From fdc85222d58eb6e19ac11181024ac766b19dc470 Mon Sep 17 00:00:00 2001 From: Daniel Xu Date: Thu, 12 Mar 2020 13:03:14 -0700 Subject: kernfs: kvmalloc xattr value instead of kmalloc xattr values have a 64k maximum size. This can result in an order 4 kmalloc request which can be difficult to fulfill. Since xattrs do not need physically contiguous memory, we can switch to kvmalloc and not have to worry about higher order allocations failing. Signed-off-by: Daniel Xu Acked-by: Chris Down Reviewed-by: Andreas Dilger Reviewed-by: Greg Kroah-Hartman Signed-off-by: Tejun Heo --- fs/xattr.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/xattr.c b/fs/xattr.c index 90dd78f0eb27..0d3c9b4d1914 100644 --- a/fs/xattr.c +++ b/fs/xattr.c @@ -817,7 +817,7 @@ struct simple_xattr *simple_xattr_alloc(const void *value, size_t size) if (len < sizeof(*new_xattr)) return NULL; - new_xattr = kmalloc(len, GFP_KERNEL); + new_xattr = kvmalloc(len, GFP_KERNEL); if (!new_xattr) return NULL; @@ -882,7 +882,7 @@ int simple_xattr_set(struct simple_xattrs *xattrs, const char *name, new_xattr->name = kstrdup(name, GFP_KERNEL); if (!new_xattr->name) { - kfree(new_xattr); + kvfree(new_xattr); return -ENOMEM; } } @@ -912,7 +912,7 @@ out: spin_unlock(&xattrs->lock); if (xattr) { kfree(xattr->name); - kfree(xattr); + kvfree(xattr); } return err; -- cgit From a46a22955bae16fc5a756af7188d3ccb25c3f797 Mon Sep 17 00:00:00 2001 From: Daniel Xu Date: Thu, 12 Mar 2020 13:03:15 -0700 Subject: kernfs: Add removed_size out param for simple_xattr_set This helps set up size accounting in the next commit. Without this out param, it's difficult to find out the removed xattr size without taking a lock for longer and walking the xattr linked list twice. Signed-off-by: Daniel Xu Acked-by: Chris Down Reviewed-by: Greg Kroah-Hartman Signed-off-by: Tejun Heo --- fs/kernfs/inode.c | 2 +- fs/xattr.c | 11 ++++++++++- 2 files changed, 11 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/kernfs/inode.c b/fs/kernfs/inode.c index d0f7a5abd9a9..5f10ae95fbfa 100644 --- a/fs/kernfs/inode.c +++ b/fs/kernfs/inode.c @@ -303,7 +303,7 @@ int kernfs_xattr_set(struct kernfs_node *kn, const char *name, if (!attrs) return -ENOMEM; - return simple_xattr_set(&attrs->xattrs, name, value, size, flags); + return simple_xattr_set(&attrs->xattrs, name, value, size, flags, NULL); } static int kernfs_vfs_xattr_get(const struct xattr_handler *handler, diff --git a/fs/xattr.c b/fs/xattr.c index 0d3c9b4d1914..e13265e65871 100644 --- a/fs/xattr.c +++ b/fs/xattr.c @@ -860,6 +860,7 @@ int simple_xattr_get(struct simple_xattrs *xattrs, const char *name, * @value: value of the xattr. If %NULL, will remove the attribute. * @size: size of the new xattr * @flags: %XATTR_{CREATE|REPLACE} + * @removed_size: returns size of the removed xattr, -1 if none removed * * %XATTR_CREATE is set, the xattr shouldn't exist already; otherwise fails * with -EEXIST. If %XATTR_REPLACE is set, the xattr should exist; @@ -868,7 +869,8 @@ int simple_xattr_get(struct simple_xattrs *xattrs, const char *name, * Returns 0 on success, -errno on failure. */ int simple_xattr_set(struct simple_xattrs *xattrs, const char *name, - const void *value, size_t size, int flags) + const void *value, size_t size, int flags, + ssize_t *removed_size) { struct simple_xattr *xattr; struct simple_xattr *new_xattr = NULL; @@ -895,8 +897,12 @@ int simple_xattr_set(struct simple_xattrs *xattrs, const char *name, err = -EEXIST; } else if (new_xattr) { list_replace(&xattr->list, &new_xattr->list); + if (removed_size) + *removed_size = xattr->size; } else { list_del(&xattr->list); + if (removed_size) + *removed_size = xattr->size; } goto out; } @@ -908,6 +914,9 @@ int simple_xattr_set(struct simple_xattrs *xattrs, const char *name, list_add(&new_xattr->list, &xattrs->head); xattr = NULL; } + + if (removed_size) + *removed_size = -1; out: spin_unlock(&xattrs->lock); if (xattr) { -- cgit From 0c47383ba3bd10877956e41149d19644fba937d1 Mon Sep 17 00:00:00 2001 From: Daniel Xu Date: Thu, 12 Mar 2020 13:03:16 -0700 Subject: kernfs: Add option to enable user xattrs User extended attributes are useful as metadata storage for kernfs consumers like cgroups. Especially in the case of cgroups, it is useful to have a central metadata store that multiple processes/services can use to coordinate actions. A concrete example is for userspace out of