[PATCH v2 04/12] cachefiles: fix slab-use-after-free in cachefiles_ondemand_daemon_read()

[Baokun Li]

On 2024/5/20 15:36, Jingbo Xu wrote:
>
> On 5/15/24 4:45 PM, libaokun at huaweicloud.com wrote:
>> From: Baokun Li <libaokun1 at huawei.com>
>>
>> We got the following issue in a fuzz test of randomly issuing the restore
>> command:
>>
>> ==================================================================
>> BUG: KASAN: slab-use-after-free in cachefiles_ondemand_daemon_read+0xb41/0xb60
>> Read of size 8 at addr ffff888122e84088 by task ondemand-04-dae/963
>>
>> CPU: 13 PID: 963 Comm: ondemand-04-dae Not tainted 6.8.0-dirty #564
>> Call Trace:
>>   kasan_report+0x93/0xc0
>>   cachefiles_ondemand_daemon_read+0xb41/0xb60
>>   vfs_read+0x169/0xb50
>>   ksys_read+0xf5/0x1e0
>>
>> Allocated by task 116:
>>   kmem_cache_alloc+0x140/0x3a0
>>   cachefiles_lookup_cookie+0x140/0xcd0
>>   fscache_cookie_state_machine+0x43c/0x1230
>>   [...]
>>
>> Freed by task 792:
>>   kmem_cache_free+0xfe/0x390
>>   cachefiles_put_object+0x241/0x480
>>   fscache_cookie_state_machine+0x5c8/0x1230
>>   [...]
>> ==================================================================
>>
>> Following is the process that triggers the issue:
>>
>>       mount  |   daemon_thread1    |    daemon_thread2
>> ------------------------------------------------------------
>> cachefiles_withdraw_cookie
>>   cachefiles_ondemand_clean_object(object)
>>    cachefiles_ondemand_send_req
>>     REQ_A = kzalloc(sizeof(*req) + data_len)
>>     wait_for_completion(&REQ_A->done)
>>
>>              cachefiles_daemon_read
>>               cachefiles_ondemand_daemon_read
>>                REQ_A = cachefiles_ondemand_select_req
>>                msg->object_id = req->object->ondemand->ondemand_id
>>                                    ------ restore ------
>>                                    cachefiles_ondemand_restore
>>                                    xas_for_each(&xas, req, ULONG_MAX)
>>                                     xas_set_mark(&xas, CACHEFILES_REQ_NEW)
>>
>>                                    cachefiles_daemon_read
>>                                     cachefiles_ondemand_daemon_read
>>                                      REQ_A = cachefiles_ondemand_select_req
>>                copy_to_user(_buffer, msg, n)
>>                 xa_erase(&cache->reqs, id)
>>                 complete(&REQ_A->done)
>>                ------ close(fd) ------
>>                cachefiles_ondemand_fd_release
>>                 cachefiles_put_object
>>   cachefiles_put_object
>>    kmem_cache_free(cachefiles_object_jar, object)
>>                                      REQ_A->object->ondemand->ondemand_id
>>                                       // object UAF !!!
>>
>> When we see the request within xa_lock, req->object must not have been
>> freed yet, so grab the reference count of object before xa_unlock to
>> avoid the above issue.
>>
>> Fixes: 0a7e54c1959c ("cachefiles: resend an open request if the read request's object is closed")
>> Signed-off-by: Baokun Li <libaokun1 at huawei.com>
>> Reviewed-by: Jia Zhu <zhujia.zj at bytedance.com>
>> ---
>>   fs/cachefiles/ondemand.c          | 2 ++
>>   include/trace/events/cachefiles.h | 6 +++++-
>>   2 files changed, 7 insertions(+), 1 deletion(-)
>>
>> diff --git a/fs/cachefiles/ondemand.c b/fs/cachefiles/ondemand.c
>> index 56d12fe4bf73..bb94ef6a6f61 100644
>> --- a/fs/cachefiles/ondemand.c
>> +++ b/fs/cachefiles/ondemand.c
>> @@ -336,6 +336,7 @@ ssize_t cachefiles_ondemand_daemon_read(struct cachefiles_cache *cache,
>>   	xas_clear_mark(&xas, CACHEFILES_REQ_NEW);
>>   	cache->req_id_next = xas.xa_index + 1;
>>   	refcount_inc(&req->ref);
>> +	cachefiles_grab_object(req->object, cachefiles_obj_get_read_req);
>>   	xa_unlock(&cache->reqs);
>>   
>>   	if (msg->opcode == CACHEFILES_OP_OPEN) {
>> @@ -355,6 +356,7 @@ ssize_t cachefiles_ondemand_daemon_read(struct cachefiles_cache *cache,
>>   			close_fd(((struct cachefiles_open *)msg->data)->fd);
>>   	}
>>   out:
>> +	cachefiles_put_object(req->object, cachefiles_obj_put_read_req);
>>   	/* Remove error request and CLOSE request has no reply */
>>   	if (ret || msg->opcode == CACHEFILES_OP_CLOSE) {
>>   		xas_reset(&xas);
>> diff --git a/include/trace/events/cachefiles.h b/include/trace/events/cachefiles.h
>> index cf4b98b9a9ed..119a823fb5a0 100644
>> --- a/include/trace/events/cachefiles.h
>> +++ b/include/trace/events/cachefiles.h
>> @@ -33,6 +33,8 @@ enum cachefiles_obj_ref_trace {
>>   	cachefiles_obj_see_withdrawal,
>>   	cachefiles_obj_get_ondemand_fd,
>>   	cachefiles_obj_put_ondemand_fd,
>> +	cachefiles_obj_get_read_req,
>> +	cachefiles_obj_put_read_req,
> How about cachefiles_obj_[get|put]_ondemand_read, so that it could be
> easily identified as ondemand mode at the first glance?
The ondemand_read tends to confuse whether it's
ondemand_daemon_read or ondemand_data_read. I think it's better
to emphasise the read request, and currently only the ondemand
mode has a cachefiles req.
>>   };
>>   
>>   enum fscache_why_object_killed {
>> @@ -127,7 +129,9 @@ enum cachefiles_error_trace {
>>   	EM(cachefiles_obj_see_lookup_cookie,	"SEE lookup_cookie")	\
>>   	EM(cachefiles_obj_see_lookup_failed,	"SEE lookup_failed")	\
>>   	EM(cachefiles_obj_see_withdraw_cookie,	"SEE withdraw_cookie")	\
>> -	E_(cachefiles_obj_see_withdrawal,	"SEE withdrawal")
>> +	EM(cachefiles_obj_see_withdrawal,	"SEE withdrawal")	\
>> +	EM(cachefiles_obj_get_read_req,		"GET read_req")		\
>> +	E_(cachefiles_obj_put_read_req,		"PUT read_req")
> Ditto.
>
>
> Otherwise, LGTM.
>
> Reviewed-by: Jingbo Xu <jefflexu at linux.alibaba.com>
>
Thank you very much for your review!

-- 
With Best Regards,
Baokun Li