6.12.80-rc1 review

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

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

From: Namjae Jeon <[email protected]>

commit beef2634f81f1c086208191f7228bce1d366493d upstream.

When a compound request consists of QUERY_DIRECTORY + QUERY_INFO (FILE_ALL_INFORMATION) and the first command consumes nearly the entire max_trans_size, get_file_all_info() would blindly call smbConvertToUTF16() with PATH_MAX, causing out-of-bounds write beyond the response buffer. In get_file_all_info(), there was a missing validation check for the client-provided OutputBufferLength before copying the filename into FileName field of the smb2_file_all_info structure.
If the filename length exceeds the available buffer space, it could lead to potential buffer overflows or memory corruption during smbConvertToUTF16 conversion. This calculating the actual free buffer size using smb2_calc_max_out_buf_len() and returning -EINVAL if the buffer is insufficient and updating smbConvertToUTF16 to use the actual filename length (clamped by PATH_MAX) to ensure a safe copy operation.

---
 fs/smb/server/smb2pdu.c |   16 ++++++++++++++--
 1 file changed, 14 insertions(+), 2 deletions(-)

@@ -4941,7 +4941,8 @@ static int get_file_all_info(struct ksmb
 	int conv_len;
 	char *filename;
 	u64 time;
-	int ret;
+	int ret, buf_free_len, filename_len;
+	struct smb2_query_info_req *req = ksmbd_req_buf_next(work);
 
 	if (!(fp->daccess & FILE_READ_ATTRIBUTES_LE)) {
 		ksmbd_debug(SMB, "no right to read the attributes : 0x%x\n",

@@ -4953,6 +4954,16 @@ static int get_file_all_info(struct ksmb
 	if (IS_ERR(filename))
 		return PTR_ERR(filename);
 
+	filename_len = strlen(filename);
+	buf_free_len = smb2_calc_max_out_buf_len(work,
+			offsetof(struct smb2_query_info_rsp, Buffer) +
+			offsetof(struct smb2_file_all_info, FileName),
+			le32_to_cpu(req->OutputBufferLength));
+	if (buf_free_len < (filename_len + 1) * 2) {
+		kfree(filename);
+		return -EINVAL;
+	}
+
 	ret = vfs_getattr(&fp->filp->f_path, &stat, STATX_BASIC_STATS,
 			  AT_STATX_SYNC_AS_STAT);
 	if (ret) {

@@ -4988,7 +4999,8 @@ static int get_file_all_info(struct ksmb
 	file_info->Mode = fp->coption;
 	file_info->AlignmentRequirement = 0;
 	conv_len = smbConvertToUTF16((__le16 *)file_info->FileName, filename,
-				     PATH_MAX, conn->local_nls, 0);
+				     min(filename_len, PATH_MAX),
+				     conn->local_nls, 0);
 	conv_len *= 2;
 	file_info->FileNameLength = cpu_to_le32(conv_len);
 	rsp->OutputBufferLength =

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

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

From: Werner Kasselman <[email protected]>

commit 309b44ed684496ed3f9c5715d10b899338623512 upstream.

smb2_lock() has three error handling issues after list_del() detaches smb_lock from lock_list at no_check_cl:

1) If vfs_lock_file() returns an unexpected error in the non-UNLOCK path, goto out leaks smb_lock and its flock because the out:
handler only iterates lock_list and rollback_list, neither of which contains the detached smb_lock.

2) If vfs_lock_file() returns -ENOENT in the UNLOCK path, goto out leaks smb_lock and flock for the same reason. The error code returned to the dispatcher is also stale.

3) In the rollback path, smb_flock_init() can return NULL on allocation failure. The result is dereferenced unconditionally, causing a kernel NULL pointer dereference. Add a NULL check to prevent the crash and clean up the bookkeeping; the VFS lock itself cannot be rolled back without the allocation and will be released at file or connection teardown.

Fix cases 1 and 2 by hoisting the locks_free_lock()/kfree() to before the if(!rc) check in the UNLOCK branch so all exit paths share one free site, and by freeing smb_lock and flock before goto out in the non-UNLOCK branch. Propagate the correct error code in both cases. Fix case 3 by wrapping the VFS unlock in an if(rlock) guard and adding a NULL check for locks_free_lock(rlock) in the shared cleanup.

Found via call-graph analysis using sqry.

---
 fs/smb/server/smb2pdu.c |   29 +++++++++++++++++++----------
 1 file changed, 19 insertions(+), 10 deletions(-)

@@ -7552,14 +7552,15 @@ retry:
 		rc = vfs_lock_file(filp, smb_lock->cmd, flock, NULL);
 skip:
 		if (smb_lock->flags & SMB2_LOCKFLAG_UNLOCK) {
+			locks_free_lock(flock);
+			kfree(smb_lock);
 			if (!rc) {
 				ksmbd_debug(SMB, "File unlocked\n");
 			} else if (rc == -ENOENT) {
 				rsp->hdr.Status = STATUS_NOT_LOCKED;
+				err = rc;
 				goto out;
 			}
-			locks_free_lock(flock);
-			kfree(smb_lock);
 		} else {
 			if (rc == FILE_LOCK_DEFERRED) {
 				void **argv;

@@ -7628,6 +7629,9 @@ skip:
 				spin_unlock(&work->conn->llist_lock);
 				ksmbd_debug(SMB, "successful in taking lock\n");
 			} else {
+				locks_free_lock(flock);
+				kfree(smb_lock);
+				err = rc;
 				goto out;
 			}
 		}

@@ -7658,13 +7662,17 @@ out:
 		struct file_lock *rlock = NULL;
 
 		rlock = smb_flock_init(filp);
-		rlock->c.flc_type = F_UNLCK;
-		rlock->fl_start = smb_lock->start;
-		rlock->fl_end = smb_lock->end;
-
-		rc = vfs_lock_file(filp, F_SETLK, rlock, NULL);
-		if (rc)
-			pr_err("rollback unlock fail : %d\n", rc);
+		if (rlock) {
+			rlock->c.flc_type = F_UNLCK;
+			rlock->fl_start = smb_lock->start;
+			rlock->fl_end = smb_lock->end;
+
+			rc = vfs_lock_file(filp, F_SETLK, rlock, NULL);
+			if (rc)
+				pr_err("rollback unlock fail : %d\n", rc);
+		} else {
+			pr_err("rollback unlock alloc failed\n");
+		}
 
 		list_del(&smb_lock->llist);
 		spin_lock(&work->conn->llist_lock);

@@ -7674,7 +7682,8 @@ out:
 		spin_unlock(&work->conn->llist_lock);
 
 		locks_free_lock(smb_lock->fl);
-		locks_free_lock(rlock);
+		if (rlock)
+			locks_free_lock(rlock);
 		kfree(smb_lock);
 	}
 out2:

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

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

From: Hyunwoo Kim <[email protected]>

commit 9bbb19d21ded7d78645506f20d8c44895e3d0fb9 upstream.

