6.1.75-rc1 review

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: John Fastabend <[email protected]>

[ Upstream commit 16b2f264983dc264c1560cc0170e760dec1bf54f ]

When sockets are added to a sockmap or sockhash we allocate and init a psock. Then update the proto ops with sock_map_init_proto the flow is

sock_hash_update_common
sock_map_link
psock = sock_map_psock_get_checked() <-returns existing psock
sock_map_init_proto(sk, psock) <- updates sk_proto

If the socket is already in a map this results in the sock_map_init_proto being called multiple times on the same socket. We do this because when a socket is added to multiple maps this might result in a new set of BPF programs being attached to the socket requiring an updated ops struct.

This creates a rule where it must be safe to call psock_update_sk_prot multiple times. When we added a fix for UAF through unix sockets in patch 4dd9a38a753fc we broke this rule by adding a sock_hold in that path to ensure the sock is not released. The result is if a af_unix stream sock is placed in multiple maps it results in a memory leak because we call sock_hold multiple times with only a single sock_put on it.

---
 net/unix/unix_bpf.c | 21 ++++++++++++++++++---
 1 file changed, 18 insertions(+), 3 deletions(-)

index 7ea7c3a0d0d0..bd84785bf8d6 100644
@@ -161,15 +161,30 @@ int unix_stream_bpf_update_proto(struct sock *sk, struct sk_psock *psock, bool r
 {
 	struct sock *sk_pair;
 
+	/* Restore does not decrement the sk_pair reference yet because we must
+	 * keep the a reference to the socket until after an RCU grace period
+	 * and any pending sends have completed.
+	 */
 	if (restore) {
 		sk->sk_write_space = psock->saved_write_space;
 		sock_replace_proto(sk, psock->sk_proto);
 		return 0;
 	}
 
-	sk_pair = unix_peer(sk);
-	sock_hold(sk_pair);
-	psock->sk_pair = sk_pair;
+	/* psock_update_sk_prot can be called multiple times if psock is
+	 * added to multiple maps and/or slots in the same map. There is
+	 * also an edge case where replacing a psock with itself can trigger
+	 * an extra psock_update_sk_prot during the insert process. So it
+	 * must be safe to do multiple calls. Here we need to ensure we don't
+	 * increment the refcnt through sock_hold many times. There will only
+	 * be a single matching destroy operation.
+	 */
+	if (!psock->sk_pair) {
+		sk_pair = unix_peer(sk);
+		sock_hold(sk_pair);
+		psock->sk_pair = sk_pair;
+	}
+
 	unix_stream_bpf_check_needs_rebuild(psock->sk_proto);
 	sock_replace_proto(sk, &unix_stream_bpf_prot);
 	return 0;

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Eric Dumazet <[email protected]>

[ Upstream commit 4746b36b1abe11ca32987b2d21e1e770deab17cc ]

For some reason sctp_poll() generates EPOLLERR if sk->sk_error_queue is not empty but recvmsg() can not drain the error queue yet.

This is needed to better support timestamping.

I had to export inet_recv_error(), since sctp
can be compiled as a module.

Stable-dep-of: a562c0a2d651 ("sctp: fix busy polling")

---
 net/ipv4/af_inet.c | 1 +
 net/sctp/socket.c  | 3 +++
 2 files changed, 4 insertions(+)

index 347c3768df6e..c13b8ed63f87 100644
@@ -1642,6 +1642,7 @@ int inet_recv_error(struct sock *sk, struct msghdr *msg, int len, int *addr_len)
 #endif
 	return -EINVAL;
 }
+EXPORT_SYMBOL(inet_recv_error);
 
 int inet_gro_complete(struct sk_buff *skb, int nhoff)
 {

index e25dc1709131..da56832179f0 100644
@@ -2098,6 +2098,9 @@ static int sctp_recvmsg(struct sock *sk, struct msghdr *msg, size_t len,
 	pr_debug("%s: sk:%p, msghdr:%p, len:%zd, flags:0x%x, addr_len:%p)\n",
 		 __func__, sk, msg, len, flags, addr_len);
 
+	if (unlikely(flags & MSG_ERRQUEUE))
+		return inet_recv_error(sk, msg, len, addr_len);
+
 	lock_sock(sk);
 
 	if (sctp_style(sk, TCP) && !sctp_sstate(sk, ESTABLISHED) &&

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Eric Dumazet <[email protected]>

[ Upstream commit a562c0a2d651e040681b0bfce9b4d229ac3b0b8c ]

Busy polling while holding the socket lock makes litle sense, because incoming packets wont reach our receive queue.

---
 net/sctp/socket.c | 10 ++++------
 1 file changed, 4 insertions(+), 6 deletions(-)

index da56832179f0..237a6b04adf6 100644
@@ -2101,6 +2101,10 @@ static int sctp_recvmsg(struct sock *sk, struct msghdr *msg, size_t len,
 	if (unlikely(flags & MSG_ERRQUEUE))
 		return inet_recv_error(sk, msg, len, addr_len);
 
+	if (sk_can_busy_loop(sk) &&
+	    skb_queue_empty_lockless(&sk->sk_receive_queue))
+		sk_busy_loop(sk, flags & MSG_DONTWAIT);
+
 	lock_sock(sk);
 
 	if (sctp_style(sk, TCP) && !sctp_sstate(sk, ESTABLISHED) &&

@@ -9041,12 +9045,6 @@ struct sk_buff *sctp_skb_recv_datagram(struct sock *sk, int flags, int *err)
 		if (sk->sk_shutdown & RCV_SHUTDOWN)
 			break;
 
-		if (sk_can_busy_loop(sk)) {
-			sk_busy_loop(sk, flags & MSG_DONTWAIT);
-
-			if (!skb_queue_empty_lockless(&sk->sk_receive_queue))
-				continue;
-		}
 
 		/* User doesn't want to wait.  */
 		error = -EAGAIN;

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Tao Liu <[email protected]>

[ Upstream commit 3f14b377d01d8357eba032b4cabc8c1149b458b6 ]

act_ct adds skb->users before defragmentation. If frags arrive in order, the last frag's reference is reset in:

inet_frag_reasm_prepare
skb_morph

which is not straightforward.

However when frags arrive out of order, nobody unref the last frag, and all frags are leaked. The situation is even worse, as initiating packet capture can lead to a crash[0] when skb has been cloned and shared at the same time.

Fix the issue by removing skb_get() before defragmentation. act_ct returns TC_ACT_CONSUMED when defrag failed or in progress.

[0]:

[  843.804823] ------------[ cut here ]------------
[  843.809659] kernel BUG at net/core/skbuff.c:2091!
[  843.814516] invalid opcode: 0000 [#1] PREEMPT SMP
[  843.819296] CPU: 7 PID: 0 Comm: swapper/7 Kdump: loaded Tainted: G S 6.7.0-rc3 #2
[  843.824107] Hardware name: XFUSION 1288H V6/BC13MBSBD, BIOS 1.29 11/25/2022
[  843.828953] RIP: 0010:pskb_expand_head+0x2ac/0x300
[  843.833805] Code: 8b 70 28 48 85 f6 74 82 48 83 c6 08 bf 01 00 00 00 e8 38 bd ff ff 8b 83 c0 00 00 00 48 03 83 c8 00 00 00 e9 62 ff ff ff 0f 0b <0f> 0b e8 8d d0 ff ff e9 b3 fd ff ff 81 7c 24 14 40 01 00 00 4c 89
[  843.843698] RSP: 0018:ffffc9000cce07c0 EFLAGS: 00010202
[  843.848524] RAX: 0000000000000002 RBX: ffff88811a211d00 RCX: 0000000000000820
[  843.853299] RDX: 0000000000000640 RSI: 0000000000000000 RDI: ffff88811a211d00
[  843.857974] RBP: ffff888127d39518 R08: 00000000bee97314 R09: 0000000000000000
[  843.862584] R10: 0000000000000000 R11: ffff8881109f0000 R12: 0000000000000880
[  843.867147] R13: ffff888127d39580 R14: 0000000000000640 R15: ffff888170f7b900
[  843.871680] FS:  0000000000000000(0000) GS:ffff889ffffc0000(0000) knlGS:0000000000000000
[  843.876242] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[  843.880778] CR2: 00007fa42affcfb8 CR3: 000000011433a002 CR4: 0000000000770ef0
[  843.885336] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[  843.889809] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[  843.894229] PKRU: 55555554
[  843.898539] Call Trace:
[  843.902772]  <IRQ>
[  843.906922]  ? __die_body+0x1e/0x60
[  843.911032]  ? die+0x3c/0x60
[  843.915037]  ? do_trap+0xe2/0x110
[  843.918911]  ? pskb_expand_head+0x2ac/0x300
[  843.922687]  ? do_error_trap+0x65/0x80
[  843.926342]  ? pskb_expand_head+0x2ac/0x300
[  843.929905]  ? exc_invalid_op+0x50/0x60
[  843.933398]  ? pskb_expand_head+0x2ac/0x300
[  843.936835]  ? asm_exc_invalid_op+0x1a/0x20
[  843.940226]  ? pskb_expand_head+0x2ac/0x300
[  843.943580]  inet_frag_reasm_prepare+0xd1/0x240
[  843.946904]  ip_defrag+0x5d4/0x870
[  843.950132]  nf_ct_handle_fragments+0xec/0x130 [nf_conntrack]
[  843.953334]  tcf_ct_act+0x252/0xd90 [act_ct]
[  843.956473]  ? tcf_mirred_act+0x516/0x5a0 [act_mirred]
[  843.959657]  tcf_action_exec+0xa1/0x160
[  843.962823]  fl_classify+0x1db/0x1f0 [cls_flower]
[  843.966010]  ? skb_clone+0x53/0xc0
[  843.969173]  tcf_classify+0x24d/0x420
[  843.972333]  tc_run+0x8f/0xf0
[  843.975465]  __netif_receive_skb_core+0x67a/0x1080
[  843.978634]  ? dev_gro_receive+0x249/0x730
[  843.981759]  __netif_receive_skb_list_core+0x12d/0x260
[  843.984869]  netif_receive_skb_list_internal+0x1cb/0x2f0
[  843.987957]  ? mlx5e_handle_rx_cqe_mpwrq_rep+0xfa/0x1a0 [mlx5_core]
[  843.991170]  napi_complete_done+0x72/0x1a0
[  843.994305]  mlx5e_napi_poll+0x28c/0x6d0 [mlx5_core]
[  843.997501]  __napi_poll+0x25/0x1b0
[  844.000627]  net_rx_action+0x256/0x330
[  844.003705]  __do_softirq+0xb3/0x29b
[  844.006718]  irq_exit_rcu+0x9e/0xc0
[  844.009672]  common_interrupt+0x86/0xa0
[  844.012537]  </IRQ>
[  844.015285]  <TASK>
[  844.017937]  asm_common_interrupt+0x26/0x40
[  844.020591] RIP: 0010:acpi_safe_halt+0x1b/0x20
[  844.023247] Code: ff 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 40 00 65 48 8b 04 25 00 18 03 00 48 8b 00 a8 08 75 0c 66 90 0f 00 2d 81 d0 44 00 fb f4 <fa> c3 0f 1f 00 89 fa ec 48 8b 05 ee 88 ed 00 a9 00 00 00 80 75 11
[  844.028900] RSP: 0018:ffffc90000533e70 EFLAGS: 00000246
[  844.031725] RAX: 0000000000004000 RBX: 0000000000000001 RCX: 0000000000000000
[  844.034553] RDX: ffff889ffffc0000 RSI: ffffffff828b7f20 RDI: ffff88a090f45c64
[  844.037368] RBP: ffff88a0901a2800 R08: ffff88a090f45c00 R09: 00000000000317c0
[  844.040155] R10: 00ec812281150475 R11: ffff889fffff0e04 R12: ffffffff828b7fa0
[  844.042962] R13: ffffffff828b7f20 R14: 0000000000000001 R15: 0000000000000000
[  844.045819]  acpi_idle_enter+0x7b/0xc0
[  844.048621]  cpuidle_enter_state+0x7f/0x430
[  844.051451]  cpuidle_enter+0x2d/0x40
[  844.054279]  do_idle+0x1d4/0x240
[  844.057096]  cpu_startup_entry+0x2a/0x30
[  844.059934]  start_secondary+0x104/0x130
[  844.062787]  secondary_startup_64_no_verify+0x16b/0x16b
[  844.065674]  </TASK>

---
 net/sched/act_ct.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

index d6d33f854050..84e15116f18c 100644
@@ -863,7 +863,6 @@ static int tcf_ct_handle_fragments(struct net *net, struct sk_buff *skb,
 	if (err || !frag)
 		return err;
 
-	skb_get(skb);
 	mru = tc_skb_cb(skb)->mru;
 
 	if (family == NFPROTO_IPV4) {

@@ -1150,12 +1149,8 @@ static int tcf_ct_act(struct sk_buff *skb, const struct tc_action *a,
 	nh_ofs = skb_network_offset(skb);
 	skb_pull_rcsum(skb, nh_ofs);
 	err = tcf_ct_handle_fragments(net, skb, family, p->zone, &defrag);
-	if (err == -EINPROGRESS) {
-		retval = TC_ACT_STOLEN;
-		goto out_clear;
-	}
 	if (err)
-		goto drop;
+		goto out_frag;
 
 	err = tcf_ct_skb_network_trim(skb, family);
 	if (err)

@@ -1226,6 +1221,11 @@ static int tcf_ct_act(struct sk_buff *skb, const struct tc_action *a,
 		qdisc_skb_cb(skb)->pkt_len = skb->len;
 	return retval;
 
+out_frag:
+	if (err != -EINPROGRESS)
+		tcf_action_inc_drop_qstats(&c->common);
+	return TC_ACT_CONSUMED;
+
 drop:
 	tcf_action_inc_drop_qstats(&c->common);
 	return TC_ACT_SHOT;

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Asmaa Mnebhi <[email protected]>

[ Upstream commit ef210ef85d5cb543ce34a57803ed856d0c8c08c2 ]

Although the link is up, there is no ip assigned on setups with high background traffic. Nothing is transmitted nor received. The RX error count keeps on increasing. After several minutes, the RX error count stagnates and the GigE interface finally gets an ip.

The issue is that mlxbf_gige_rx_init() is called before phy_start(). As soon as the RX DMA is enabled in mlxbf_gige_rx_init(), the RX CI reaches the max of 128, and becomes equal to RX PI. RX CI doesn't decrease since the code hasn't ran phy_start yet.
Bring the PHY up before starting the RX.

---
 .../ethernet/mellanox/mlxbf_gige/mlxbf_gige_main.c | 14 +++++++-------
 .../ethernet/mellanox/mlxbf_gige/mlxbf_gige_rx.c   |  6 +++---
 2 files changed, 10 insertions(+), 10 deletions(-)

index 2292d63a279c..eafc0d3313fd 100644
@@ -147,14 +147,14 @@ static int mlxbf_gige_open(struct net_device *netdev)
 	 */
 	priv->valid_polarity = 0;
 
-	err = mlxbf_gige_rx_init(priv);
+	phy_start(phydev);
+
+	err = mlxbf_gige_tx_init(priv);
 	if (err)
 		goto free_irqs;
-	err = mlxbf_gige_tx_init(priv);
+	err = mlxbf_gige_rx_init(priv);
 	if (err)
-		goto rx_deinit;
-
-	phy_start(phydev);
+		goto tx_deinit;
 
 	netif_napi_add(netdev, &priv->napi, mlxbf_gige_poll);
 	napi_enable(&priv->napi);

@@ -176,8 +176,8 @@ static int mlxbf_gige_open(struct net_device *netdev)
 
 	return 0;
 
-rx_deinit:
-	mlxbf_gige_rx_deinit(priv);
+tx_deinit:
+	mlxbf_gige_tx_deinit(priv);
 
 free_irqs:
 	mlxbf_gige_free_irqs(priv);

index 227d01cace3f..699984358493 100644
@@ -142,6 +142,9 @@ int mlxbf_gige_rx_init(struct mlxbf_gige *priv)
 	writeq(MLXBF_GIGE_RX_MAC_FILTER_COUNT_PASS_EN,
 	       priv->base + MLXBF_GIGE_RX_MAC_FILTER_COUNT_PASS);
 
+	writeq(ilog2(priv->rx_q_entries),
+	       priv->base + MLXBF_GIGE_RX_WQE_SIZE_LOG2);
+
 	/* Clear MLXBF_GIGE_INT_MASK 'receive pkt' bit to
 	 * indicate readiness to receive interrupts
 	 */

@@ -154,9 +157,6 @@ int mlxbf_gige_rx_init(struct mlxbf_gige *priv)
 	data |= MLXBF_GIGE_RX_DMA_EN;
 	writeq(data, priv->base + MLXBF_GIGE_RX_DMA);
 
-	writeq(ilog2(priv->rx_q_entries),
-	       priv->base + MLXBF_GIGE_RX_WQE_SIZE_LOG2);
-
 	return 0;
 
 free_wqe_and_skb:

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Asmaa Mnebhi <[email protected]>

[ Upstream commit a460f4a684511e007bbf1700758a41f05d9981e6 ]

At the moment, the GigE port is enabled in the mlxbf_gige_probe function. If the mlxbf_gige_open is not executed, this could cause pause frames to increase in the case where there is high backgroud traffic. This results in clogging the port.
So move enabling the OOB port to mlxbf_gige_open.

