Viewing: gss_svc_upcall.c
// SPDX-License-Identifier: GPL-2.0
/*
* Modifications for Lustre
*
* Copyright (c) 2007, 2010, Oracle and/or its affiliates. All rights reserved.
*
* Copyright (c) 2012, 2014, Intel Corporation.
*
* Author: Eric Mei <ericm@clusterfs.com>
*/
/*
* Neil Brown <neilb@cse.unsw.edu.au>
* J. Bruce Fields <bfields@umich.edu>
* Andy Adamson <andros@umich.edu>
* Dug Song <dugsong@monkey.org>
*
* RPCSEC_GSS server authentication.
* This implements RPCSEC_GSS as defined in rfc2203 (rpcsec_gss) and rfc2078
* (gssapi)
*
* The RPCSEC_GSS involves three stages:
* 1/ context creation
* 2/ data exchange
* 3/ context destruction
*
* Context creation is handled largely by upcalls to user-space.
* In particular, GSS_Accept_sec_context is handled by an upcall
* Data exchange is handled entirely within the kernel
* In particular, GSS_GetMIC, GSS_VerifyMIC, GSS_Seal, GSS_Unseal are in-kernel.
* Context destruction is handled in-kernel
* GSS_Delete_sec_context is in-kernel
*
* Context creation is initiated by a RPCSEC_GSS_INIT request arriving.
* The context handle and gss_token are used as a key into the rpcsec_init cache.
* The content of this cache includes some of the outputs of GSS_Accept_sec_context,
* being major_status, minor_status, context_handle, reply_token.
* These are sent back to the client.
* Sequence window management is handled by the kernel. The window size if currently
* a compile time constant.
*
* When user-space is happy that a context is established, it places an entry
* in the rpcsec_context cache. The key for this cache is the context_handle.
* The content includes:
* uid/gidlist - for determining access rights
* mechanism type
* mechanism specific information, such as a key
*/
#define DEBUG_SUBSYSTEM S_SEC
#include <linux/types.h>
#include <linux/init.h>
#include <linux/module.h>
#include <linux/random.h>
#include <linux/slab.h>
#include <linux/mutex.h>
#include <linux/binfmts.h>
#include <net/sock.h>
#include <linux/un.h>
#include <linux/hash.h>
#include <obd.h>
#include <obd_class.h>
#include <obd_support.h>
#include <lustre_import.h>
#include <lustre_net.h>
#include <lustre_nodemap.h>
#include <lustre_sec.h>
#include "gss_err.h"
#include "gss_internal.h"
#include "gss_api.h"
#include "gss_crypto.h"
static DEFINE_SPINLOCK(__ctx_index_lock);
static __u64 __ctx_index;
unsigned int krb5_allow_old_client_csum;
__u64 gss_get_next_ctx_index(void)
{
__u64 idx;
spin_lock(&__ctx_index_lock);
idx = __ctx_index++;
spin_unlock(&__ctx_index_lock);
return idx;
}
static inline unsigned long hash_mem(char *buf, int length, int bits)
{
unsigned long hash = 0;
unsigned long l = 0;
int len = 0;
unsigned char c;
do {
if (len == length) {
c = (char) len;
len = -1;
} else
c = *buf++;
l = (l << 8) | c;
len++;
if ((len & (BITS_PER_LONG/8-1)) == 0)
hash = hash_long(hash^l, BITS_PER_LONG);
} while (len);
return hash >> (BITS_PER_LONG - bits);
}
/* This is a little bit of a concern but we need to make our own hash64 function
* as the one from the kernel seems to be buggy by returning a u32:
* static __always_inline u32 hash_64_generic(u64 val, unsigned int bits)
*/
#if BITS_PER_LONG == 64
static __always_inline __u64 gss_hash_64(__u64 val, unsigned int bits)
{
__u64 hash = val;
/* Sigh, gcc can't optimise this alone like it does for 32 bits. */
__u64 n = hash;
n <<= 18;
hash -= n;
n <<= 33;
hash -= n;
n <<= 3;
hash += n;
n <<= 3;
hash -= n;
n <<= 4;
hash += n;
n <<= 2;
hash += n;
/* High bits are more random, so use them. */
return hash >> (64 - bits);
}
static inline unsigned long hash_mem_64(char *buf, int length, int bits)
{
__u64 hash = 0;
__u64 l = 0;
int len = 0;