When a multichannel session binding request fails (e.g. wrong password), the error path unconditionally sets sess->state = SMB2_SESSION_EXPIRED. However, during binding, sess points to the target session looked up via ksmbd_session_lookup_slowpath() -- which belongs to another connection's user. This allows a remote attacker to invalidate any active session by simply sending a binding request with a wrong password (DoS).

Fix this by skipping session expiration when the failed request was a binding attempt, since the session does not belong to the current connection. The reference taken by ksmbd_session_lookup_slowpath() is still correctly released via ksmbd_user_session_put().

---
 fs/smb/server/smb2pdu.c |   10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)

@@ -1945,8 +1945,14 @@ out_err:
 			if (sess->user && sess->user->flags & KSMBD_USER_FLAG_DELAY_SESSION)
 				try_delay = true;
 
-			sess->last_active = jiffies;
-			sess->state = SMB2_SESSION_EXPIRED;
+			/*
+			 * For binding requests, session belongs to another
+			 * connection. Do not expire it.
+			 */
+			if (!(req->Flags & SMB2_SESSION_REQ_FLAG_BINDING)) {
+				sess->last_active = jiffies;
+				sess->state = SMB2_SESSION_EXPIRED;
+			}
 			ksmbd_user_session_put(sess);
 			work->sess = NULL;
 			if (try_delay) {

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

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

From: Alexey Nepomnyashih <[email protected]>

commit bb120ad57def62e3f23e3d999c5fbed11f610993 upstream.

Similar to commit d8dc8720468a ("ALSA: firewire-lib: fix uninitialized local variable"), the local variable curr_cycle_time in process_rx_packets() is declared without initialization.

When the tracepoint event is not probed, the variable may appear to be used without being initialized. In practice the value is only relevant when the tracepoint is enabled, however initializing it avoids potential use of an uninitialized value and improves code safety.

Initialize curr_cycle_time to zero.

---
 sound/firewire/amdtp-stream.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -1179,7 +1179,7 @@ static void process_rx_packets(struct fw
 	struct pkt_desc *desc = s->packet_descs_cursor;
 	unsigned int pkt_header_length;
 	unsigned int packets;
-	u32 curr_cycle_time;
+	u32 curr_cycle_time = 0;
 	bool need_hw_irq;
 	int i;
 

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

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

From: Peter Ujfalusi <[email protected]>

commit d40a198e2b7821197c5c77b89d0130cc90f400f5 upstream.

It is unexpected, but allowed to have no initial payload for a bytes control and the code is prepared to handle this case, but the size check missed this corner case.

Update the check for minimal size to allow the initial size to be 0.

---
 sound/soc/sof/ipc4-topology.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -2527,7 +2527,7 @@ static int sof_ipc4_control_load_bytes(s
 		return -EINVAL;
 	}
 
-	if (scontrol->priv_size < sizeof(struct sof_abi_hdr)) {
+	if (scontrol->priv_size && scontrol->priv_size < sizeof(struct sof_abi_hdr)) {
 		dev_err(sdev->dev,
 			"bytes control %s initial data size %zu is insufficient.\n",
 			scontrol->name, scontrol->priv_size);

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

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

From: Ali Norouzi <[email protected]>

commit b9c310d72783cc2f30d103eed83920a5a29c671a upstream.

cgw_csum_crc8_rel() correctly computes bounds-safe indices via calc_idx():

    int from = calc_idx(crc8->from_idx, cf->len);
    int to   = calc_idx(crc8->to_idx,   cf->len);
    int res  = calc_idx(crc8->result_idx, cf->len);

    if (from < 0 || to < 0 || res < 0)
        return;

However, the loop and the result write then use the raw s8 fields directly instead of the computed variables:

    for (i = crc8->from_idx; ...)        /* BUG: raw negative index */
    cf->data[crc8->result_idx] = ...;    /* BUG: raw negative index */

With from_idx = to_idx = result_idx = -64 on a 64-byte CAN FD frame, calc_idx(-64, 64) = 0 so the guard passes, but the loop iterates with i = -64, reading cf->data[-64], and the write goes to cf->data[-64]. This write might end up to 56 (7.0-rc) or 40 (<= 6.19) bytes before the start of the canfd_frame on the heap.

The companion function cgw_csum_xor_rel() uses from/to/res correctly throughout; fix cgw_csum_crc8_rel() to match.

Confirmed with KASAN on linux-7.0-rc2:

  BUG: KASAN: slab-out-of-bounds in cgw_csum_crc8_rel+0x515/0x5b0
  Read of size 1 at addr ffff8880076619c8 by task poc_cgw_oob/62

To configure the can-gw crc8 checksums CAP_NET_ADMIN is needed.

---
 net/can/gw.c |    6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

@@ -374,10 +374,10 @@ static void cgw_csum_crc8_rel(struct can
 		return;
 
 	if (from <= to) {
-		for (i = crc8->from_idx; i <= crc8->to_idx; i++)
+		for (i = from; i <= to; i++)
 			crc = crc8->crctab[crc ^ cf->data[i]];
 	} else {
-		for (i = crc8->from_idx; i >= crc8->to_idx; i--)
+		for (i = from; i >= to; i--)
 			crc = crc8->crctab[crc ^ cf->data[i]];
 	}
 

@@ -396,7 +396,7 @@ static void cgw_csum_crc8_rel(struct can
 		break;
 	}
 
-	cf->data[crc8->result_idx] = crc ^ crc8->final_xor_val;
+	cf->data[res] = crc ^ crc8->final_xor_val;
 }
 
 static void cgw_csum_crc8_pos(struct canfd_frame *cf,

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

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

From: Oliver Hartkopp <[email protected]>

commit 424e95d62110cdbc8fd12b40918f37e408e35a92 upstream.

isotp_sendmsg() uses only cmpxchg() on so->tx.state to serialize access to so->tx.buf. isotp_release() waits for ISOTP_IDLE via wait_event_interruptible() and then calls kfree(so->tx.buf).

If a signal interrupts the wait_event_interruptible() inside close() while tx.state is ISOTP_SENDING, the loop exits early and release proceeds to force ISOTP_SHUTDOWN and continues to kfree(so->tx.buf) while sendmsg may still be reading so->tx.buf for the final CAN frame in isotp_fill_dataframe().

The so->tx.buf can be allocated once when the standard tx.buf length needs to be extended. Move the kfree() of this potentially extended tx.buf to sk_destruct time when either isotp_sendmsg() and isotp_release() are done.