---
 .../ethernet/mellanox/mlxbf_gige/mlxbf_gige_main.c   | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

index eafc0d3313fd..83c4659390fd 100644
@@ -130,9 +130,15 @@ static int mlxbf_gige_open(struct net_device *netdev)
 {
 	struct mlxbf_gige *priv = netdev_priv(netdev);
 	struct phy_device *phydev = netdev->phydev;
+	u64 control;
 	u64 int_en;
 	int err;
 
+	/* Perform general init of GigE block */
+	control = readq(priv->base + MLXBF_GIGE_CONTROL);
+	control |= MLXBF_GIGE_CONTROL_PORT_EN;
+	writeq(control, priv->base + MLXBF_GIGE_CONTROL);
+
 	err = mlxbf_gige_request_irqs(priv);
 	if (err)
 		return err;

@@ -279,7 +285,6 @@ static int mlxbf_gige_probe(struct platform_device *pdev)
 	void __iomem *plu_base;
 	void __iomem *base;
 	int addr, phy_irq;
-	u64 control;
 	int err;
 
 	base = devm_platform_ioremap_resource(pdev, MLXBF_GIGE_RES_MAC);

@@ -294,11 +299,6 @@ static int mlxbf_gige_probe(struct platform_device *pdev)
 	if (IS_ERR(plu_base))
 		return PTR_ERR(plu_base);
 
-	/* Perform general init of GigE block */
-	control = readq(base + MLXBF_GIGE_CONTROL);
-	control |= MLXBF_GIGE_CONTROL_PORT_EN;
-	writeq(control, base + MLXBF_GIGE_CONTROL);
-
 	netdev = devm_alloc_etherdev(&pdev->dev, sizeof(*priv));
 	if (!netdev)
 		return -ENOMEM;

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Eric Dumazet <[email protected]>

[ Upstream commit d375b98e0248980681e5e56b712026174d617198 ]

syzbot pointed out [1] that NEXTHDR_FRAGMENT handling is broken.

Reading frag_off can only be done if we pulled enough bytes to skb->head. Currently we might access garbage.

[1]

BUG: KMSAN: uninit-value in ip6_tnl_parse_tlv_enc_lim+0x94f/0xbb0
ip6_tnl_parse_tlv_enc_lim+0x94f/0xbb0
ipxip6_tnl_xmit net/ipv6/ip6_tunnel.c:1326 [inline]
ip6_tnl_start_xmit+0xab2/0x1a70 net/ipv6/ip6_tunnel.c:1432
__netdev_start_xmit include/linux/netdevice.h:4940 [inline]
netdev_start_xmit include/linux/netdevice.h:4954 [inline]
xmit_one net/core/dev.c:3548 [inline]
dev_hard_start_xmit+0x247/0xa10 net/core/dev.c:3564
__dev_queue_xmit+0x33b8/0x5130 net/core/dev.c:4349
dev_queue_xmit include/linux/netdevice.h:3134 [inline]
neigh_connected_output+0x569/0x660 net/core/neighbour.c:1592
neigh_output include/net/neighbour.h:542 [inline]
ip6_finish_output2+0x23a9/0x2b30 net/ipv6/ip6_output.c:137
ip6_finish_output+0x855/0x12b0 net/ipv6/ip6_output.c:222
NF_HOOK_COND include/linux/netfilter.h:303 [inline]
ip6_output+0x323/0x610 net/ipv6/ip6_output.c:243
dst_output include/net/dst.h:451 [inline]
ip6_local_out+0xe9/0x140 net/ipv6/output_core.c:155
ip6_send_skb net/ipv6/ip6_output.c:1952 [inline]
ip6_push_pending_frames+0x1f9/0x560 net/ipv6/ip6_output.c:1972
rawv6_push_pending_frames+0xbe8/0xdf0 net/ipv6/raw.c:582
rawv6_sendmsg+0x2b66/0x2e70 net/ipv6/raw.c:920
inet_sendmsg+0x105/0x190 net/ipv4/af_inet.c:847
sock_sendmsg_nosec net/socket.c:730 [inline]
__sock_sendmsg net/socket.c:745 [inline]
____sys_sendmsg+0x9c2/0xd60 net/socket.c:2584
___sys_sendmsg+0x28d/0x3c0 net/socket.c:2638
__sys_sendmsg net/socket.c:2667 [inline]
__do_sys_sendmsg net/socket.c:2676 [inline]
__se_sys_sendmsg net/socket.c:2674 [inline]
__x64_sys_sendmsg+0x307/0x490 net/socket.c:2674
do_syscall_x64 arch/x86/entry/common.c:52 [inline]
do_syscall_64+0x44/0x110 arch/x86/entry/common.c:83
entry_SYSCALL_64_after_hwframe+0x63/0x6b

Uninit was created at:

slab_post_alloc_hook+0x129/0xa70 mm/slab.h:768
slab_alloc_node mm/slub.c:3478 [inline]
__kmem_cache_alloc_node+0x5c9/0x970 mm/slub.c:3517
__do_kmalloc_node mm/slab_common.c:1006 [inline]
__kmalloc_node_track_caller+0x118/0x3c0 mm/slab_common.c:1027
kmalloc_reserve+0x249/0x4a0 net/core/skbuff.c:582
pskb_expand_head+0x226/0x1a00 net/core/skbuff.c:2098
__pskb_pull_tail+0x13b/0x2310 net/core/skbuff.c:2655
pskb_may_pull_reason include/linux/skbuff.h:2673 [inline]
pskb_may_pull include/linux/skbuff.h:2681 [inline]
ip6_tnl_parse_tlv_enc_lim+0x901/0xbb0 net/ipv6/ip6_tunnel.c:408
ipxip6_tnl_xmit net/ipv6/ip6_tunnel.c:1326 [inline]
ip6_tnl_start_xmit+0xab2/0x1a70 net/ipv6/ip6_tunnel.c:1432
__netdev_start_xmit include/linux/netdevice.h:4940 [inline]
netdev_start_xmit include/linux/netdevice.h:4954 [inline]
xmit_one net/core/dev.c:3548 [inline]
dev_hard_start_xmit+0x247/0xa10 net/core/dev.c:3564
__dev_queue_xmit+0x33b8/0x5130 net/core/dev.c:4349
dev_queue_xmit include/linux/netdevice.h:3134 [inline]
neigh_connected_output+0x569/0x660 net/core/neighbour.c:1592
neigh_output include/net/neighbour.h:542 [inline]
ip6_finish_output2+0x23a9/0x2b30 net/ipv6/ip6_output.c:137
ip6_finish_output+0x855/0x12b0 net/ipv6/ip6_output.c:222
NF_HOOK_COND include/linux/netfilter.h:303 [inline]
ip6_output+0x323/0x610 net/ipv6/ip6_output.c:243
dst_output include/net/dst.h:451 [inline]
ip6_local_out+0xe9/0x140 net/ipv6/output_core.c:155
ip6_send_skb net/ipv6/ip6_output.c:1952 [inline]
ip6_push_pending_frames+0x1f9/0x560 net/ipv6/ip6_output.c:1972
rawv6_push_pending_frames+0xbe8/0xdf0 net/ipv6/raw.c:582
rawv6_sendmsg+0x2b66/0x2e70 net/ipv6/raw.c:920
inet_sendmsg+0x105/0x190 net/ipv4/af_inet.c:847
sock_sendmsg_nosec net/socket.c:730 [inline]
__sock_sendmsg net/socket.c:745 [inline]
____sys_sendmsg+0x9c2/0xd60 net/socket.c:2584
___sys_sendmsg+0x28d/0x3c0 net/socket.c:2638
__sys_sendmsg net/socket.c:2667 [inline]
__do_sys_sendmsg net/socket.c:2676 [inline]
__se_sys_sendmsg net/socket.c:2674 [inline]
__x64_sys_sendmsg+0x307/0x490 net/socket.c:2674
do_syscall_x64 arch/x86/entry/common.c:52 [inline]
do_syscall_64+0x44/0x110 arch/x86/entry/common.c:83
entry_SYSCALL_64_after_hwframe+0x63/0x6b

CPU: 0 PID: 7345 Comm: syz-executor.3 Not tainted 6.7.0-rc8-syzkaller-00024-gac865f00af29 #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 11/17/2023

---
 net/ipv6/ip6_tunnel.c | 26 +++++++++++++-------------
 1 file changed, 13 insertions(+), 13 deletions(-)

index afc922c88d17..9125e92d9917 100644
@@ -399,7 +399,7 @@ __u16 ip6_tnl_parse_tlv_enc_lim(struct sk_buff *skb, __u8 *raw)
 	const struct ipv6hdr *ipv6h = (const struct ipv6hdr *)raw;
 	unsigned int nhoff = raw - skb->data;
 	unsigned int off = nhoff + sizeof(*ipv6h);
-	u8 next, nexthdr = ipv6h->nexthdr;
+	u8 nexthdr = ipv6h->nexthdr;
 
 	while (ipv6_ext_hdr(nexthdr) && nexthdr != NEXTHDR_NONE) {
 		struct ipv6_opt_hdr *hdr;

@@ -410,25 +410,25 @@ __u16 ip6_tnl_parse_tlv_enc_lim(struct sk_buff *skb, __u8 *raw)
 
 		hdr = (struct ipv6_opt_hdr *)(skb->data + off);
 		if (nexthdr == NEXTHDR_FRAGMENT) {
-			struct frag_hdr *frag_hdr = (struct frag_hdr *) hdr;
-			if (frag_hdr->frag_off)
-				break;
 			optlen = 8;
 		} else if (nexthdr == NEXTHDR_AUTH) {
 			optlen = ipv6_authlen(hdr);
 		} else {
 			optlen = ipv6_optlen(hdr);
 		}
-		/* cache hdr->nexthdr, since pskb_may_pull() might
-		 * invalidate hdr
-		 */
-		next = hdr->nexthdr;
-		if (nexthdr == NEXTHDR_DEST) {
-			u16 i = 2;
 
-			/* Remember : hdr is no longer valid at this point. */
-			if (!pskb_may_pull(skb, off + optlen))
+		if (!pskb_may_pull(skb, off + optlen))
+			break;
+
+		hdr = (struct ipv6_opt_hdr *)(skb->data + off);
+		if (nexthdr == NEXTHDR_FRAGMENT) {
+			struct frag_hdr *frag_hdr = (struct frag_hdr *)hdr;
+
+			if (frag_hdr->frag_off)
 				break;
+		}
+		if (nexthdr == NEXTHDR_DEST) {
+			u16 i = 2;
 
 			while (1) {
 				struct ipv6_tlv_tnl_enc_lim *tel;

@@ -449,7 +449,7 @@ __u16 ip6_tnl_parse_tlv_enc_lim(struct sk_buff *skb, __u8 *raw)
 					i++;
 			}
 		}
-		nexthdr = next;
+		nexthdr = hdr->nexthdr;
 		off += optlen;
 	}
 	return 0;

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Arnd Bergmann <[email protected]>

[ Upstream commit 40974ee421b4d1fc74ac733d86899ce1b83d8f65 ]

The select was lost by accident during the multiplatform conversion. Any davinci-only

arm-linux-gnueabi-ld: arch/arm/mach-davinci/sleep.o: in function `CACHE_FLUSH':
(.text+0x168): undefined reference to `arm926_flush_kern_cache_all'

---
 arch/arm/mach-davinci/Kconfig | 1 +
 1 file changed, 1 insertion(+)

index c8cbd9a30791..0b54ca56555b 100644
@@ -4,6 +4,7 @@ menuconfig ARCH_DAVINCI
 	bool "TI DaVinci"
 	depends on ARCH_MULTI_V5
 	depends on CPU_LITTLE_ENDIAN
+	select CPU_ARM926T
 	select DAVINCI_TIMER
 	select ZONE_DMA
 	select PM_GENERIC_DOMAINS if PM

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Tomi Valkeinen <[email protected]>

[ Upstream commit ca34d816558c3e4c3f8fe037b5a6b16c944693de ]

This reverts commit 4d56a4f08391857ba93465de489707b66adad114.

The DMA-fence annotations cause a lockdep warning (see below). As per
https://patchwork.freedesktop.org/patch/462170/ it sounds like the annotations don't work correctly.

======================================================
WARNING: possible circular locking dependency detected
6.6.0-rc2+ #1 Not tainted
------------------------------------------------------
kmstest/733 is trying to acquire lock:
ffff8000819377f0 (fs_reclaim){+.+.}-{0:0}, at: __kmem_cache_alloc_node+0x58/0x2d4

but task is already holding lock:
ffff800081a06aa0 (dma_fence_map){++++}-{0:0}, at: tidss_atomic_commit_tail+0x20/0xc0 [tidss]

which lock already depends on the new lock.

the existing dependency chain (in reverse order) is:

-> #2 (dma_fence_map){++++}-{0:0}:

       __dma_fence_might_wait+0x5c/0xd0
       dma_resv_lockdep+0x1a4/0x32c
       do_one_initcall+0x84/0x2fc
       kernel_init_freeable+0x28c/0x4c4
       kernel_init+0x24/0x1dc
       ret_from_fork+0x10/0x20

-> #1 (mmu_notifier_invalidate_range_start){+.+.}-{0:0}:

       fs_reclaim_acquire+0x70/0xe4
       __kmem_cache_alloc_node+0x58/0x2d4
       kmalloc_trace+0x38/0x78
       __kthread_create_worker+0x3c/0x150
       kthread_create_worker+0x64/0x8c
       workqueue_init+0x1e8/0x2f0
       kernel_init_freeable+0x11c/0x4c4
       kernel_init+0x24/0x1dc
       ret_from_fork+0x10/0x20

-> #0 (fs_reclaim){+.+.}-{0:0}:

       __lock_acquire+0x1370/0x20d8
       lock_acquire+0x1e8/0x308
       fs_reclaim_acquire+0xd0/0xe4
       __kmem_cache_alloc_node+0x58/0x2d4
       __kmalloc_node_track_caller+0x58/0xf0
       kmemdup+0x34/0x60
       regmap_bulk_write+0x64/0x2c0
       tc358768_bridge_pre_enable+0x8c/0x12d0 [tc358768]
       drm_atomic_bridge_call_pre_enable+0x68/0x80 [drm]
       drm_atomic_bridge_chain_pre_enable+0x50/0x158 [drm]
       drm_atomic_helper_commit_modeset_enables+0x164/0x264 [drm_kms_helper]
       tidss_atomic_commit_tail+0x58/0xc0 [tidss]
       commit_tail+0xa0/0x188 [drm_kms_helper]
       drm_atomic_helper_commit+0x1a8/0x1c0 [drm_kms_helper]
       drm_atomic_commit+0xa8/0xe0 [drm]
       drm_mode_atomic_ioctl+0x9ec/0xc80 [drm]
       drm_ioctl_kernel+0xc4/0x170 [drm]
       drm_ioctl+0x234/0x4b0 [drm]
       drm_compat_ioctl+0x110/0x12c [drm]
       __arm64_compat_sys_ioctl+0x128/0x150
       invoke_syscall+0x48/0x110
       el0_svc_common.constprop.0+0x40/0xe0
       do_el0_svc_compat+0x1c/0x38
       el0_svc_compat+0x48/0xb4
       el0t_32_sync_handler+0xb0/0x138
       el0t_32_sync+0x194/0x198

other info that might help us debug this:

Chain exists of:
  fs_reclaim --> mmu_notifier_invalidate_range_start --> dma_fence_map

 Possible unsafe locking scenario:

       CPU0                    CPU1
       ----                    ----
  rlock(dma_fence_map);
                               lock(mmu_notifier_invalidate_range_start);
                               lock(dma_fence_map);

lock(fs_reclaim);

