- Kernel - mailweb.openeuler.org

openEuler Kernel SIG双周例会
by openEuler conference 19 Aug '26

19 Aug '26

您好！ Kernel 邀请您参加 2026-08-21 14:00 召开的WeLink会议(自动录制) 会议主题：openEuler Kernel SIG双周例会会议发起人： LiaoTao_Wave 会议内容： 1. 进展update 2. 议题征集中（新增议题请申报至会议纪要看板）会议链接：https://meeting.huaweicloud.com:36443/#/j/986322224 会议纪要&签到链接：https://etherpad.openeuler.org/p/Kernel-meetings 更多资讯尽在：https://www.openeuler.org/zh/ Hello! Kernel invites you to attend the WeLink conference(auto recording) will be held at 2026-08-21 14:00 The subject of the conference is openEuler Kernel SIG双周例会 The sponsor of the conference is LiaoTao_Wave Summary: 1. 进展update 2. 议题征集中（新增议题请申报至会议纪要看板） You can join the meeting at https://meeting.huaweicloud.com:36443/#/j/986322224 Add topics at https://etherpad.openeuler.org/p/Kernel-meetings More information: https://www.openeuler.org/en/

1 0

[PATCH OLK-6.6 v4 0/8] irqbypass: Cleanups and a perf improvement
by Qiqi Li 19 Aug '26

19 Aug '26

From: Qiqi Li <liqiqi23(a)huawei.com> The two primary goals of this series are to make the irqbypass concept easier to understand, and to address the terrible performance that can result from using a list to track connections. Provides additional adaptation for the ub vfio device in openEuler. Sean Christopherson (8): irqbypass: Drop pointless and misleading THIS_MODULE get/put irqbypass: Drop superfluous might_sleep() annotations irqbypass: Take ownership of producer/consumer token tracking irqbypass: Explicitly track producer and consumer bindings irqbypass: Use paired consumer/producer to disconnect during unregister irqbypass: Use guard(mutex) in lieu of manual lock+unlock irqbypass: Use xarray to track producers and consumers irqbypass: Require producers to pass in Linux IRQ number during registration arch/x86/kvm/x86.c | 4 +- drivers/vfio/pci/vfio_pci_intrs.c | 10 +- drivers/vfio/ubus/vfio_ub_intrs.c | 12 +- drivers/vhost/vdpa.c | 10 +- include/linux/irqbypass.h | 46 +++---- virt/kvm/eventfd.c | 7 +- virt/lib/irqbypass.c | 192 +++++++++++------------------- 7 files changed, 112 insertions(+), 169 deletions(-) -- 2.43.0

2 9

[PATCH OLK-6.6 v4 0/8] irqbypass: Cleanups and a perf improvement
by Qiqi Li 19 Aug '26

19 Aug '26

From: Qiqi Li <liqiqi23(a)huawei.com> The two primary goals of this series are to make the irqbypass concept easier to understand, and to address the terrible performance that can result from using a list to track connections. Provides additional adaptation for the ub vfio device in openEuler. Sean Christopherson (8): irqbypass: Drop pointless and misleading THIS_MODULE get/put irqbypass: Drop superfluous might_sleep() annotations irqbypass: Take ownership of producer/consumer token tracking irqbypass: Explicitly track producer and consumer bindings irqbypass: Use paired consumer/producer to disconnect during unregister irqbypass: Use guard(mutex) in lieu of manual lock+unlock irqbypass: Use xarray to track producers and consumers irqbypass: Require producers to pass in Linux IRQ number during registration arch/x86/kvm/x86.c | 4 +- drivers/vfio/pci/vfio_pci_intrs.c | 10 +- drivers/vfio/ubus/vfio_ub_intrs.c | 12 +- drivers/vhost/vdpa.c | 10 +- include/linux/irqbypass.h | 46 +++---- virt/kvm/eventfd.c | 7 +- virt/lib/irqbypass.c | 192 +++++++++++------------------- 7 files changed, 112 insertions(+), 169 deletions(-) -- 2.43.0

2 9

[PATCH OLK-5.10] ceph: fix pre-auth out-of-bounds read on snaptrace in ceph_handle_caps()
by Pan Taixi 19 Aug '26