---
 net/can/isotp.c |   24 ++++++++++++++++++------
 1 file changed, 18 insertions(+), 6 deletions(-)

@@ -1229,12 +1229,6 @@ static int isotp_release(struct socket *
 	so->ifindex = 0;
 	so->bound = 0;
 
-	if (so->rx.buf != so->rx.sbuf)
-		kfree(so->rx.buf);
-
-	if (so->tx.buf != so->tx.sbuf)
-		kfree(so->tx.buf);
-
 	sock_orphan(sk);
 	sock->sk = NULL;
 

@@ -1602,6 +1596,21 @@ static int isotp_notifier(struct notifie
 	return NOTIFY_DONE;
 }
 
+static void isotp_sock_destruct(struct sock *sk)
+{
+	struct isotp_sock *so = isotp_sk(sk);
+
+	/* do the standard CAN sock destruct work */
+	can_sock_destruct(sk);
+
+	/* free potential extended PDU buffers */
+	if (so->rx.buf != so->rx.sbuf)
+		kfree(so->rx.buf);
+
+	if (so->tx.buf != so->tx.sbuf)
+		kfree(so->tx.buf);
+}
+
 static int isotp_init(struct sock *sk)
 {
 	struct isotp_sock *so = isotp_sk(sk);

@@ -1648,6 +1657,9 @@ static int isotp_init(struct sock *sk)
 	list_add_tail(&so->notifier, &isotp_notifier_list);
 	spin_unlock(&isotp_notifier_lock);
 
+	/* re-assign default can_sock_destruct() reference */
+	sk->sk_destruct = isotp_sock_destruct;
+
 	return 0;
 }
 

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

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

From: Viresh Kumar <[email protected]>

commit 6a28fb8cb28b9eb39a392e531d938a889eacafc5 upstream.

A recently reported issue highlighted that the cached requested_freq is not guaranteed to stay in sync with policy->cur. If the platform changes the actual CPU frequency after the governor sets one (e.g. due to platform-specific frequency scaling) and a re-sync occurs later, policy->cur may diverge from requested_freq.

This can lead to incorrect behavior in the conservative governor. For example, the governor may assume the CPU is already running at the maximum frequency and skip further increases even though there is still headroom.

Avoid this by resetting the cached requested_freq to policy->cur on detecting a change in policy limits.

---
 drivers/cpufreq/cpufreq_conservative.c |   12 ++++++++++++
 drivers/cpufreq/cpufreq_governor.c     |    3 +++
 drivers/cpufreq/cpufreq_governor.h     |    1 +
 3 files changed, 16 insertions(+)

@@ -313,6 +313,17 @@ static void cs_start(struct cpufreq_poli
 	dbs_info->requested_freq = policy->cur;
 }
 
+static void cs_limits(struct cpufreq_policy *policy)
+{
+	struct cs_policy_dbs_info *dbs_info = to_dbs_info(policy->governor_data);
+
+	/*
+	 * The limits have changed, so may have the current frequency. Reset
+	 * requested_freq to avoid any unintended outcomes due to the mismatch.
+	 */
+	dbs_info->requested_freq = policy->cur;
+}
+
 static struct dbs_governor cs_governor = {
 	.gov = CPUFREQ_DBS_GOVERNOR_INITIALIZER("conservative"),
 	.kobj_type = { .default_groups = cs_groups },

@@ -322,6 +333,7 @@ static struct dbs_governor cs_governor =
 	.init = cs_init,
 	.exit = cs_exit,
 	.start = cs_start,
+	.limits = cs_limits,
 };
 
 #define CPU_FREQ_GOV_CONSERVATIVE	(cs_governor.gov)

@@ -563,6 +563,7 @@ EXPORT_SYMBOL_GPL(cpufreq_dbs_governor_s
 
 void cpufreq_dbs_governor_limits(struct cpufreq_policy *policy)
 {
+	struct dbs_governor *gov = dbs_governor_of(policy);
 	struct policy_dbs_info *policy_dbs;
 
 	/* Protect gov->gdbs_data against cpufreq_dbs_governor_exit() */

@@ -574,6 +575,8 @@ void cpufreq_dbs_governor_limits(struct
 	mutex_lock(&policy_dbs->update_mutex);
 	cpufreq_policy_apply_limits(policy);
 	gov_update_sample_delay(policy_dbs, 0);
+	if (gov->limits)
+		gov->limits(policy);
 	mutex_unlock(&policy_dbs->update_mutex);
 
 out:

@@ -138,6 +138,7 @@ struct dbs_governor {
 	int (*init)(struct dbs_data *dbs_data);
 	void (*exit)(struct dbs_data *dbs_data);
 	void (*start)(struct cpufreq_policy *policy);
+	void (*limits)(struct cpufreq_policy *policy);
 };
 
 static inline struct dbs_governor *dbs_governor_of(struct cpufreq_policy *policy)

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

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

From: Srinivas Pandruvada <[email protected]>

commit fbddf68d7b4e1e6da7a78dd7fbd8ec376536584a upstream.

SST-PP locked bit width is set to three bits. It should be only one bit. Use SST_PP_LOCK_WIDTH define instead of SST_PP_LEVEL_WIDTH.

---
 drivers/platform/x86/intel/speed_select_if/isst_tpmi_core.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -866,7 +866,7 @@ static int isst_if_get_perf_level(void _
 	_read_pp_info("current_level", perf_level.current_level, SST_PP_STATUS_OFFSET,
 		      SST_PP_LEVEL_START, SST_PP_LEVEL_WIDTH, SST_MUL_FACTOR_NONE)
 	_read_pp_info("locked", perf_level.locked, SST_PP_STATUS_OFFSET,
-		      SST_PP_LOCK_START, SST_PP_LEVEL_WIDTH, SST_MUL_FACTOR_NONE)
+		      SST_PP_LOCK_START, SST_PP_LOCK_WIDTH, SST_MUL_FACTOR_NONE)
 	_read_pp_info("feature_state", perf_level.feature_state, SST_PP_STATUS_OFFSET,
 		      SST_PP_FEATURE_STATE_START, SST_PP_FEATURE_STATE_WIDTH, SST_MUL_FACTOR_NONE)
 	perf_level.enabled = !!(power_domain_info->sst_header.cap_mask & BIT(1));

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

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

From: Marc Zyngier <[email protected]>

commit 1744a6ef48b9a48f017e3e1a0d05de0a6978396e upstream.

Our vcpu reset suffers from a particularly interesting flaw, as it does not correctly deal with state that will have an effect on the execution flow out of reset.

Take the following completely random example, never seen in the wild and that never resulted in a couple of sleepless nights: /s

- vcpu-A issues a PSCI_CPU_OFF using the SMC conduit

- SMC being a trapped instruction (as opposed to HVC which is always normally executed), we annotate the vcpu as needing to skip the next instruction, which is the SMC itself

- vcpu-A is now safely off

- vcpu-B issues a PSCI_CPU_ON for vcpu-A, providing a starting PC

- vcpu-A gets reset, get the new PC, and is sent on its merry way

- right at the point of entering the guest, we notice that a PC increment is pending (remember the earlier SMC?)