*** DEADLOCK ***

3 locks held by kmstest/733:
#0: ffff800082e5bba0 (crtc_ww_class_acquire){+.+.}-{0:0}, at: drm_mode_atomic_ioctl+0x118/0xc80 [drm] #1: ffff000004224c88 (crtc_ww_class_mutex){+.+.}-{3:3}, at: modeset_lock+0xdc/0x1a0 [drm] #2: ffff800081a06aa0 (dma_fence_map){++++}-{0:0}, at: tidss_atomic_commit_tail+0x20/0xc0 [tidss]

stack backtrace:
CPU: 0 PID: 733 Comm: kmstest Not tainted 6.6.0-rc2+ #1
Hardware name: Toradex Verdin AM62 on Verdin Development Board (DT)
Call trace:
 dump_backtrace+0x98/0x118
 show_stack+0x18/0x24
 dump_stack_lvl+0x60/0xac
 dump_stack+0x18/0x24
 print_circular_bug+0x288/0x368
 check_noncircular+0x168/0x17c
 __lock_acquire+0x1370/0x20d8
 lock_acquire+0x1e8/0x308
 fs_reclaim_acquire+0xd0/0xe4
 __kmem_cache_alloc_node+0x58/0x2d4
 __kmalloc_node_track_caller+0x58/0xf0
 kmemdup+0x34/0x60
 regmap_bulk_write+0x64/0x2c0
 tc358768_bridge_pre_enable+0x8c/0x12d0 [tc358768]
 drm_atomic_bridge_call_pre_enable+0x68/0x80 [drm]
 drm_atomic_bridge_chain_pre_enable+0x50/0x158 [drm]
 drm_atomic_helper_commit_modeset_enables+0x164/0x264 [drm_kms_helper]
 tidss_atomic_commit_tail+0x58/0xc0 [tidss]
 commit_tail+0xa0/0x188 [drm_kms_helper]
 drm_atomic_helper_commit+0x1a8/0x1c0 [drm_kms_helper]
 drm_atomic_commit+0xa8/0xe0 [drm]
 drm_mode_atomic_ioctl+0x9ec/0xc80 [drm]
 drm_ioctl_kernel+0xc4/0x170 [drm]
 drm_ioctl+0x234/0x4b0 [drm]
 drm_compat_ioctl+0x110/0x12c [drm]
 __arm64_compat_sys_ioctl+0x128/0x150
 invoke_syscall+0x48/0x110
 el0_svc_common.constprop.0+0x40/0xe0
 do_el0_svc_compat+0x1c/0x38
 el0_svc_compat+0x48/0xb4
 el0t_32_sync_handler+0xb0/0x138
 el0t_32_sync+0x194/0x198

---
 drivers/gpu/drm/tidss/tidss_kms.c | 4 ----
 1 file changed, 4 deletions(-)

index afb2879980c6..995bac488392 100644
@@ -4,8 +4,6 @@
  * Author: Tomi Valkeinen <[email protected]>
  */
 
-#include <linux/dma-fence.h>
-
 #include <drm/drm_atomic.h>
 #include <drm/drm_atomic_helper.h>
 #include <drm/drm_bridge.h>

@@ -27,7 +25,6 @@ static void tidss_atomic_commit_tail(struct drm_atomic_state *old_state)
 {
 	struct drm_device *ddev = old_state->dev;
 	struct tidss_device *tidss = to_tidss(ddev);
-	bool fence_cookie = dma_fence_begin_signalling();
 
 	dev_dbg(ddev->dev, "%s\n", __func__);
 

@@ -38,7 +35,6 @@ static void tidss_atomic_commit_tail(struct drm_atomic_state *old_state)
 	drm_atomic_helper_commit_modeset_enables(ddev, old_state);
 
 	drm_atomic_helper_commit_hw_done(old_state);
-	dma_fence_end_signalling(fence_cookie);
 	drm_atomic_helper_wait_for_flip_done(ddev, old_state);
 
 	drm_atomic_helper_cleanup_planes(ddev, old_state);

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Tomi Valkeinen <[email protected]>

[ Upstream commit 9d7c8c066916f231ca0ed4e4fce6c4b58ca3e451 ]

This reverts commit 250aa22920cd5d956a5d3e9c6a43d671c2bae217.

The DMA-fence annotations cause a lockdep warning (see below). As per
https://patchwork.freedesktop.org/patch/462170/ it sounds like the annotations don't work correctly.

======================================================
WARNING: possible circular locking dependency detected 6.5.0-rc2+ #2 Not tainted
------------------------------------------------------
kmstest/219 is trying to acquire lock:
c4705838 (&hdmi->lock){+.+.}-{3:3}, at: hdmi5_bridge_mode_set+0x1c/0x50

but task is already holding lock:
c11e1128 (dma_fence_map){++++}-{0:0}, at: omap_atomic_commit_tail+0x14/0xbc

which lock already depends on the new lock.

the existing dependency chain (in reverse order) is:

-> #2 (dma_fence_map){++++}-{0:0}:

       __dma_fence_might_wait+0x48/0xb4
       dma_resv_lockdep+0x1b8/0x2bc
       do_one_initcall+0x68/0x3b0
       kernel_init_freeable+0x260/0x34c
       kernel_init+0x14/0x140
       ret_from_fork+0x14/0x28

-> #1 (fs_reclaim){+.+.}-{0:0}:

       fs_reclaim_acquire+0x70/0xa8
       __kmem_cache_alloc_node+0x3c/0x368
       kmalloc_trace+0x28/0x58
       _drm_do_get_edid+0x7c/0x35c
       hdmi5_bridge_get_edid+0xc8/0x1ac
       drm_bridge_connector_get_modes+0x64/0xc0
       drm_helper_probe_single_connector_modes+0x170/0x528
       drm_client_modeset_probe+0x208/0x1334
       __drm_fb_helper_initial_config_and_unlock+0x30/0x548
       omap_fbdev_client_hotplug+0x3c/0x6c
       drm_client_register+0x58/0x94
       pdev_probe+0x544/0x6b0
       platform_probe+0x58/0xbc
       really_probe+0xd8/0x3fc
       __driver_probe_device+0x94/0x1f4
       driver_probe_device+0x2c/0xc4
       __device_attach_driver+0xa4/0x11c
       bus_for_each_drv+0x84/0xdc
       __device_attach+0xac/0x20c
       bus_probe_device+0x8c/0x90
       device_add+0x588/0x7e0
       platform_device_add+0x110/0x24c
       platform_device_register_full+0x108/0x15c
       dss_bind+0x90/0xc0
       try_to_bring_up_aggregate_device+0x1e0/0x2c8
       __component_add+0xa4/0x174
       hdmi5_probe+0x1c8/0x270
       platform_probe+0x58/0xbc
       really_probe+0xd8/0x3fc
       __driver_probe_device+0x94/0x1f4
       driver_probe_device+0x2c/0xc4
       __device_attach_driver+0xa4/0x11c
       bus_for_each_drv+0x84/0xdc
       __device_attach+0xac/0x20c
       bus_probe_device+0x8c/0x90
       deferred_probe_work_func+0x8c/0xd8
       process_one_work+0x2ac/0x6e4
       worker_thread+0x30/0x4ec
       kthread+0x100/0x124
       ret_from_fork+0x14/0x28

-> #0 (&hdmi->lock){+.+.}-{3:3}:

       __lock_acquire+0x145c/0x29cc
       lock_acquire.part.0+0xb4/0x258
       __mutex_lock+0x90/0x950
       mutex_lock_nested+0x1c/0x24
       hdmi5_bridge_mode_set+0x1c/0x50
       drm_bridge_chain_mode_set+0x48/0x5c
       crtc_set_mode+0x188/0x1d0
       omap_atomic_commit_tail+0x2c/0xbc
       commit_tail+0x9c/0x188
       drm_atomic_helper_commit+0x158/0x18c
       drm_atomic_commit+0xa4/0xe8
       drm_mode_atomic_ioctl+0x9a4/0xc38
       drm_ioctl+0x210/0x4a8
       sys_ioctl+0x138/0xf00
       ret_fast_syscall+0x0/0x1c

other info that might help us debug this:

Chain exists of:
  &hdmi->lock --> fs_reclaim --> dma_fence_map

 Possible unsafe locking scenario:

       CPU0                    CPU1
       ----                    ----
  rlock(dma_fence_map);
                               lock(fs_reclaim);
                               lock(dma_fence_map);

lock(&hdmi->lock);

*** DEADLOCK ***

3 locks held by kmstest/219:
#0: f1011de4 (crtc_ww_class_acquire){+.+.}-{0:0}, at: drm_mode_atomic_ioctl+0xf0/0xc38 #1: c47059c8 (crtc_ww_class_mutex){+.+.}-{3:3}, at: modeset_lock+0xf8/0x230 #2: c11e1128 (dma_fence_map){++++}-{0:0}, at: omap_atomic_commit_tail+0x14/0xbc

stack backtrace:
CPU: 1 PID: 219 Comm: kmstest Not tainted 6.5.0-rc2+ #2
Hardware name: Generic DRA74X (Flattened Device Tree)
 unwind_backtrace from show_stack+0x10/0x14
 show_stack from dump_stack_lvl+0x58/0x70
 dump_stack_lvl from check_noncircular+0x164/0x198
 check_noncircular from __lock_acquire+0x145c/0x29cc
 __lock_acquire from lock_acquire.part.0+0xb4/0x258
 lock_acquire.part.0 from __mutex_lock+0x90/0x950
 __mutex_lock from mutex_lock_nested+0x1c/0x24
 mutex_lock_nested from hdmi5_bridge_mode_set+0x1c/0x50
 hdmi5_bridge_mode_set from drm_bridge_chain_mode_set+0x48/0x5c
 drm_bridge_chain_mode_set from crtc_set_mode+0x188/0x1d0
 crtc_set_mode from omap_atomic_commit_tail+0x2c/0xbc
 omap_atomic_commit_tail from commit_tail+0x9c/0x188
 commit_tail from drm_atomic_helper_commit+0x158/0x18c
 drm_atomic_helper_commit from drm_atomic_commit+0xa4/0xe8
 drm_atomic_commit from drm_mode_atomic_ioctl+0x9a4/0xc38
 drm_mode_atomic_ioctl from drm_ioctl+0x210/0x4a8
 drm_ioctl from sys_ioctl+0x138/0xf00
 sys_ioctl from ret_fast_syscall+0x0/0x1c
Exception stack(0xf1011fa8 to 0xf1011ff0)

1fa0: 00466d58 be9ab510 00000003 c03864bc be9ab510 be9ab4e0
1fc0: 00466d58 be9ab510 c03864bc 00000036 00466ef0 00466fc0 00467020 00466f20 1fe0: b6bc7ef4 be9ab4d0 b6bbbb00 b6cb2cc0

---
 drivers/gpu/drm/omapdrm/omap_drv.c | 9 ++++-----
 1 file changed, 4 insertions(+), 5 deletions(-)

index eaf67b9e5f12..5b6d1668f405 100644
@@ -68,7 +68,6 @@ static void omap_atomic_commit_tail(struct drm_atomic_state *old_state)
 {
 	struct drm_device *dev = old_state->dev;
 	struct omap_drm_private *priv = dev->dev_private;
-	bool fence_cookie = dma_fence_begin_signalling();
 
 	dispc_runtime_get(priv->dispc);
 

@@ -91,6 +90,8 @@ static void omap_atomic_commit_tail(struct drm_atomic_state *old_state)
 		omap_atomic_wait_for_completion(dev, old_state);
 
 		drm_atomic_helper_commit_planes(dev, old_state, 0);
+
+		drm_atomic_helper_commit_hw_done(old_state);
 	} else {
 		/*
 		 * OMAP3 DSS seems to have issues with the work-around above,

@@ -100,11 +101,9 @@ static void omap_atomic_commit_tail(struct drm_atomic_state *old_state)
 		drm_atomic_helper_commit_planes(dev, old_state, 0);
 
 		drm_atomic_helper_commit_modeset_enables(dev, old_state);
-	}
 
-	drm_atomic_helper_commit_hw_done(old_state);
-
-	dma_fence_end_signalling(fence_cookie);
+		drm_atomic_helper_commit_hw_done(old_state);
+	}
 
 	/*
 	 * Wait for completion of the page flips to ensure that old buffers

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: AngeloGioacchino Del Regno <[email protected]>

[ Upstream commit 22aa1a209018dc2eca78745f7666db63637cd5dc ]

The layout of the registers {TILER,SHADER,L2}_PWROFF_LO, used to request powering off cores, is the same as the {TILER,SHADER,L2}_PWRON_LO ones:
this means that in order to request poweroff of cores, we are supposed to write a bitmask of cores that should be powered off!
This means that the panfrost_gpu_power_off() function has always been doing nothing.

Fix powering off the GPU by writing a bitmask of the cores to poweroff to the relevant PWROFF_LO registers and then check that the transition (from ON to OFF) has finished by polling the relevant PWRTRANS_LO registers.

While at it, in order to avoid code duplication, move the core mask logic from panfrost_gpu_power_on() to a new panfrost_get_core_mask() function, used in both poweron and poweroff.

---
 drivers/gpu/drm/panfrost/panfrost_gpu.c | 64 ++++++++++++++++++-------
 1 file changed, 46 insertions(+), 18 deletions(-)

index 6452e4e900dd..c08715f033c5 100644
@@ -313,28 +313,38 @@ static void panfrost_gpu_init_features(struct panfrost_device *pfdev)
 		 pfdev->features.shader_present, pfdev->features.l2_present);
 }
 
+static u64 panfrost_get_core_mask(struct panfrost_device *pfdev)
+{
+	u64 core_mask;
+
+	if (pfdev->features.l2_present == 1)
+		return U64_MAX;
+
+	/*
+	 * Only support one core group now.
+	 * ~(l2_present - 1) unsets all bits in l2_present except
+	 * the bottom bit. (l2_present - 2) has all the bits in
+	 * the first core group set. AND them together to generate
+	 * a mask of cores in the first core group.
+	 */
+	core_mask = ~(pfdev->features.l2_present - 1) &
+		     (pfdev->features.l2_present - 2);
+	dev_info_once(pfdev->dev, "using only 1st core group (%lu cores from %lu)\n",
+		      hweight64(core_mask),
+		      hweight64(pfdev->features.shader_present));
+
+	return core_mask;
+}
+
 void panfrost_gpu_power_on(struct panfrost_device *pfdev)
 {
 	int ret;
 	u32 val;
-	u64 core_mask = U64_MAX;
+	u64 core_mask;
 
 	panfrost_gpu_init_quirks(pfdev);
+	core_mask = panfrost_get_core_mask(pfdev);
 
-	if (pfdev->features.l2_present != 1) {
-		/*
-		 * Only support one core group now.
-		 * ~(l2_present - 1) unsets all bits in l2_present except
-		 * the bottom bit. (l2_present - 2) has all the bits in
-		 * the first core group set. AND them together to generate
-		 * a mask of cores in the first core group.
-		 */
-		core_mask = ~(pfdev->features.l2_present - 1) &
-			     (pfdev->features.l2_present - 2);
-		dev_info_once(pfdev->dev, "using only 1st core group (%lu cores from %lu)\n",
-			      hweight64(core_mask),
-			      hweight64(pfdev->features.shader_present));
-	}
 	gpu_write(pfdev, L2_PWRON_LO, pfdev->features.l2_present & core_mask);
 	ret = readl_relaxed_poll_timeout(pfdev->iomem + L2_READY_LO,
 		val, val == (pfdev->features.l2_present & core_mask),

@@ -359,9 +369,27 @@ void panfrost_gpu_power_on(struct panfrost_device *pfdev)
 
 void panfrost_gpu_power_off(struct panfrost_device *pfdev)
 {
-	gpu_write(pfdev, TILER_PWROFF_LO, 0);
-	gpu_write(pfdev, SHADER_PWROFF_LO, 0);
-	gpu_write(pfdev, L2_PWROFF_LO, 0);
+	u64 core_mask = panfrost_get_core_mask(pfdev);
+	int ret;
+	u32 val;
+
+	gpu_write(pfdev, SHADER_PWROFF_LO, pfdev->features.shader_present & core_mask);
+	ret = readl_relaxed_poll_timeout(pfdev->iomem + SHADER_PWRTRANS_LO,
+					 val, !val, 1, 1000);
+	if (ret)
+		dev_err(pfdev->dev, "shader power transition timeout");
+
+	gpu_write(pfdev, TILER_PWROFF_LO, pfdev->features.tiler_present);
+	ret = readl_relaxed_poll_timeout(pfdev->iomem + TILER_PWRTRANS_LO,
+					 val, !val, 1, 1000);
+	if (ret)
+		dev_err(pfdev->dev, "tiler power transition timeout");
+
+	gpu_write(pfdev, L2_PWROFF_LO, pfdev->features.l2_present & core_mask);
+	ret = readl_poll_timeout(pfdev->iomem + L2_PWRTRANS_LO,
+				 val, !val, 0, 1000);
+	if (ret)
+		dev_err(pfdev->dev, "l2 power transition timeout");
 }
 
 int panfrost_gpu_init(struct panfrost_device *pfdev)

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Leon Romanovsky <[email protected]>

[ Upstream commit b9a85e5eec126d6ae6c362f94b447c223e8fe6e4 ]

The patch 1da177e4c3f4: "Linux-2.6.12-rc2" from Apr 16, 2005 (linux-next), leads to the following Smatch static checker warning:

drivers/infiniband/hw/mthca/mthca_cmd.c:644 mthca_SYS_EN()
error: uninitialized symbol 'out'.

drivers/infiniband/hw/mthca/mthca_cmd.c

    636 int mthca_SYS_EN(struct mthca_dev *dev)
    637 {
    638         u64 out;
    639         int ret;
    640
    641         ret = mthca_cmd_imm(dev, 0, &out, 0, 0, CMD_SYS_EN, CMD_TIME_CLASS_D);

We pass out here and it gets used without being initialized.

        err = mthca_cmd_post(dev, in_param,
                             out_param ? *out_param : 0,
                                         ^^^^^^^^^^
                             in_modifier, op_modifier,
                             op, context->token, 1);

It's the same in mthca_cmd_wait() and mthca_cmd_poll().