19 Aug '26

From: Bryam Vargas <hexlabsecurity(a)proton.me> mainline inclusion from mainline-v7.2-rc5 commit 4dbc71bcaf9a30abf3920a4e2cc4ed33bba78c02 category: bugfix bugzilla: https://atomgit.com/src-openeuler/kernel/issues/16967 CVE: CVE-2026-68160 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… -------------------------------- ceph_handle_caps() reads snap_trace_len from the wire-format ceph_mds_caps header and uses it unconditionally to build a fake end pointer (snaptrace + snaptrace_len) that is later handed to ceph_update_snap_trace() in the CEPH_CAP_OP_IMPORT case: snaptrace = h + 1; snaptrace_len = le32_to_cpu(h->snap_trace_len); p = snaptrace + snaptrace_len; ... case CEPH_CAP_OP_IMPORT: if (snaptrace_len) { ... if (ceph_update_snap_trace(mdsc, snaptrace, snaptrace + snaptrace_len, false, &realm)) { ... } ceph_update_snap_trace() then decodes a struct ceph_mds_snap_realm from snaptrace using ceph_decode_need(&p, e, sizeof(*ri), bad) with the attacker-supplied fake end e == snaptrace + snaptrace_len. With snaptrace_len == 0xFFFFFFFF the bound check is trivially satisfied, ri = p reads sizeof(struct ceph_mds_snap_realm) past the legitimate msg->front buffer, and ri->num_snaps / ri->num_prior_parent_snaps then drive further out-of-bounds reads of the encoded snap arrays. The eleven msg_version >= 2 .. msg_version >= 12 decoder blocks above the op switch each catch this OOB through their ceph_decode_*_safe() / ceph_decode_need() helpers, but they sit behind a hdr.version-gated if, so a malicious or compromised MDS that sets msg->hdr.version = 1 reaches the IMPORT path with no version-gated decoder having validated snap_trace_len. The shape has been present since ceph_handle_caps() was introduced. Validate snap_trace_len against the message front buffer before consuming it, using the canonical ceph_decode_need() / ceph_has_room() helper. The helper bounds the length with subtraction (n <= end - p, guarded by end >= p) rather than pointer addition, so it is wrap-safe for the attacker-controlled u32 length on 32-bit builds where p + snap_trace_len could overflow the address space. This matches the rest of the ceph decode path (e.g. the pool_ns_len check a few lines below), and the existing goto bad cleanup already covers this exit path. Cc: stable(a)vger.kernel.org Fixes: a8599bd821d0 ("ceph: capability management") Signed-off-by: Bryam Vargas <hexlabsecurity(a)proton.me> Reviewed-by: Viacheslav Dubeyko <Slava.Dubeyko(a)ibm.com> Signed-off-by: Ilya Dryomov <idryomov(a)gmail.com> Conflicts: fs/ceph/caps.c [Context conflicts unmerged commit 8ea412e18131 ("ceph: improve caps debugging output")] Signed-off-by: Pan Taixi <pantaixi1(a)huawei.com> --- fs/ceph/caps.c | 1 + 1 file changed, 1 insertion(+) diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c index 8c1dd9e80e7e..31e5d87ec945 100644 --- a/fs/ceph/caps.c +++ b/fs/ceph/caps.c @@ -4124,10 +4124,11 @@ void ceph_handle_caps(struct ceph_mds_session *session, seq = le32_to_cpu(h->seq); mseq = le32_to_cpu(h->migrate_seq); snaptrace = h + 1; snaptrace_len = le32_to_cpu(h->snap_trace_len); + ceph_decode_need(&snaptrace, end, snaptrace_len, bad); p = snaptrace + snaptrace_len; if (msg_version >= 2) { u32 flock_len; ceph_decode_32_safe(&p, end, flock_len, bad); -- 2.34.1

2 1

[PATCH openEuler-1.0-LTS] ceph: fix pre-auth out-of-bounds read on snaptrace in ceph_handle_caps()
by Pan Taixi 19 Aug '26

19 Aug '26

From: Bryam Vargas <hexlabsecurity(a)proton.me> mainline inclusion from mainline-v7.2-rc5 commit 4dbc71bcaf9a30abf3920a4e2cc4ed33bba78c02 category: bugfix bugzilla: https://atomgit.com/src-openeuler/kernel/issues/16967 CVE: CVE-2026-68160 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… -------------------------------- ceph_handle_caps() reads snap_trace_len from the wire-format ceph_mds_caps header and uses it unconditionally to build a fake end pointer (snaptrace + snaptrace_len) that is later handed to ceph_update_snap_trace() in the CEPH_CAP_OP_IMPORT case: snaptrace = h + 1; snaptrace_len = le32_to_cpu(h->snap_trace_len); p = snaptrace + snaptrace_len; ... case CEPH_CAP_OP_IMPORT: if (snaptrace_len) { ... if (ceph_update_snap_trace(mdsc, snaptrace, snaptrace + snaptrace_len, false, &realm)) { ... } ceph_update_snap_trace() then decodes a struct ceph_mds_snap_realm from snaptrace using ceph_decode_need(&p, e, sizeof(*ri), bad) with the attacker-supplied fake end e == snaptrace + snaptrace_len. With snaptrace_len == 0xFFFFFFFF the bound check is trivially satisfied, ri = p reads sizeof(struct ceph_mds_snap_realm) past the legitimate msg->front buffer, and ri->num_snaps / ri->num_prior_parent_snaps then drive further out-of-bounds reads of the encoded snap arrays. The eleven msg_version >= 2 .. msg_version >= 12 decoder blocks above the op switch each catch this OOB through their ceph_decode_*_safe() / ceph_decode_need() helpers, but they sit behind a hdr.version-gated if, so a malicious or compromised MDS that sets msg->hdr.version = 1 reaches the IMPORT path with no version-gated decoder having validated snap_trace_len. The shape has been present since ceph_handle_caps() was introduced. Validate snap_trace_len against the message front buffer before consuming it, using the canonical ceph_decode_need() / ceph_has_room() helper. The helper bounds the length with subtraction (n <= end - p, guarded by end >= p) rather than pointer addition, so it is wrap-safe for the attacker-controlled u32 length on 32-bit builds where p + snap_trace_len could overflow the address space. This matches the rest of the ceph decode path (e.g. the pool_ns_len check a few lines below), and the existing goto bad cleanup already covers this exit path. Cc: stable(a)vger.kernel.org Fixes: a8599bd821d0 ("ceph: capability management") Signed-off-by: Bryam Vargas <hexlabsecurity(a)proton.me> Reviewed-by: Viacheslav Dubeyko <Slava.Dubeyko(a)ibm.com> Signed-off-by: Ilya Dryomov <idryomov(a)gmail.com> Conflicts: fs/ceph/caps.c [Context conflicts unmerged commit 8ea412e18131 ("ceph: improve caps debugging output")] Signed-off-by: Pan Taixi <pantaixi1(a)huawei.com> --- fs/ceph/caps.c | 1 + 1 file changed, 1 insertion(+) diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c index b631719941c2..80591f37dd77 100644 --- a/fs/ceph/caps.c +++ b/fs/ceph/caps.c @@ -3765,10 +3765,11 @@ void ceph_handle_caps(struct ceph_mds_session *session, seq = le32_to_cpu(h->seq); mseq = le32_to_cpu(h->migrate_seq); snaptrace = h + 1; snaptrace_len = le32_to_cpu(h->snap_trace_len); + ceph_decode_need(&snaptrace, end, snaptrace_len, bad); p = snaptrace + snaptrace_len; if (msg_version >= 2) { u32 flock_len; ceph_decode_32_safe(&p, end, flock_len, bad); -- 2.34.1