- vcpu-A skips its first instruction...

What could possibly go wrong?

Well, I'm glad you asked. For pKVM as a NV guest, that first instruction is extremely significant, as it indicates whether the CPU is booting or resuming. Having skipped that instruction, nothing makes any sense anymore, and CPU hotplugging fails.

This is all caused by the decoupling of PC update from the handling of an exception that triggers such update, making it non-obvious what affects what when.

Fix this train wreck by discarding all the PC-affecting state on vcpu reset.

---
 arch/arm64/kvm/reset.c |   14 ++++++++++++++
 1 file changed, 14 insertions(+)

@@ -257,6 +257,20 @@ void kvm_reset_vcpu(struct kvm_vcpu *vcp
 			kvm_vcpu_set_be(vcpu);
 
 		*vcpu_pc(vcpu) = target_pc;
+
+		/*
+		 * We may come from a state where either a PC update was
+		 * pending (SMC call resulting in PC being increpented to
+		 * skip the SMC) or a pending exception. Make sure we get
+		 * rid of all that, as this cannot be valid out of reset.
+		 *
+		 * Note that clearing the exception mask also clears PC
+		 * updates, but that's an implementation detail, and we
+		 * really want to make it explicit.
+		 */
+		vcpu_clear_flag(vcpu, PENDING_EXCEPTION);
+		vcpu_clear_flag(vcpu, EXCEPT_MASK);
+		vcpu_clear_flag(vcpu, INCREMENT_PC);
 		vcpu_set_reg(vcpu, 0, reset_state.r0);
 	}
 

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

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

From: Sanman Pradhan <[email protected]>

commit 3075a3951f7708da5a8ab47b0b7d068a32f69e58 upstream.

The custom avs0_enable and avs1_enable sysfs attributes access PMBus registers through the exported API helpers (pmbus_read_byte_data, pmbus_read_word_data, pmbus_write_word_data, pmbus_update_byte_data) without holding the PMBus update_lock mutex. These exported helpers do not acquire the mutex internally, unlike the core's internal callers which hold the lock before invoking them.

The store callback is especially vulnerable: it performs a multi-step read-modify-write sequence (read VOUT_COMMAND, write VOUT_COMMAND, then update OPERATION) where concurrent access from another thread could interleave and corrupt the register state.

Add pmbus_lock_interruptible()/pmbus_unlock() around both the show and store callbacks to serialize PMBus register access with the rest of the driver.

---
 drivers/hwmon/pmbus/isl68137.c |   21 ++++++++++++++++++---
 1 file changed, 18 insertions(+), 3 deletions(-)

@@ -78,7 +78,15 @@ static ssize_t isl68137_avs_enable_show_
 					     int page,
 					     char *buf)
 {
-	int val = pmbus_read_byte_data(client, page, PMBUS_OPERATION);
+	int val;
+
+	val = pmbus_lock_interruptible(client);
+	if (val)
+		return val;
+
+	val = pmbus_read_byte_data(client, page, PMBUS_OPERATION);
+
+	pmbus_unlock(client);
 
 	if (val < 0)
 		return val;

@@ -100,6 +108,10 @@ static ssize_t isl68137_avs_enable_store
 
 	op_val = result ? ISL68137_VOUT_AVS : 0;
 
+	rc = pmbus_lock_interruptible(client);
+	if (rc)
+		return rc;
+
 	/*
 	 * Writes to VOUT setpoint over AVSBus will persist after the VRM is
 	 * switched to PMBus control. Switching back to AVSBus control

@@ -111,17 +123,20 @@ static ssize_t isl68137_avs_enable_store
 		rc = pmbus_read_word_data(client, page, 0xff,
 					  PMBUS_VOUT_COMMAND);
 		if (rc < 0)
-			return rc;
+			goto unlock;
 
 		rc = pmbus_write_word_data(client, page, PMBUS_VOUT_COMMAND,
 					   rc);
 		if (rc < 0)
-			return rc;
+			goto unlock;
 	}
 
 	rc = pmbus_update_byte_data(client, page, PMBUS_OPERATION,
 				    ISL68137_VOUT_AVS, op_val);
 
+unlock:
+	pmbus_unlock(client);
+
 	return (rc < 0) ? rc : count;
 }
 

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

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

From: Sanman Pradhan <[email protected]>

commit 0adc752b4f7d82af7bd14f7cad3091b3b5d702ba upstream.

The hwmon sysfs ABI expects tempN_crit_hyst to report the temperature at which the critical condition clears, not the hysteresis delta from the critical limit.

The peci cputemp driver currently returns tjmax - tcontrol for crit_hyst_type, which is the hysteresis margin rather than the corresponding absolute temperature.

Return tcontrol directly, and update the documentation accordingly.

---
 Documentation/hwmon/peci-cputemp.rst |   10 ++++++----
 drivers/hwmon/peci/cputemp.c         |    2 +-
 2 files changed, 7 insertions(+), 5 deletions(-)

@@ -51,8 +51,9 @@ temp1_max		Provides thermal control temp
 temp1_crit		Provides shutdown temperature of the CPU package which
 			is also known as the maximum processor junction
 			temperature, Tjmax or Tprochot.
-temp1_crit_hyst		Provides the hysteresis value from Tcontrol to Tjmax of
-			the CPU package.
+temp1_crit_hyst		Provides the hysteresis temperature of the CPU
+			package. Returns Tcontrol, the temperature at which
+			the critical condition clears.
 
 temp2_label		"DTS"
 temp2_input		Provides current temperature of the CPU package scaled

@@ -62,8 +63,9 @@ temp2_max		Provides thermal control temp
 temp2_crit		Provides shutdown temperature of the CPU package which
 			is also known as the maximum processor junction
 			temperature, Tjmax or Tprochot.
-temp2_crit_hyst		Provides the hysteresis value from Tcontrol to Tjmax of
-			the CPU package.
+temp2_crit_hyst		Provides the hysteresis temperature of the CPU
+			package. Returns Tcontrol, the temperature at which
+			the critical condition clears.
 
 temp3_label		"Tcontrol"
 temp3_input		Provides current Tcontrol temperature of the CPU

@@ -133,7 +133,7 @@ static int get_temp_target(struct peci_c
 		*val = priv->temp.target.tjmax;
 		break;
 	case crit_hyst_type:
-		*val = priv->temp.target.tjmax - priv->temp.target.tcontrol;
+		*val = priv->temp.target.tcontrol;
 		break;
 	default:
 		ret = -EOPNOTSUPP;

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

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

From: Sanman Pradhan <[email protected]>

commit b0c9d8ae71509f25690d57f2efddebf7f4b12194 upstream.

cputemp_is_visible() validates the channel index against CPUTEMP_CHANNEL_NUMS, but currently uses '>' instead of '>='. As a result, channel == CPUTEMP_CHANNEL_NUMS is not rejected even though valid indices are 0 .. CPUTEMP_CHANNEL_NUMS - 1.