---
 drivers/infiniband/hw/mthca/mthca_cmd.c  | 4 ++--
 drivers/infiniband/hw/mthca/mthca_main.c | 2 +-
 2 files changed, 3 insertions(+), 3 deletions(-)

index f330ce895d88..8fe0cef7e2be 100644
@@ -635,7 +635,7 @@ void mthca_free_mailbox(struct mthca_dev *dev, struct mthca_mailbox *mailbox)
 
 int mthca_SYS_EN(struct mthca_dev *dev)
 {
-	u64 out;
+	u64 out = 0;
 	int ret;
 
 	ret = mthca_cmd_imm(dev, 0, &out, 0, 0, CMD_SYS_EN, CMD_TIME_CLASS_D);

@@ -1955,7 +1955,7 @@ int mthca_WRITE_MGM(struct mthca_dev *dev, int index,
 int mthca_MGID_HASH(struct mthca_dev *dev, struct mthca_mailbox *mailbox,
 		    u16 *hash)
 {
-	u64 imm;
+	u64 imm = 0;
 	int err;
 
 	err = mthca_cmd_imm(dev, mailbox->dma, &imm, 0, 0, CMD_MGID_HASH,

index b54bc8865dae..1ab268b77096 100644
@@ -382,7 +382,7 @@ static int mthca_init_icm(struct mthca_dev *mdev,
 			  struct mthca_init_hca_param *init_hca,
 			  u64 icm_size)
 {
-	u64 aux_pages;
+	u64 aux_pages = 0;
 	int err;
 
 	err = mthca_SET_ICM_SIZE(mdev, icm_size, &aux_pages);

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Junxian Huang <[email protected]>

[ Upstream commit f45b83ad39f8033e717b1eee57e81811113d5a84 ]

EOPNOTSUPP is more situable than EINVAL for allocating XRCD while XRC is not supported and unsupported resizing SRQ.

---
 drivers/infiniband/hw/hns/hns_roce_hw_v2.c | 2 +-
 drivers/infiniband/hw/hns/hns_roce_pd.c    | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

index c2ee80546d12..280a3458bb53 100644
@@ -5791,7 +5791,7 @@ static int hns_roce_v2_modify_srq(struct ib_srq *ibsrq,
 
 	/* Resizing SRQs is not supported yet */
 	if (srq_attr_mask & IB_SRQ_MAX_WR)
-		return -EINVAL;
+		return -EOPNOTSUPP;
 
 	if (srq_attr_mask & IB_SRQ_LIMIT) {
 		if (srq_attr->srq_limit > srq->wqe_cnt)

index 783e71852c50..bd1fe89ca205 100644
@@ -150,7 +150,7 @@ int hns_roce_alloc_xrcd(struct ib_xrcd *ib_xrcd, struct ib_udata *udata)
 	int ret;
 
 	if (!(hr_dev->caps.flags & HNS_ROCE_CAP_FLAG_XRC))
-		return -EINVAL;
+		return -EOPNOTSUPP;
 
 	ret = hns_roce_xrcd_alloc(hr_dev, &xrcd->xrcdn);
 	if (ret)

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Chris Morgan <[email protected]>

[ Upstream commit 03c5b2a5f6c39fe4e090346536cf1c14ee18b61e ]

For devices like the Anbernic RG351M and RG351P the panel is wired to an always on regulator. When the device suspends and wakes up, there are some slight artifacts on the screen that go away over time. If instead we hold the panel in reset status after it is unprepared, this does not happen.

---
 drivers/gpu/drm/panel/panel-elida-kd35t133.c | 2 ++
 1 file changed, 2 insertions(+)

index eee714cf3f49..3a7fc3ca6a6f 100644
@@ -112,6 +112,8 @@ static int kd35t133_unprepare(struct drm_panel *panel)
 		return ret;
 	}
 
+	gpiod_set_value_cansleep(ctx->reset_gpio, 1);
+
 	regulator_disable(ctx->iovcc);
 	regulator_disable(ctx->vdd);
 

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Abhinav Singh <[email protected]>

[ Upstream commit 5f35a624c1e30b5bae5023b3c256e94e0ad4f806 ]

Fix a sparse warning with this message
"warning:dereference of noderef expression". In this context it means we are dereferencing a __rcu tagged pointer directly.

We should not be directly dereferencing a rcu pointer. To get a normal (non __rcu tagged pointer) from a __rcu tagged pointer we are using the function unrcu_pointer(...). The non __rcu tagged pointer then can be dereferenced just like a normal pointer.

I tested with qemu with this command

qemu-system-x86_64 \
	-m 2G \
	-smp 2 \
	-kernel bzImage \
	-append "console=ttyS0 root=/dev/sda earlyprintk=serial net.ifnames=0" \
	-drive file=bullseye.img,format=raw \
	-net user,host=10.0.2.10,hostfwd=tcp:127.0.0.1:10021-:22 \
	-net nic,model=e1000 \
	-enable-kvm \
	-nographic \
	-pidfile vm.pid \
	2>&1 | tee vm.log

with lockdep enabled.

---
 drivers/gpu/drm/nouveau/nv04_fence.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

index 5b71a5a5cd85..cdbc75e3d1f6 100644
@@ -39,7 +39,7 @@ struct nv04_fence_priv {
 static int
 nv04_fence_emit(struct nouveau_fence *fence)
 {
-	struct nvif_push *push = fence->channel->chan.push;
+	struct nvif_push *push = unrcu_pointer(fence->channel)->chan.push;
 	int ret = PUSH_WAIT(push, 2);
 	if (ret == 0) {
 		PUSH_NVSQ(push, NV_SW, 0x0150, fence->base.seqno);

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Uwe Kleine-König <[email protected]>

[ Upstream commit ce3e112e7ae854249d8755906acc5f27e1542114 ]

With tpd12s015_remove() marked with __exit this function is discarded when the driver is compiled as a built-in. The result is that when the driver unbinds there is no cleanup done which results in resource leakage or worse.

---
 drivers/gpu/drm/bridge/ti-tpd12s015.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

index e0e015243a60..b588fea12502 100644
@@ -179,7 +179,7 @@ static int tpd12s015_probe(struct platform_device *pdev)
 	return 0;
 }
 
-static int __exit tpd12s015_remove(struct platform_device *pdev)
+static int tpd12s015_remove(struct platform_device *pdev)
 {
 	struct tpd12s015_device *tpd = platform_get_drvdata(pdev);
 

@@ -197,7 +197,7 @@ MODULE_DEVICE_TABLE(of, tpd12s015_of_match);
 
 static struct platform_driver tpd12s015_driver = {
 	.probe	= tpd12s015_probe,
-	.remove	= __exit_p(tpd12s015_remove),
+	.remove = tpd12s015_remove,
 	.driver	= {
 		.name	= "tpd12s015",
 		.of_match_table = tpd12s015_of_match,

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Tomi Valkeinen <[email protected]>

[ Upstream commit 38360bf96d816e175bc602c4ee76953cd303b71d ]

The driver only frees the reserved irq if priv->irq_enabled is set to true. However, the driver mistakenly sets priv->irq_enabled to false, instead of true, in tilcdc_irq_install(), and thus the driver never frees the irq, causing issues on loading the driver a second time.

---
 drivers/gpu/drm/tilcdc/tilcdc_drv.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

index f72755b8ea14..86d34b77b37d 100644
@@ -138,7 +138,7 @@ static int tilcdc_irq_install(struct drm_device *dev, unsigned int irq)
 	if (ret)
 		return ret;
 
-	priv->irq_enabled = false;
+	priv->irq_enabled = true;
 
 	return 0;
 }

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Ricardo B. Marliere <[email protected]>

[ Upstream commit ded85b0c0edd8f45fec88783d7555a5b982449c1 ]

Upon module load, a kthread is created targeting the pvr2_context_thread_func function, which may call pvr2_context_destroy and thus call kfree() on the context object. However, that might happen before the usb hub_event handler is able to notify the driver. This patch adds a sanity check before the invalid read reported by syzbot, within the context disconnection call stack.

---
 drivers/media/usb/pvrusb2/pvrusb2-context.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

index 14170a5d72b3..1764674de98b 100644
@@ -268,7 +268,8 @@ void pvr2_context_disconnect(struct pvr2_context *mp)
 {
 	pvr2_hdw_disconnect(mp->hdw);
 	mp->disconnect_flag = !0;
-	pvr2_context_notify(mp);
+	if (!pvr2_context_shutok())
+		pvr2_context_notify(mp);
 }
 
 

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: kyrie wu <[email protected]>

[ Upstream commit 08d530a8da706f157e9dcb4d9b7b4f0eff908ab9 ]

mtk jpeg decoder is built as a module, export some functions to make them visible by other modules.

Stable-dep-of: d8212c5c87c1 ("media: mtk-jpeg: Remove cancel worker in mtk_jpeg_remove to avoid the crash of multi-core JPEG devices")

---
 drivers/media/platform/mediatek/jpeg/mtk_jpeg_dec_hw.c | 6 ++++++
 1 file changed, 6 insertions(+)

index afbbfd5d02bc..6d200e23754e 100644
@@ -188,6 +188,7 @@ int mtk_jpeg_dec_fill_param(struct mtk_jpeg_dec_param *param)
 
 	return 0;
 }
+EXPORT_SYMBOL_GPL(mtk_jpeg_dec_fill_param);
 
 u32 mtk_jpeg_dec_get_int_status(void __iomem *base)
 {

@@ -199,6 +200,7 @@ u32 mtk_jpeg_dec_get_int_status(void __iomem *base)
 
 	return ret;
 }
+EXPORT_SYMBOL_GPL(mtk_jpeg_dec_get_int_status);
 
 u32 mtk_jpeg_dec_enum_result(u32 irq_result)
 {

@@ -215,11 +217,13 @@ u32 mtk_jpeg_dec_enum_result(u32 irq_result)
 
 	return MTK_JPEG_DEC_RESULT_ERROR_UNKNOWN;
 }
+EXPORT_SYMBOL_GPL(mtk_jpeg_dec_enum_result);
 
 void mtk_jpeg_dec_start(void __iomem *base)
 {
 	writel(0, base + JPGDEC_REG_TRIG);
 }
+EXPORT_SYMBOL_GPL(mtk_jpeg_dec_start);
 
 static void mtk_jpeg_dec_soft_reset(void __iomem *base)
 {

@@ -239,6 +243,7 @@ void mtk_jpeg_dec_reset(void __iomem *base)
 	mtk_jpeg_dec_soft_reset(base);
 	mtk_jpeg_dec_hard_reset(base);
 }
+EXPORT_SYMBOL_GPL(mtk_jpeg_dec_reset);
 
 static void mtk_jpeg_dec_set_brz_factor(void __iomem *base, u8 yscale_w,
 					u8 yscale_h, u8 uvscale_w, u8 uvscale_h)

@@ -407,3 +412,4 @@ void mtk_jpeg_dec_set_config(void __iomem *base,
 				   config->dma_last_mcu);
 	mtk_jpeg_dec_set_pause_mcu_idx(base, config->total_mcu);
 }
+EXPORT_SYMBOL_GPL(mtk_jpeg_dec_set_config);

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Zheng Wang <[email protected]>

[ Upstream commit d8212c5c87c143ca01b78f6bf61244af07e0058e ]

This patch reverts commit c677d7ae8314
("media: mtk-jpeg: Fix use after free bug due to uncanceled work"). The job_timeout_work is initialized only for
the single-core JPEG device so it will cause the crash for multi-core JPEG devices.

Fix it by removing the cancel_delayed_work_sync function.

---
 drivers/media/platform/mediatek/jpeg/mtk_jpeg_core.c | 1 -
 1 file changed, 1 deletion(-)

index e9a4f8abd21c..3071b61946c3 100644
@@ -1412,7 +1412,6 @@ static int mtk_jpeg_remove(struct platform_device *pdev)
 {
 	struct mtk_jpeg_dev *jpeg = platform_get_drvdata(pdev);
 
-	cancel_delayed_work_sync(&jpeg->job_timeout_work);
 	pm_runtime_disable(&pdev->dev);
 	video_unregister_device(jpeg->vdev);
 	v4l2_m2m_release(jpeg->m2m_dev);

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Paul Kocialkowski <[email protected]>

[ Upstream commit 6c0d9e12b1d12bbd95484e4b99f63feeb423765f ]

The (TRY_)DECODER_CMD ioctls are used to support flushing when holding capture buffers is supported. This is the case of this driver but the ioctls were never hooked to the ioctl ops.

Add them to correctly support flushing.

---
 drivers/media/platform/verisilicon/hantro_drv.c  | 2 ++
 drivers/media/platform/verisilicon/hantro_v4l2.c | 3 +++
 2 files changed, 5 insertions(+)

index 08840ba313e7..69a2442f3122 100644
@@ -813,6 +813,8 @@ static int hantro_add_func(struct hantro_dev *vpu, unsigned int funcid)
 
 	if (funcid == MEDIA_ENT_F_PROC_VIDEO_ENCODER) {
 		vpu->encoder = func;
+		v4l2_disable_ioctl(vfd, VIDIOC_TRY_DECODER_CMD);
+		v4l2_disable_ioctl(vfd, VIDIOC_DECODER_CMD);
 	} else {
 		vpu->decoder = func;
 		v4l2_disable_ioctl(vfd, VIDIOC_TRY_ENCODER_CMD);

index 30e650edaea8..b2da48936e3f 100644
@@ -759,6 +759,9 @@ const struct v4l2_ioctl_ops hantro_ioctl_ops = {
 	.vidioc_g_selection = vidioc_g_selection,
 	.vidioc_s_selection = vidioc_s_selection,
 
+	.vidioc_decoder_cmd = v4l2_m2m_ioctl_stateless_decoder_cmd,
+	.vidioc_try_decoder_cmd = v4l2_m2m_ioctl_stateless_try_decoder_cmd,
+
 	.vidioc_try_encoder_cmd = v4l2_m2m_ioctl_try_encoder_cmd,
 	.vidioc_encoder_cmd = vidioc_encoder_cmd,
 };

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Paul Kocialkowski <[email protected]>

[ Upstream commit 1fb7b5ab62113b29ce331464048d8c39e58fd08a ]

The (TRY_)DECODER_CMD ioctls are used to support flushing when holding capture buffers is supported. This is the case of this driver but the ioctls were never hooked to the ioctl ops.

Add them to correctly support flushing.

---
 drivers/staging/media/rkvdec/rkvdec.c | 3 +++
 1 file changed, 3 insertions(+)

index a9bd1e71ea48..d16cf4115d03 100644
@@ -461,6 +461,9 @@ static const struct v4l2_ioctl_ops rkvdec_ioctl_ops = {
 
 	.vidioc_streamon = v4l2_m2m_ioctl_streamon,
 	.vidioc_streamoff = v4l2_m2m_ioctl_streamoff,
+
+	.vidioc_decoder_cmd = v4l2_m2m_ioctl_stateless_decoder_cmd,
+	.vidioc_try_decoder_cmd = v4l2_m2m_ioctl_stateless_try_decoder_cmd,
 };
 
 static int rkvdec_queue_setup(struct vb2_queue *vq, unsigned int *num_buffers,

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Dario Binacchi <[email protected]>

[ Upstream commit 288b039db225676e0c520c981a1b5a2562d893a3 ]

s/singals/signals/

---
 include/drm/drm_bridge.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

index 6b656ea23b96..a76f4103d48b 100644
@@ -191,7 +191,7 @@ struct drm_bridge_funcs {
 	 * or &drm_encoder_helper_funcs.dpms hook.
 	 *
 	 * The bridge must assume that the display pipe (i.e. clocks and timing
-	 * singals) feeding it is no longer running when this callback is
+	 * signals) feeding it is no longer running when this callback is
 	 * called.
 	 *
 	 * The @post_disable callback is optional.

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Chao Yu <[email protected]>

[ Upstream commit 53edb549565f55ccd0bdf43be3d66ce4c2d48b28 ]

As Al reported in link[1]:

f2fs_rename()
...

	if (old_dir != new_dir && !whiteout)
		f2fs_set_link(old_inode, old_dir_entry,
					old_dir_page, new_dir);
	else
		f2fs_put_page(old_dir_page, 0);

You want correct inumber in the ".." link. And cross-directory rename does move the source to new parent, even if you'd been asked to leave a whiteout in the old place.

[1] https://lore.kernel.org/all/20231017055040.GN800259@ZenIV/

With below testcase, it may cause dirent corruption, due to it missed to call f2fs_set_link() to update ".." link to new directory.
- mkdir -p dir/foo
- renameat2 -w dir/foo bar

[ASSERT] (__chk_dots_dentries:1421) --> Bad inode number[0x4] for '..', parent parent ino is [0x3]
[FSCK] other corrupted bugs [Fail]

---
 fs/f2fs/namei.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

index bd020a992c2e..328cd20b16a5 100644
@@ -1099,7 +1099,7 @@ static int f2fs_rename(struct user_namespace *mnt_userns, struct inode *old_dir,
 	}
 
 	if (old_dir_entry) {
-		if (old_dir != new_dir && !whiteout)
+		if (old_dir != new_dir)
 			f2fs_set_link(old_inode, old_dir_entry,
 						old_dir_page, new_dir);
 		else

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Nikita Zhandarovich <[email protected]>

[ Upstream commit 39c960bbf9d9ea862398759e75736cfb68c3446f ]

While improbable, there may be a chance of hitting integer overflow when the result of radeon_get_ib_value() gets shifted left.