unsigned char c;
do {
if (len == length) {
c = (char) len;
len = -1;
} else
c = *buf++;
l = (l << 8) | c;
len++;
if ((len & (BITS_PER_LONG / 8 - 1)) == 0)
hash = gss_hash_64(hash ^ l, BITS_PER_LONG);
} while (len);
return hash >> (BITS_PER_LONG - bits);
}
#endif /* BITS_PER_LONG == 64 */
/****************************************
* rpc sec init (rsi) cache *
****************************************/
#define RSI_HASHBITS (6)
static void rsi_entry_init(struct upcall_cache_entry *entry,
void *args)
{
struct gss_rsi *rsi = &entry->u.rsi;
struct gss_rsi *tmp = args;
rsi->si_uc_entry = entry;
rawobj_dup(&rsi->si_in_handle, &tmp->si_in_handle);
rawobj_dup(&rsi->si_in_token, &tmp->si_in_token);
rsi->si_out_handle = RAWOBJ_EMPTY;
rsi->si_out_token = RAWOBJ_EMPTY;
rsi->si_lustre_svc = tmp->si_lustre_svc;
rsi->si_nid4 = tmp->si_nid4;
memcpy(rsi->si_nm_name, tmp->si_nm_name, sizeof(tmp->si_nm_name));
}
static void __rsi_free(struct gss_rsi *rsi)
{
rawobj_free(&rsi->si_in_handle);
rawobj_free(&rsi->si_in_token);
rawobj_free(&rsi->si_out_handle);
rawobj_free(&rsi->si_out_token);
}
static void rsi_entry_free(struct upcall_cache *cache,
struct upcall_cache_entry *entry)
{
struct gss_rsi *rsi = &entry->u.rsi;
__rsi_free(rsi);
}
static inline int rsi_entry_hash(struct gss_rsi *rsi)
{
#if BITS_PER_LONG == 64
return hash_mem_64((char *)rsi->si_in_handle.data,
rsi->si_in_handle.len, RSI_HASHBITS) ^
hash_mem_64((char *)rsi->si_in_token.data,
rsi->si_in_token.len, RSI_HASHBITS);
#else
return hash_mem((char *)rsi->si_in_handle.data, rsi->si_in_handle.len,
RSI_HASHBITS) ^
hash_mem((char *)rsi->si_in_token.data, rsi->si_in_token.len,
RSI_HASHBITS);
#endif
}
static inline int __rsi_entry_match(rawobj_t *h1, rawobj_t *h2,
rawobj_t *t1, rawobj_t *t2)
{
return !(rawobj_equal(h1, h2) && rawobj_equal(t1, t2));
}
static inline int rsi_entry_match(struct gss_rsi *rsi, struct gss_rsi *tmp)
{
return __rsi_entry_match(&rsi->si_in_handle, &tmp->si_in_handle,
&rsi->si_in_token, &tmp->si_in_token);
}
/* Returns 0 to tell this is a match */
static inline int rsi_upcall_compare(struct upcall_cache *cache,
struct upcall_cache_entry *entry,
__u64 key, void *args)
{
struct gss_rsi *rsi1 = &entry->u.rsi;
struct gss_rsi *rsi2 = args;
return rsi_entry_match(rsi1, rsi2);
}
/* See handle_channel_request() userspace for where the upcall data is read */
static int rsi_do_upcall(struct upcall_cache *cache,
struct upcall_cache_entry *entry)
{
int size, len, *blen;
char *buffer, *bp, **bpp;
char *argv[] = {
[0] = cache->uc_upcall,
[1] = "-c",
[2] = cache->uc_name,
[3] = "-r",
[4] = NULL,
[5] = NULL
};
char *envp[] = {
[0] = "HOME=/",
[1] = "PATH=/sbin:/usr/sbin",
[2] = NULL
};
ktime_t start, end;
struct gss_rsi *rsi = &entry->u.rsi;
__u64 index = 0;
int rc;
ENTRY;
CDEBUG(D_SEC, "rsi upcall '%s' on '%s'\n",
cache->uc_upcall, cache->uc_name);
size = 24 + 1 + /* ue_key is uint64_t */
12 + 1 + /* si_lustre_svc is __u32*/
18 + 1 + /* si_nid4 is lnet_nid_t, hex with leading 0x */
18 + 1 + /* index is __u64, hex with leading 0x */
strlen(rsi->si_nm_name) + 1 +
BASE64URL_CHARS(rsi->si_in_handle.len) + 1 +
BASE64URL_CHARS(rsi->si_in_token.len) + 1 +
1 + 1; /* eol */
if (size > MAX_ARG_STRLEN)
RETURN(-E2BIG);
OBD_ALLOC_LARGE(buffer, size);
if (!buffer)
RETURN(-ENOMEM);
bp = buffer;
bpp = &bp;
len = size;
blen = &len;
/* if in_handle is null, provide kernel suggestion */
if (rsi->si_in_handle.len == 0)
index = gss_get_next_ctx_index();
/* entry->ue_key is put into args sent via upcall, so that it can be
* returned by userspace. This will help find cache entry at downcall,
* without unnecessary recomputation of the hash.