Fix the bounds check by using '>=' so invalid channel indices are rejected before indexing the core bitmap.

---
 drivers/hwmon/peci/cputemp.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -339,7 +339,7 @@ static umode_t cputemp_is_visible(const
 {
 	const struct peci_cputemp *priv = data;
 
-	if (channel > CPUTEMP_CHANNEL_NUMS)
+	if (channel >= CPUTEMP_CHANNEL_NUMS)
 		return 0;
 
 	if (channel < channel_core)

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

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

From: Yuchan Nam <[email protected]>

commit bef4f4a88b73e4cc550d25f665b8a9952af22773 upstream.

MEDIA_REQUEST_IOC_REINIT can run concurrently with VIDIOC_REQBUFS(0) queue teardown paths. This can race request object cleanup against vb2 queue cancellation and lead to use-after-free reports.

We already serialize request queueing against STREAMON/OFF with req_queue_mutex. Extend that serialization to REQBUFS, and also take the same mutex in media_request_ioctl_reinit() so REINIT is in the same exclusion domain.

This keeps request cleanup and queue cancellation from running in parallel for request-capable devices.

---
 drivers/media/mc/mc-request.c        |    5 +++++
 drivers/media/v4l2-core/v4l2-ioctl.c |    5 +++--
 2 files changed, 8 insertions(+), 2 deletions(-)

@@ -190,6 +190,8 @@ static long media_request_ioctl_reinit(s
 	struct media_device *mdev = req->mdev;
 	unsigned long flags;
 
+	mutex_lock(&mdev->req_queue_mutex);
+
 	spin_lock_irqsave(&req->lock, flags);
 	if (req->state != MEDIA_REQUEST_STATE_IDLE &&
 	    req->state != MEDIA_REQUEST_STATE_COMPLETE) {

@@ -197,6 +199,7 @@ static long media_request_ioctl_reinit(s
 			"request: %s not in idle or complete state, cannot reinit\n",
 			req->debug_str);
 		spin_unlock_irqrestore(&req->lock, flags);
+		mutex_unlock(&mdev->req_queue_mutex);
 		return -EBUSY;
 	}
 	if (req->access_count) {

@@ -204,6 +207,7 @@ static long media_request_ioctl_reinit(s
 			"request: %s is being accessed, cannot reinit\n",
 			req->debug_str);
 		spin_unlock_irqrestore(&req->lock, flags);
+		mutex_unlock(&mdev->req_queue_mutex);
 		return -EBUSY;
 	}
 	req->state = MEDIA_REQUEST_STATE_CLEANING;

@@ -214,6 +218,7 @@ static long media_request_ioctl_reinit(s
 	spin_lock_irqsave(&req->lock, flags);
 	req->state = MEDIA_REQUEST_STATE_IDLE;
 	spin_unlock_irqrestore(&req->lock, flags);
+	mutex_unlock(&mdev->req_queue_mutex);
 
 	return 0;
 }

@@ -3069,13 +3069,14 @@ static long __video_do_ioctl(struct file
 		vfh = file->private_data;
 
 	/*
-	 * We need to serialize streamon/off with queueing new requests.
+	 * We need to serialize streamon/off/reqbufs with queueing new requests.
 	 * These ioctls may trigger the cancellation of a streaming
 	 * operation, and that should not be mixed with queueing a new
 	 * request at the same time.
 	 */
 	if (v4l2_device_supports_requests(vfd->v4l2_dev) &&
-	    (cmd == VIDIOC_STREAMON || cmd == VIDIOC_STREAMOFF)) {
+	    (cmd == VIDIOC_STREAMON || cmd == VIDIOC_STREAMOFF ||
+	     cmd == VIDIOC_REQBUFS)) {
 		req_queue_lock = &vfd->v4l2_dev->mdev->req_queue_mutex;
 
 		if (mutex_lock_interruptible(req_queue_lock))

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

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

From: xietangxin <[email protected]>

commit ba8bda9a0896746053aa97ac6c3e08168729172c upstream.

A UAF issue occurs when the virtio_net driver is configured with napi_tx=N and the device's IFF_XMIT_DST_RELEASE flag is cleared
(e.g., during the configuration of tc route filter rules).

When IFF_XMIT_DST_RELEASE is removed from the net_device, the network stack expects the driver to hold the reference to skb->dst until the packet is fully transmitted and freed. In virtio_net with napi_tx=N, skbs may remain in the virtio transmit ring for an extended period.

If the network namespace is destroyed while these skbs are still pending, the corresponding dst_ops structure has freed. When a subsequent packet is transmitted, free_old_xmit() is triggered to clean up old skbs. It then calls dst_release() on the skb associated with the stale dst_entry. Since the dst_ops (referenced by the dst_entry) has already been freed, a UAF kernel paging request occurs.

fix it by adds skb_dst_drop(skb) in start_xmit to explicitly release the dst reference before the skb is queued in virtio_net.

Call Trace:
 Unable to handle kernel paging request at virtual address ffff80007e150000
 CPU: 2 UID: 0 PID: 6236 Comm: ping Kdump: loaded Not tainted 7.0.0-rc1+ #6 PREEMPT
  ...
  percpu_counter_add_batch+0x3c/0x158 lib/percpu_counter.c:98 (P)
  dst_release+0xe0/0x110  net/core/dst.c:177
  skb_release_head_state+0xe8/0x108 net/core/skbuff.c:1177
  sk_skb_reason_drop+0x54/0x2d8 net/core/skbuff.c:1255
  dev_kfree_skb_any_reason+0x64/0x78 net/core/dev.c:3469
  napi_consume_skb+0x1c4/0x3a0 net/core/skbuff.c:1527
  __free_old_xmit+0x164/0x230  drivers/net/virtio_net.c:611 [virtio_net]
  free_old_xmit drivers/net/virtio_net.c:1081 [virtio_net]
  start_xmit+0x7c/0x530 drivers/net/virtio_net.c:3329 [virtio_net]
  ...