Avoid it by casting one of the operands to larger data type (u64).

Found by Linux Verification Center (linuxtesting.org) with static analysis tool SVACE.

---
 drivers/gpu/drm/radeon/r600_cs.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

index 638f861af80f..6cf54a747749 100644
@@ -1275,7 +1275,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 			return -EINVAL;
 		}
 		tmp = (reg - CB_COLOR0_BASE) / 4;
-		track->cb_color_bo_offset[tmp] = radeon_get_ib_value(p, idx) << 8;
+		track->cb_color_bo_offset[tmp] = (u64)radeon_get_ib_value(p, idx) << 8;
 		ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff);
 		track->cb_color_base_last[tmp] = ib[idx];
 		track->cb_color_bo[tmp] = reloc->robj;

@@ -1302,7 +1302,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 					"0x%04X\n", reg);
 			return -EINVAL;
 		}
-		track->htile_offset = radeon_get_ib_value(p, idx) << 8;
+		track->htile_offset = (u64)radeon_get_ib_value(p, idx) << 8;
 		ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff);
 		track->htile_bo = reloc->robj;
 		track->db_dirty = true;

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Nikita Zhandarovich <[email protected]>

[ Upstream commit b5c5baa458faa5430c445acd9a17481274d77ccf ]

It may be possible, albeit unlikely, to encounter integer overflow during the multiplication of several unsigned int variables, the result being assigned to a variable 'size' of wider type.

Prevent this potential behaviour by converting one of the multiples to unsigned long.

Found by Linux Verification Center (linuxtesting.org) with static analysis tool SVACE.

---
 drivers/gpu/drm/radeon/r100.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

index d4f09ecc3d22..f336b5b3b11f 100644
@@ -2321,7 +2321,7 @@ int r100_cs_track_check(struct radeon_device *rdev, struct r100_cs_track *track)
 	switch (prim_walk) {
 	case 1:
 		for (i = 0; i < track->num_arrays; i++) {
-			size = track->arrays[i].esize * track->max_indx * 4;
+			size = track->arrays[i].esize * track->max_indx * 4UL;
 			if (track->arrays[i].robj == NULL) {
 				DRM_ERROR("(PW %u) Vertex array %u no buffer "
 					  "bound\n", prim_walk, i);

@@ -2340,7 +2340,7 @@ int r100_cs_track_check(struct radeon_device *rdev, struct r100_cs_track *track)
 		break;
 	case 2:
 		for (i = 0; i < track->num_arrays; i++) {
-			size = track->arrays[i].esize * (nverts - 1) * 4;
+			size = track->arrays[i].esize * (nverts - 1) * 4UL;
 			if (track->arrays[i].robj == NULL) {
 				DRM_ERROR("(PW %u) Vertex array %u no buffer "
 					  "bound\n", prim_walk, i);

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Nikita Zhandarovich <[email protected]>

[ Upstream commit 71225e1c930942cb1e042fc08c5cc0c4ef30e95e ]

In the unlikely event of radeon_ring_lock() failing, its errno return value should be processed. This patch checks said return value and prints a debug message in case of an error.

Found by Linux Verification Center (linuxtesting.org) with static analysis tool SVACE.

---
 drivers/gpu/drm/radeon/si.c | 4 ++++
 1 file changed, 4 insertions(+)

index a91012447b56..85e9cba49cec 100644
@@ -3611,6 +3611,10 @@ static int si_cp_start(struct radeon_device *rdev)
 	for (i = RADEON_RING_TYPE_GFX_INDEX; i <= CAYMAN_RING_TYPE_CP2_INDEX; ++i) {
 		ring = &rdev->ring[i];
 		r = radeon_ring_lock(rdev, ring, 2);
+		if (r) {
+			DRM_ERROR("radeon: cp failed to lock ring (%d).\n", r);
+			return r;
+		}
 
 		/* clear the compute context state */
 		radeon_ring_write(ring, PACKET3_COMPUTE(PACKET3_CLEAR_STATE, 0));

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Tomi Valkeinen <[email protected]>

[ Upstream commit 36d1e0852680aa038e2428d450673390111b165c ]

We do a DSS reset in the middle of the dispc_init(). While that happens to work now, we should really make sure that e..g the fclk, which is acquired only later in the function, is enabled when doing a reset. This will be handled in a later patch, but for now, let's move the dispc_softreset() call to the end of dispc_init(), which is a sensible place for it anyway.

Stable-dep-of: bc288a927815 ("drm/tidss: Fix dss reset")

---
 drivers/gpu/drm/tidss/tidss_dispc.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

index 16301bdfead1..9ce452288c9e 100644
@@ -2726,10 +2726,6 @@ int dispc_init(struct tidss_device *tidss)
 			return r;
 	}
 
-	/* K2G display controller does not support soft reset */
-	if (feat->subrev != DISPC_K2G)
-		dispc_softreset(dispc);
-
 	for (i = 0; i < dispc->feat->num_vps; i++) {
 		u32 gamma_size = dispc->feat->vp_feat.color.gamma_size;
 		u32 *gamma_table;

@@ -2778,6 +2774,10 @@ int dispc_init(struct tidss_device *tidss)
 	of_property_read_u32(dispc->dev->of_node, "max-memory-bandwidth",
 			     &dispc->memory_bandwidth_limit);
 
+	/* K2G display controller does not support soft reset */
+	if (feat->subrev != DISPC_K2G)
+		dispc_softreset(dispc);
+
 	tidss->dispc = dispc;
 
 	return 0;

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Tomi Valkeinen <[email protected]>

[ Upstream commit aceafbb5035c4bfc75a321863ed1e393d644d2d2 ]

Return an error value from dispc_softreset() so that the caller can handle the errors.

Stable-dep-of: bc288a927815 ("drm/tidss: Fix dss reset")

---
 drivers/gpu/drm/tidss/tidss_dispc.c | 17 ++++++++++++-----
 1 file changed, 12 insertions(+), 5 deletions(-)

index 9ce452288c9e..591f0606f7f8 100644
@@ -2653,7 +2653,7 @@ static void dispc_init_errata(struct dispc_device *dispc)
 	}
 }
 
-static void dispc_softreset(struct dispc_device *dispc)
+static int dispc_softreset(struct dispc_device *dispc)
 {
 	u32 val;
 	int ret = 0;

@@ -2663,8 +2663,12 @@ static void dispc_softreset(struct dispc_device *dispc)
 	/* Wait for reset to complete */
 	ret = readl_poll_timeout(dispc->base_common + DSS_SYSSTATUS,
 				 val, val & 1, 100, 5000);
-	if (ret)
-		dev_warn(dispc->dev, "failed to reset dispc\n");
+	if (ret) {
+		dev_err(dispc->dev, "failed to reset dispc\n");
+		return ret;
+	}
+
+	return 0;
 }
 
 int dispc_init(struct tidss_device *tidss)

@@ -2775,8 +2779,11 @@ int dispc_init(struct tidss_device *tidss)
 			     &dispc->memory_bandwidth_limit);
 
 	/* K2G display controller does not support soft reset */
-	if (feat->subrev != DISPC_K2G)
-		dispc_softreset(dispc);
+	if (feat->subrev != DISPC_K2G) {
+		r = dispc_softreset(dispc);
+		if (r)
+			return r;
+	}
 
 	tidss->dispc = dispc;
 

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Tomi Valkeinen <[email protected]>

[ Upstream commit 151825150cf9c2e9fb90763d35b9dff3783628ac ]

K2G doesn't have softreset feature. Instead of having every caller of dispc_softreset() check for K2G, move the check into dispc_softreset(), and make dispc_softreset() return 0 in case of K2G.

Stable-dep-of: bc288a927815 ("drm/tidss: Fix dss reset")

---
 drivers/gpu/drm/tidss/tidss_dispc.c | 13 +++++++------
 1 file changed, 7 insertions(+), 6 deletions(-)

index 591f0606f7f8..4bdd4c7b4991 100644
@@ -2658,6 +2658,10 @@ static int dispc_softreset(struct dispc_device *dispc)
 	u32 val;
 	int ret = 0;
 
+	/* K2G display controller does not support soft reset */
+	if (dispc->feat->subrev == DISPC_K2G)
+		return 0;
+
 	/* Soft reset */
 	REG_FLD_MOD(dispc, DSS_SYSCONFIG, 1, 1, 1);
 	/* Wait for reset to complete */

@@ -2778,12 +2782,9 @@ int dispc_init(struct tidss_device *tidss)
 	of_property_read_u32(dispc->dev->of_node, "max-memory-bandwidth",
 			     &dispc->memory_bandwidth_limit);
 
-	/* K2G display controller does not support soft reset */
-	if (feat->subrev != DISPC_K2G) {
-		r = dispc_softreset(dispc);
-		if (r)
-			return r;
-	}
+	r = dispc_softreset(dispc);
+	if (r)
+		return r;
 
 	tidss->dispc = dispc;
 

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Tomi Valkeinen <[email protected]>

[ Upstream commit bc288a927815efcf9d7f4a54d4d89c5df478c635 ]

The probe function calls dispc_softreset() before runtime PM is enabled and without enabling any of the DSS clocks. This happens to work by luck, and we need to make sure the DSS HW is active and the fclk is enabled.

To fix the above, add a new function, dispc_init_hw(), which does:

- pm_runtime_set_active()
- clk_prepare_enable(fclk)
- dispc_softreset().

This ensures that the reset can be successfully accomplished.

Note that we use pm_runtime_set_active(), not the normal pm_runtime_get(). The reason for this is that at this point we haven't enabled the runtime PM yet and also we don't want the normal resume callback to be called: the dispc resume callback does some initial HW setup, and it expects that the HW was off (no video ports are streaming). If the bootloader has enabled the DSS and has set up a boot time splash-screen, the DSS would be enabled and streaming which might lead to issues with the normal resume callback.

---
 drivers/gpu/drm/tidss/tidss_dispc.c | 45 ++++++++++++++++++++++++++++-
 1 file changed, 44 insertions(+), 1 deletion(-)

index 4bdd4c7b4991..95b75236fe5e 100644
@@ -2675,6 +2675,49 @@ static int dispc_softreset(struct dispc_device *dispc)
 	return 0;
 }
 
+static int dispc_init_hw(struct dispc_device *dispc)
+{
+	struct device *dev = dispc->dev;
+	int ret;
+
+	ret = pm_runtime_set_active(dev);
+	if (ret) {
+		dev_err(dev, "Failed to set DSS PM to active\n");
+		return ret;
+	}
+
+	ret = clk_prepare_enable(dispc->fclk);
+	if (ret) {
+		dev_err(dev, "Failed to enable DSS fclk\n");
+		goto err_runtime_suspend;
+	}
+
+	ret = dispc_softreset(dispc);
+	if (ret)
+		goto err_clk_disable;
+
+	clk_disable_unprepare(dispc->fclk);
+	ret = pm_runtime_set_suspended(dev);
+	if (ret) {
+		dev_err(dev, "Failed to set DSS PM to suspended\n");
+		return ret;
+	}
+
+	return 0;
+
+err_clk_disable:
+	clk_disable_unprepare(dispc->fclk);
+
+err_runtime_suspend:
+	ret = pm_runtime_set_suspended(dev);
+	if (ret) {
+		dev_err(dev, "Failed to set DSS PM to suspended\n");
+		return ret;
+	}
+
+	return ret;
+}
+
 int dispc_init(struct tidss_device *tidss)
 {
 	struct device *dev = tidss->dev;

@@ -2782,7 +2825,7 @@ int dispc_init(struct tidss_device *tidss)
 	of_property_read_u32(dispc->dev->of_node, "max-memory-bandwidth",
 			     &dispc->memory_bandwidth_limit);
 
-	r = dispc_softreset(dispc);
+	r = dispc_init_hw(dispc);
 	if (r)
 		return r;
 

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Linus Walleij <[email protected]>

[ Upstream commit 50678d339d670a92658e5538ebee30447c88ccb3 ]

This driver includes the legacy GPIO APIs <linux/gpio.h> and <linux/of_gpio.h> but does not use any symbols from any of them.

Drop the includes.

Further the driver is requesting "reset-gpios" rather than just "reset" from the GPIO framework. This is wrong because the gpiolib core will add "-gpios" before processing the request from e.g. device tree. Drop the suffix.

The last problem means that the optional RESET GPIO has never been properly retrieved and used even if it existed, but nobody noticed.

---
 sound/soc/codecs/cs35l33.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

index 15e79168d256..c3b7046fd29b 100644
@@ -22,13 +22,11 @@
 #include <sound/soc-dapm.h>
 #include <sound/initval.h>
 #include <sound/tlv.h>
-#include <linux/gpio.h>
 #include <linux/gpio/consumer.h>
 #include <sound/cs35l33.h>
 #include <linux/pm_runtime.h>
 #include <linux/regulator/consumer.h>
 #include <linux/regulator/machine.h>
-#include <linux/of_gpio.h>
 #include <linux/of.h>
 #include <linux/of_device.h>
 #include <linux/of_irq.h>

@@ -1167,7 +1165,7 @@ static int cs35l33_i2c_probe(struct i2c_client *i2c_client)
 
 	/* We could issue !RST or skip it based on AMP topology */
 	cs35l33->reset_gpio = devm_gpiod_get_optional(&i2c_client->dev,
-			"reset-gpios", GPIOD_OUT_HIGH);
+			"reset", GPIOD_OUT_HIGH);
 	if (IS_ERR(cs35l33->reset_gpio)) {
 		dev_err(&i2c_client->dev, "%s ERROR: Can't get reset GPIO\n",
 			__func__);

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Linus Walleij <[email protected]>

[ Upstream commit a6122b0b4211d132934ef99e7b737910e6d54d2f ]

This driver includes the legacy GPIO APIs <linux/gpio.h> and <linux/of_gpio.h> but does not use any symbols from any of them.

Drop the includes.

Further the driver is requesting "reset-gpios" rather than just "reset" from the GPIO framework. This is wrong because the gpiolib core will add "-gpios" before processing the request from e.g. device tree. Drop the suffix.

The last problem means that the optional RESET GPIO has never been properly retrieved and used even if it existed, but nobody noticed.

---
 sound/soc/codecs/cs35l34.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

index b3f98023e6a7..57d125c03fad 100644
@@ -20,14 +20,12 @@
 #include <linux/regulator/machine.h>
 #include <linux/pm_runtime.h>
 #include <linux/of_device.h>
-#include <linux/of_gpio.h>
 #include <linux/of_irq.h>
 #include <sound/core.h>
 #include <sound/pcm.h>
 #include <sound/pcm_params.h>
 #include <sound/soc.h>
 #include <sound/soc-dapm.h>
-#include <linux/gpio.h>
 #include <linux/gpio/consumer.h>
 #include <sound/initval.h>
 #include <sound/tlv.h>

@@ -1061,7 +1059,7 @@ static int cs35l34_i2c_probe(struct i2c_client *i2c_client)
 		dev_err(&i2c_client->dev, "Failed to request IRQ: %d\n", ret);
 
 	cs35l34->reset_gpio = devm_gpiod_get_optional(&i2c_client->dev,
-				"reset-gpios", GPIOD_OUT_LOW);
+				"reset", GPIOD_OUT_LOW);
 	if (IS_ERR(cs35l34->reset_gpio)) {
 		ret = PTR_ERR(cs35l34->reset_gpio);
 		goto err_regulator;

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Dmitry Baryshkov <[email protected]>

[ Upstream commit c6721b3c6423d8a348ae885a0f4c85e14f9bf85c ]

Flush queued events when disabling the crtc. This avoids timeouts when we come back and wait for dependencies (like the previous frame's flip_done).

Patchwork: https://patchwork.freedesktop.org/patch/569127/

---
 drivers/gpu/drm/msm/disp/mdp4/mdp4_crtc.c | 9 +++++++++
 1 file changed, 9 insertions(+)

index 169f9de4a12a..3100957225a7 100644
@@ -269,6 +269,7 @@ static void mdp4_crtc_atomic_disable(struct drm_crtc *crtc,
 {
 	struct mdp4_crtc *mdp4_crtc = to_mdp4_crtc(crtc);
 	struct mdp4_kms *mdp4_kms = get_kms(crtc);
+	unsigned long flags;
 
 	DBG("%s", mdp4_crtc->name);
 

@@ -281,6 +282,14 @@ static void mdp4_crtc_atomic_disable(struct drm_crtc *crtc,
 	mdp_irq_unregister(&mdp4_kms->base, &mdp4_crtc->err);
 	mdp4_disable(mdp4_kms);
 
+	if (crtc->state->event && !crtc->state->active) {
+		WARN_ON(mdp4_crtc->event);
+		spin_lock_irqsave(&mdp4_kms->dev->event_lock, flags);
+		drm_crtc_send_vblank_event(crtc, crtc->state->event);
+		crtc->state->event = NULL;
+		spin_unlock_irqrestore(&mdp4_kms->dev->event_lock, flags);
+	}
+
 	mdp4_crtc->enabled = false;
 }
 

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Konrad Dybcio <[email protected]>

[ Upstream commit 3d07a411b4faaf2b498760ccf12888f8de529de0 ]

This helper has been introduced to avoid programmer errors (missing _put calls leading to dangling refcnt) when using pm_runtime_get, use it.