*/
gss_u64_write_string(bpp, blen, entry->ue_key);
gss_u64_write_string(bpp, blen, rsi->si_lustre_svc);
gss_u64_write_hex_string(bpp, blen, rsi->si_nid4);
gss_u64_write_hex_string(bpp, blen, index);
gss_string_write(bpp, blen, (char *) rsi->si_nm_name);
gss_base64url_encode(bpp, blen, rsi->si_in_handle.data,
rsi->si_in_handle.len);
gss_base64url_encode(bpp, blen, rsi->si_in_token.data,
rsi->si_in_token.len);
(*bpp)[-1] = '\n';
(*bpp)[0] = '\0';
argv[4] = buffer;
down_read(&cache->uc_upcall_rwsem);
start = ktime_get();
rc = call_usermodehelper(argv[0], argv, envp, UMH_WAIT_EXEC);
end = ktime_get();
up_read(&cache->uc_upcall_rwsem);
if (rc < 0) {
CERROR("%s: error invoking upcall %s %s (time %ldus): rc = %d\n",
cache->uc_name, argv[0], argv[2],
(long)ktime_us_delta(end, start), rc);
} else {
CDEBUG(D_SEC, "%s: invoked upcall %s %s (time %ldus)\n",
cache->uc_name, argv[0], argv[2],
(long)ktime_us_delta(end, start));
rc = 0;
}
OBD_FREE_LARGE(buffer, size);
RETURN(rc);
}
static inline int rsi_downcall_compare(struct upcall_cache *cache,
struct upcall_cache_entry *entry,
__u64 key, void *args)
{
struct gss_rsi *rsi = &entry->u.rsi;
struct rsi_downcall_data *sid = args;
char *mesg = sid->sid_val;
rawobj_t handle, token;
char *p = mesg;
int len;
/* sid_val starts with handle and token */
/* First, handle */
len = gss_buffer_get(&mesg, &handle.len, &handle.data);
sid->sid_offset = mesg - p;
p = mesg;
/* Second, token */
len = gss_buffer_get(&mesg, &token.len, &token.data);
sid->sid_offset += mesg - p;
return __rsi_entry_match(&rsi->si_in_handle, &handle,
&rsi->si_in_token, &token);
}
static int rsi_parse_downcall(struct upcall_cache *cache,
struct upcall_cache_entry *entry,
void *args)
{
struct gss_rsi *rsi = &entry->u.rsi;
struct rsi_downcall_data *sid = args;
int mlen = sid->sid_len;
char *mesg = sid->sid_val + sid->sid_offset;
char *buf = sid->sid_val;
int status = -EINVAL;
int len;
ENTRY;
if (mlen <= 0)
goto out;
rsi->si_major_status = sid->sid_maj_stat;
rsi->si_minor_status = sid->sid_min_stat;
/* in_handle and in_token have already been consumed in
* rsi_downcall_compare(). sid_offset gives next field.
*/
/* out_handle */
len = gss_buffer_read(&mesg, buf, mlen);
if (len < 0)
goto out;
if (rawobj_alloc(&rsi->si_out_handle, buf, len)) {
status = -ENOMEM;
goto out;
}
/* out_token */
len = gss_buffer_read(&mesg, buf, mlen);
if (len < 0)
goto out;
if (rawobj_alloc(&rsi->si_out_token, buf, len)) {
status = -ENOMEM;
goto out;
}
entry->ue_expire = 0;
status = 0;
out:
CDEBUG(D_OTHER, "rsi parse %p: %d\n", rsi, status);
RETURN(status);
}
struct gss_rsi *rsi_entry_get(struct upcall_cache *cache, struct gss_rsi *rsi)
{
struct upcall_cache_entry *entry;
int hash = rsi_entry_hash(rsi);
if (!cache)
return ERR_PTR(-ENOENT);
entry = upcall_cache_get_entry(cache, (__u64)hash, rsi);
if (unlikely(!entry))
return ERR_PTR(-ENOENT);
if (IS_ERR(entry))
return ERR_CAST(entry);
return &entry->u.rsi;
}
void rsi_entry_put(struct upcall_cache *cache, struct gss_rsi *rsi)
{
if (!cache || !rsi)
return;
upcall_cache_put_entry(cache, rsi->si_uc_entry);
}
struct upcall_cache_ops rsi_upcall_cache_ops = {
.init_entry = rsi_entry_init,
.free_entry = rsi_entry_free,
.upcall_compare = rsi_upcall_compare,
.downcall_compare = rsi_downcall_compare,
.do_upcall = rsi_do_upcall,
.parse_downcall = rsi_parse_downcall,
};
struct upcall_cache *rsicache;
/****************************************
* rpc sec context (rsc) cache *
****************************************/
#define RSC_HASHBITS (10)
