > On Fri, 2018-05-11 at 16:12 -0700, Matthew Garrett wrote: > > > + > > > > + if (strcmp(xattr->name, ".") == 0) { > > > > + evm_xattrs_locked = 1; > > > > + inode = evm_xattrs->d_inode; > > > > + inode_lock(inode); > > > > + newattrs.ia_mode = S_IFREG | 0440; > > > > + newattrs.ia_valid = ATTR_MODE; > > > > + err = notify_change(evm_xattrs, &newattrs, NULL); > > > > + inode_unlock(inode); > > > > + audit_log_format(ab, "locked"); > > > > + if (!err) > > > > + err = count; > > > > + goto out; > > > > + } > > > > + > > > > + audit_log_format(ab, "xattr="); > > > > + audit_log_untrustedstring(ab, xattr->name); > > > > + > > > > + if (strncmp(xattr->name, XATTR_SECURITY_PREFIX, > > > > + XATTR_SECURITY_PREFIX_LEN) != 0) { > > > > + err = -EINVAL; > > > > + goto out; > > > > + } > > > > > This test now prevents locking the xattr names list. Making this an > > > else clause will fix it. > > > > Are you sure? The check for "." happens before this, and jumps over the > > rest of the function. > > Oh! It did work, but the messages are confusing. > > # cat /sys/kernel/security/integrity/evm/evm_xattrs > security.selinux > security.ima > security.capability > security.foo > # echo foo > /sys/kernel/security/integrity/evm/evm_xattrs > bash: echo: write error: Operation not permitted > > # echo security.foo > /sys/kernel/security/integrity/evm/evm_xattrs > bash: echo: write error: Operation not permitted This makes sense, as it was already added. > > # cat /sys/kernel/security/integrity/evm/evm_xattrs > security.selinux > security.ima > security.capability > security.foo > > # echo . > /sys/kernel/security/integrity/evm/evm_xattrs > bash: echo: write error: Operation not permitted I'm still seeing this message. Mimi
On Mon, May 14, 2018 at 10:35 AM Mimi Zohar <zohar@linux.vnet.ibm.com> wrote: > > # echo . > /sys/kernel/security/integrity/evm/evm_xattrs > > bash: echo: write error: Operation not permitted > I'm still seeing this message. Looking into it.
On Mon, May 14, 2018 at 10:36 AM Matthew Garrett <mjg59@google.com> wrote: > On Mon, May 14, 2018 at 10:35 AM Mimi Zohar <zohar@linux.vnet.ibm.com> > wrote: > > > # echo . > /sys/kernel/security/integrity/evm/evm_xattrs > > > bash: echo: write error: Operation not permitted > > I'm still seeing this message. > Looking into it. I can't reproduce this - the only way you should be getting EPERM is if the list is already locked or if you don't have CAP_SYS_ADMIN.