memory killers. We want to let delegated cgroup subtree owners (running as non-root) to be able to say "please avoid killing this cgroup". This is especially important for desktop linux as delegated subtrees owners are less likely to run as root. This patch introduces a new flag, KERNFS_ROOT_SUPPORT_USER_XATTR, that lets kernfs consumers enable user xattr support. An initial limit of 128 entries or 128KB -- whichever is hit first -- is placed per cgroup because xattrs come from kernel memory and we don't want to let unprivileged users accidentally eat up too much kernel memory. Signed-off-by: Daniel Xu Acked-by: Chris Down Reviewed-by: Greg Kroah-Hartman Signed-off-by: Tejun Heo --- fs/kernfs/inode.c | 89 +++++++++++++++++++++++++++++++++++++++++++++ fs/kernfs/kernfs-internal.h | 2 + 2 files changed, 91 insertions(+) (limited to 'fs') diff --git a/fs/kernfs/inode.c b/fs/kernfs/inode.c index 5f10ae95fbfa..fc2469a20fed 100644 --- a/fs/kernfs/inode.c +++ b/fs/kernfs/inode.c @@ -53,6 +53,8 @@ static struct kernfs_iattrs *__kernfs_iattrs(struct kernfs_node *kn, int alloc) kn->iattr->ia_ctime = kn->iattr->ia_atime; simple_xattrs_init(&kn->iattr->xattrs); + atomic_set(&kn->iattr->nr_user_xattrs, 0); + atomic_set(&kn->iattr->user_xattr_size, 0); out_unlock: ret = kn->iattr; mutex_unlock(&iattr_mutex); @@ -327,6 +329,86 @@ static int kernfs_vfs_xattr_set(const struct xattr_handler *handler, return kernfs_xattr_set(kn, name, value, size, flags); } +static int kernfs_vfs_user_xattr_add(struct kernfs_node *kn, + const char *full_name, + struct simple_xattrs *xattrs, + const void *value, size_t size, int flags) +{ + atomic_t *sz = &kn->iattr->user_xattr_size; + atomic_t *nr = &kn->iattr->nr_user_xattrs; + ssize_t removed_size; + int ret; + + if (atomic_inc_return(nr) > KERNFS_MAX_USER_XATTRS) { + ret = -ENOSPC; + goto dec_count_out; + } + + if (atomic_add_return(size, sz) > KERNFS_USER_XATTR_SIZE_LIMIT) { + ret = -ENOSPC; + goto dec_size_out; + } + + ret = simple_xattr_set(xattrs, full_name, value, size, flags, + &removed_size); + + if (!ret && removed_size >= 0) + size = removed_size; + else if (!ret) + return 0; +dec_size_out: + atomic_sub(size, sz); +dec_count_out: + atomic_dec(nr); + return ret; +} + +static int kernfs_vfs_user_xattr_rm(struct kernfs_node *kn, + const char *full_name, + struct simple_xattrs *xattrs, + const void *value, size_t size, int flags) +{ + atomic_t *sz = &kn->iattr->user_xattr_size; + atomic_t *nr = &kn->iattr->nr_user_xattrs; + ssize_t removed_size; + int ret; + + ret = simple_xattr_set(xattrs, full_name, value, size, flags, + &removed_size); + + if (removed_size >= 0) { + atomic_sub(removed_size, sz); + atomic_dec(nr); + } + + return ret; +} + +static int kernfs_vfs_user_xattr_set(const struct xattr_handler *handler, + struct dentry *unused, struct inode *inode, + const char *suffix, const void *value, + size_t size, int flags) +{ + const char *full_name = xattr_full_name(handler, suffix); + struct kernfs_node *kn = inode->i_private; + struct kernfs_iattrs *attrs; + + if (!(kernfs_root(kn)->flags & KERNFS_ROOT_SUPPORT_USER_XATTR)) + return -EOPNOTSUPP; + + attrs = kernfs_iattrs(kn); + if (!attrs) + return -ENOMEM; + + if (value) + return kernfs_vfs_user_xattr_add(kn, full_name, &attrs->xattrs, + value, size, flags); + else + return kernfs_vfs_user_xattr_rm(kn, full_name, &attrs->xattrs, + value, size, flags); + +} + static const struct xattr_handler kernfs_trusted_xattr_handler = { .prefix = XATTR_TRUSTED_PREFIX, .get = kernfs_vfs_xattr_get, @@ -339,8 +421,15 @@ static const struct xattr_handler kernfs_security_xattr_handler = { .set = kernfs_vfs_xattr_set, }; +static const struct xattr_handler kernfs_user_xattr_handler = { + .prefix = XATTR_USER_PREFIX, + .get = kernfs_vfs_xattr_get, + .set = kernfs_vfs_user_xattr_set, +}; + const struct xattr_handler *kernfs_xattr_handlers[] = { &kernfs_trusted_xattr_handler, &kernfs_security_xattr_handler, + &kernfs_user_xattr_handler, NULL }; diff --git a/fs/kernfs/kernfs-internal.h b/fs/kernfs/kernfs-internal.h index 2f3c51d55261..7ee97ef59184 100644 --- a/fs/kernfs/kernfs-internal.h +++ b/fs/kernfs/kernfs-internal.h @@ -26,6 +26,8 @@ struct kernfs_iattrs { struct timespec64 ia_ctime; struct simple_xattrs xattrs; + atomic_t nr_user_xattrs; + atomic_t user_xattr_size; }; /* +1 to avoid triggering overflow warning when negating it */ -- cgit