CVE-2025-22025

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> nfsd: put dl_stid if fail to queue dl_recall<br /> <br /> Before calling nfsd4_run_cb to queue dl_recall to the callback_wq, we<br /> increment the reference count of dl_stid.<br /> We expect that after the corresponding work_struct is processed, the<br /> reference count of dl_stid will be decremented through the callback<br /> function nfsd4_cb_recall_release.<br /> However, if the call to nfsd4_run_cb fails, the incremented reference<br /> count of dl_stid will not be decremented correspondingly, leading to the<br /> following nfs4_stid leak:<br /> unreferenced object 0xffff88812067b578 (size 344):<br /> comm "nfsd", pid 2761, jiffies 4295044002 (age 5541.241s)<br /> hex dump (first 32 bytes):<br /> 01 00 00 00 6b 6b 6b 6b b8 02 c0 e2 81 88 ff ff ....kkkk........<br /> 00 6b 6b 6b 6b 6b 6b 6b 00 00 00 00 ad 4e ad de .kkkkkkk.....N..<br /> backtrace:<br /> kmem_cache_alloc+0x4b9/0x700<br /> nfsd4_process_open1+0x34/0x300<br /> nfsd4_open+0x2d1/0x9d0<br /> nfsd4_proc_compound+0x7a2/0xe30<br /> nfsd_dispatch+0x241/0x3e0<br /> svc_process_common+0x5d3/0xcc0<br /> svc_process+0x2a3/0x320<br /> nfsd+0x180/0x2e0<br /> kthread+0x199/0x1d0<br /> ret_from_fork+0x30/0x50<br /> ret_from_fork_asm+0x1b/0x30<br /> unreferenced object 0xffff8881499f4d28 (size 368):<br /> comm "nfsd", pid 2761, jiffies 4295044005 (age 5541.239s)<br /> hex dump (first 32 bytes):<br /> 01 00 00 00 00 00 00 00 30 4d 9f 49 81 88 ff ff ........0M.I....<br /> 30 4d 9f 49 81 88 ff ff 20 00 00 00 01 00 00 00 0M.I.... .......<br /> backtrace:<br /> kmem_cache_alloc+0x4b9/0x700<br /> nfs4_alloc_stid+0x29/0x210<br /> alloc_init_deleg+0x92/0x2e0<br /> nfs4_set_delegation+0x284/0xc00<br /> nfs4_open_delegation+0x216/0x3f0<br /> nfsd4_process_open2+0x2b3/0xee0<br /> nfsd4_open+0x770/0x9d0<br /> nfsd4_proc_compound+0x7a2/0xe30<br /> nfsd_dispatch+0x241/0x3e0<br /> svc_process_common+0x5d3/0xcc0<br /> svc_process+0x2a3/0x320<br /> nfsd+0x180/0x2e0<br /> kthread+0x199/0x1d0<br /> ret_from_fork+0x30/0x50<br /> ret_from_fork_asm+0x1b/0x30<br /> Fix it by checking the result of nfsd4_run_cb and call nfs4_put_stid if<br /> fail to queue dl_recall.