On Mon, 2018-05-14 at 11:50 -0700, Matthew Garrett wrote: > On Mon, May 14, 2018 at 10:36 AM Matthew Garrett <mjg59@google.com> wrote: > > > On Mon, May 14, 2018 at 10:35 AM Mimi Zohar <zohar@linux.vnet.ibm.com> > > wrote: > > > > # echo . > /sys/kernel/security/integrity/evm/evm_xattrs > > > > bash: echo: write error: Operation not permitted > > > > I'm still seeing this message. > > > Looking into it. > > I can't reproduce this - the only way you should be getting EPERM is if the > list is already locked or if you don't have CAP_SYS_ADMIN. The call to notify_change() calls security_inode_setattr(), which is failing, because there is no security.evm xattr. It's failing with -EPERM. Mimi
On Mon, May 14, 2018 at 2:02 PM Mimi Zohar <zohar@linux.vnet.ibm.com> wrote: > The call to notify_change() calls security_inode_setattr(), which is > failing, because there is no security.evm xattr. It's failing with > -EPERM. Ah! Ok yes that would explain it. Let me figure out how to do this reasonably…
diff --git a/Documentation/ABI/testing/evm b/Documentation/ABI/testing/evm index d12cb2eae9ee..fa31df7fd30b 100644 --- a/Documentation/ABI/testing/evm +++ b/Documentation/ABI/testing/evm @@ -57,3 +57,16 @@ Description: dracut (via 97masterkey and 98integrity) and systemd (via core/ima-setup) have support for loading keys at boot time. + +What: security/evm_xattrs +Date: April 2018 +Contact: Matthew Garrett <mjg59@google.com> +Description: + Shows the set of extended attributes used to calculate or + validate the EVM signature, and allows additional attributes + to be added at runtime. Any signatures generated after + additional attributes are added (and on files posessing those + additional attributes) will only be valid if the same + additional attributes are configured on system boot. Writing + a single period (.) will lock the xattr list from any further + modification. diff --git a/include/uapi/linux/audit.h b/include/uapi/linux/audit.h index 4e61a9e05132..65d9293f1fb8 100644 --- a/include/uapi/linux/audit.h +++ b/include/uapi/linux/audit.h @@ -147,6 +147,7 @@ #define AUDIT_INTEGRITY_HASH 1803 /* Integrity HASH type */ #define AUDIT_INTEGRITY_PCR 1804 /* PCR invalidation msgs */ #define AUDIT_INTEGRITY_RULE 1805 /* policy rule */ +#define AUDIT_INTEGRITY_EVM_XATTR 1806 /* New EVM-covered xattr */ #define AUDIT_KERNEL 2000 /* Asynchronous audit record. NOT A REQUEST. */ diff --git a/security/integrity/evm/Kconfig b/security/integrity/evm/Kconfig index e825e0ae78e7..54adb3f9ad1d 100644 --- a/security/integrity/evm/Kconfig +++ b/security/integrity/evm/Kconfig @@ -42,6 +42,17 @@ config EVM_EXTRA_SMACK_XATTRS additional info to the calculation, requires existing EVM labeled file systems to be relabeled. +config EVM_ADD_XATTRS + bool "Add additional EVM extended attributes at runtime" + depends on EVM + default n + help + Allow userland to provide additional xattrs for HMAC calculation. + + When this option is enabled, root can add additional xattrs to the + list used by EVM by writing them into + /sys/kernel/security/evm_xattrs. + config EVM_LOAD_X509 bool "Load an X509 certificate onto the '.evm' trusted keyring" depends on EVM && INTEGRITY_TRUSTED_KEYRING diff --git a/security/integrity/evm/evm_crypto.c b/security/integrity/evm/evm_crypto.c index caeea20670cc..494da5fcc092 100644 --- a/security/integrity/evm/evm_crypto.c +++ b/security/integrity/evm/evm_crypto.c @@ -208,7 +208,7 @@ static int evm_calc_hmac_or_hash(struct dentry *dentry, return PTR_ERR(desc); error = -ENODATA; - list_for_each_entry(xattr, &evm_config_xattrnames, list) { + list_for_each_entry_rcu(xattr, &evm_config_xattrnames, list) { bool is_ima = false; if (strcmp(xattr->name, XATTR_NAME_IMA) == 0) diff --git a/security/integrity/evm/evm_main.c b/security/integrity/evm/evm_main.c index 09582d4fc4a8..f9eff5041e4c 100644 --- a/security/integrity/evm/evm_main.c +++ b/security/integrity/evm/evm_main.c @@ -35,7 +35,7 @@ static const char * const integrity_status_msg[] = { }; int evm_hmac_attrs; -static struct xattr_list evm_config_default_xattrnames[] __ro_after_init = { +static struct xattr_list evm_config_default_xattrnames[] = { #ifdef CONFIG_SECURITY_SELINUX {.name = XATTR_NAME_SELINUX}, #endif @@ -101,7 +101,7 @@ static int evm_find_protected_xattrs(struct dentry *dentry) if (!