linux/fs/nfsd/export.h
Yang Erkun f8c989a0c8 nfsd: release svc_expkey/svc_export with rcu_work
The last reference for `cache_head` can be reduced to zero in `c_show`
and `e_show`(using `rcu_read_lock` and `rcu_read_unlock`). Consequently,
`svc_export_put` and `expkey_put` will be invoked, leading to two
issues:

1. The `svc_export_put` will directly free ex_uuid. However,
   `e_show`/`c_show` will access `ex_uuid` after `cache_put`, which can
   trigger a use-after-free issue, shown below.

   ==================================================================
   BUG: KASAN: slab-use-after-free in svc_export_show+0x362/0x430 [nfsd]
   Read of size 1 at addr ff11000010fdc120 by task cat/870

   CPU: 1 UID: 0 PID: 870 Comm: cat Not tainted 6.12.0-rc3+ #1
   Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS
   1.16.1-2.fc37 04/01/2014
   Call Trace:
    <TASK>
    dump_stack_lvl+0x53/0x70
    print_address_description.constprop.0+0x2c/0x3a0
    print_report+0xb9/0x280
    kasan_report+0xae/0xe0
    svc_export_show+0x362/0x430 [nfsd]
    c_show+0x161/0x390 [sunrpc]
    seq_read_iter+0x589/0x770
    seq_read+0x1e5/0x270
    proc_reg_read+0xe1/0x140
    vfs_read+0x125/0x530
    ksys_read+0xc1/0x160
    do_syscall_64+0x5f/0x170
    entry_SYSCALL_64_after_hwframe+0x76/0x7e

   Allocated by task 830:
    kasan_save_stack+0x20/0x40
    kasan_save_track+0x14/0x30
    __kasan_kmalloc+0x8f/0xa0
    __kmalloc_node_track_caller_noprof+0x1bc/0x400
    kmemdup_noprof+0x22/0x50
    svc_export_parse+0x8a9/0xb80 [nfsd]
    cache_do_downcall+0x71/0xa0 [sunrpc]
    cache_write_procfs+0x8e/0xd0 [sunrpc]
    proc_reg_write+0xe1/0x140
    vfs_write+0x1a5/0x6d0
    ksys_write+0xc1/0x160
    do_syscall_64+0x5f/0x170
    entry_SYSCALL_64_after_hwframe+0x76/0x7e

   Freed by task 868:
    kasan_save_stack+0x20/0x40
    kasan_save_track+0x14/0x30
    kasan_save_free_info+0x3b/0x60
    __kasan_slab_free+0x37/0x50
    kfree+0xf3/0x3e0
    svc_export_put+0x87/0xb0 [nfsd]
    cache_purge+0x17f/0x1f0 [sunrpc]
    nfsd_destroy_serv+0x226/0x2d0 [nfsd]
    nfsd_svc+0x125/0x1e0 [nfsd]
    write_threads+0x16a/0x2a0 [nfsd]
    nfsctl_transaction_write+0x74/0xa0 [nfsd]
    vfs_write+0x1a5/0x6d0
    ksys_write+0xc1/0x160
    do_syscall_64+0x5f/0x170
    entry_SYSCALL_64_after_hwframe+0x76/0x7e

2. We cannot sleep while using `rcu_read_lock`/`rcu_read_unlock`.
   However, `svc_export_put`/`expkey_put` will call path_put, which
   subsequently triggers a sleeping operation due to the following
   `dput`.

   =============================
   WARNING: suspicious RCU usage
   5.10.0-dirty #141 Not tainted
   -----------------------------
   ...
   Call Trace:
   dump_stack+0x9a/0xd0
   ___might_sleep+0x231/0x240
   dput+0x39/0x600
   path_put+0x1b/0x30
   svc_export_put+0x17/0x80
   e_show+0x1c9/0x200
   seq_read_iter+0x63f/0x7c0
   seq_read+0x226/0x2d0
   vfs_read+0x113/0x2c0
   ksys_read+0xc9/0x170
   do_syscall_64+0x33/0x40
   entry_SYSCALL_64_after_hwframe+0x67/0xd1

Fix these issues by using `rcu_work` to help release
`svc_expkey`/`svc_export`. This approach allows for an asynchronous
context to invoke `path_put` and also facilitates the freeing of
`uuid/exp/key` after an RCU grace period.