2 1

[PATCH OLK-6.6] nfs: use nfsi->rwsem to protect traversal of the file lock list
by Zizhi Wo 18 Aug '26

18 Aug '26

From: Yang Erkun <yangerkun(a)huawei.com> hulk inclusion category: bugfix bugzilla: https://atomgit.com/openeuler/kernel/issues/9787 -------------------------------- Lingfeng identified a bug and suggested two solutions, but both appear to have issues. Generally, we cannot release flc_lock while iterating over the file lock list to avoid use-after-free (UAF) problems with file locks. However, functions like nfs_delegation_claim_locks and nfs4_reclaim_locks cannot adhere to this rule because recover_lock or nfs4_lock_delegation_recall may take a long time. To resolve this, NFS switches to using nfsi->rwsem for the same protection, and nfs_reclaim_locks follows this approach. Although nfs_delegation_claim_locks uses so_delegreturn_mutex instead, this is inadequate since a single inode can have multiple nfs4_state instances. Therefore, the fix is to also use nfsi->rwsem in this case. Furthermore, after commit c69899a17ca4 ("NFSv4: Update of VFS byte range lock must be atomic with the stateid update"), the functions nfs4_locku_done and nfs4_lock_done also break this rule because they call locks_lock_inode_wait without holding nfsi->rwsem. Simply adding this protection could cause many deadlocks, so instead, the call to locks_lock_inode_wait is moved into _nfs4_proc_setlk. Regarding the bug fixed by commit c69899a17ca4 ("NFSv4: Update of VFS byte range lock must be atomic with the stateid update"), it has been resolved after commit 0460253913e5 ("NFSv4: nfs4_do_open() is incorrectly triggering state recovery") because all slots are drained before calling nfs4_do_reclaim, which prevents concurrent stateid changes along this path. Also, nfs_delegation_claim_locks does not cause this concurrency either since when _nfs4_proc_setlk is called with NFS_DELEGATED_STATE, no RPC is sent, so nfs4_lock_done is not called. Therefore, nfs4_lock_delegation_recall from nfs_delegation_claim_locks is the first time the stateid is set. Reported-by: Li Lingfeng <lilingfeng3(a)huawei.com> Closes: https://lore.kernel.org/all/20250419085709.1452492-1-lilingfeng3@huawei.com/ Closes: https://lore.kernel.org/all/20250715030559.2906634-1-lilingfeng3@huawei.com/ Fixes: c69899a17ca4 ("NFSv4: Update of VFS byte range lock must be atomic with the stateid update") Signed-off-by: Yang Erkun <yangerkun(a)huawei.com> Signed-off-by: Li Lingfeng <lilingfeng3(a)huawei.com> Signed-off-by: Zizhi Wo <wozizhi(a)huawei.com> --- fs/nfs/delegation.c | 9 ++++++++- fs/nfs/nfs4proc.c | 22 +++++++++++----------- include/linux/nfs_xdr.h | 1 - 3 files changed, 19 insertions(+), 13 deletions(-) diff --git a/fs/nfs/delegation.c b/fs/nfs/delegation.c index bbd582d8a7dc..fb93b9edb39b 100644 --- a/fs/nfs/delegation.c +++ b/fs/nfs/delegation.c @@ -143,35 +143,42 @@ int nfs4_check_delegation(struct inode *inode, fmode_t flags) } static int nfs_delegation_claim_locks(struct nfs4_state *state, const nfs4_stateid *stateid) { struct inode *inode = state->inode; + struct nfs_inode *nfsi = NFS_I(inode); struct file_lock *fl; struct file_lock_context *flctx = locks_inode_context(inode); struct list_head *list; int status = 0; if (flctx == NULL) goto out; list = &flctx->flc_posix; + + /* Guard against delegation returns and new lock/unlock calls */ + down_write(&nfsi->rwsem); spin_lock(&flctx->flc_lock); restart: list_for_each_entry(fl, list, fl_list) { if (nfs_file_open_context(fl->fl_file)->state != state) continue; spin_unlock(&flctx->flc_lock); status = nfs4_lock_delegation_recall(fl, state, stateid); - if (status < 0) + if (status < 0) { + up_write(&nfsi->rwsem); goto out; + } spin_lock(&flctx->flc_lock); } if (list == &flctx->flc_posix) { list = &flctx->flc_flock; goto restart; } spin_unlock(&flctx->flc_lock); + up_write(&nfsi->rwsem); out: return status; } static int nfs_delegation_claim_opens(struct inode *inode, diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 58739de4e793..f7163d9de438 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -6953,11 +6953,10 @@ static void nfs4_locku_done(struct rpc_task *task, void *data) if (!nfs4_sequence_done(task, &calldata->res.seq_res)) return; switch (task->tk_status) { case 0: renew_lease(calldata->server, calldata->timestamp); - locks_lock_inode_wait(calldata->lsp->ls_state->inode, &calldata->fl); if (nfs4_update_lock_stateid(calldata->lsp, &calldata->res.stateid)) break; fallthrough; case -NFS4ERR_ADMIN_REVOKED: @@ -7217,15 +7216,10 @@ static void nfs4_lock_done(struct rpc_task *task, void *calldata) data->rpc_status = task->tk_status; switch (task->tk_status) { case 0: renew_lease(NFS_SERVER(d_inode(data->ctx->dentry)), data->timestamp); - if (data->arg.new_lock && !data->cancelled) { - data->fl.fl_flags &= ~(FL_SLEEP | FL_ACCESS); - if (locks_lock_inode_wait(lsp->ls_state->inode, &data->fl) < 0) - goto out_restart; - } if (data->arg.new_lock_owner != 0) { nfs_confirm_seqid(&lsp->ls_seqid, 0); nfs4_stateid_copy(&lsp->ls_stateid, &data->res.stateid); set_bit(NFS_LOCK_INITIALIZED, &lsp->ls_flags); } else if (!nfs4_update_lock_stateid(lsp, &data->res.stateid)) @@ -7330,15 +7324,14 @@ static int _nfs4_do_setlk(struct nfs4_state *state, int cmd, struct file_lock *f nfs4_init_sequence(&data->arg.seq_args, &data->res.seq_res, 1, recovery_type > NFS_LOCK_NEW); msg.rpc_argp = &data->arg; msg.rpc_resp = &data->res; task_setup_data.callback_data = data; - if (recovery_type > NFS_LOCK_NEW) { - if (recovery_type == NFS_LOCK_RECLAIM) - data->arg.reclaim = NFS_LOCK_RECLAIM; - } else - data->arg.new_lock = 1; + + if (recovery_type == NFS_LOCK_RECLAIM) + data->arg.reclaim = NFS_LOCK_RECLAIM; + task = rpc_run_task(&task_setup_data); if (IS_ERR(task)) return PTR_ERR(task); ret = rpc_wait_for_completion_task(task); if (ret == 0) { @@ -7446,10 +7439,17 @@ static int _nfs4_proc_setlk(struct nfs4_state *state, int cmd, struct file_lock goto out; } up_read(&nfsi->rwsem); mutex_unlock(&sp->so_delegreturn_mutex); status = _nfs4_do_setlk(state, cmd, request, NFS_LOCK_NEW); + if (status) + goto out; + + down_read(&nfsi->rwsem); + request->fl_flags &= ~(FL_SLEEP | FL_ACCESS); + status = locks_lock_inode_wait(state->inode, request); + up_read(&nfsi->rwsem); out: request->fl_flags = fl_flags; return status; } diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h index 5d2ad7bd92c3..ed27379ff302 100644 --- a/include/linux/nfs_xdr.h +++ b/include/linux/nfs_xdr.h @@ -561,11 +561,10 @@ struct nfs_lock_args { struct nfs_seqid * open_seqid; nfs4_stateid open_stateid; struct nfs_lowner lock_owner; unsigned char block : 1; unsigned char reclaim : 1; - unsigned char new_lock : 1; unsigned char new_lock_owner : 1; }; struct nfs_lock_res { struct nfs4_sequence_res seq_res; -- 2.52.0

2 1

[PATCH OLK-6.6 V1] IB/mad: Drop unmatched RMPP responses before reassembly
by Yao Yiqi 18 Aug '26