static void rsc_entry_init(struct upcall_cache_entry *entry,
void *args)
{
struct gss_rsc *rsc = &entry->u.rsc;
struct gss_rsc *tmp = args;
rsc->sc_uc_entry = entry;
rawobj_dup(&rsc->sc_handle, &tmp->sc_handle);
rsc->sc_target = NULL;
memset(&rsc->sc_ctx, 0, sizeof(rsc->sc_ctx));
rsc->sc_ctx.gsc_rvs_hdl = RAWOBJ_EMPTY;
memset(&rsc->sc_ctx.gsc_seqdata, 0, sizeof(rsc->sc_ctx.gsc_seqdata));
spin_lock_init(&rsc->sc_ctx.gsc_seqdata.ssd_lock);
rsc->sc_ctx.gsc_nm_name = NULL;
}
void __rsc_free(struct gss_rsc *rsc)
{
rawobj_free(&rsc->sc_handle);
rawobj_free(&rsc->sc_ctx.gsc_rvs_hdl);
OBD_FREE(rsc->sc_ctx.gsc_nm_name, LUSTRE_NODEMAP_NAME_LENGTH + 1);
lgss_delete_sec_context(&rsc->sc_ctx.gsc_mechctx);
}
static void rsc_entry_free(struct upcall_cache *cache,
struct upcall_cache_entry *entry)
{
struct gss_rsc *rsc = &entry->u.rsc;
__rsc_free(rsc);
}
static inline int rsc_entry_hash(struct gss_rsc *rsc)
{
#if BITS_PER_LONG == 64
return hash_mem_64((char *)rsc->sc_handle.data,
rsc->sc_handle.len, RSC_HASHBITS);
#else
return hash_mem((char *)rsc->sc_handle.data,
rsc->sc_handle.len, RSC_HASHBITS);
#endif
}
static inline int __rsc_entry_match(rawobj_t *h1, rawobj_t *h2)
{
return !(rawobj_equal(h1, h2));
}
static inline int rsc_entry_match(struct gss_rsc *rsc, struct gss_rsc *tmp)
{
return __rsc_entry_match(&rsc->sc_handle, &tmp->sc_handle);
}
/* Returns 0 to tell this is a match */
static inline int rsc_upcall_compare(struct upcall_cache *cache,
struct upcall_cache_entry *entry,
__u64 key, void *args)
{
struct gss_rsc *rsc1 = &entry->u.rsc;
struct gss_rsc *rsc2 = args;
return rsc_entry_match(rsc1, rsc2);
}
/* rsc upcall is a no-op, we just need a valid entry */
static inline int rsc_do_upcall(struct upcall_cache *cache,
struct upcall_cache_entry *entry)
{
upcall_cache_update_entry(cache, entry,
ktime_get_seconds() + cache->uc_entry_expire,
0);
wake_up(&entry->ue_waitq);
return 0;
}
static inline int rsc_downcall_compare(struct upcall_cache *cache,
struct upcall_cache_entry *entry,
__u64 key, void *args)
{
struct gss_rsc *rsc = &entry->u.rsc;
struct rsc_downcall_data *scd = args;
char *mesg = scd->scd_val;
rawobj_t handle;
int len;
/* scd_val starts with handle */
len = gss_buffer_get(&mesg, &handle.len, &handle.data);
scd->scd_offset = mesg - scd->scd_val;
return __rsc_entry_match(&rsc->sc_handle, &handle);
}
static int rsc_parse_downcall(struct upcall_cache *cache,
struct upcall_cache_entry *entry,
void *args)
{
struct gss_api_mech *gm = NULL;
struct gss_rsc *rsc = &entry->u.rsc;
struct rsc_downcall_data *scd = args;
int mlen = scd->scd_len;
char *mesg = scd->scd_val + scd->scd_offset;
char *buf = scd->scd_val;
int status = -EINVAL;
time64_t ctx_expiry;
rawobj_t tmp_buf;
int len;
ENTRY;
if (mlen <= 0)
goto out;
rsc->sc_ctx.gsc_remote = !!(scd->scd_flags & RSC_DATA_FLAG_REMOTE);
rsc->sc_ctx.gsc_usr_root = !!(scd->scd_flags & RSC_DATA_FLAG_ROOT);
rsc->sc_ctx.gsc_usr_mds = !!(scd->scd_flags & RSC_DATA_FLAG_MDS);
rsc->sc_ctx.gsc_usr_oss = !!(scd->scd_flags & RSC_DATA_FLAG_OSS);
rsc->sc_ctx.gsc_mapped_uid = scd->scd_mapped_uid;
rsc->sc_ctx.gsc_uid = scd->scd_uid;
if (strlen(scd->scd_nmname)) {
OBD_ALLOC(rsc->sc_ctx.gsc_nm_name,
LUSTRE_NODEMAP_NAME_LENGTH + 1);
if (!rsc->sc_ctx.gsc_nm_name) {
status = -ENOMEM;
goto out;
}
strscpy(rsc->sc_ctx.gsc_nm_name, scd->scd_nmname,
LUSTRE_NODEMAP_NAME_LENGTH + 1);
}
rsc->sc_ctx.gsc_gid = scd->scd_gid;
gm = lgss_name_to_mech(scd->scd_mechname);
if (!gm) {
status = -EOPNOTSUPP;
goto out;
}
/* handle has already been consumed in rsc_downcall_compare().
* scd_offset gives next field.