While at it, start checking the return value.

Patchwork: https://patchwork.freedesktop.org/patch/543350/

---
 drivers/gpu/drm/msm/dsi/phy/dsi_phy.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

index 7fc0975cb869..62bc3756f2e2 100644
@@ -512,7 +512,9 @@ static int dsi_phy_enable_resource(struct msm_dsi_phy *phy)
 	struct device *dev = &phy->pdev->dev;
 	int ret;
 
-	pm_runtime_get_sync(dev);
+	ret = pm_runtime_resume_and_get(dev);
+	if (ret)
+		return ret;
 
 	ret = clk_prepare_enable(phy->ahb_clk);
 	if (ret) {

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Dmitry Baryshkov <[email protected]>

[ Upstream commit 5f8dec200923a76dc57187965fd59c1136f5d085 ]

In case the drm_modeset_register_all() function fails, its error code will be ignored. Instead make the drm_dev_register() bail out in case of such an error.

---
 drivers/gpu/drm/drm_drv.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)

index 203bf8d6c34c..d41a5eaa3e89 100644
@@ -895,8 +895,11 @@ int drm_dev_register(struct drm_device *dev, unsigned long flags)
 			goto err_minors;
 	}
 
-	if (drm_core_check_feature(dev, DRIVER_MODESET))
-		drm_modeset_register_all(dev);
+	if (drm_core_check_feature(dev, DRIVER_MODESET)) {
+		ret = drm_modeset_register_all(dev);
+		if (ret)
+			goto err_unload;
+	}
 
 	DRM_INFO("Initialized %s %d.%d.%d %s for %s on minor %d\n",
 		 driver->name, driver->major, driver->minor,

@@ -906,6 +909,9 @@ int drm_dev_register(struct drm_device *dev, unsigned long flags)
 
 	goto out_unlock;
 
+err_unload:
+	if (dev->driver->unload)
+		dev->driver->unload(dev);
 err_minors:
 	remove_compat_control_link(dev);
 	drm_minor_unregister(dev, DRM_MINOR_PRIMARY);

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Brent Lu <[email protected]>

[ Upstream commit 486ede0df82dd74472c6f5651e38ff48f7f766c1 ]

The drv_name in enumeration table for ALC5682I-VS codec does not match the board id string in machine driver. Modify the entry of "10EC5682" to enumerate "RTL5682" as well and remove invalid entry.

---
 sound/soc/intel/common/soc-acpi-intel-glk-match.c | 14 ++++++--------
 1 file changed, 6 insertions(+), 8 deletions(-)

index 387e73100884..8911c90bbaf6 100644
@@ -19,6 +19,11 @@ static const struct snd_soc_acpi_codecs glk_codecs = {
 	.codecs = {"MX98357A"}
 };
 
+static const struct snd_soc_acpi_codecs glk_rt5682_rt5682s_hp = {
+	.num_codecs = 2,
+	.codecs = {"10EC5682", "RTL5682"},
+};
+
 struct snd_soc_acpi_mach snd_soc_acpi_intel_glk_machines[] = {
 	{
 		.id = "INT343A",

@@ -35,20 +40,13 @@ struct snd_soc_acpi_mach snd_soc_acpi_intel_glk_machines[] = {
 		.sof_tplg_filename = "sof-glk-da7219.tplg",
 	},
 	{
-		.id = "10EC5682",
+		.comp_ids = &glk_rt5682_rt5682s_hp,
 		.drv_name = "glk_rt5682_mx98357a",
 		.fw_filename = "intel/dsp_fw_glk.bin",
 		.machine_quirk = snd_soc_acpi_codec_list,
 		.quirk_data = &glk_codecs,
 		.sof_tplg_filename = "sof-glk-rt5682.tplg",
 	},
-	{
-		.id = "RTL5682",
-		.drv_name = "glk_rt5682_max98357a",
-		.machine_quirk = snd_soc_acpi_codec_list,
-		.quirk_data = &glk_codecs,
-		.sof_tplg_filename = "sof-glk-rt5682.tplg",
-	},
 	{
 		.id = "10134242",
 		.drv_name = "glk_cs4242_mx98357a",

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: AngeloGioacchino Del Regno <[email protected]>

[ Upstream commit a4f5892914ca7709ea6d191f0edace93a5935966 ]

Some SoCs may be equipped with a GPU containing two core groups and this is exactly the case of Samsung's Exynos 5422 featuring an ARM Mali-T628 MP6 GPU: the support for this GPU in Panfrost is partial, as this driver currently supports using only one core group and that's reflected on all parts of it, including the power on (and power off, previously to this patch) function.

The issue with this is that even though executing the soft reset operation should power off all cores unconditionally, on at least one platform we're seeing a crash that seems to be happening due to an interrupt firing which may be because we are calling power transition only on the first core group, leaving the second one unchanged, or because ISR execution was pending before entering the panfrost_gpu_power_off() function and executed after powering off the GPU cores, or all of the above.

Finally, solve this by:
- Avoid to enable the power transition interrupt on reset; and
- Ignoring the core_mask and ask the GPU to poweroff both core groups

---
 drivers/gpu/drm/panfrost/panfrost_gpu.c | 12 ++++++++----
 1 file changed, 8 insertions(+), 4 deletions(-)

index c08715f033c5..55d243048516 100644
@@ -71,7 +71,12 @@ int panfrost_gpu_soft_reset(struct panfrost_device *pfdev)
 	}
 
 	gpu_write(pfdev, GPU_INT_CLEAR, GPU_IRQ_MASK_ALL);
-	gpu_write(pfdev, GPU_INT_MASK, GPU_IRQ_MASK_ALL);
+
+	/* Only enable the interrupts we care about */
+	gpu_write(pfdev, GPU_INT_MASK,
+		  GPU_IRQ_MASK_ERROR |
+		  GPU_IRQ_PERFCNT_SAMPLE_COMPLETED |
+		  GPU_IRQ_CLEAN_CACHES_COMPLETED);
 
 	return 0;
 }

@@ -369,11 +374,10 @@ void panfrost_gpu_power_on(struct panfrost_device *pfdev)
 
 void panfrost_gpu_power_off(struct panfrost_device *pfdev)
 {
-	u64 core_mask = panfrost_get_core_mask(pfdev);
 	int ret;
 	u32 val;
 
-	gpu_write(pfdev, SHADER_PWROFF_LO, pfdev->features.shader_present & core_mask);
+	gpu_write(pfdev, SHADER_PWROFF_LO, pfdev->features.shader_present);
 	ret = readl_relaxed_poll_timeout(pfdev->iomem + SHADER_PWRTRANS_LO,
 					 val, !val, 1, 1000);
 	if (ret)

@@ -385,7 +389,7 @@ void panfrost_gpu_power_off(struct panfrost_device *pfdev)
 	if (ret)
 		dev_err(pfdev->dev, "tiler power transition timeout");
 
-	gpu_write(pfdev, L2_PWROFF_LO, pfdev->features.l2_present & core_mask);
+	gpu_write(pfdev, L2_PWROFF_LO, pfdev->features.l2_present);
 	ret = readl_poll_timeout(pfdev->iomem + L2_PWRTRANS_LO,
 				 val, !val, 0, 1000);
 	if (ret)

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Yang Yingliang <[email protected]>

[ Upstream commit 7a2464fac80d42f6f8819fed97a553e9c2f43310 ]

check the alloc_workqueue return value in radeon_crtc_init() to avoid null-ptr-deref.

---
 drivers/gpu/drm/radeon/radeon_display.c | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)

index ca5598ae8bfc..1814bb8e14f1 100644
@@ -687,11 +687,16 @@ static void radeon_crtc_init(struct drm_device *dev, int index)
 	if (radeon_crtc == NULL)
 		return;
 
+	radeon_crtc->flip_queue = alloc_workqueue("radeon-crtc", WQ_HIGHPRI, 0);
+	if (!radeon_crtc->flip_queue) {
+		kfree(radeon_crtc);
+		return;
+	}
+
 	drm_crtc_init(dev, &radeon_crtc->base, &radeon_crtc_funcs);
 
 	drm_mode_crtc_set_gamma_size(&radeon_crtc->base, 256);
 	radeon_crtc->crtc_id = index;
-	radeon_crtc->flip_queue = alloc_workqueue("radeon-crtc", WQ_HIGHPRI, 0);
 	rdev->mode_info.crtcs[index] = radeon_crtc;
 
 	if (rdev->family >= CHIP_BONAIRE) {

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Zhipeng Lu <[email protected]>

[ Upstream commit 0737df9ed0997f5b8addd6e2b9699a8c6edba2e4 ]

The rdev->pm.dpm.ps allocated by kcalloc should be freed in every following error-handling path. However, in the error-handling of rdev->pm.power_state[i].clock_info the rdev->pm.dpm.ps is not freed, resulting in a memleak in this function.

---
 drivers/gpu/drm/radeon/sumo_dpm.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

index f74f381af05f..d49c145db437 100644
@@ -1493,8 +1493,10 @@ static int sumo_parse_power_table(struct radeon_device *rdev)
 		non_clock_array_index = power_state->v2.nonClockInfoIndex;
 		non_clock_info = (struct _ATOM_PPLIB_NONCLOCK_INFO *)
 			&non_clock_info_array->nonClockInfo[non_clock_array_index];
-		if (!rdev->pm.power_state[i].clock_info)
+		if (!rdev->pm.power_state[i].clock_info) {
+			kfree(rdev->pm.dpm.ps);
 			return -EINVAL;
+		}
 		ps = kzalloc(sizeof(struct sumo_ps), GFP_KERNEL);
 		if (ps == NULL) {
 			kfree(rdev->pm.dpm.ps);

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Zhipeng Lu <[email protected]>

[ Upstream commit 28c28d7f77c06ac2c0b8f9c82bc04eba22912b3b ]

The rdev->pm.dpm.ps allocated by kcalloc should be freed in every following error-handling path. However, in the error-handling of rdev->pm.power_state[i].clock_info the rdev->pm.dpm.ps is not freed, resulting in a memleak in this function.

---
 drivers/gpu/drm/radeon/trinity_dpm.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

index 08ea1c864cb2..ef1cc7bad20a 100644
@@ -1726,8 +1726,10 @@ static int trinity_parse_power_table(struct radeon_device *rdev)
 		non_clock_array_index = power_state->v2.nonClockInfoIndex;
 		non_clock_info = (struct _ATOM_PPLIB_NONCLOCK_INFO *)
 			&non_clock_info_array->nonClockInfo[non_clock_array_index];
-		if (!rdev->pm.power_state[i].clock_info)
+		if (!rdev->pm.power_state[i].clock_info) {
+			kfree(rdev->pm.dpm.ps);
 			return -EINVAL;
+		}
 		ps = kzalloc(sizeof(struct sumo_ps), GFP_KERNEL);
 		if (ps == NULL) {
 			kfree(rdev->pm.dpm.ps);

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Tomi Valkeinen <[email protected]>

[ Upstream commit 155d6fb61270dd297f128731cd155080deee8f3a ]

'ret' could be uninitialized at the end of the function, although it's not clear if that can happen in practice.

---
 drivers/gpu/drm/bridge/cadence/cdns-mhdp8546-hdcp.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

index 946212a95598..5e3b8edcf794 100644
@@ -403,7 +403,8 @@ static int _cdns_mhdp_hdcp_disable(struct cdns_mhdp_device *mhdp)
 
 static int _cdns_mhdp_hdcp_enable(struct cdns_mhdp_device *mhdp, u8 content_type)
 {
-	int ret, tries = 3;
+	int ret = -EINVAL;
+	int tries = 3;
 	u32 i;
 
 	for (i = 0; i < tries; i++) {

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Tomi Valkeinen <[email protected]>

[ Upstream commit 32bd29b619638256c5b75fb021d6d9f12fc4a984 ]

If the hpd_pin is invalid, the driver returns 'ret'. But 'ret' contains 0, instead of an error value.

Return -EINVAL instead.

---
 drivers/gpu/drm/bridge/tc358767.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

index 7ef78283e3d3..926ab5c3c31a 100644
@@ -2097,7 +2097,7 @@ static int tc_probe(struct i2c_client *client, const struct i2c_device_id *id)
 	} else {
 		if (tc->hpd_pin < 0 || tc->hpd_pin > 1) {
 			dev_err(dev, "failed to parse HPD number\n");
-			return ret;
+			return -EINVAL;
 		}
 	}
 

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Zhipeng Lu <[email protected]>

[ Upstream commit 5d3c8990e2bbf929cb211563dadd70708f42e4e6 ]

The dma_q->p_left_data alloced by kzalloc should be freed in all the following error handling paths. However, it hasn't been freed in the allocation error paths of dev->video_mode.isoc_ctl.urb and dev->video_mode.isoc_ctl.transfer_buffer.

On the other hand, the dma_q->p_left_data did be freed in the error-handling paths after that of dev->video_mode.isoc_ctl.urb and dev->video_mode.isoc_ctl.transfer_buffer, by calling cx231xx_uninit_isoc(dev). So the same free operation should be done in error-handling paths of those two allocation.

---
 drivers/media/usb/cx231xx/cx231xx-core.c | 2 ++
 1 file changed, 2 insertions(+)

index 727e6268567f..f1feccc28bf0 100644
@@ -1024,6 +1024,7 @@ int cx231xx_init_isoc(struct cx231xx *dev, int max_packets,
 	if (!dev->video_mode.isoc_ctl.urb) {
 		dev_err(dev->dev,
 			"cannot alloc memory for usb buffers\n");
+		kfree(dma_q->p_left_data);
 		return -ENOMEM;
 	}
 

@@ -1033,6 +1034,7 @@ int cx231xx_init_isoc(struct cx231xx *dev, int max_packets,
 		dev_err(dev->dev,
 			"cannot allocate memory for usbtransfer\n");
 		kfree(dev->video_mode.isoc_ctl.urb);
+		kfree(dma_q->p_left_data);
 		return -ENOMEM;
 	}
 

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Chengchang Tang <[email protected]>

[ Upstream commit 288f535951aa81ed674f5e5477ab11b9d9351b8c ]

When a reserved QP fails to be created, the memory of the remaining created reserved QPs is leaked.

---
 drivers/infiniband/hw/hns/hns_roce_hw_v2.c | 4 ++++
 1 file changed, 4 insertions(+)

index 280a3458bb53..58fbb1d3b7f4 100644
@@ -2819,6 +2819,10 @@ static int free_mr_alloc_res(struct hns_roce_dev *hr_dev)
 	return 0;
 
 create_failed_qp:
+	for (i--; i >= 0; i--) {
+		hns_roce_v2_destroy_qp(&free_mr->rsv_qp[i]->ibqp, NULL);
+		kfree(free_mr->rsv_qp[i]);
+	}
 	hns_roce_destroy_cq(cq, NULL);
 	kfree(cq);
 

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Satya Priya Kakitapalli <[email protected]>

[ Upstream commit 6ebd9a4f8b8d2b35cf965a04849c4ba763722f13 ]

Update the test_ctl_hi_val and test_ctl_hi1_val of gpu_cc_pll1 as per latest HW recommendation.

---
 drivers/clk/qcom/gpucc-sm8150.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

index 8422fd047493..c89a5b59ddb7 100644
@@ -37,8 +37,8 @@ static struct alpha_pll_config gpu_cc_pll1_config = {
 	.config_ctl_hi_val = 0x00002267,
 	.config_ctl_hi1_val = 0x00000024,
 	.test_ctl_val = 0x00000000,
-	.test_ctl_hi_val = 0x00000002,
-	.test_ctl_hi1_val = 0x00000000,
+	.test_ctl_hi_val = 0x00000000,
+	.test_ctl_hi1_val = 0x00000020,
 	.user_ctl_val = 0x00000000,
 	.user_ctl_hi_val = 0x00000805,
 	.user_ctl_hi1_val = 0x000000d0,

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Tomi Valkeinen <[email protected]>

[ Upstream commit 5705b0e0eb550ff834125a46a4ef99b62093d83d ]

The driver always calls mipi_csis_runtime_suspend() and mipi_csis_clk_disable() in remove(). This causes multiple WARNs from the kernel, as the clocks get disabled too many times.

Fix the remove() to call mipi_csis_runtime_suspend() and mipi_csis_clk_disable() in a way that reverses what is done in probe().

---
 drivers/media/platform/nxp/imx-mipi-csis.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

index 905072871ed2..196f2bba419f 100644
@@ -1553,8 +1553,10 @@ static int mipi_csis_remove(struct platform_device *pdev)
 	v4l2_async_nf_cleanup(&csis->notifier);
 	v4l2_async_unregister_subdev(&csis->sd);
 
+	if (!pm_runtime_enabled(&pdev->dev))
+		mipi_csis_runtime_suspend(&pdev->dev);
+
 	pm_runtime_disable(&pdev->dev);
-	mipi_csis_runtime_suspend(&pdev->dev);
 	mipi_csis_clk_disable(csis);
 	media_entity_cleanup(&csis->sd.entity);
 	fwnode_handle_put(csis->sd.fwnode);

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Mehdi Djait <[email protected]>

[ Upstream commit 25bf28b25a2afa1864b7143259443160d9163ea0 ]

The bus-type belongs to the endpoint's properties and should therefore be moved.