18 Aug '26

From: Michael Bommarito <michael.bommarito(a)gmail.com> stable inclusion from stable-v6.6.148 commit dfa535c94406c03d3f0c869ef3ba5528e395737c category: bugfix bugzilla: https://atomgit.com/src-openeuler/kernel/issues/17033 CVE: CVE-2026-68425 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- [ Upstream commit d2e52d610b9b09694261632340b801a421e0b0c5 ] Kernel-handled RMPP receive processing starts reassembly for active DATA responses before the response is matched to an outstanding send. The normal match happens later, after ib_process_rmpp_recv_wc() has either assembled a complete message or consumed the segment. That ordering lets an unsolicited response that routes to a kernel RMPP agent by the high TID bits allocate or extend RMPP receive state before the full TID and source address are checked against a real request. A reordered burst can therefore reach the receive-side insertion path even though the response would not match any send. For kernel-handled RMPP DATA responses, require the existing ib_find_send_mad() match before entering RMPP reassembly. The matcher already checks the full TID, management class and source address/GID against the agent wait, backlog and in-flight send lists. If there is no match, drop the response without creating RMPP state. This leaves the RMPP window behavior unchanged and only rejects responses that have no corresponding request. Fixes: fa619a77046b ("[PATCH] IB: Add RMPP implementation") Assisted-by: Codex:gpt-5-5-xhigh Signed-off-by: Michael Bommarito <michael.bommarito(a)gmail.com> Link: https://patch.msgid.link/3170ff3bc389a930bb1641f2caa394a0b2241579.178077490… Signed-off-by: Leon Romanovsky <leon(a)kernel.org> Signed-off-by: Sasha Levin <sashal(a)kernel.org> Conflicts: drivers/infiniband/core/mad.c [ context conflict ] Signed-off-by: Yao Yiqi <yaoyiqi3(a)huawei.com> --- drivers/infiniband/core/mad.c | 30 ++++++++++++++++++++++++++++++ 1 file changed, 30 insertions(+) diff --git a/drivers/infiniband/core/mad.c b/drivers/infiniband/core/mad.c index 242434c09e8d..bddb1c607aff 100644 --- a/drivers/infiniband/core/mad.c +++ b/drivers/infiniband/core/mad.c @@ -1778,6 +1778,24 @@ void ib_mark_mad_done(struct ib_mad_send_wr_private *mad_send_wr) &mad_send_wr->mad_agent_priv->done_list); } +static bool is_kernel_rmpp_data_response(struct ib_mad_agent_private *agent, + struct ib_mad_recv_wc *mad_recv_wc) +{ + const struct ib_mad_hdr *mad_hdr = &mad_recv_wc->recv_buf.mad->mad_hdr; + struct ib_rmpp_mad *rmpp_mad; + + if (!ib_mad_kernel_rmpp_agent(&agent->agent) || + !ib_response_mad(mad_hdr) || + !ib_is_mad_class_rmpp(mad_hdr->mgmt_class)) + return false; + + rmpp_mad = (struct ib_rmpp_mad *)mad_recv_wc->recv_buf.mad; + + return (ib_get_rmpp_flags(&rmpp_mad->rmpp_hdr) & + IB_MGMT_RMPP_FLAG_ACTIVE) && + rmpp_mad->rmpp_hdr.rmpp_type == IB_MGMT_RMPP_TYPE_DATA; +} + static void ib_mad_complete_recv(struct ib_mad_agent_private *mad_agent_priv, struct ib_mad_recv_wc *mad_recv_wc) { @@ -1796,6 +1814,18 @@ static void ib_mad_complete_recv(struct ib_mad_agent_private *mad_agent_priv, } list_add(&mad_recv_wc->recv_buf.list, &mad_recv_wc->rmpp_list); + if (is_kernel_rmpp_data_response(mad_agent_priv, mad_recv_wc)) { + spin_lock_irqsave(&mad_agent_priv->lock, flags); + mad_send_wr = ib_find_send_mad(mad_agent_priv, mad_recv_wc); + spin_unlock_irqrestore(&mad_agent_priv->lock, flags); + + if (!mad_send_wr) { + ib_free_recv_mad(mad_recv_wc); + deref_mad_agent(mad_agent_priv); + return; + } + } + if (ib_mad_kernel_rmpp_agent(&mad_agent_priv->agent)) { mad_recv_wc = ib_process_rmpp_recv_wc(mad_agent_priv, mad_recv_wc); -- 2.34.1

2 1

[PATCH OLK-6.6] nfs: use nfsi->rwsem to protect traversal of the file lock list
by Zizhi Wo 18 Aug '26

