| Message ID | 20200812191525.1120850-5-dburgener@linux.microsoft.com (mailing list archive) |
|---|---|
| State | Changes Requested |
| Delegated to: | Paul Moore |
| Headers | show |
| Series | Update SELinuxfs out of tree and then swapover | expand |
On 8/12/20 3:15 PM, Daniel Burgener wrote: > In order to avoid concurrency issues around selinuxfs resource availability > during policy load, we first create new directories out of tree for > reloaded resources, then swap them in, and finally delete the old versions. > > This fix focuses on concurrency in each of the three subtrees swapped, and > not concurrency across the three trees. This means that it is still possible > that subsequent reads to eg the booleans directory and the class directory > during a policy load could see the old state for one and the new for the other. > The problem of ensuring that policy loads are fully atomic from the perspective > of userspace is larger than what is dealt with here. This commit focuses on > ensuring that the directories contents always match either the new or the old > policy state from the perspective of userspace. > > In the previous implementation, on policy load /sys/fs/selinux is updated > by deleting the previous contents of > /sys/fs/selinux/{class,booleans} and then recreating them. This means > that there is a period of time when the contents of these directories do not > exist which can cause race conditions as userspace relies on them for > information about the policy. In addition, it means that error recovery in > the event of failure is challenging. > > In order to demonstrate the race condition that this series fixes, you > can use the following commands: > > while true; do cat /sys/fs/selinux/class/service/perms/status >> /dev/null; done & > while true; do load_policy; done; > > In the existing code, this will display errors fairly often as the class > lookup fails. (In normal operation from systemd, this would result in a > permission check which would be allowed or denied based on policy settings > around unknown object classes.) After applying this patch series you > should expect to no longer see such error messages. > > Signed-off-by: Daniel Burgener <dburgener@linux.microsoft.com> > --- > security/selinux/selinuxfs.c | 145 +++++++++++++++++++++++++++++------ > 1 file changed, 120 insertions(+), 25 deletions(-) > > diff --git a/security/selinux/selinuxfs.c b/security/selinux/selinuxfs.c > index f09afdb90ddd..d3a19170210a 100644 > --- a/security/selinux/selinuxfs.c > +++ b/security/selinux/selinuxfs.c > + tmp_policycap_dir = sel_make_dir(tmp_parent, POLICYCAP_DIR_NAME, &fsi->last_ino); > + if (IS_ERR(tmp_policycap_dir)) { > + ret = PTR_ERR(tmp_policycap_dir); > + goto out; > + } No need to re-create this one. > - return 0; > + // booleans > + old_dentry = fsi->bool_dir; > + lock_rename(tmp_bool_dir, old_dentry); > + ret = vfs_rename(tmp_parent->d_inode, tmp_bool_dir, fsi->sb->s_root->d_inode, > + fsi->bool_dir, NULL, RENAME_EXCHANGE); One issue with using vfs_rename() is that it will trigger all of the permission checks associated with renaming, and previously this was never required for selinuxfs and therefore might not be allowed in some policies even to a process allowed to reload policy. So if you need to do this, you may want to override creds around this call to use the init cred (which will still require allowing it to the kernel domain but not necessarily to the process that is performing the policy load). The other issue is that you then have to implement a rename inode operation and thus technically it is possible for userspace to also attempt renames on selinuxfs to the extent allowed by policy. I see that debugfs has a debugfs_rename() that internally uses simple_rename() but I guess that doesn't cover the RENAME_EXCHANGE case. > + // Since the other temporary dirs are children of tmp_parent > + // this will handle all the cleanup in the case of a failure before > + // the swapover Don't use // style comments please, especially not for multi-line comments. I think they are only used in selinux for the script-generated license lines. > +static struct dentry *sel_make_disconnected_dir(struct super_block *sb, > + unsigned long *ino) > +{ > + struct inode *inode = sel_make_inode(sb, S_IFDIR | S_IRUGO | S_IXUGO); > + > + if (!inode) > + return ERR_PTR(-ENOMEM); > + > + inode->i_op = &sel_dir_inode_operations; > + inode->i_fop = &simple_dir_operations; > + inode->i_ino = ++(*ino); > + /* directory inodes start off with i_nlink == 2 (for "." entry) */ > + inc_nlink(inode); > + return d_obtain_alias(inode); > +} > + Since you are always incrementing the last_ino counter and never reusing the ones for the removed inodes, you could technically eventually end up with one of these directories have the same inode number as one of the inodes whose inode numbers are generated in specific ranges (i.e. for initial_contexts, booleans, classes, and policy capabilities). Optimally we'd just reuse the inode number for the inode we are replacing?