*/
/* context token */
len = gss_buffer_read(&mesg, buf, mlen);
if (len < 0)
goto out;
tmp_buf.len = len;
tmp_buf.data = (unsigned char *)buf;
if (lgss_import_sec_context(&tmp_buf, gm,
&rsc->sc_ctx.gsc_mechctx))
goto out;
if (lgss_inquire_context(rsc->sc_ctx.gsc_mechctx, &ctx_expiry))
goto out;
/* ctx_expiry is the number of seconds since Jan 1 1970.
* We just want the number of seconds into the future.
*/
entry->ue_expire = ktime_get_seconds() +
(ctx_expiry - ktime_get_real_seconds());
status = 0;
out:
if (gm)
lgss_mech_put(gm);
CDEBUG(D_OTHER, "rsc parse %p: %d\n", rsc, status);
RETURN(status);
}
/* Returns 1 to tell the expired entry is acceptable */
static inline int rsc_accept_expired(struct upcall_cache *cache,
struct upcall_cache_entry *entry)
{
struct gss_rsc *rsc;
time64_t now = ktime_get_seconds();
if (!entry)
return 0;
rsc = &entry->u.rsc;
/* entry not expired? */
if (now < entry->ue_expire)
return 0;
/* We want to accept an expired entry in the following case:
* the client received an ldlm callback request to release a lock,
* and the server used an expired reverse context to send this request.
* The server cannot be blamed for that, as it only has a reverse
* context and cannot refresh it explicitly. And the client cannot
* refuse to use the associated gss context, otherwise it fails to reply
* to the ldlm callback request and gets evicted. The client, which is
* responsible for the context, cannot refresh it immediately, as it
* would not match the reverse context used by the server. But the
* client context is going to be refreshed right after that, along with
* the subsequent ldlm cancel request.
* The way to make sure we are presently dealing with a client-side
* rpc sec context is to check that sc_target is not NULL and
* gsc_rvs_hdl is empty. On server side gsc_rvs_hdl (the reverse handle)
* is always set.
*/
if (rsc->sc_target && rawobj_empty(&rsc->sc_ctx.gsc_rvs_hdl))
return 1;
return 0;
}
struct gss_rsc *rsc_entry_get(struct upcall_cache *cache, struct gss_rsc *rsc)
{
struct upcall_cache_entry *entry;
int hash = rsc_entry_hash(rsc);
if (!cache)
return ERR_PTR(-ENOENT);
entry = upcall_cache_get_entry(cache, (__u64)hash, rsc);
if (unlikely(!entry))
return ERR_PTR(-ENOENT);
if (IS_ERR(entry))
return ERR_CAST(entry);
return &entry->u.rsc;
}
void rsc_entry_put(struct upcall_cache *cache, struct gss_rsc *rsc)
{
if (!cache || !rsc)
return;
upcall_cache_put_entry(cache, rsc->sc_uc_entry);
}
struct upcall_cache_ops rsc_upcall_cache_ops = {
.init_entry = rsc_entry_init,
.free_entry = rsc_entry_free,
.upcall_compare = rsc_upcall_compare,
.downcall_compare = rsc_downcall_compare,
.do_upcall = rsc_do_upcall,
.parse_downcall = rsc_parse_downcall,
.accept_expired = rsc_accept_expired,
};
struct upcall_cache *rsccache;
/****************************************
* rsc cache flush *
****************************************/
static struct gss_rsc *gss_svc_searchbyctx(rawobj_t *handle)
{
struct gss_rsc rsc;
struct gss_rsc *found;
int rc;
memset(&rsc, 0, sizeof(rsc));
rc = rawobj_dup(&rsc.sc_handle, handle);
if (rc)
return ERR_PTR(rc);
found = rsc_entry_get(rsccache, &rsc);
__rsc_free(&rsc);
if (IS_ERR_OR_NULL(found))
return found;
if (!found->sc_ctx.gsc_mechctx) {
CWARN("ctx hdl %#llx does not have mech ctx: rc = %d\n",
gss_handle_to_u64(handle), -ENOENT);