18 Aug '26

From: Yang Erkun <yangerkun(a)huawei.com> hulk inclusion category: bugfix bugzilla: http://openeuler.huawei.com/bugzilla/show_bug.cgi?id=190626 -------------------------------- Lingfeng identified a bug and suggested two solutions, but both appear to have issues. Generally, we cannot release flc_lock while iterating over the file lock list to avoid use-after-free (UAF) problems with file locks. However, functions like nfs_delegation_claim_locks and nfs4_reclaim_locks cannot adhere to this rule because recover_lock or nfs4_lock_delegation_recall may take a long time. To resolve this, NFS switches to using nfsi->rwsem for the same protection, and nfs_reclaim_locks follows this approach. Although nfs_delegation_claim_locks uses so_delegreturn_mutex instead, this is inadequate since a single inode can have multiple nfs4_state instances. Therefore, the fix is to also use nfsi->rwsem in this case. Furthermore, after commit c69899a17ca4 ("NFSv4: Update of VFS byte range lock must be atomic with the stateid update"), the functions nfs4_locku_done and nfs4_lock_done also break this rule because they call locks_lock_inode_wait without holding nfsi->rwsem. Simply adding this protection could cause many deadlocks, so instead, the call to locks_lock_inode_wait is moved into _nfs4_proc_setlk. Regarding the bug fixed by commit c69899a17ca4 ("NFSv4: Update of VFS byte range lock must be atomic with the stateid update"), it has been resolved after commit 0460253913e5 ("NFSv4: nfs4_do_open() is incorrectly triggering state recovery") because all slots are drained before calling nfs4_do_reclaim, which prevents concurrent stateid changes along this path. Also, nfs_delegation_claim_locks does not cause this concurrency either since when _nfs4_proc_setlk is called with NFS_DELEGATED_STATE, no RPC is sent, so nfs4_lock_done is not called. Therefore, nfs4_lock_delegation_recall from nfs_delegation_claim_locks is the first time the stateid is set. Reported-by: Li Lingfeng <lilingfeng3(a)huawei.com> Closes: https://lore.kernel.org/all/20250419085709.1452492-1-lilingfeng3@huawei.com/ Closes: https://lore.kernel.org/all/20250715030559.2906634-1-lilingfeng3@huawei.com/ Fixes: c69899a17ca4 ("NFSv4: Update of VFS byte range lock must be atomic with the stateid update") Signed-off-by: Yang Erkun <yangerkun(a)huawei.com> Signed-off-by: Li Lingfeng <lilingfeng3(a)huawei.com> Signed-off-by: Zizhi Wo <wozizhi(a)huawei.com> --- fs/nfs/delegation.c | 9 ++++++++- fs/nfs/nfs4proc.c | 22 +++++++++++----------- include/linux/nfs_xdr.h | 1 - 3 files changed, 19 insertions(+), 13 deletions(-) diff --git a/fs/nfs/delegation.c b/fs/nfs/delegation.c index bbd582d8a7dc..fb93b9edb39b 100644 --- a/fs/nfs/delegation.c +++ b/fs/nfs/delegation.c @@ -143,35 +143,42 @@ int nfs4_check_delegation(struct inode *inode, fmode_t flags) } static int nfs_delegation_claim_locks(struct nfs4_state *state, const nfs4_stateid *stateid) { struct inode *inode = state->inode; + struct nfs_inode *nfsi = NFS_I(inode); struct file_lock *fl; struct file_lock_context *flctx = locks_inode_context(inode); struct list_head *list; int status = 0; if (flctx == NULL) goto out; list = &flctx->flc_posix; + + /* Guard against delegation returns and new lock/unlock calls */ + down_write(&nfsi->rwsem); spin_lock(&flctx->flc_lock); restart: list_for_each_entry(fl, list, fl_list) { if (nfs_file_open_context(fl->fl_file)->state != state) continue; spin_unlock(&flctx->flc_lock); status = nfs4_lock_delegation_recall(fl, state, stateid); - if (status < 0) + if (status < 0) { + up_write(&nfsi->rwsem); goto out; + } spin_lock(&flctx->flc_lock); } if (list == &flctx->flc_posix) { list = &flctx->flc_flock; goto restart; } spin_unlock(&flctx->flc_lock); + up_write(&nfsi->rwsem); out: return status; } static int nfs_delegation_claim_opens(struct inode *inode, diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 58739de4e793..f7163d9de438 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -6953,11 +6953,10 @@ static void nfs4_locku_done(struct rpc_task *task, void *data) if (!nfs4_sequence_done(task, &calldata->res.seq_res)) return; switch (task->tk_status) { case 0: renew_lease(calldata->server, calldata->timestamp); - locks_lock_inode_wait(calldata->lsp->ls_state->inode, &calldata->fl); if (nfs4_update_lock_stateid(calldata->lsp, &calldata->res.stateid)) break; fallthrough; case -NFS4ERR_ADMIN_REVOKED: @@ -7217,15 +7216,10 @@ static void nfs4_lock_done(struct rpc_task *task, void *calldata) data->rpc_status = task->tk_status; switch (task->tk_status) { case 0: renew_lease(NFS_SERVER(d_inode(data->ctx->dentry)), data->timestamp); - if (data->arg.new_lock && !data->cancelled) { - data->fl.fl_flags &= ~(FL_SLEEP | FL_ACCESS); - if (locks_lock_inode_wait(lsp->ls_state->inode, &data->fl) < 0) - goto out_restart; - } if (data->arg.new_lock_owner != 0) { nfs_confirm_seqid(&lsp->ls_seqid, 0); nfs4_stateid_copy(&lsp->ls_stateid, &data->res.stateid); set_bit(NFS_LOCK_INITIALIZED, &lsp->ls_flags); } else if (!nfs4_update_lock_stateid(lsp, &data->res.stateid)) @@ -7330,15 +7324,14 @@ static int _nfs4_do_setlk(struct nfs4_state *state, int cmd, struct file_lock *f nfs4_init_sequence(&data->arg.seq_args, &data->res.seq_res, 1, recovery_type > NFS_LOCK_NEW); msg.rpc_argp = &data->arg; msg.rpc_resp = &data->res; task_setup_data.callback_data = data; - if (recovery_type > NFS_LOCK_NEW) { - if (recovery_type == NFS_LOCK_RECLAIM) - data->arg.reclaim = NFS_LOCK_RECLAIM; - } else - data->arg.new_lock = 1; + + if (recovery_type == NFS_LOCK_RECLAIM) + data->arg.reclaim = NFS_LOCK_RECLAIM; + task = rpc_run_task(&task_setup_data); if (IS_ERR(task)) return PTR_ERR(task); ret = rpc_wait_for_completion_task(task); if (ret == 0) { @@ -7446,10 +7439,17 @@ static int _nfs4_proc_setlk(struct nfs4_state *state, int cmd, struct file_lock goto out; } up_read(&nfsi->rwsem); mutex_unlock(&sp->so_delegreturn_mutex); status = _nfs4_do_setlk(state, cmd, request, NFS_LOCK_NEW); + if (status) + goto out; + + down_read(&nfsi->rwsem); + request->fl_flags &= ~(FL_SLEEP | FL_ACCESS); + status = locks_lock_inode_wait(state->inode, request); + up_read(&nfsi->rwsem); out: request->fl_flags = fl_flags; return status; } diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h index 5d2ad7bd92c3..ed27379ff302 100644 --- a/include/linux/nfs_xdr.h +++ b/include/linux/nfs_xdr.h @@ -561,11 +561,10 @@ struct nfs_lock_args { struct nfs_seqid * open_seqid; nfs4_stateid open_stateid; struct nfs_lowner lock_owner; unsigned char block : 1; unsigned char reclaim : 1; - unsigned char new_lock : 1; unsigned char new_lock_owner : 1; }; struct nfs_lock_res { struct nfs4_sequence_res seq_res; -- 2.52.0