---
 .../devicetree/bindings/media/rockchip-isp1.yaml      | 11 ++++++-----
 1 file changed, 6 insertions(+), 5 deletions(-)

index b3661d7d4357..2a0ad332f5ce 100644
@@ -90,15 +90,16 @@ properties:
         description: connection point for input on the parallel interface
 
         properties:
-          bus-type:
-            enum: [5, 6]
-
           endpoint:
             $ref: video-interfaces.yaml#
             unevaluatedProperties: false
 
-        required:
-          - bus-type
+            properties:
+              bus-type:
+                enum: [5, 6]
+
+            required:
+              - bus-type
 
     anyOf:
       - required:

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Tomi Valkeinen <[email protected]>

[ Upstream commit 452f604a4683654f4d9472b3126d8da61d748443 ]

Add missing calls to media_device_cleanup() to fix memory leak.

---
 drivers/media/platform/rockchip/rkisp1/rkisp1-dev.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

index f2475c6235ea..2b76339f9381 100644
@@ -582,7 +582,7 @@ static int rkisp1_probe(struct platform_device *pdev)
 
 	ret = v4l2_device_register(rkisp1->dev, &rkisp1->v4l2_dev);
 	if (ret)
-		goto err_pm_runtime_disable;
+		goto err_media_dev_cleanup;
 
 	ret = media_device_register(&rkisp1->media_dev);
 	if (ret) {

@@ -617,6 +617,8 @@ static int rkisp1_probe(struct platform_device *pdev)
 	media_device_unregister(&rkisp1->media_dev);
 err_unreg_v4l2_dev:
 	v4l2_device_unregister(&rkisp1->v4l2_dev);
+err_media_dev_cleanup:
+	media_device_cleanup(&rkisp1->media_dev);
 err_pm_runtime_disable:
 	pm_runtime_disable(&pdev->dev);
 	return ret;

@@ -637,6 +639,8 @@ static int rkisp1_remove(struct platform_device *pdev)
 	media_device_unregister(&rkisp1->media_dev);
 	v4l2_device_unregister(&rkisp1->v4l2_dev);
 
+	media_device_cleanup(&rkisp1->media_dev);
+
 	pm_runtime_disable(&pdev->dev);
 
 	return 0;

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Chris Morgan <[email protected]>

[ Upstream commit 799825aa87200ade1ba21db853d1c2ff720dcfe0 ]

The AVCL register, according to the datasheet, comes in increments of -0.2v between -4.4v (represented by 0x0) to -5.0v (represented by 0x3). The current calculation is done by adding the defined AVCL value in mV to -4400 and then dividing by 200 to get the register value. Unfortunately if I subtract -4400 from -4400 I get -8800, which divided by 200 gives me -44. If I instead subtract -4400 from -4400 I get 0, which divided by 200 gives me 0. Based on the datasheet this is the expected register value.

---
 drivers/gpu/drm/panel/panel-sitronix-st7701.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

index 225b9884f61a..54b28992db5d 100644
@@ -288,7 +288,7 @@ static void st7701_init_sequence(struct st7701 *st7701)
 		   FIELD_PREP(DSI_CMD2_BK1_PWRCTRL2_AVDD_MASK,
 			      DIV_ROUND_CLOSEST(desc->avdd_mv - 6200, 200)) |
 		   FIELD_PREP(DSI_CMD2_BK1_PWRCTRL2_AVCL_MASK,
-			      DIV_ROUND_CLOSEST(-4400 + desc->avcl_mv, 200)));
+			      DIV_ROUND_CLOSEST(-4400 - desc->avcl_mv, 200)));
 
 	/* T2D = 0.2us * T2D[3:0] */
 	ST7701_DSI(st7701, DSI_CMD2_BK1_SPD1,

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Chao Yu <[email protected]>

[ Upstream commit 55fdc1c24a1d6229fe0ecf31335fb9a2eceaaa00 ]

If inode is compressed, but not encrypted, it missed to call f2fs_wait_on_block_writeback() to wait for GCed page writeback in IPU write path.

Thread A GC-Thread
- f2fs_gc
- do_garbage_collect
- gc_data_segment
- move_data_block
- f2fs_submit_page_write migrate normal cluster's block via meta_inode's page cache
- f2fs_write_single_data_page
- f2fs_do_write_data_page
- f2fs_inplace_write_data
- f2fs_submit_page_bio

IRQ
- f2fs_read_end_io
IRQ
old data overrides new data due to out-of-order GC and common IO.
- f2fs_read_end_io

---
 fs/f2fs/data.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

index 3666c1fd77a6..8b561af37974 100644
@@ -2490,9 +2490,6 @@ int f2fs_encrypt_one_page(struct f2fs_io_info *fio)
 
 	page = fio->compressed_page ? fio->compressed_page : fio->page;
 
-	/* wait for GCed page writeback via META_MAPPING */
-	f2fs_wait_on_block_writeback(inode, fio->old_blkaddr);
-
 	if (fscrypt_inode_uses_inline_crypto(inode))
 		return 0;
 

@@ -2681,6 +2678,10 @@ int f2fs_do_write_data_page(struct f2fs_io_info *fio)
 		goto out_writepage;
 	}
 
+	/* wait for GCed page writeback via META_MAPPING */
+	if (fio->post_read)
+		f2fs_wait_on_block_writeback(inode, fio->old_blkaddr);
+
 	/*
 	 * If current allocation needs SSR,
 	 * it had better in-place writes for updated data.

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Chao Yu <[email protected]>

[ Upstream commit fb9b65340c818875ea86464faf3c744bdce0055c ]

f2fs_move_file_range() doesn't support migrating compressed cluster data, let's add the missing check condition and return -EOPNOTSUPP for the case until we support it.

---
 fs/f2fs/file.c | 5 +++++
 1 file changed, 5 insertions(+)

index 9b9fb3c57ec6..3f2c55b9aa8a 100644
@@ -2821,6 +2821,11 @@ static int f2fs_move_file_range(struct file *file_in, loff_t pos_in,
 			goto out;
 	}
 
+	if (f2fs_compressed_file(src) || f2fs_compressed_file(dst)) {
+		ret = -EOPNOTSUPP;
+		goto out_unlock;
+	}
+
 	ret = -EINVAL;
 	if (pos_in + len > src->i_size || pos_in + len < pos_in)
 		goto out_unlock;

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Chao Yu <[email protected]>

[ Upstream commit bb34cc6ca87ff78f9fb5913d7619dc1389554da6 ]

In f2fs_filemap_fault(), it fixes to update iostat info only if VM_FAULT_LOCKED is tagged in return value of filemap_fault().

---
 fs/f2fs/file.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

index 3f2c55b9aa8a..fd22854dbeae 100644
@@ -42,7 +42,7 @@ static vm_fault_t f2fs_filemap_fault(struct vm_fault *vmf)
 	vm_fault_t ret;
 
 	ret = filemap_fault(vmf);
-	if (!ret)
+	if (ret & VM_FAULT_LOCKED)
 		f2fs_update_iostat(F2FS_I_SB(inode), inode,
 					APP_MAPPED_READ_IO, F2FS_BLKSIZE);
 

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Dan Carpenter <[email protected]>

[ Upstream commit a2dd235df435a05d389240be748909ada91201d2 ]

If call to file->f_op->open() fails, then call dvb_device_put(dvbdev).

---
 drivers/media/dvb-core/dvbdev.c | 2 ++
 1 file changed, 2 insertions(+)

index 9f9a97652708..d352e028491a 100644
@@ -104,6 +104,8 @@ static int dvb_device_open(struct inode *inode, struct file *file)
 			err = file->f_op->open(inode, file);
 		up_read(&minor_rwsem);
 		mutex_unlock(&dvbdev_mutex);
+		if (err)
+			dvb_device_put(dvbdev);
 		return err;
 	}
 fail:

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Christophe JAILLET <[email protected]>

[ Upstream commit 5b2f885e2f6f482d05c23f04c8240f7b4fc5bdb5 ]

If an error occurs after a successful i2c_mux_add_adapter(), then i2c_mux_del_adapters() should be called to free some resources, as already done in the remove function.

---
 drivers/media/dvb-frontends/m88ds3103.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

index 9a0d43c7ba9e..ce99f7dfb5a5 100644
@@ -1894,7 +1894,7 @@ static int m88ds3103_probe(struct i2c_client *client,
 		/* get frontend address */
 		ret = regmap_read(dev->regmap, 0x29, &utmp);
 		if (ret)
-			goto err_kfree;
+			goto err_del_adapters;
 		dev->dt_addr = ((utmp & 0x80) == 0) ? 0x42 >> 1 : 0x40 >> 1;
 		dev_dbg(&client->dev, "dt addr is 0x%02x\n", dev->dt_addr);
 

@@ -1902,11 +1902,14 @@ static int m88ds3103_probe(struct i2c_client *client,
 						      dev->dt_addr);
 		if (IS_ERR(dev->dt_client)) {
 			ret = PTR_ERR(dev->dt_client);
-			goto err_kfree;
+			goto err_del_adapters;
 		}
 	}
 
 	return 0;
+
+err_del_adapters:
+	i2c_mux_del_adapters(dev->muxc);
 err_kfree:
 	kfree(dev);
 err:

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Claudiu Beznea <[email protected]>

[ Upstream commit 5f9e29b9159a41fcf6733c3b59fa46a90ce3ae20 ]

Code in rzg2l_cpg_reset() is equivalent with the combined code of rzg2l_cpg_assert() and rzg2l_cpg_deassert(). There is no need to have different versions thus re-use rzg2l_cpg_assert() and rzg2l_cpg_deassert().

Stable-dep-of: da235d2fac21 ("clk: renesas: rzg2l: Check reset monitor registers")

---
 drivers/clk/renesas/rzg2l-cpg.c | 38 +++++++++++++--------------------
 1 file changed, 15 insertions(+), 23 deletions(-)

index 84767cfc1e73..93720f319409 100644
@@ -1115,29 +1115,6 @@ rzg2l_cpg_register_mod_clk(const struct rzg2l_mod_clk *mod,
 
 #define rcdev_to_priv(x)	container_of(x, struct rzg2l_cpg_priv, rcdev)
 
-static int rzg2l_cpg_reset(struct reset_controller_dev *rcdev,
-			   unsigned long id)
-{
-	struct rzg2l_cpg_priv *priv = rcdev_to_priv(rcdev);
-	const struct rzg2l_cpg_info *info = priv->info;
-	unsigned int reg = info->resets[id].off;
-	u32 dis = BIT(info->resets[id].bit);
-	u32 we = dis << 16;
-
-	dev_dbg(rcdev->dev, "reset id:%ld offset:0x%x\n", id, CLK_RST_R(reg));
-
-	/* Reset module */
-	writel(we, priv->base + CLK_RST_R(reg));
-
-	/* Wait for at least one cycle of the RCLK clock (@ ca. 32 kHz) */
-	udelay(35);
-
-	/* Release module from reset state */
-	writel(we | dis, priv->base + CLK_RST_R(reg));
-
-	return 0;
-}
-
 static int rzg2l_cpg_assert(struct reset_controller_dev *rcdev,
 			    unsigned long id)
 {

@@ -1168,6 +1145,21 @@ static int rzg2l_cpg_deassert(struct reset_controller_dev *rcdev,
 	return 0;
 }
 
+static int rzg2l_cpg_reset(struct reset_controller_dev *rcdev,
+			   unsigned long id)
+{
+	int ret;
+
+	ret = rzg2l_cpg_assert(rcdev, id);
+	if (ret)
+		return ret;
+
+	/* Wait for at least one cycle of the RCLK clock (@ ca. 32 kHz) */
+	udelay(35);
+
+	return rzg2l_cpg_deassert(rcdev, id);
+}
+
 static int rzg2l_cpg_status(struct reset_controller_dev *rcdev,
 			    unsigned long id)
 {

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Claudiu Beznea <[email protected]>

[ Upstream commit da235d2fac212d0add570e755feb1167a830bc99 ]

The hardware manual of both RZ/G2L and RZ/G3S specifies that the reset monitor registers need to be interrogated when the reset signals are toggled (chapters "Procedures for Supplying and Stopping Reset Signals" and "Procedure for Activating Modules"). Without this, there is a chance that different modules (e.g. Ethernet) are not ready after their reset signal is toggled, leading to failures (on probe or resume from deep sleep states).

The same indications are available for RZ/V2M for TYPE-B reset controls.

---
 drivers/clk/renesas/rzg2l-cpg.c | 59 ++++++++++++++++++++++++---------
 1 file changed, 44 insertions(+), 15 deletions(-)

index 93720f319409..473feb36a38f 100644
@@ -1121,12 +1121,27 @@ static int rzg2l_cpg_assert(struct reset_controller_dev *rcdev,
 	struct rzg2l_cpg_priv *priv = rcdev_to_priv(rcdev);
 	const struct rzg2l_cpg_info *info = priv->info;
 	unsigned int reg = info->resets[id].off;
-	u32 value = BIT(info->resets[id].bit) << 16;
+	u32 mask = BIT(info->resets[id].bit);
+	s8 monbit = info->resets[id].monbit;
+	u32 value = mask << 16;
 
 	dev_dbg(rcdev->dev, "assert id:%ld offset:0x%x\n", id, CLK_RST_R(reg));
 
 	writel(value, priv->base + CLK_RST_R(reg));
-	return 0;
+
+	if (info->has_clk_mon_regs) {
+		reg = CLK_MRST_R(reg);
+	} else if (monbit >= 0) {
+		reg = CPG_RST_MON;
+		mask = BIT(monbit);
+	} else {
+		/* Wait for at least one cycle of the RCLK clock (@ ca. 32 kHz) */
+		udelay(35);
+		return 0;
+	}
+
+	return readl_poll_timeout_atomic(priv->base + reg, value,
+					 value & mask, 10, 200);
 }
 
 static int rzg2l_cpg_deassert(struct reset_controller_dev *rcdev,

@@ -1135,14 +1150,28 @@ static int rzg2l_cpg_deassert(struct reset_controller_dev *rcdev,
 	struct rzg2l_cpg_priv *priv = rcdev_to_priv(rcdev);
 	const struct rzg2l_cpg_info *info = priv->info;
 	unsigned int reg = info->resets[id].off;
-	u32 dis = BIT(info->resets[id].bit);
-	u32 value = (dis << 16) | dis;
+	u32 mask = BIT(info->resets[id].bit);
+	s8 monbit = info->resets[id].monbit;
+	u32 value = (mask << 16) | mask;
 
 	dev_dbg(rcdev->dev, "deassert id:%ld offset:0x%x\n", id,
 		CLK_RST_R(reg));
 
 	writel(value, priv->base + CLK_RST_R(reg));
-	return 0;
+
+	if (info->has_clk_mon_regs) {
+		reg = CLK_MRST_R(reg);
+	} else if (monbit >= 0) {
+		reg = CPG_RST_MON;
+		mask = BIT(monbit);
+	} else {
+		/* Wait for at least one cycle of the RCLK clock (@ ca. 32 kHz) */
+		udelay(35);
+		return 0;
+	}
+
+	return readl_poll_timeout_atomic(priv->base + reg, value,
+					 !(value & mask), 10, 200);
 }
 
 static int rzg2l_cpg_reset(struct reset_controller_dev *rcdev,

@@ -1154,9 +1183,6 @@ static int rzg2l_cpg_reset(struct reset_controller_dev *rcdev,
 	if (ret)
 		return ret;
 
-	/* Wait for at least one cycle of the RCLK clock (@ ca. 32 kHz) */
-	udelay(35);
-
 	return rzg2l_cpg_deassert(rcdev, id);
 }
 

@@ -1165,18 +1191,21 @@ static int rzg2l_cpg_status(struct reset_controller_dev *rcdev,
 {
 	struct rzg2l_cpg_priv *priv = rcdev_to_priv(rcdev);
 	const struct rzg2l_cpg_info *info = priv->info;
-	unsigned int reg = info->resets[id].off;
-	u32 bitmask = BIT(info->resets[id].bit);
 	s8 monbit = info->resets[id].monbit;
+	unsigned int reg;
+	u32 bitmask;
 
 	if (info->has_clk_mon_regs) {
-		return !!(readl(priv->base + CLK_MRST_R(reg)) & bitmask);
+		reg = CLK_MRST_R(info->resets[id].off);
+		bitmask = BIT(info->resets[id].bit);
 	} else if (monbit >= 0) {
-		u32 monbitmask = BIT(monbit);
-
-		return !!(readl(priv->base + CPG_RST_MON) & monbitmask);
+		reg = CPG_RST_MON;
+		bitmask = BIT(monbit);
+	} else {
+		return -ENOTSUPP;
 	}
-	return -ENOTSUPP;
+
+	return !!(readl(priv->base + reg) & bitmask);
 }
 
 static const struct reset_control_ops rzg2l_cpg_reset_ops = {

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Jessica Zhang <[email protected]>

[ Upstream commit 980fffd0c69e5df0f67ee089d405899d532aeeab ]

Set the input_sel bit for encoders as it was missed in the initial implementation.