Reproduction Steps:
NETDEV="enp3s0"

config_qdisc_route_filter() {
tc qdisc del dev $NETDEV root
tc qdisc add dev $NETDEV root handle 1: prio tc filter add dev $NETDEV parent 1:0 \
protocol ip prio 100 route to 100 flowid 1:1
ip route add 192.168.1.100/32 dev $NETDEV realm 100
}

test_ns() {
ip netns add testns
ip link set $NETDEV netns testns
ip netns exec testns ifconfig $NETDEV 10.0.32.46/24 ip netns exec testns ping -c 1 10.0.32.1
ip netns del testns
}

config_qdisc_route_filter

test_ns
sleep 2
test_ns

---
 drivers/net/virtio_net.c |    1 +
 1 file changed, 1 insertion(+)

@@ -3178,6 +3178,7 @@ static netdev_tx_t start_xmit(struct sk_
 	/* Don't wait up for transmitted skbs to be freed. */
 	if (!use_napi) {
 		skb_orphan(skb);
+		skb_dst_drop(skb);
 		nf_reset_ct(skb);
 	}
 

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

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

From: Jiucheng Xu <[email protected]>

commit c23df30915f83e7257c8625b690a1cece94142a0 upstream.

The bio completion path in the process context (e.g. dm-verity) will directly call into decompression rather than trigger another workqueue context for minimal scheduling latencies, which can then call vm_map_ram() with GFP_KERNEL.

Due to insufficient memory, vm_map_ram() may generate memory swapping I/O, which can cause submit_bio_wait to deadlock in some scenarios.

Trimmed down the call stack, as follows:

f2fs_submit_read_io
submit_bio //bio_list is initialized.
mmc_blk_mq_recovery
z_erofs_endio
vm_map_ram
__pte_alloc_kernel
__alloc_pages_direct_reclaim
shrink_folio_list
__swap_writepage
submit_bio_wait //bio_list is non-NULL, hang!!!

Use memalloc_noio_{save,restore}() to wrap up this path.

---
 fs/erofs/zdata.c |    3 +++
 1 file changed, 3 insertions(+)

@@ -1425,6 +1425,7 @@ static void z_erofs_decompress_kickoff(s
 				       int bios)
 {
 	struct erofs_sb_info *const sbi = EROFS_SB(io->sb);
+	int gfp_flag;
 
 	/* wake up the caller thread for sync decompression */
 	if (io->sync) {

@@ -1457,7 +1458,9 @@ static void z_erofs_decompress_kickoff(s
 			sbi->opt.sync_decompress = EROFS_SYNC_DECOMPRESS_FORCE_ON;
 		return;
 	}
+	gfp_flag = memalloc_noio_save();
 	z_erofs_decompressqueue_work(&io->u.work);
+	memalloc_noio_restore(gfp_flag);
 }
 
 static void z_erofs_fill_bio_vec(struct bio_vec *bvec,

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

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

From: Zhan Xusheng <[email protected]>

commit 5d16467ae56343b9205caedf85e3a131e0914ad8 upstream.

alarm_timer_forward() passes arguments to alarm_forward() in the wrong order:

  alarm_forward(alarm, timr->it_interval, now);

However, alarm_forward() is defined as:

  u64 alarm_forward(struct alarm *alarm, ktime_t now, ktime_t interval);

and uses the second argument as the current time:

  delta = ktime_sub(now, alarm->node.expires);

Passing the interval as "now" results in incorrect delta computation, which can lead to missed expirations or incorrect overrun accounting.

This issue has been present since the introduction of alarm_timer_forward().

Fix this by swapping the arguments.

---
 kernel/time/alarmtimer.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -614,7 +614,7 @@ static s64 alarm_timer_forward(struct k_
 {
 	struct alarm *alarm = &timr->it.alarm.alarmtimer;
 
-	return alarm_forward(alarm, timr->it_interval, now);
+	return alarm_forward(alarm, now, timr->it_interval);
 }
 
 /**

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

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

From: Nikunj A Dadhania <[email protected]>

commit 05243d490bb7852a8acca7b5b5658019c7797a52 upstream.

Move FSGSBASE enablement from identify_cpu() to cpu_init_exception_handling() to ensure it is enabled before any exceptions can occur on both boot and secondary CPUs.

== Background ==

Exception entry code (paranoid_entry()) uses ALTERNATIVE patching based on X86_FEATURE_FSGSBASE to decide whether to use RDGSBASE/WRGSBASE instructions or the slower RDMSR/SWAPGS sequence for saving/restoring GSBASE.

On boot CPU, ALTERNATIVE patching happens after enabling FSGSBASE in CR4. When the feature is available, the code is permanently patched to use RDGSBASE/WRGSBASE, which require CR4.FSGSBASE=1 to execute without triggering

== Boot Sequence ==

Boot CPU (with CR pinning enabled):
trap_init()
cpu_init() <- Uses unpatched code (RDMSR/SWAPGS)
x2apic_setup()
...
arch_cpu_finalize_init()
identify_boot_cpu()
identify_cpu()
cr4_set_bits(X86_CR4_FSGSBASE) # Enables the feature # This becomes part of cr4_pinned_bits
...
alternative_instructions() <- Patches code to use RDGSBASE/WRGSBASE

Secondary CPUs (with CR pinning enabled):
start_secondary()
cr4_init() <- Code already patched, CR4.FSGSBASE=1
set implicitly via cr4_pinned_bits

cpu_init() <- exceptions work because FSGSBASE is
already enabled

Secondary CPU (with CR pinning disabled):
start_secondary()
cr4_init() <- Code already patched, CR4.FSGSBASE=0
cpu_init()
x2apic_setup()
rdmsrq(MSR_IA32_APICBASE) <- Triggers #VC in SNP guests
exc_vmm_communication()
paranoid_entry() <- Uses RDGSBASE with CR4.FSGSBASE=0
(patched code)
...
ap_starting()
identify_secondary_cpu()
identify_cpu()
cr4_set_bits(X86_CR4_FSGSBASE) <- Enables the feature, which is
too late

== CR Pinning ==

Currently, for secondary CPUs, CR4.FSGSBASE is set implicitly through
CR-pinning: the boot CPU sets it during identify_cpu(), it becomes part of cr4_pinned_bits, and cr4_init() applies those pinned bits to secondary CPUs. This works but creates an undocumented dependency between cr4_init() and the pinning mechanism.

== Problem ==

Secondary CPUs boot after alternatives have been applied globally. They execute already-patched paranoid_entry() code that uses RDGSBASE/WRGSBASE instructions, which require CR4.FSGSBASE=1. Upcoming changes to CR pinning behavior will break the implicit dependency, causing secondary CPUs to generate #UD.

This issue manifests itself on AMD SEV-SNP guests, where the rdmsrq() in x2apic_setup() triggers a #VC exception early during cpu_init(). The #VC handler (exc_vmm_communication()) executes the patched paranoid_entry() path. Without CR4.FSGSBASE enabled, RDGSBASE instructions trigger #UD.