2 1

[PATCH OLK-6.6 v3 0/8] irqbypass: Cleanups and a perf improvement
by Qiqi Li 18 Aug '26

18 Aug '26

From: Qiqi Li <liqiqi23(a)huawei.com> The two primary goals of this series are to make the irqbypass concept easier to understand, and to address the terrible performance that can result from using a list to track connections. Provides additional adaptation for the ub vfio device in openEuler. Sean Christopherson (8): irqbypass: Drop pointless and misleading THIS_MODULE get/put irqbypass: Drop superfluous might_sleep() annotations irqbypass: Take ownership of producer/consumer token tracking irqbypass: Explicitly track producer and consumer bindings irqbypass: Use paired consumer/producer to disconnect during unregister irqbypass: Use guard(mutex) in lieu of manual lock+unlock irqbypass: Use xarray to track producers and consumers irqbypass: Require producers to pass in Linux IRQ number during registration arch/x86/kvm/x86.c | 4 +- drivers/vfio/pci/vfio_pci_intrs.c | 10 +- drivers/vfio/ubus/vfio_ub_intrs.c | 12 +- drivers/vhost/vdpa.c | 10 +- include/linux/irqbypass.h | 46 +++---- virt/kvm/eventfd.c | 7 +- virt/lib/irqbypass.c | 192 +++++++++++------------------- 7 files changed, 112 insertions(+), 169 deletions(-) -- 2.43.0

2 9

[PATCH OLK-5.10] tty: serial: kgdboc: Fix 8250_* kgdb over serial
by Yi Yang 18 Aug '26