Patchwork: https://patchwork.freedesktop.org/patch/572007/

---
 drivers/gpu/drm/msm/disp/dpu1/dpu_hw_intf.c | 2 +-
 drivers/gpu/drm/msm/disp/dpu1/dpu_hw_lm.c   | 2 +-
 drivers/gpu/drm/msm/disp/dpu1/dpu_hw_util.c | 9 +++++++--
 drivers/gpu/drm/msm/disp/dpu1/dpu_hw_util.h | 3 ++-
 4 files changed, 11 insertions(+), 5 deletions(-)

index b9dddf576c02..7e210ba0b104 100644
@@ -324,7 +324,7 @@ static u32 dpu_hw_intf_get_line_count(struct dpu_hw_intf *intf)
 
 static void dpu_hw_intf_setup_misr(struct dpu_hw_intf *intf, bool enable, u32 frame_count)
 {
-	dpu_hw_setup_misr(&intf->hw, INTF_MISR_CTRL, enable, frame_count);
+	dpu_hw_setup_misr(&intf->hw, INTF_MISR_CTRL, enable, frame_count, 0x1);
 }
 
 static int dpu_hw_intf_collect_misr(struct dpu_hw_intf *intf, u32 *misr_value)

index f5120ea91ede..2dd9f9185cfc 100644
@@ -101,7 +101,7 @@ static void dpu_hw_lm_setup_border_color(struct dpu_hw_mixer *ctx,
 
 static void dpu_hw_lm_setup_misr(struct dpu_hw_mixer *ctx, bool enable, u32 frame_count)
 {
-	dpu_hw_setup_misr(&ctx->hw, LM_MISR_CTRL, enable, frame_count);
+	dpu_hw_setup_misr(&ctx->hw, LM_MISR_CTRL, enable, frame_count, 0x0);
 }
 
 static int dpu_hw_lm_collect_misr(struct dpu_hw_mixer *ctx, u32 *misr_value)

index 8062228eada6..119dc07d6ab5 100644
@@ -450,9 +450,13 @@ u64 _dpu_hw_get_qos_lut(const struct dpu_qos_lut_tbl *tbl,
 	return 0;
 }
 
+/*
+ * note: Aside from encoders, input_sel should be set to 0x0 by default
+ */
 void dpu_hw_setup_misr(struct dpu_hw_blk_reg_map *c,
 		u32 misr_ctrl_offset,
-		bool enable, u32 frame_count)
+		bool enable, u32 frame_count,
+		u8 input_sel)
 {
 	u32 config = 0;
 

@@ -463,7 +467,8 @@ void dpu_hw_setup_misr(struct dpu_hw_blk_reg_map *c,
 
 	if (enable) {
 		config = (frame_count & MISR_FRAME_COUNT_MASK) |
-			MISR_CTRL_ENABLE | MISR_CTRL_FREE_RUN_MASK;
+			MISR_CTRL_ENABLE | MISR_CTRL_FREE_RUN_MASK |
+			((input_sel & 0xF) << 24);
 
 		DPU_REG_WRITE(c, misr_ctrl_offset, config);
 	} else {

index 27f4c39e35ab..dffad0a83781 100644
@@ -352,7 +352,8 @@ u64 _dpu_hw_get_qos_lut(const struct dpu_qos_lut_tbl *tbl,
 void dpu_hw_setup_misr(struct dpu_hw_blk_reg_map *c,
 		u32 misr_ctrl_offset,
 		bool enable,
-		u32 frame_count);
+		u32 frame_count,
+		u8 input_sel);
 
 int dpu_hw_collect_misr(struct dpu_hw_blk_reg_map *c,
 		u32 misr_ctrl_offset,

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Jessica Zhang <[email protected]>

[ Upstream commit 3313c23f3eab698bc6b904520ee608fc0f7b03d0 ]

Drop the enable and frame_count parameters from dpu_hw_setup_misr() as they are always set to the same values.

In addition, replace MISR_FRAME_COUNT_MASK with MISR_FRAME_COUNT as frame_count is always set to the same value.

Patchwork: https://patchwork.freedesktop.org/patch/572009/

---
 drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.c    |  4 ++--
 drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c |  4 ++--
 drivers/gpu/drm/msm/disp/dpu1/dpu_hw_intf.c |  6 +++---
 drivers/gpu/drm/msm/disp/dpu1/dpu_hw_intf.h |  4 ++--
 drivers/gpu/drm/msm/disp/dpu1/dpu_hw_lm.c   |  6 +++---
 drivers/gpu/drm/msm/disp/dpu1/dpu_hw_lm.h   |  3 ++-
 drivers/gpu/drm/msm/disp/dpu1/dpu_hw_util.c | 19 +++++--------------
 drivers/gpu/drm/msm/disp/dpu1/dpu_hw_util.h |  9 +++------
 8 files changed, 22 insertions(+), 33 deletions(-)

index 6c0ffe8e4adb..5a5821e59dc1 100644
@@ -1,6 +1,6 @@
 // SPDX-License-Identifier: GPL-2.0-only
 /*
- * Copyright (c) 2022 Qualcomm Innovation Center, Inc. All rights reserved.
+ * Copyright (c) 2022-2023 Qualcomm Innovation Center, Inc. All rights reserved.
  * Copyright (c) 2014-2021 The Linux Foundation. All rights reserved.
  * Copyright (C) 2013 Red Hat
  * Author: Rob Clark <[email protected]>

@@ -124,7 +124,7 @@ static void dpu_crtc_setup_lm_misr(struct dpu_crtc_state *crtc_state)
 			continue;
 
 		/* Calculate MISR over 1 frame */
-		m->hw_lm->ops.setup_misr(m->hw_lm, true, 1);
+		m->hw_lm->ops.setup_misr(m->hw_lm);
 	}
 }
 

index 547f9f2b9fcb..b0eb881f8af1 100644
@@ -2,7 +2,7 @@
 /*
  * Copyright (C) 2013 Red Hat
  * Copyright (c) 2014-2018, 2020-2021 The Linux Foundation. All rights reserved.
- * Copyright (c) 2022 Qualcomm Innovation Center, Inc. All rights reserved.
+ * Copyright (c) 2022-2023 Qualcomm Innovation Center, Inc. All rights reserved.
  *
  * Author: Rob Clark <[email protected]>
  */

@@ -257,7 +257,7 @@ void dpu_encoder_setup_misr(const struct drm_encoder *drm_enc)
 		if (!phys->hw_intf || !phys->hw_intf->ops.setup_misr)
 			continue;
 
-		phys->hw_intf->ops.setup_misr(phys->hw_intf, true, 1);
+		phys->hw_intf->ops.setup_misr(phys->hw_intf);
 	}
 }
 

index 7e210ba0b104..384558d2f960 100644
@@ -1,6 +1,6 @@
 // SPDX-License-Identifier: GPL-2.0-only
 /*
- * Copyright (c) 2022 Qualcomm Innovation Center, Inc. All rights reserved.
+ * Copyright (c) 2022-2023 Qualcomm Innovation Center, Inc. All rights reserved.
  * Copyright (c) 2015-2018, The Linux Foundation. All rights reserved.
  */
 

@@ -322,9 +322,9 @@ static u32 dpu_hw_intf_get_line_count(struct dpu_hw_intf *intf)
 	return DPU_REG_READ(c, INTF_LINE_COUNT);
 }
 
-static void dpu_hw_intf_setup_misr(struct dpu_hw_intf *intf, bool enable, u32 frame_count)
+static void dpu_hw_intf_setup_misr(struct dpu_hw_intf *intf)
 {
-	dpu_hw_setup_misr(&intf->hw, INTF_MISR_CTRL, enable, frame_count, 0x1);
+	dpu_hw_setup_misr(&intf->hw, INTF_MISR_CTRL, 0x1);
 }
 
 static int dpu_hw_intf_collect_misr(struct dpu_hw_intf *intf, u32 *misr_value)

index 643dd10bc030..e75339b96a1d 100644
@@ -1,6 +1,6 @@
 /* SPDX-License-Identifier: GPL-2.0-only */
 /*
- * Copyright (c) 2022 Qualcomm Innovation Center, Inc. All rights reserved.
+ * Copyright (c) 2022-2023 Qualcomm Innovation Center, Inc. All rights reserved.
  * Copyright (c) 2015-2018, The Linux Foundation. All rights reserved.
  */
 

@@ -80,7 +80,7 @@ struct dpu_hw_intf_ops {
 	void (*bind_pingpong_blk)(struct dpu_hw_intf *intf,
 			bool enable,
 			const enum dpu_pingpong pp);
-	void (*setup_misr)(struct dpu_hw_intf *intf, bool enable, u32 frame_count);
+	void (*setup_misr)(struct dpu_hw_intf *intf);
 	int (*collect_misr)(struct dpu_hw_intf *intf, u32 *misr_value);
 };
 

index 2dd9f9185cfc..cc04fb979fb5 100644
@@ -1,6 +1,6 @@
 // SPDX-License-Identifier: GPL-2.0-only
 /*
- * Copyright (c) 2022 Qualcomm Innovation Center, Inc. All rights reserved.
+ * Copyright (c) 2022-2023 Qualcomm Innovation Center, Inc. All rights reserved.
  * Copyright (c) 2015-2021, The Linux Foundation. All rights reserved.
  */
 

@@ -99,9 +99,9 @@ static void dpu_hw_lm_setup_border_color(struct dpu_hw_mixer *ctx,
 	}
 }
 
-static void dpu_hw_lm_setup_misr(struct dpu_hw_mixer *ctx, bool enable, u32 frame_count)
+static void dpu_hw_lm_setup_misr(struct dpu_hw_mixer *ctx)
 {
-	dpu_hw_setup_misr(&ctx->hw, LM_MISR_CTRL, enable, frame_count, 0x0);
+	dpu_hw_setup_misr(&ctx->hw, LM_MISR_CTRL, 0x0);
 }
 
 static int dpu_hw_lm_collect_misr(struct dpu_hw_mixer *ctx, u32 *misr_value)

index 652ddfdedec3..0a050eb247b9 100644
@@ -1,5 +1,6 @@
 /* SPDX-License-Identifier: GPL-2.0-only */
 /*
+ * Copyright (c) 2022-2023 Qualcomm Innovation Center, Inc. All rights reserved.
  * Copyright (c) 2015-2021, The Linux Foundation. All rights reserved.
  */
 

@@ -57,7 +58,7 @@ struct dpu_hw_lm_ops {
 	/**
 	 * setup_misr: Enable/disable MISR
 	 */
-	void (*setup_misr)(struct dpu_hw_mixer *ctx, bool enable, u32 frame_count);
+	void (*setup_misr)(struct dpu_hw_mixer *ctx);
 
 	/**
 	 * collect_misr: Read MISR signature

index 119dc07d6ab5..1b7439ae686a 100644
@@ -1,6 +1,6 @@
 // SPDX-License-Identifier: GPL-2.0-only
 /*
- * Copyright (c) 2022 Qualcomm Innovation Center, Inc. All rights reserved.
+ * Copyright (c) 2022-2023 Qualcomm Innovation Center, Inc. All rights reserved.
  * Copyright (c) 2015-2018, The Linux Foundation. All rights reserved.
  */
 #define pr_fmt(fmt)	"[drm:%s:%d] " fmt, __func__, __LINE__

@@ -454,9 +454,7 @@ u64 _dpu_hw_get_qos_lut(const struct dpu_qos_lut_tbl *tbl,
  * note: Aside from encoders, input_sel should be set to 0x0 by default
  */
 void dpu_hw_setup_misr(struct dpu_hw_blk_reg_map *c,
-		u32 misr_ctrl_offset,
-		bool enable, u32 frame_count,
-		u8 input_sel)
+		u32 misr_ctrl_offset, u8 input_sel)
 {
 	u32 config = 0;
 

@@ -465,16 +463,9 @@ void dpu_hw_setup_misr(struct dpu_hw_blk_reg_map *c,
 	/* Clear old MISR value (in case it's read before a new value is calculated)*/
 	wmb();
 
-	if (enable) {
-		config = (frame_count & MISR_FRAME_COUNT_MASK) |
-			MISR_CTRL_ENABLE | MISR_CTRL_FREE_RUN_MASK |
-			((input_sel & 0xF) << 24);
-
-		DPU_REG_WRITE(c, misr_ctrl_offset, config);
-	} else {
-		DPU_REG_WRITE(c, misr_ctrl_offset, 0);
-	}
-
+	config = MISR_FRAME_COUNT | MISR_CTRL_ENABLE | MISR_CTRL_FREE_RUN_MASK |
+		((input_sel & 0xF) << 24);
+	DPU_REG_WRITE(c, misr_ctrl_offset, config);
 }
 
 int dpu_hw_collect_misr(struct dpu_hw_blk_reg_map *c,

index dffad0a83781..4ae2a434372c 100644
@@ -1,6 +1,6 @@
 /* SPDX-License-Identifier: GPL-2.0-only */
 /*
- * Copyright (c) 2022 Qualcomm Innovation Center, Inc. All rights reserved.
+ * Copyright (c) 2022-2023 Qualcomm Innovation Center, Inc. All rights reserved.
  * Copyright (c) 2015-2021, The Linux Foundation. All rights reserved.
  */
 

@@ -13,7 +13,7 @@
 #include "dpu_hw_catalog.h"
 
 #define REG_MASK(n)                     ((BIT(n)) - 1)
-#define MISR_FRAME_COUNT_MASK           0xFF
+#define MISR_FRAME_COUNT                0x1
 #define MISR_CTRL_ENABLE                BIT(8)
 #define MISR_CTRL_STATUS                BIT(9)
 #define MISR_CTRL_STATUS_CLEAR          BIT(10)

@@ -350,10 +350,7 @@ u64 _dpu_hw_get_qos_lut(const struct dpu_qos_lut_tbl *tbl,
 		u32 total_fl);
 
 void dpu_hw_setup_misr(struct dpu_hw_blk_reg_map *c,
-		u32 misr_ctrl_offset,
-		bool enable,
-		u32 frame_count,
-		u8 input_sel);
+		u32 misr_ctrl_offset, u8 input_sel);
 
 int dpu_hw_collect_misr(struct dpu_hw_blk_reg_map *c,
 		u32 misr_ctrl_offset,

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Hsiao Chien Sung <[email protected]>

[ Upstream commit 21b287146adf39304193e4c49198021e06a28ded ]

Return the result of clk_prepare_enable() instead of always returns 0.

---
 drivers/gpu/drm/mediatek/mtk_mdp_rdma.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

index eecfa98ff52e..b288bb6eeecc 100644
@@ -223,8 +223,7 @@ int mtk_mdp_rdma_clk_enable(struct device *dev)
 {
 	struct mtk_mdp_rdma *rdma = dev_get_drvdata(dev);
 
-	clk_prepare_enable(rdma->clk);
-	return 0;
+	return clk_prepare_enable(rdma->clk);
 }
 
 void mtk_mdp_rdma_clk_disable(struct device *dev)

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Hsiao Chien Sung <[email protected]>

[ Upstream commit 73b5ab27ab2ee616f2709dc212c2b0007894a12e ]

Do not reset Merge while using CMDQ because reset API doesn't wait for frame done event as CMDQ does and could lead to underrun when the layer is switching off.

---
 drivers/gpu/drm/mediatek/mtk_disp_merge.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

index 6428b6203ffe..211140e87568 100644
@@ -104,7 +104,7 @@ void mtk_merge_stop_cmdq(struct device *dev, struct cmdq_pkt *cmdq_pkt)
 	mtk_ddp_write(cmdq_pkt, 0, &priv->cmdq_reg, priv->regs,
 		      DISP_REG_MERGE_CTRL);
 
-	if (priv->async_clk)
+	if (!cmdq_pkt && priv->async_clk)
 		reset_control_reset(priv->reset_ctl);
 }
 

2.43.0

6.1-stable review patch. If anyone has any objections, please let me know.

------------------

From: Alex Deucher <[email protected]>

[ Upstream commit afe58346d5d3887b3e49ff623d2f2e471f232a8d ]

Should be -EOPNOTSUPP.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

index 8123feb1a116..06ab6066da61 100644
@@ -596,7 +596,7 @@ static ssize_t amdgpu_debugfs_regs_smc_read(struct file *f, char __user *buf,
 	int r;
 
 	if (!adev->smc_rreg)
-		return -EPERM;
+		return -EOPNOTSUPP;
 
 	if (size & 0x3 || *pos & 0x3)
 		return -EINVAL;

@@ -655,7 +655,7 @@ static ssize_t amdgpu_debugfs_regs_smc_write(struct file *f, const char __user *
 	int r;
 
 	if (!adev->smc_wreg)
-		return -EPERM;
+		return -EOPNOTSUPP;
 
 	if (size & 0x3 || *pos & 0x3)
 		return -EINVAL;