== Fix ==

Enable FSGSBASE explicitly in cpu_init_exception_handling() before loading exception handlers. This makes the dependency explicit and ensures both boot and secondary CPUs have FSGSBASE enabled before paranoid_entry() executes.

---
 arch/x86/kernel/cpu/common.c |   18 ++++++++++++------
 1 file changed, 12 insertions(+), 6 deletions(-)

@@ -1938,12 +1938,6 @@ static void identify_cpu(struct cpuinfo_
 	setup_smap(c);
 	setup_umip(c);
 
-	/* Enable FSGSBASE instructions if available. */
-	if (cpu_has(c, X86_FEATURE_FSGSBASE)) {
-		cr4_set_bits(X86_CR4_FSGSBASE);
-		elf_hwcap2 |= HWCAP2_FSGSBASE;
-	}
-
 	/*
 	 * The vendor-specific functions might have changed features.
 	 * Now we do "generic changes."

@@ -2286,6 +2280,18 @@ void cpu_init_exception_handling(bool bo
 	/* GHCB needs to be setup to handle #VC. */
 	setup_ghcb();
 
+	/*
+	 * On CPUs with FSGSBASE support, paranoid_entry() uses
+	 * ALTERNATIVE-patched RDGSBASE/WRGSBASE instructions. Secondary CPUs
+	 * boot after alternatives are patched globally, so early exceptions
+	 * execute patched code that depends on FSGSBASE. Enable the feature
+	 * before any exceptions occur.
+	 */
+	if (cpu_feature_enabled(X86_FEATURE_FSGSBASE)) {
+		cr4_set_bits(X86_CR4_FSGSBASE);
+		elf_hwcap2 |= HWCAP2_FSGSBASE;
+	}
+
 	if (cpu_feature_enabled(X86_FEATURE_FRED)) {
 		/* The boot CPU has enabled FRED during early boot */
 		if (!boot_cpu)

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

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

From: Borislav Petkov (AMD) <[email protected]>

commit 411df123c017169922cc767affce76282b8e6c85 upstream.

Commit in Fixes added the FRED CR4 bit to the CR4 pinned bits mask so that whenever something else modifies CR4, that bit remains set. Which in itself is a perfectly fine idea.

However, there's an issue when during boot FRED is initialized: first on the BSP and later on the APs. Thus, there's a window in time when exceptions cannot be handled.

This becomes particularly nasty when running as SEV-{ES,SNP} or TDX guests which, when they manage to trigger exceptions during that short window described above, triple fault due to FRED MSRs not being set up yet.

See Link tag below for a much more detailed explanation of the situation.

So, as a result, the commit in that Link URL tried to address this shortcoming by temporarily disabling CR4 pinning when an AP is not online yet.

However, that is a problem in itself because in this case, an attack on the kernel needs to only modify the online bit - a single bit in RW memory - and then disable CR4 pinning and then disable SM*P, leading to more and worse things to happen to the system.

So, instead, remove the FRED bit from the CR4 pinning mask, thus obviating the need to temporarily disable CR4 pinning.

If someone manages to disable FRED when poking at CR4, then idt_invalidate() would make sure the system would crash'n'burn on the first exception triggered, which is a much better outcome security-wise.

---
 arch/x86/kernel/cpu/common.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -405,7 +405,7 @@ out:
 
 /* These bits should not change their value after CPU init is finished. */
 static const unsigned long cr4_pinned_mask = X86_CR4_SMEP | X86_CR4_SMAP | X86_CR4_UMIP |
-					     X86_CR4_FSGSBASE | X86_CR4_CET | X86_CR4_FRED;
+					     X86_CR4_FSGSBASE | X86_CR4_CET;
 static DEFINE_STATIC_KEY_FALSE_RO(cr_pinning);
 static unsigned long cr4_pinned_bits __ro_after_init;
 

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

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

From: Abel Vesa <[email protected]>

commit 81af9e40e2e4e1aa95f09fb34811760be6742c58 upstream.

According to internal documentation, on SM8650, when the PHY is configured in Gear 4, the QPHY_V6_PCS_UFS_PLL_CNTL register needs to have the same value as for Gear 5.

At the moment, there is no board that comes with a UFS 3.x device, so this issue doesn't show up, but with the new Eliza SoC, which uses the same init sequence as SM8650, on the MTP board, the link startup fails with the current Gear 4 PCS table.

So fix that by moving the entry into the PCS generic table instead, while keeping the value from Gear 5 configuration.

---
 drivers/phy/qualcomm/phy-qcom-qmp-ufs.c |    3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

@@ -927,6 +927,7 @@ static const struct qmp_phy_init_tbl sm8
 	QMP_PHY_INIT_CFG(QPHY_V6_PCS_UFS_MULTI_LANE_CTRL1, 0x02),
 	QMP_PHY_INIT_CFG(QPHY_V6_PCS_UFS_TX_MID_TERM_CTRL1, 0x43),
 	QMP_PHY_INIT_CFG(QPHY_V6_PCS_UFS_PCS_CTRL1, 0xc1),
+	QMP_PHY_INIT_CFG(QPHY_V6_PCS_UFS_PLL_CNTL, 0x33),
 	QMP_PHY_INIT_CFG(QPHY_V6_PCS_UFS_TX_LARGE_AMP_DRV_LVL, 0x0f),
 	QMP_PHY_INIT_CFG(QPHY_V6_PCS_UFS_RX_SIGDET_CTRL2, 0x68),
 	QMP_PHY_INIT_CFG(QPHY_V6_PCS_UFS_TX_POST_EMP_LVL_S4, 0x0e),

@@ -936,13 +937,11 @@ static const struct qmp_phy_init_tbl sm8
 };
 
 static const struct qmp_phy_init_tbl sm8650_ufsphy_g4_pcs[] = {
-	QMP_PHY_INIT_CFG(QPHY_V6_PCS_UFS_PLL_CNTL, 0x13),
 	QMP_PHY_INIT_CFG(QPHY_V6_PCS_UFS_TX_HSGEAR_CAPABILITY, 0x04),
 	QMP_PHY_INIT_CFG(QPHY_V6_PCS_UFS_RX_HSGEAR_CAPABILITY, 0x04),
 };
 
 static const struct qmp_phy_init_tbl sm8650_ufsphy_g5_pcs[] = {
-	QMP_PHY_INIT_CFG(QPHY_V6_PCS_UFS_PLL_CNTL, 0x33),
 	QMP_PHY_INIT_CFG(QPHY_V6_PCS_UFS_TX_HSGEAR_CAPABILITY, 0x05),
 	QMP_PHY_INIT_CFG(QPHY_V6_PCS_UFS_RX_HSGEAR_CAPABILITY, 0x05),
 	QMP_PHY_INIT_CFG(QPHY_V6_PCS_UFS_RX_HS_G5_SYNC_LENGTH_CAPABILITY, 0x4d),

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

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

From: Amir Goldstein <[email protected]>

commit 53a7c171e9dd833f0a96b545adcb89bd57387239 upstream.

