Re: [PATCH 0/6] SUNRPC: Address remaining cache_check_rcu() UAF in cache content files

[Chuck Lever <cel@kernel.org>]

On 5/5/26 12:49 PM, Calum Mackay wrote:
> On 01/05/2026 3:51 pm, Chuck Lever wrote:
>> Misbah Anjum reported a use-after-free in cache_check_rcu()
>> reached through e_show() while sosreport was reading
>> /proc/fs/nfsd/exports on ppc64le.  Two fixes for that report
>> landed in v7.0:
>>
>>    48db892356d6 ("NFSD: Defer sub-object cleanup in export put
>> callbacks")
>>    e7fcf179b82d ("NFSD: Hold net reference for the lifetime of /proc/
>> fs/nfs/exports fd")
>>
>> The original e_show() repro is now fixed.  However, the same
>> sosreport workload still reproduces a closely related fault on
>> post-v7.0 mainline (Misbah, ppc64le) and on master.20260424
>> (internal report, aarch64).  In both cases the fault is in
>> cache_check_rcu() reached through c_show() rather than e_show(),
>> and the cache_head pointer is plain garbage:
>>
>>    pc : cache_check_rcu+0x40 [sunrpc]
>>    lr : c_show+0x60 [sunrpc]
>>    ...faulting on h->flags off h = 0x0000000200000000
>>
>> c_show() is the generic show callback used by
>> /proc/net/rpc/<cd>/content for every per-net cache_detail
>> (auth.unix.ip, auth.unix.gid, nfsd.fh, nfsd.export).  Two
>> bugs combine in that path:
>>
>> 1. cache_unregister_net() / cache_destroy_net() free cd and
>>     cd->hash_table synchronously when the namespace exits.  The
>>     /proc/net/rpc/.../content open path takes only a module
>>     reference, so a fd kept open across a netns exit walks a
>>     freed hash_table and returns garbage cache_head pointers.
>>     This is the same hazard that e7fcf179b82d closed for the
>>     /proc/fs/nfs/exports file alone.
>>
>> 2. ip_map_put() drops auth_domain_put() before kfree_rcu(), so
>>     sub-objects can be freed before the RCU grace period -- the
>>     same hazard that 48db892356d6 fixed for svc_export_put() and
>>     expkey_put().  unix_gid_put() does not have this bug
>>     structurally (its put_group_info() runs inside the call_rcu()
>>     callback) but it uses a separate idiom from the other three
>>     caches.
>>
>> This series replaces the v1 narrow fixes with shared
>> infrastructure that covers all four cache_detail .put paths
>> and all three per-cache file types:
>>
>> Patch 1 hoists nfsd_export_wq up to the sunrpc layer as
>> sunrpc_cache_wq, exposed through sunrpc_cache_queue_release()
>> and sunrpc_cache_drain() so all four put callbacks share one
>> workqueue and one drain primitive.
>>
>> Patch 2 converts ip_map_put() to the queue_rcu_work() pattern,
>> moving auth_domain_put() into a deferred ip_map_release() that
>> runs after the RCU grace period.
>>
>> Patch 3 unifies unix_gid_put() onto the same pattern for
>> consistency (not a bug fix on its own).
>>
>> Patch 4 takes a get_net(cd->net) in content_open(), cache_open(),
>> and open_flush() and drops it in the matching release helpers,
>> so cache_destroy_net() cannot run while a sunrpc cache fd is
>> open.
>>
>> Series has been compile-tested only.
>>
>> ---
>> Chuck Lever (6):
>>        SUNRPC: Move cache_initialize() declaration to sunrpc-private
>> header
>>        SUNRPC: Provide a shared workqueue for cache release callbacks
>>        SUNRPC: Defer ip_map sub-object cleanup past RCU grace period
>>        SUNRPC: Use shared release pattern for the unix_gid cache
>>        SUNRPC: Hold cd->net for the lifetime of cache files
>>        NFSD: Convert nfsd_export_shutdown() to sunrpc_cache_destroy_net()
>>
>>   fs/nfsd/export.c             | 45 ++--------------------
>>   fs/nfsd/export.h             |  2 -
>>   fs/nfsd/nfsctl.c             |  8 +---
>>   include/linux/sunrpc/cache.h |  3 +-
>>   net/sunrpc/cache.c           | 90 ++++++++++++++++++++++++++++++++++
>> ++++++++--
>>   net/sunrpc/sunrpc.h          |  2 +
>>   net/sunrpc/sunrpc_syms.c     | 23 ++++++-----
>>   net/sunrpc/svcauth_unix.c    | 46 ++++++++++++----------
>>   8 files changed, 135 insertions(+), 84 deletions(-)
>> ---
>> base-commit: f3a313ecd1fdab1f5da119db355363b13af6fcac
>> change-id: 20260430-cache-uaf-fix-a13000f67c37
>>
>> Best regards,
>> -- 
>> Chuck Lever
>>
>>
> 
> Looks good Chuck, thanks very much.
> 
> With these patches, testing shows no crashes, sosreport no longer hangs,
> no seq_file errors.
> 
> Tested-by: Alexandr Alexandrov <alexandr.alexandrov@oracle.com>
> 
> cheers,
> c.
> 

Excellent; pushed with Jeff's R-b and Alexandr's T-b.


-- 
Chuck Lever