(inode->i_opflags & IOP_XATTR)) return -EOPNOTSUPP; - list_for_each_entry(xattr, &evm_config_xattrnames, list) { + list_for_each_entry_rcu(xattr, &evm_config_xattrnames, list) { error = __vfs_getxattr(dentry, inode, xattr->name, NULL, 0); if (error < 0) { if (error == -ENODATA) @@ -228,7 +228,7 @@ static int evm_protected_xattr(const char *req_xattr_name) struct xattr_list *xattr; namelen = strlen(req_xattr_name); - list_for_each_entry(xattr, &evm_config_xattrnames, list) { + list_for_each_entry_rcu(xattr, &evm_config_xattrnames, list) { if ((strlen(xattr->name) == namelen) && (strncmp(req_xattr_name, xattr->name, namelen) == 0)) { found = 1; diff --git a/security/integrity/evm/evm_secfs.c b/security/integrity/evm/evm_secfs.c index e44380f0cb45..97e975d91241 100644 --- a/security/integrity/evm/evm_secfs.c +++ b/security/integrity/evm/evm_secfs.c @@ -15,14 +15,22 @@ #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt +#include <linux/audit.h> #include <linux/uaccess.h> #include <linux/module.h> +#include <linux/mutex.h> #include "evm.h" static struct dentry *evm_dir; static struct dentry *evm_init_tpm; static struct dentry *evm_symlink; +#ifdef CONFIG_EVM_ADD_XATTRS +static struct dentry *evm_xattrs; +static DEFINE_MUTEX(xattr_list_mutex); +static int evm_xattrs_locked; +#endif + /** * evm_read_key - read() for <securityfs>/evm * @@ -109,6 +117,166 @@ static const struct file_operations evm_key_ops = { .write = evm_write_key, }; +#ifdef CONFIG_EVM_ADD_XATTRS +/** + * evm_read_xattrs - read() for <securityfs>/evm_xattrs + * + * @filp: file pointer, not actually used + * @buf: where to put the result + * @count: maximum to send along + * @ppos: where to start + * + * Returns number of bytes read or error code, as appropriate + */ +static ssize_t evm_read_xattrs(struct file *filp, char __user *buf, + size_t count, loff_t *ppos) +{ + char *temp; + int offset = 0; + ssize_t rc, size = 0; + struct xattr_list *xattr; + + if (*ppos != 0) + return 0; + + rc = mutex_lock_interruptible(&xattr_list_mutex); + if (rc) + return -ERESTARTSYS; + + list_for_each_entry(xattr, &evm_config_xattrnames, list) + size += strlen(xattr->name) + 1; + + temp = kmalloc(size + 1, GFP_KERNEL); + if (!temp) + return -ENOMEM; + + list_for_each_entry(xattr, &evm_config_xattrnames, list) { + sprintf(temp + offset, "%s\n", xattr->name); + offset += strlen(xattr->name) + 1; + } + + mutex_unlock(&xattr_list_mutex); + rc = simple_read_from_buffer(buf, count, ppos, temp, strlen(temp)); + + return rc; +} + +/** + * evm_write_xattrs - write() for <securityfs>/evm_xattrs + * @file: file pointer, not actually used + * @buf: where to get the data from + * @count: bytes sent + * @ppos: where to start + * + * Returns number of bytes written or error code, as appropriate + */ +static ssize_t evm_write_xattrs(struct file *file, const char __user *buf, + size_t count, loff_t *ppos) +{ + int len, err; + struct xattr_list *xattr, *tmp; + struct audit_buffer *ab; + struct iattr newattrs; + struct inode *inode; + + if (!capable(CAP_SYS_ADMIN) || evm_xattrs_locked) + return -EPERM; + + if (*ppos != 0) + return -EINVAL; + + if (count > XATTR_NAME_MAX) + return -E2BIG; + + ab = audit_log_start(NULL, GFP_KERNEL, AUDIT_INTEGRITY_EVM_XATTR); + if (IS_ERR(ab)) + return PTR_ERR(ab); + + xattr = kmalloc(sizeof(struct xattr_list), GFP_KERNEL); + if (!xattr) { + err = -ENOMEM; + goto out; + } + + xattr->name = memdup_user_nul(buf, count); + if (IS_ERR(xattr->name)) { + err = PTR_ERR(xattr->name); + xattr->name = NULL; + goto out; + } + + /* Remove any trailing newline */ + len = strlen(xattr->name); + if (xattr->name[len-1] == '\n') + xattr->name[len-1] = '\0'; + + if (strcmp(xattr->name, ".") == 0) { + evm_xattrs_locked = 1; + inode = evm_xattrs->d_inode; + inode_lock(inode); + newattrs.ia_mode = S_IFREG | 0440; + newattrs.ia_valid = ATTR_MODE; + err = notify_change(evm_xattrs, &newattrs, NULL); + inode_unlock(inode); + audit_log_format(ab, "locked"); + if (!err) + err = count; + goto out; + } + + audit_log_format(ab, "xattr="); + audit_log_untrustedstring(ab, xattr->name); + + if (strncmp(xattr->name, XATTR_SECURITY_PREFIX, + XATTR_SECURITY_PREFIX_LEN) != 0) { + err = -EINVAL; + goto out; + } + + /* Guard against races in evm_read_xattrs */ + mutex_lock(&xattr_list_mutex); + list_for_each_entry(tmp, &evm_config_xattrnames, list) { + if (strcmp(xattr->name, tmp->name) == 0) { + err = -EEXIST; + mutex_unlock(&xattr_list_mutex); + goto out; + } + } + list_add_tail_rcu(&xattr->list, &evm_config_xattrnames); + mutex_unlock(&xattr_list_mutex); + + audit_log_format(ab, " res=0"); + audit_log_end(ab); + return count; +out: + audit_log_format(ab, " res=%d", err); + audit_log_end(ab); + kfree(xattr->name); + kfree(xattr); + return err; +} + +static const struct file_operations evm_xattr_ops = { + .read = evm_read_xattrs, + .write = evm_write_xattrs, +}; + +static int evm_init_xattrs(void) +{ + evm_xattrs = securityfs_create_file("evm_xattrs", 0660, evm_dir, NULL, + &evm_xattr_ops); + if (!evm_xattrs || IS_ERR(evm_xattrs)) + return -EFAULT; + + return 0; +} +#else +static int evm_init_xattrs(void) +{ + return 0; +} +#endif + int __init evm_init_secfs(void) { int error = 0; @@ -131,6 +299,11 @@ int __init evm_init_secfs(void) goto out; } + if (evm_init_xattrs() != 0) { + error = -EFAULT; + goto out; + } + return 0; out: securityfs_remove(evm_symlink);
Sites may wish to provide additional metadata alongside files in order to make more fine-grained security decisions[1]. The security of this is enhanced if this metadata is protected, something that EVM makes possible. However, the kernel cannot know about the set of extended attributes that local admins may wish to protect, and hardcoding this policy in the kernel makes it difficult to change over time and less convenient for distributions to enable. This patch adds a new /sys/kernel/security/evm_xattrs node, which can be read to obtain the current set of EVM-protected extended attributes or written to in order to add new entries. Extending this list will not change the validity of any existing signatures provided that the file in question does not have any of the additional extended attributes - missing xattrs are skipped when calculating the EVM hash. [1] For instance, a package manager could install information about the package uploader in an additional extended attribute. Local LSM policy could then be associated with that extended attribute in order to restrict the privileges available to packages from less trusted uploaders. Signed-off-by: Matthew Garrett <mjg59@google.com> --- I think this covers the review comments. Documentation/ABI/testing/evm | 13 +++ include/uapi/linux/audit.h | 1 + security/integrity/evm/Kconfig | 11 ++ security/integrity/evm/evm_crypto.c | 2 +- security/integrity/evm/evm_main.c | 6 +- security/integrity/evm/evm_secfs.c | 173 ++++++++++++++++++++++++++++ 6 files changed, 202 insertions(+), 4 deletions(-)