Fixes: 9ceddd9da1 ("knfsd: Allow lockless lookups of the exports")
Signed-off-by: Yang Erkun <yangerkun@huawei.com>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
2024-11-18 20:23:05 -05:00

137 lines
3.5 KiB
C

/* SPDX-License-Identifier: GPL-2.0 */
/*
* Copyright (C) 1995-1997 Olaf Kirch <okir@monad.swb.de>
*/
#ifndef NFSD_EXPORT_H
#define NFSD_EXPORT_H
#include <linux/sunrpc/cache.h>
#include <linux/percpu_counter.h>
#include <uapi/linux/nfsd/export.h>
#include <linux/nfs4.h>
struct knfsd_fh;
struct svc_fh;
struct svc_rqst;
/*
* FS Locations
*/
#define MAX_FS_LOCATIONS 128
struct nfsd4_fs_location {
char *hosts; /* colon separated list of hosts */
char *path; /* slash separated list of path components */
};
struct nfsd4_fs_locations {
uint32_t locations_count;
struct nfsd4_fs_location *locations;
/* If we're not actually serving this data ourselves (only providing a
* list of replicas that do serve it) then we set "migrated": */
int migrated;
};
/*
* We keep an array of pseudoflavors with the export, in order from most
* to least preferred. For the foreseeable future, we don't expect more
* than the eight pseudoflavors null, unix, krb5, krb5i, krb5p, skpm3,
* spkm3i, and spkm3p (and using all 8 at once should be rare).
*/
#define MAX_SECINFO_LIST 8
#define EX_UUID_LEN 16
struct exp_flavor_info {
u32 pseudoflavor;
u32 flags;
};
/* Per-export stats */
enum {
EXP_STATS_FH_STALE,
EXP_STATS_IO_READ,
EXP_STATS_IO_WRITE,
EXP_STATS_COUNTERS_NUM
};
struct export_stats {
time64_t start_time;
struct percpu_counter counter[EXP_STATS_COUNTERS_NUM];
};
struct svc_export {
struct cache_head h;
struct auth_domain * ex_client;
int ex_flags;
int ex_fsid;
struct path ex_path;
kuid_t ex_anon_uid;
kgid_t ex_anon_gid;
unsigned char * ex_uuid; /* 16 byte fsid */
struct nfsd4_fs_locations ex_fslocs;
uint32_t ex_nflavors;
struct exp_flavor_info ex_flavors[MAX_SECINFO_LIST];
u32 ex_layout_types;
struct nfsd4_deviceid_map *ex_devid_map;
struct cache_detail *cd;
struct rcu_work ex_rcu_work;
unsigned long ex_xprtsec_modes;
struct export_stats *ex_stats;
};
/* an "export key" (expkey) maps a filehandlefragement to an
* svc_export for a given client. There can be several per export,
* for the different fsid types.
*/
struct svc_expkey {
struct cache_head h;
struct auth_domain * ek_client;
int ek_fsidtype;
u32 ek_fsid[6];
struct path ek_path;
struct rcu_work ek_rcu_work;
};
#define EX_ISSYNC(exp) (!((exp)->ex_flags & NFSEXP_ASYNC))
#define EX_NOHIDE(exp) ((exp)->ex_flags & NFSEXP_NOHIDE)
#define EX_WGATHER(exp) ((exp)->ex_flags & NFSEXP_GATHERED_WRITES)
struct svc_cred;
int nfsexp_flags(struct svc_cred *cred, struct svc_export *exp);
__be32 check_nfsd_access(struct svc_export *exp, struct svc_rqst *rqstp,
bool may_bypass_gss);
/*
* Function declarations
*/
int nfsd_export_init(struct net *);
void nfsd_export_shutdown(struct net *);
void nfsd_export_flush(struct net *);
struct svc_export * rqst_exp_get_by_name(struct svc_rqst *,
struct path *);
struct svc_export * rqst_exp_parent(struct svc_rqst *,
struct path *);
struct svc_export * rqst_find_fsidzero_export(struct svc_rqst *);
int exp_rootfh(struct net *, struct auth_domain *,
char *path, struct knfsd_fh *, int maxsize);
__be32 exp_pseudoroot(struct svc_rqst *, struct svc_fh *);
static inline void exp_put(struct svc_export *exp)
{
cache_put(&exp->h, exp->cd);
}
static inline struct svc_export *exp_get(struct svc_export *exp)
{
cache_get(&exp->h);
return exp;
}
struct svc_export *rqst_exp_find(struct cache_req *reqp, struct net *net,
struct auth_domain *cl, struct auth_domain *gsscl,
int fsid_type, u32 *fsidv);
#endif /* NFSD_EXPORT_H */