rsc_entry_put(rsccache, found);
return ERR_PTR(-ENOENT);
}
return found;
}
int gss_svc_upcall_install_rvs_ctx(struct obd_import *imp,
struct gss_sec *gsec,
struct gss_cli_ctx *gctx)
{
struct gss_rsc rsc, *rscp = NULL;
time64_t ctx_expiry;
__u32 major;
int rc;
ENTRY;
memset(&rsc, 0, sizeof(rsc));
if (!imp || !imp->imp_obd) {
CERROR("invalid imp, drop\n");
RETURN(-EPROTO);
}
if (rawobj_alloc(&rsc.sc_handle, (char *)&gsec->gs_rvs_hdl,
sizeof(gsec->gs_rvs_hdl)))
GOTO(out, rc = -ENOMEM);
rscp = rsc_entry_get(rsccache, &rsc);
__rsc_free(&rsc);
if (IS_ERR_OR_NULL(rscp))
GOTO(out, rc = -ENOMEM);
major = lgss_copy_reverse_context(gctx->gc_mechctx,
&rscp->sc_ctx.gsc_mechctx);
if (major != GSS_S_COMPLETE)
GOTO(out, rc = -ENOMEM);
if (lgss_inquire_context(rscp->sc_ctx.gsc_mechctx, &ctx_expiry)) {
CERROR("%s: unable to get expire time, drop\n",
imp->imp_obd->obd_name);
GOTO(out, rc = -EINVAL);
}
rscp->sc_uc_entry->ue_expire = ktime_get_seconds() +
(ctx_expiry - ktime_get_real_seconds());
switch (imp->imp_obd->u.cli.cl_sp_to) {
case LUSTRE_SP_MDT:
rscp->sc_ctx.gsc_usr_mds = 1;
break;
case LUSTRE_SP_OST:
rscp->sc_ctx.gsc_usr_oss = 1;
break;
case LUSTRE_SP_CLI:
rscp->sc_ctx.gsc_usr_root = 1;
break;
case LUSTRE_SP_MGS:
/* by convention, all 3 set to 1 means MGS */
rscp->sc_ctx.gsc_usr_mds = 1;
rscp->sc_ctx.gsc_usr_oss = 1;
rscp->sc_ctx.gsc_usr_root = 1;
break;
default:
break;
}
rscp->sc_target = imp->imp_obd;
rawobj_dup(&gctx->gc_svc_handle, &rscp->sc_handle);
CDEBUG(D_SEC, "%s: create reverse svc ctx %p to %s: idx %#llx\n",
imp->imp_obd->obd_name, &rscp->sc_ctx, obd2cli_tgt(imp->imp_obd),
gsec->gs_rvs_hdl);
rc = 0;
out:
if (!IS_ERR_OR_NULL(rscp))
rsc_entry_put(rsccache, rscp);
if (rc)
CERROR("%s: can't create reverse svc ctx idx %#llx: rc = %d\n",
imp->imp_obd->obd_name, gsec->gs_rvs_hdl, rc);
RETURN(rc);
}
int gss_svc_upcall_expire_rvs_ctx(rawobj_t *handle)
{
const time64_t expire = 20;
struct gss_rsc *rscp;
rscp = gss_svc_searchbyctx(handle);
if (!IS_ERR_OR_NULL(rscp)) {
CDEBUG(D_SEC,
"reverse svcctx %p (rsc %p) expire in %lld seconds\n",
&rscp->sc_ctx, rscp, expire);
rscp->sc_uc_entry->ue_expire = ktime_get_seconds() + expire;
rsc_entry_put(rsccache, rscp);
}
return 0;
}
int gss_svc_upcall_dup_handle(rawobj_t *handle, struct gss_svc_ctx *ctx)
{
struct gss_rsc *rscp = container_of(ctx, struct gss_rsc, sc_ctx);
return rawobj_dup(handle, &rscp->sc_handle);
}
int gss_svc_upcall_update_sequence(rawobj_t *handle, __u32 seq)
{
struct gss_rsc *rscp;
rscp = gss_svc_searchbyctx(handle);
if (!IS_ERR_OR_NULL(rscp)) {
CDEBUG(D_SEC, "reverse svcctx %p (rsc %p) update seq to %u\n",
&rscp->sc_ctx, rscp, seq + 1);
rscp->sc_ctx.gsc_rvs_seq = seq + 1;
rsc_entry_put(rsccache, rscp);
}
return 0;
}
int gss_svc_upcall_handle_init(struct ptlrpc_request *req,
struct gss_svc_reqctx *grctx,
struct gss_wire_ctx *gw,
struct obd_device *target,
__u32 lustre_svc,
rawobj_t *rvs_hdl,
rawobj_t *in_token)
{
struct gss_rsi rsi = { 0 }, *rsip = NULL;
struct ptlrpc_reply_state *rs;
struct gss_rsc *rscp = NULL;
int replen = sizeof(struct ptlrpc_body);
struct gss_rep_header *rephdr;
int rc, rc2;
ENTRY;
rsi.si_lustre_svc = lustre_svc;
/* In case of MR, rq_peer is not the NID from which request is received,
* but primary NID of peer.
* So we need LNetPrimaryNID(rq_source) to match what the clients uses.