On 8/13/20 12:25 PM, Stephen Smalley wrote: > On 8/12/20 3:15 PM, Daniel Burgener wrote: > >> In order to avoid concurrency issues around selinuxfs resource >> availability >> during policy load, we first create new directories out of tree for >> reloaded resources, then swap them in, and finally delete the old >> versions. >> >> This fix focuses on concurrency in each of the three subtrees >> swapped, and >> not concurrency across the three trees. This means that it is still >> possible >> that subsequent reads to eg the booleans directory and the class >> directory >> during a policy load could see the old state for one and the new for >> the other. >> The problem of ensuring that policy loads are fully atomic from the >> perspective >> of userspace is larger than what is dealt with here. This commit >> focuses on >> ensuring that the directories contents always match either the new or >> the old >> policy state from the perspective of userspace. >> >> In the previous implementation, on policy load /sys/fs/selinux is >> updated >> by deleting the previous contents of >> /sys/fs/selinux/{class,booleans} and then recreating them. This means >> that there is a period of time when the contents of these directories >> do not >> exist which can cause race conditions as userspace relies on them for >> information about the policy. In addition, it means that error >> recovery in >> the event of failure is challenging. >> >> In order to demonstrate the race condition that this series fixes, you >> can use the following commands: >> >> while true; do cat /sys/fs/selinux/class/service/perms/status >>> /dev/null; done & >> while true; do load_policy; done; >> >> In the existing code, this will display errors fairly often as the class >> lookup fails. (In normal operation from systemd, this would result in a >> permission check which would be allowed or denied based on policy >> settings >> around unknown object classes.) After applying this patch series you >> should expect to no longer see such error messages. >> >> Signed-off-by: Daniel Burgener <dburgener@linux.microsoft.com> >> --- >> security/selinux/selinuxfs.c | 145 +++++++++++++++++++++++++++++------ >> 1 file changed, 120 insertions(+), 25 deletions(-) >> >> diff --git a/security/selinux/selinuxfs.c b/security/selinux/selinuxfs.c >> index f09afdb90ddd..d3a19170210a 100644 >> --- a/security/selinux/selinuxfs.c >> +++ b/security/selinux/selinuxfs.c >> + tmp_policycap_dir = sel_make_dir(tmp_parent, POLICYCAP_DIR_NAME, >> &fsi->last_ino); >> + if (IS_ERR(tmp_policycap_dir)) { >> + ret = PTR_ERR(tmp_policycap_dir); >> + goto out; >> + } > > No need to re-create this one. > >> - return 0; >> + // booleans >> + old_dentry = fsi->bool_dir; >> + lock_rename(tmp_bool_dir, old_dentry); >> + ret = vfs_rename(tmp_parent->d_inode, tmp_bool_dir, >> fsi->sb->s_root->d_inode, >> + fsi->bool_dir, NULL, RENAME_EXCHANGE); > > One issue with using vfs_rename() is that it will trigger all of the > permission checks associated with renaming, and previously this was > never required for selinuxfs and therefore might not be allowed in > some policies even to a process allowed to reload policy. So if you > need to do this, you may want to override creds around this call to > use the init cred (which will still require allowing it to the kernel > domain but not necessarily to the process that is performing the > policy load). The other issue is that you then have to implement a > rename inode operation and thus technically it is possible for > userspace to also attempt renames on selinuxfs to the extent allowed > by policy. I see that debugfs has a debugfs_rename() that internally > uses simple_rename() but I guess that doesn't cover the > RENAME_EXCHANGE