18 Aug '26

From: Michael Trimarchi <michael(a)amarulasolutions.com> mainline inclusion from mainline-v6.8-rc1 commit 788aeef392d27545ae99af2875068a9dd0531d5f category: bugfix bugzilla: https://atomgit.com/openeuler/kernel/issues/9772 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… -------------------------------- Check if port type is not PORT_UNKNOWN during poll_init. The kgdboc calls the tty_find_polling_driver that check if the serial is able to use poll_init. The poll_init calls the uart uart_poll_init that try to configure the uart with the selected boot parameters. The uart must be ready before setting parameters. Seems that PORT_UNKNOWN is already used by other functions in serial_core to detect uart status, so use the same to avoid to use it in invalid state. The crash happen for instance in am62x architecture where the 8250 register the platform driver after the 8250 core is initialized. Follow the report crash coming from KGDB Thread 2 received signal SIGSEGV, Segmentation fault. [Switching to Thread 1] _outb (addr=<optimized out>, value=<optimized out>) at ./include/asm-generic/io.h:584 584 __raw_writeb(value, PCI_IOBASE + addr); (gdb) bt This section of the code is too early because in this case the omap serial is not probed Thread 2 received signal SIGSEGV, Segmentation fault. [Switching to Thread 1] _outb (addr=<optimized out>, value=<optimized out>) at ./include/asm-generic/io.h:584 584 __raw_writeb(value, PCI_IOBASE + addr); (gdb) bt Thread 2 received signal SIGSEGV, Segmentation fault. [Switching to Thread 1] _outb (addr=<optimized out>, value=<optimized out>) at ./include/asm-generic/io.h:584 584 __raw_writeb(value, PCI_IOBASE + addr); (gdb) bt 0 _outb (addr=<optimized out>, value=<optimized out>) at ./include/asm-generic/io.h:584 1 logic_outb (value=0 '\000', addr=18446739675637874689) at lib/logic_pio.c:299 2 0xffff80008082dfcc in io_serial_out (p=0x0, offset=16760830, value=0) at drivers/tty/serial/8250/8250_port.c:416 3 0xffff80008082fe34 in serial_port_out (value=<optimized out>, offset=<optimized out>, up=<optimized out>) at ./include/linux/serial_core.h:677 4 serial8250_do_set_termios (port=0xffff8000828ee940 <serial8250_ports+1568>, termios=0xffff80008292b93c, old=0x0) at drivers/tty/serial/8250/8250_port.c:2860 5 0xffff800080830064 in serial8250_set_termios (port=0xfffffbfffe800000, termios=0xffbffe, old=0x0) at drivers/tty/serial/8250/8250_port.c:2912 6 0xffff80008082571c in uart_set_options (port=0xffff8000828ee940 <serial8250_ports+1568>, co=0x0, baud=115200, parity=110, bits=8, flow=110) at drivers/tty/serial/serial_core.c:2285 7 0xffff800080828434 in uart_poll_init (driver=0xfffffbfffe800000, line=16760830, options=0xffff8000828f7506 <config+6> "115200n8") at drivers/tty/serial/serial_core.c:2656 8 0xffff800080801690 in tty_find_polling_driver (name=0xffff8000828f7500 <config> "ttyS2,115200n8", line=0xffff80008292ba90) at drivers/tty/tty_io.c:410 9 0xffff80008086c0b0 in configure_kgdboc () at drivers/tty/serial/kgdboc.c:194 10 0xffff80008086c1ec in kgdboc_probe (pdev=0xfffffbfffe800000) at drivers/tty/serial/kgdboc.c:249 11 0xffff8000808b399c in platform_probe (_dev=0xffff000000ebb810) at drivers/base/platform.c:1404 12 0xffff8000808b0b44 in call_driver_probe (drv=<optimized out>, dev=<optimized out>) at drivers/base/dd.c:579 13 really_probe (dev=0xffff000000ebb810, drv=0xffff80008277f138 <kgdboc_platform_driver+48>) at drivers/base/dd.c:658 14 0xffff8000808b0d2c in __driver_probe_device (drv=0xffff80008277f138 <kgdboc_platform_driver+48>, dev=0xffff000000ebb810) at drivers/base/dd.c:800 15 0xffff8000808b0eb8 in driver_probe_device (drv=0xfffffbfffe800000, dev=0xffff000000ebb810) at drivers/base/dd.c:830 16 0xffff8000808b0ff4 in __device_attach_driver (drv=0xffff80008277f138 <kgdboc_platform_driver+48>, _data=0xffff80008292bc48) at drivers/base/dd.c:958 17 0xffff8000808ae970 in bus_for_each_drv (bus=0xfffffbfffe800000, start=0x0, data=0xffff80008292bc48, fn=0xffff8000808b0f3c <__device_attach_driver>) at drivers/base/bus.c:457 18 0xffff8000808b1408 in __device_attach (dev=0xffff000000ebb810, allow_async=true) at drivers/base/dd.c:1030 19 0xffff8000808b16d8 in device_initial_probe (dev=0xfffffbfffe800000) at drivers/base/dd.c:1079 20 0xffff8000808af9f4 in bus_probe_device (dev=0xffff000000ebb810) at drivers/base/bus.c:532 21 0xffff8000808ac77c in device_add (dev=0xfffffbfffe800000) at drivers/base/core.c:3625 22 0xffff8000808b3428 in platform_device_add (pdev=0xffff000000ebb800) at drivers/base/platform.c:716 23 0xffff800081b5dc0c in init_kgdboc () at drivers/tty/serial/kgdboc.c:292 24 0xffff800080014db0 in do_one_initcall (fn=0xffff800081b5dba4 <init_kgdboc>) at init/main.c:1236 25 0xffff800081b0114c in do_initcall_level (command_line=<optimized out>, level=<optimized out>) at init/main.c:1298 26 do_initcalls () at init/main.c:1314 27 do_basic_setup () at init/main.c:1333 28 kernel_init_freeable () at init/main.c:1551 29 0xffff8000810271ec in kernel_init (unused=0xfffffbfffe800000) at init/main.c:1441 30 0xffff800080015e80 in ret_from_fork () at arch/arm64/kernel/entry.S:857 Reviewed-by: Douglas Anderson <dianders(a)chromium.org> Signed-off-by: Michael Trimarchi <michael(a)amarulasolutions.com> Link: https://lore.kernel.org/r/20231224131200.266224-1-michael@amarulasolutions.… Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> Conflicts: drivers/tty/serial/serial_core.c [Context conflicts.] Signed-off-by: Yi Yang <yiyang13(a)huawei.com> --- drivers/tty/serial/serial_core.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/tty/serial/serial_core.c b/drivers/tty/serial/serial_core.c index 86e028023b66..8207e4dec81d 100644 --- a/drivers/tty/serial/serial_core.c +++ b/drivers/tty/serial/serial_core.c @@ -2512,7 +2512,8 @@ static int uart_poll_init(struct tty_driver *driver, int line, char *options) mutex_lock(&tport->mutex); port = uart_port_check(state); - if (!port || !(port->ops->poll_get_char && port->ops->poll_put_char)) { + if (!port || port->type == PORT_UNKNOWN || + !(port->ops->poll_get_char && port->ops->poll_put_char)) { ret = -1; goto out; } -- 2.25.1

2 1