*/
LNetPrimaryNID(&req->rq_source.nid);
rsi.si_nid4 = lnet_nid_to_nid4(&req->rq_source.nid);
nodemap_test_nid(&req->rq_peer.nid, rsi.si_nm_name,
sizeof(rsi.si_nm_name));
/* Note that context handle is always 0 for for INIT. */
rc2 = rawobj_dup(&rsi.si_in_handle, &gw->gw_handle);
if (rc2) {
CERROR("%s: failed to duplicate context handle: rc = %d\n",
target->obd_name, rc2);
GOTO(out, rc = SECSVC_DROP);
}
rc2 = rawobj_dup(&rsi.si_in_token, in_token);
if (rc2) {
CERROR("%s: failed to duplicate token: rc = %d\n",
target->obd_name, rc2);
rawobj_free(&rsi.si_in_handle);
GOTO(out, rc = SECSVC_DROP);
}
rsip = rsi_entry_get(rsicache, &rsi);
__rsi_free(&rsi);
if (IS_ERR_OR_NULL(rsip)) {
if (IS_ERR(rsip))
rc2 = PTR_ERR(rsip);
else
rc2 = -EINVAL;
CERROR("%s: failed to get entry from rsi cache (nid %s): rc = %d\n",
target->obd_name,
libcfs_nid2str(lnet_nid_to_nid4(&req->rq_source.nid)),
rc2);
if (!gss_pack_err_notify(req, GSS_S_FAILURE, 0))
rc = SECSVC_COMPLETE;
else
rc = SECSVC_DROP;
GOTO(out, rc);
}
rscp = gss_svc_searchbyctx(&rsip->si_out_handle);
if (IS_ERR_OR_NULL(rscp)) {
/* gss mechanism returned major and minor code so we return
* those in error message */
if (!gss_pack_err_notify(req, rsip->si_major_status,
rsip->si_minor_status))
rc = SECSVC_COMPLETE;
else
rc = SECSVC_DROP;
CERROR("%s: authentication failed: rc = %d\n",
target->obd_name, rc);
GOTO(out, rc);
} else {
/* we need to take an extra ref on the cache entry,
* as a pointer to sc_ctx is stored in grctx
*/
upcall_cache_get_entry_raw(rscp->sc_uc_entry);
grctx->src_ctx = &rscp->sc_ctx;
}
if (gw->gw_flags & LUSTRE_GSS_PACK_KCSUM) {
grctx->src_ctx->gsc_mechctx->hash_func = gss_digest_hash;
} else if (!strcmp(grctx->src_ctx->gsc_mechctx->mech_type->gm_name,
"krb5") &&
!krb5_allow_old_client_csum) {
CWARN("%s: deny connection from '%s' due to missing 'krb_csum' feature, set 'sptlrpc.gss.krb5_allow_old_client_csum=1' to allow, but recommend client upgrade: rc = %d\n",
target->obd_name, libcfs_nidstr(&req->rq_peer.nid),
-EPROTO);
GOTO(out, rc = SECSVC_DROP);
} else {
grctx->src_ctx->gsc_mechctx->hash_func =
gss_digest_hash_compat;
}
if (rawobj_dup(&rscp->sc_ctx.gsc_rvs_hdl, rvs_hdl)) {
CERROR("%s: failed duplicate reverse handle\n",
target->obd_name);
GOTO(out, rc = SECSVC_DROP);
}
rscp->sc_target = target;
CDEBUG(D_SEC, "%s: server create rsc %p(%u->%s)\n",
target->obd_name, rscp,
rscp->sc_ctx.gsc_mapped_uid != -1 ?
rscp->sc_ctx.gsc_mapped_uid :
rscp->sc_ctx.gsc_uid,
libcfs_nidstr(&req->rq_peer.nid));
if (rsip->si_out_handle.len > PTLRPC_GSS_MAX_HANDLE_SIZE) {
CERROR("%s: handle size %u too large\n",
target->obd_name, rsip->si_out_handle.len);
GOTO(out, rc = SECSVC_DROP);
}
grctx->src_init = 1;
grctx->src_reserve_len = round_up(rsip->si_out_token.len, 4);
rc = lustre_pack_reply_v2(req, 1, &replen, NULL, 0);
if (rc) {
CERROR("%s: failed to pack reply: rc = %d\n",
target->obd_name, rc);
GOTO(out, rc = SECSVC_DROP);
}
rs = req->rq_reply_state;
LASSERT(rs->rs_repbuf->lm_bufcount == 3);
LASSERT(rs->rs_repbuf->lm_buflens[0] >=
sizeof(*rephdr) + rsip->si_out_handle.len);
LASSERT(rs->rs_repbuf->lm_buflens[2] >= rsip->si_out_token.len);
rephdr = lustre_msg_buf(rs->rs_repbuf, 0, 0);
rephdr->gh_version = PTLRPC_GSS_VERSION;
rephdr->gh_flags = 0;
rephdr->gh_proc = PTLRPC_GSS_PROC_ERR;
rephdr->gh_major = rsip->si_major_status;
rephdr->gh_minor = rsip->si_minor_status;
rephdr->gh_seqwin = GSS_SEQ_WIN;
rephdr->gh_handle.len = rsip->si_out_handle.len;
memcpy(rephdr->gh_handle.data, rsip->si_out_handle.data,
rsip->si_out_handle.len);
memcpy(lustre_msg_buf(rs->rs_repbuf, 2, 0), rsip->si_out_token.data,
rsip->si_out_token.len);
rs->rs_repdata_len = lustre_shrink_msg(rs->rs_repbuf, 2,
rsip->si_out_token.len, 0);
rc = SECSVC_OK;
out:
if (!IS_ERR_OR_NULL(rsip)) {
/* After rpcsec init request has been handled,
* no need to keep rsi entry in cache, no matter the result.