case. Those are good points. Do you see any problems with just calling d_exchange() directly? It seems to work fine in very limited initial testing on my end. That should hopefully address all the problems you mentioned here. -Daniel > >> + // Since the other temporary dirs are children of tmp_parent >> + // this will handle all the cleanup in the case of a failure before >> + // the swapover > > Don't use // style comments please, especially not for multi-line > comments. I think they are only used in selinux for the > script-generated license lines. > >> +static struct dentry *sel_make_disconnected_dir(struct super_block *sb, >> + unsigned long *ino) >> +{ >> + struct inode *inode = sel_make_inode(sb, S_IFDIR | S_IRUGO | >> S_IXUGO); >> + >> + if (!inode) >> + return ERR_PTR(-ENOMEM); >> + >> + inode->i_op = &sel_dir_inode_operations; >> + inode->i_fop = &simple_dir_operations; >> + inode->i_ino = ++(*ino); >> + /* directory inodes start off with i_nlink == 2 (for "." entry) */ >> + inc_nlink(inode); >> + return d_obtain_alias(inode); >> +} >> + > > Since you are always incrementing the last_ino counter and never > reusing the ones for the removed inodes, you could technically > eventually end up with one of these directories have the same inode > number as one of the inodes whose inode numbers are generated in > specific ranges (i.e. for initial_contexts, booleans, classes, and > policy capabilities). Optimally we'd just reuse the inode number for > the inode we are replacing? >
On Tue, Aug 18, 2020 at 9:49 AM Daniel Burgener <dburgener@linux.microsoft.com> wrote: > > On 8/13/20 12:25 PM, Stephen Smalley wrote: > > On 8/12/20 3:15 PM, Daniel Burgener wrote: > > > >> In order to avoid concurrency issues around selinuxfs resource > >> availability > >> during policy load, we first create new directories out of tree for > >> reloaded resources, then swap them in, and finally delete the old > >> versions. > >> > >> This fix focuses on concurrency in each of the three subtrees > >> swapped, and > >> not concurrency across the three trees. This means that it is still > >> possible > >> that subsequent reads to eg the booleans directory and the class > >> directory > >> during a policy load could see the old state for one and the new for > >> the other. > >> The problem of ensuring that policy loads are fully atomic from the > >> perspective > >> of userspace is larger than what is dealt with here. This commit > >> focuses on > >> ensuring that the directories contents always match either the new or > >> the old > >> policy state from the perspective of userspace. > >> > >> In the previous implementation, on policy load /sys/fs/selinux is > >> updated > >> by deleting the previous contents of > >> /sys/fs/selinux/{class,booleans} and then recreating them. This means > >> that there is a period of time when the contents of these directories > >> do not > >> exist which can cause race conditions as userspace relies on them for > >> information about the policy. In addition, it means that error > >> recovery in > >> the event of failure is challenging. > >> > >> In order to demonstrate the race condition that this series fixes, you > >> can use the following commands: > >> > >> while true; do cat /sys/fs/selinux/class/service/perms/status > >>> /dev/null; done & > >> while true; do load_policy; done; > >> > >> In the existing code, this will display errors fairly often as the class > >> lookup fails. (In normal operation from systemd, this would result in a > >> permission check which would be allowed or denied based on policy > >> settings > >> around unknown object classes.) After applying this patch series you > >> should expect to no longer see such error messages. > >> > >> Signed-off-by: Daniel Burgener <dburgener@linux.microsoft.com> > >> --- > >> security/selinux/selinuxfs.c | 145 +++++++++++++++++++++++++++++------ > >> 1 file changed, 120 insertions(+), 25 deletions(-) > >> > >> diff --git a/security/selinux/selinuxfs.c b/security/selinux/selinuxfs.c > >> index f09afdb90ddd..d3a19170210a 100644 > >> --- a/security/selinux/selinuxfs.c > >> +++ b/security/selinux/selinuxfs.c > >> + tmp_policycap_dir = sel_make_dir(tmp_parent, POLICYCAP_DIR_NAME, > >> &fsi->last_ino); > >> + if (IS_ERR(tmp_policycap_dir)) { > >> + ret = PTR_ERR(tmp_policycap_dir); > >> + goto out; > >> + } > > > > No need to re-create this one. > > > >> - return 0; > >> + // booleans > >> + old_dentry = fsi->bool_dir; > >> + lock_rename(tmp_bool_dir, old_dentry); > >> + ret = vfs_rename(tmp_parent->d_inode, tmp_bool_dir, > >> fsi->sb->s_root->d_inode, > >> + fsi->bool_dir, NULL, RENAME_EXCHANGE); > > > > One issue with using vfs_rename() is that it will trigger all of the > > permission checks associated with renaming, and previously this was > > never required for selinuxfs and therefore might not be allowed in > > some policies even to a process allowed to reload policy. So if you > > need to do this, you may want to override creds around this call to > > use the init cred (which will still require allowing it to the kernel > > domain but not necessarily to the process that is performing the > > policy load). The other issue is that you then have to implement a > > rename inode operation and thus technically it is possible for > > userspace to also attempt renames on selinuxfs to the extent allowed > > by policy. I see that debugfs has a debugfs_rename() that internally > > uses simple_rename() but I guess that doesn't cover the > > RENAME_EXCHANGE case. > > Those are good points. Do you see any problems with just calling > d_exchange() directly? It seems to work fine in very limited initial > testing on my end. That should hopefully address all the problems you > mentioned here. I was hoping the vfs folks would chime in but you may have to pose a more direct question to viro and linux-fsdevel to get a response. Possibly there should be a lower-level vfs helper that could be used internally by vfs_rename() and by things like debugfs_rename and a potential selinuxfs_rename.
On 8/18/20 9:55 AM, Stephen Smalley wrote: > On Tue, Aug 18, 2020 at 9:49 AM Daniel Burgener > <dburgener@linux.microsoft.com> wrote: >> On 8/13/20 12:25 PM, Stephen Smalley wrote: >>> On 8/12/20 3:15 PM, Daniel Burgener wrote: >>> >>>> In order to avoid concurrency issues around selinuxfs resource >>>> availability >>>> during policy load, we first create new directories out of tree for >>>> reloaded resources, then swap them in, and finally delete the old >>>> versions. >>>> >>>> This fix focuses on concurrency in each of the three subtrees >>>> swapped, and >>>> not concurrency across the three trees. This means that it is still >>>> possible >>>> that subsequent reads to eg the booleans directory and the class >>>> directory >>>> during a policy load could see the old state for one and the new for >>>> the other. >>>> The problem of ensuring that policy loads are fully atomic from the >>>> perspective >>>> of userspace is larger than what is dealt with here. This commit >>>> focuses on >>>> ensuring that the directories contents always match either the new or >>>> the old >>>> policy state from the perspective of userspace. >>>> >>>> In the previous implementation, on policy load /sys/fs/selinux is >>>> updated >>>> by deleting the previous contents of >>>> /sys/fs/selinux/{class,booleans} and then recreating them. This means >>>> that there is a period of time when the contents of these directories >>>> do not >>>> exist which can cause race conditions as userspace relies on them for >>>> information about the policy. In addition, it means that error >>>> recovery in >>>> the event of failure is challenging. >>>> >>>> In order to demonstrate the race condition that this series fixes, you >>>> can use the following commands: >>>> >>>> while true; do cat /sys/fs/selinux/class/service/perms/status >>>>> /dev/null; done & >>>> while true; do load_policy; done; >>>> >>>> In the existing code, this will display errors fairly often as the class >>>> lookup fails. (In normal operation from systemd, this would result in a >>>> permission check which would be allowed or denied based on policy >>>> settings >>>> around unknown object classes.) After applying this patch series you >>>> should expect to no longer see such error messages. >>>> >>>> Signed-off-by: Daniel Burgener <dburgener@linux.microsoft.com> >>>> --- >>>> security/selinux/selinuxfs.c | 145 +++++++++++++++++++++++++++++------ >>>> 1 file changed, 120 insertions(+), 25 deletions(-) >>>> >>>> diff --git a/security/selinux/selinuxfs.c b/security/selinux/selinuxfs.c >>>> index f09afdb90ddd..d3a19170210a 100644 >>>> --- a/security/selinux/selinuxfs.c >>>> +++ b/security/selinux/selinuxfs.c >>>> + tmp_policycap_dir = sel_make_dir(tmp_parent, POLICYCAP_DIR_NAME, >>>> &fsi->last_ino); >>>> + if (IS_ERR(tmp_policycap_dir)) { >>>> + ret = PTR_ERR(tmp_policycap_dir); >>>> + goto out; >>>> + } >>> No need to re-create this one. >>> >>>> - return 0; >>>> + // booleans >>>> + old_dentry = fsi->bool_dir; >>>> + lock_rename(tmp_bool_dir, old_dentry); >>>> + ret = vfs_rename(tmp_parent->d_inode, tmp_bool_dir, >>>> fsi->sb->s_root->d_inode, >>>> + fsi->bool_dir, NULL, RENAME_EXCHANGE); >>> One issue with using vfs_rename() is that it will trigger all of the >>> permission checks associated with renaming, and previously this was >>> never required for selinuxfs and therefore might not be allowed in >>> some policies even to a process allowed to reload policy. So if you >>> need to do this, you may want to override creds around this call to >>> use the init cred (which will still require allowing it to the kernel >>> domain but not necessarily to the process that is performing the >>> policy load). The other issue is that you then have to implement a >>> rename inode operation and thus technically it is possible for >>> userspace to also attempt renames on selinuxfs to the extent allowed >>> by policy. I see that debugfs has a debugfs_rename() that internally >>> uses simple_rename() but I guess that doesn't cover the >>> RENAME_EXCHANGE case. >> Those are good points. Do you see any problems with just calling >> d_exchange() directly? It seems to work fine in very limited initial >> testing on my end. That should hopefully address all the problems you >> mentioned here. > I was hoping the vfs folks would chime in but you may have to pose a > more direct question to viro and linux-fsdevel to get a response. > Possibly there should be a lower-level vfs helper that could be used > internally by vfs_rename() and by things like debugfs_rename and a > potential selinuxfs_rename. I'll send a v3 with this switched to d_exchange() and the other issues you mentioned fixed first. That way they can at least look at the latest version of things to help focus any conversation and save people time. -Daniel
diff --git a/security/selinux/selinuxfs.c b/security/selinux/selinuxfs.c index f09afdb90ddd..d3a19170210a 100644 --- a/security/selinux/selinuxfs.c +++ b/security/selinux/selinuxfs.c @@ -20,6 +20,7 @@ #include <linux/fs_context.h> #include <linux/mount.h> #include <linux/mutex.h> +#include <linux/namei.h> #include <linux/init.h> #include <linux/string.h> #include <linux/security.h> @@ -361,7 +362,11 @@ static int sel_make_classes(struct selinux_policy *newpolicy, static struct dentry *sel_make_dir(struct dentry *dir, const char *name, unsigned long *ino); -/* declaration for sel_remove_old_policy_nodes */ +/* declaration for sel_make_policy_nodes */ +static struct dentry *sel_make_disconnected_dir(struct super_block *sb, + unsigned long *ino); + +/* declaration for sel_make_policy_nodes */ static void sel_remove_entries(struct dentry *de); static ssize_t sel_read_mls(struct file *filp, char __user *buf, @@ -518,50 +523,108 @@ static const struct file_operations sel_policy_ops = { .llseek = generic_file_llseek, }; -static void sel_remove_old_policy_nodes(struct selinux_fs_info *fsi) +static void sel_remove_old_bool_data(unsigned int bool_num, char **bool_names, + unsigned int *bool_values) { u32 i; /* bool_dir cleanup */ - for (i = 0; i < fsi->bool_num; i++) - kfree(fsi->bool_pending_names[i]); - kfree(fsi->bool_pending_names); - kfree(fsi->bool_pending_values); - fsi->bool_num = 0; - fsi->bool_pending_names = NULL; - fsi->bool_pending_values = NULL; - - sel_remove_entries(fsi->bool_dir); - - /* class_dir cleanup */ - sel_remove_entries(fsi->class_dir); - - /* policycap_dir cleanup */ - sel_remove_entries(fsi->policycap_dir); + for (i = 0; i < bool_num; i++) + kfree(bool_names[i]); + kfree(bool_names); + kfree(bool_values); } static int sel_make_policy_nodes(struct selinux_fs_info *fsi, struct selinux_policy *newpolicy) { - int ret; + int ret = 0; + struct dentry *tmp_parent, *tmp_bool_dir, *tmp_class_dir, *tmp_policycap_dir, *old_dentry; + unsigned int tmp_bool_num, old_bool_num; + char **tmp_bool_names, **old_bool_names; + unsigned int *tmp_bool_values, *old_bool_values; + + tmp_parent = sel_make_disconnected_dir(fsi->sb, &fsi->last_ino); + if (IS_ERR(tmp_parent)) + return PTR_ERR(tmp_parent); - sel_remove_old_policy_nodes(fsi); + tmp_bool_dir = sel_make_dir(tmp_parent, BOOL_DIR_NAME, &fsi->last_ino); + if (IS_ERR(tmp_bool_dir)) { + ret = PTR_ERR(tmp_bool_dir); + goto out; + } + + tmp_class_dir = sel_make_dir(tmp_parent, CLASS_DIR_NAME, &fsi->last_ino); + if (IS_ERR(tmp_class_dir)) { + ret = PTR_ERR(tmp_class_dir); + goto out; + } + + tmp_policycap_dir = sel_make_dir(tmp_parent, POLICYCAP_DIR_NAME, &fsi->last_ino); + if (IS_ERR(tmp_policycap_dir)) { + ret = PTR_ERR(tmp_policycap_dir); + goto out; + } - ret = sel_make_bools(newpolicy, fsi->bool_dir, &fsi->bool_num, - &fsi->bool_pending_names, &fsi->bool_pending_values); + ret = sel_make_bools(newpolicy, tmp_bool_dir, &tmp_bool_num, + &tmp_bool_names, &tmp_bool_values); if (ret) { pr_err("SELinux: failed to load policy booleans\n"); - return ret; + goto out; } - ret = sel_make_classes(newpolicy, fsi->class_dir, + ret = sel_make_classes(newpolicy, tmp_class_dir, &fsi->last_class_ino); if (ret) { pr_err("SELinux: failed to load policy classes\n"); - return ret; + goto out; } - return 0; + // booleans + old_dentry = fsi->bool_dir; + lock_rename(tmp_bool_dir, old_dentry); + ret = vfs_rename(tmp_parent->d_inode, tmp_bool_dir, fsi->sb->s_root->d_inode, + fsi->bool_dir, NULL, RENAME_EXCHANGE); + if (ret) { + pr_err("Failed to update boolean directory\n"); + unlock_rename(tmp_bool_dir, old_dentry); + goto out; + } + + old_bool_num = fsi->bool_num; + old_bool_names = fsi->bool_pending_names; + old_bool_values = fsi->bool_pending_values; + + fsi->bool_num = tmp_bool_num; + fsi->bool_pending_names = tmp_bool_names; + fsi->bool_pending_values = tmp_bool_values; + + sel_remove_old_bool_data(old_bool_num, old_bool_names, old_bool_values); + + fsi->bool_dir = tmp_bool_dir; + unlock_rename(tmp_bool_dir, old_dentry); + + // classes + old_dentry = fsi->class_dir; + lock_rename(tmp_class_dir, old_dentry); + ret = vfs_rename(tmp_parent->d_inode, tmp_class_dir, fsi->sb->s_root->d_inode, + fsi->class_dir, NULL, RENAME_EXCHANGE); + if (ret) { + pr_err("Failed to update class directory\n"); + unlock_rename(tmp_class_dir, old_dentry); + goto out; + } + fsi->class_dir = tmp_class_dir; + unlock_rename(tmp_class_dir, old_dentry); + +out: + // Since the other temporary dirs are children of tmp_parent + // this will handle all the cleanup in the case of a failure before + // the swapover + sel_remove_entries(tmp_parent); + dput(tmp_parent); // d_genocide() only handles the children + + return ret; } static ssize_t sel_write_load(struct file *file, const char __user *buf, @@ -1984,6 +2047,38 @@ static struct dentry *sel_make_dir(struct dentry *dir, const char *name, return dentry; } +// vfs_rename() requires a rename operation to be defined. +int sel_rename(struct inode *old_dir, struct dentry *old_dentry, + struct inode *new_dir, struct dentry *new_dentry, + unsigned int flags) +{ + if (!(flags & RENAME_EXCHANGE)) + return -EINVAL; + + return 0; +} + +const struct inode_operations sel_dir_inode_operations = { + .lookup = simple_lookup, + .rename = sel_rename, +}; + +static struct dentry *sel_make_disconnected_dir(struct super_block *sb, + unsigned long *ino) +{ + struct inode *inode = sel_make_inode(sb, S_IFDIR | S_IRUGO | S_IXUGO); + + if (!inode) + return ERR_PTR(-ENOMEM); + + inode->i_op = &sel_dir_inode_operations; + inode->i_fop = &simple_dir_operations; + inode->i_ino = ++(*ino); + /* directory inodes start off with i_nlink == 2 (for "." entry) */ + inc_nlink(inode); + return d_obtain_alias(inode); +} + #define NULL_FILE_NAME "null" static int sel_fill_super(struct super_block *sb, struct fs_context *fc)
In order to avoid concurrency issues around selinuxfs resource availability during policy load, we first create new directories out of tree for reloaded resources, then swap them in, and finally delete the old versions. This fix focuses on concurrency in each of the three subtrees swapped, and not concurrency across the three trees. This means that it is still possible that subsequent reads to eg the booleans directory and the class directory during a policy load could see the old state for one and the new for the other. The problem of ensuring that policy loads are fully atomic from the perspective of userspace is larger than what is dealt with here. This commit focuses on ensuring that the directories contents always match either the new or the old policy state from the perspective of userspace. In the previous implementation, on policy load /sys/fs/selinux is updated by deleting the previous contents of /sys/fs/selinux/{class,booleans} and then recreating them. This means that there is a period of time when the contents of these directories do not exist which can cause race conditions as userspace relies on them for information about the policy. In addition, it means that error recovery in the event of failure is challenging. In order to demonstrate the race condition that this series fixes, you can use the following commands: while true; do cat /sys/fs/selinux/class/service/perms/status >/dev/null; done & while true; do load_policy; done; In the existing code, this will display errors fairly often as the class lookup fails. (In normal operation from systemd, this would result in a permission check which would be allowed or denied based on policy settings around unknown object classes.) After applying this patch series you should expect to no longer see such error messages. Signed-off-by: Daniel Burgener <dburgener@linux.microsoft.com> --- security/selinux/selinuxfs.c | 145 +++++++++++++++++++++++++++++------ 1 file changed, 120 insertions(+), 25 deletions(-)