* So mark it invalid now.
*/
UC_CACHE_SET_INVALID(rsip->si_uc_entry);
rsi_entry_put(rsicache, rsip);
}
if (!IS_ERR_OR_NULL(rscp)) {
/* if anything went wrong, we don't keep the context too */
if (rc != SECSVC_OK)
UC_CACHE_SET_INVALID(rscp->sc_uc_entry);
else
CDEBUG(D_SEC, "%s: create rsc with idx %#llx\n",
target->obd_name,
gss_handle_to_u64(&rscp->sc_handle));
rsc_entry_put(rsccache, rscp);
}
RETURN(rc);
}
struct gss_svc_ctx *gss_svc_upcall_get_ctx(struct ptlrpc_request *req,
struct gss_wire_ctx *gw)
{
struct gss_rsc *rscp;
rscp = gss_svc_searchbyctx(&gw->gw_handle);
if (IS_ERR_OR_NULL(rscp)) {
CWARN("Invalid gss ctx hdl %#llx from %s: rc = %ld\n",
gss_handle_to_u64(&gw->gw_handle),
libcfs_nidstr(&req->rq_peer.nid),
rscp ? PTR_ERR(rscp) : -1);
return NULL;
}
return &rscp->sc_ctx;
}
void gss_svc_upcall_put_ctx(struct gss_svc_ctx *ctx)
{
struct gss_rsc *rscp = container_of(ctx, struct gss_rsc, sc_ctx);
rsc_entry_put(rsccache, rscp);
}
void gss_svc_upcall_destroy_ctx(struct gss_svc_ctx *ctx)
{
struct gss_rsc *rscp = container_of(ctx, struct gss_rsc, sc_ctx);
UC_CACHE_SET_INVALID(rscp->sc_uc_entry);
rscp->sc_uc_entry->ue_expire = 1;
}
/* Wait for userspace daemon to open socket, approx 1.5 s.
* If socket is not open, upcall requests might fail.
*/
static int check_gssd_socket(void)
{
struct sockaddr_storage sstorage = {0};
struct sockaddr_un *sun = (struct sockaddr_un *)&sstorage;
struct socket *sock;
int tries = 0;
int err;
err = sock_create_kern(current->nsproxy->net_ns,
AF_UNIX, SOCK_STREAM, 0, &sock);
if (err < 0) {
CDEBUG(D_SEC, "Failed to create socket: %d\n", err);
return err;
}
sun->sun_family = AF_UNIX;
strncpy(sun->sun_path, GSS_SOCKET_PATH, sizeof(sun->sun_path));
/* Try to connect to the socket */
while (tries++ < 6) {
err = kernel_connect(sock, (struct sockaddr_unsized *)&sstorage,
sizeof(sstorage), 0);
if (!err)
break;
schedule_timeout_uninterruptible(cfs_time_seconds(1) / 4);
}
if (err < 0)
CDEBUG(D_SEC, "Failed to connect to socket: %d\n", err);
else
kernel_sock_shutdown(sock, SHUT_RDWR);
sock_release(sock);
return err;
}
int __init gss_init_svc_upcall(void)
{
int rc;
/*
* this helps reducing context index confliction. after server reboot,
* conflicting request from clients might be filtered out by initial
* sequence number checking, thus no chance to sent error notification
* back to clients.
*/
get_random_bytes(&__ctx_index, sizeof(__ctx_index));
rsicache = upcall_cache_init(RSI_CACHE_NAME, RSI_UPCALL_PATH,
UC_RSICACHE_HASH_SIZE,
600, /* entry expire: 10 mn */
30, /* acquire expire: 30 s */
false, /* can't replay acquire */
&rsi_upcall_cache_ops);
if (IS_ERR(rsicache)) {
rc = PTR_ERR(rsicache);
rsicache = NULL;
return rc;
}
rsccache = upcall_cache_init(RSC_CACHE_NAME, RSC_UPCALL_PATH,
UC_RSCCACHE_HASH_SIZE,
3600, /* replaced with one from mech */
100, /* arbitrary, not used */
false, /* can't replay acquire */
&rsc_upcall_cache_ops);
if (IS_ERR(rsccache)) {
upcall_cache_cleanup(rsicache);
rsicache = NULL;
rc = PTR_ERR(rsccache);
rsccache = NULL;
return rc;
}
if (check_gssd_socket())
CDEBUG(D_SEC,
"Init channel not opened by lsvcgssd, GSS might not work on server side until daemon is active\n");
return 0;
}
void gss_exit_svc_upcall(void)
{
upcall_cache_cleanup(rsicache);
upcall_cache_cleanup(rsccache);
}