The implicit FILEID_INO32_GEN encoder was changed to be explicit, so we need to fix the detection.

When mounting overlayfs with upperdir and lowerdir on different ext4 filesystems, the expected kmsg log is:

overlayfs: "xino" feature enabled using 32 upper inode bits.

But instead, since the regressing commit, the kmsg log was:

overlayfs: "xino" feature enabled using 2 upper inode bits.

---
 fs/overlayfs/util.c |    5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

@@ -84,7 +84,10 @@ int ovl_can_decode_fh(struct super_block
 	if (!exportfs_can_decode_fh(sb->s_export_op))
 		return 0;
 
-	return sb->s_export_op->encode_fh ? -1 : FILEID_INO32_GEN;
+	if (sb->s_export_op->encode_fh == generic_encode_ino32_fh)
+		return FILEID_INO32_GEN;
+
+	return -1;
 }
 
 struct dentry *ovl_indexdir(struct super_block *sb)

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

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

From: Tyllis Xu <[email protected]>

commit 61d099ac4a7a8fb11ebdb6e2ec8d77f38e77362f upstream.

A malicious or compromised VIO server can return a num_written value in the discover targets MAD response that exceeds max_targets. This value is stored directly in vhost->num_targets without validation, and is then used as the loop bound in ibmvfc_alloc_targets() to index into disc_buf[], which is only allocated for max_targets entries. Indices at or beyond max_targets access kernel memory outside the DMA-coherent allocation. The out-of-bounds data is subsequently embedded in Implicit Logout and PLOGI MADs that are sent back to the VIO server, leaking kernel memory.

Fix by clamping num_written to max_targets before storing it.

---
 drivers/scsi/ibmvscsi/ibmvfc.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

@@ -4963,7 +4963,8 @@ static void ibmvfc_discover_targets_done
 	switch (mad_status) {
 	case IBMVFC_MAD_SUCCESS:
 		ibmvfc_dbg(vhost, "Discover Targets succeeded\n");
-		vhost->num_targets = be32_to_cpu(rsp->num_written);
+		vhost->num_targets = min_t(u32, be32_to_cpu(rsp->num_written),
+					   max_targets);
 		ibmvfc_set_host_action(vhost, IBMVFC_HOST_ACTION_ALLOC_TGTS);
 		break;
 	case IBMVFC_MAD_FAILED:

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

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

From: Greg Kroah-Hartman <[email protected]>

commit 7a9f448d44127217fabc4065c5ba070d4e0b5d37 upstream.

ses_recv_diag() can return a positive value, which also means that an error happened, so do not only test for negative values.

---
 drivers/scsi/ses.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -216,7 +216,7 @@ static unsigned char *ses_get_page2_desc
 	unsigned char *type_ptr = ses_dev->page1_types;
 	unsigned char *desc_ptr = ses_dev->page2 + 8;
 
-	if (ses_recv_diag(sdev, 2, ses_dev->page2, ses_dev->page2_len) < 0)
+	if (ses_recv_diag(sdev, 2, ses_dev->page2, ses_dev->page2_len))
 		return NULL;
 
 	for (i = 0; i < ses_dev->page1_num_types; i++, type_ptr += 4) {

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

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

From: Kevin Hao <[email protected]>

commit 317e49358ebbf6390fa439ef3c142f9239dd25fb upstream.

The devm_free_irq() and devm_request_irq() functions should not be executed in an atomic context.

During device suspend, all userspace processes and most kernel threads are frozen. Additionally, we flush all tx/rx status, disable all macb interrupts, and halt rx operations. Therefore, it is safe to split the region protected by bp->lock into two independent sections, allowing devm_free_irq() and devm_request_irq() to run in a non-atomic context. This modification resolves the following lockdep warning:

  BUG: sleeping function called from invalid context at kernel/locking/mutex.c:591
  in_atomic(): 1, irqs_disabled(): 1, non_block: 0, pid: 501, name: rtcwake
  preempt_count: 1, expected: 0
  RCU nest depth: 1, expected: 0
  7 locks held by rtcwake/501:
   #0: ffff0008038c3408 (sb_writers#5){.+.+}-{0:0}, at: vfs_write+0xf8/0x368
   #1: ffff0008049a5e88 (&of->mutex#2){+.+.}-{4:4}, at: kernfs_fop_write_iter+0xbc/0x1c8
   #2: ffff00080098d588 (kn->active#70){.+.+}-{0:0}, at: kernfs_fop_write_iter+0xcc/0x1c8
   #3: ffff800081c84888 (system_transition_mutex){+.+.}-{4:4}, at: pm_suspend+0x1ec/0x290
   #4: ffff0008009ba0f8 (&dev->mutex){....}-{4:4}, at: device_suspend+0x118/0x4f0
   #5: ffff800081d00458 (rcu_read_lock){....}-{1:3}, at: rcu_lock_acquire+0x4/0x48
   #6: ffff0008031fb9e0 (&bp->lock){-.-.}-{3:3}, at: macb_suspend+0x144/0x558
  irq event stamp: 8682
  hardirqs last  enabled at (8681): [<ffff8000813c7d7c>] _raw_spin_unlock_irqrestore+0x44/0x88
  hardirqs last disabled at (8682): [<ffff8000813c7b58>] _raw_spin_lock_irqsave+0x38/0x98
  softirqs last  enabled at (7322): [<ffff8000800f1b4c>] handle_softirqs+0x52c/0x588
  softirqs last disabled at (7317): [<ffff800080010310>] __do_softirq+0x20/0x2c
  CPU: 1 UID: 0 PID: 501 Comm: rtcwake Not tainted 7.0.0-rc3-next-20260310-yocto-standard+ #125 PREEMPT
  Hardware name: ZynqMP ZCU102 Rev1.1 (DT)
  Call trace:
   show_stack+0x24/0x38 (C)
   __dump_stack+0x28/0x38
   dump_stack_lvl+0x64/0x88
   dump_stack+0x18/0x24
   __might_resched+0x200/0x218
   __might_sleep+0x38/0x98
   __mutex_lock_common+0x7c/0x1378
   mutex_lock_nested+0x38/0x50
   free_irq+0x68/0x2b0
   devm_irq_release+0x24/0x38
   devres_release+0x40/0x80
   devm_free_irq+0x48/0x88
   macb_suspend+0x298/0x558
   device_suspend+0x218/0x4f0
   dpm_suspend+0x244/0x3a0
   dpm_suspend_start+0x50/0x78
   suspend_devices_and_enter+0xec/0x560
   pm_suspend+0x194/0x290
   state_store+0x110/0x158
   kobj_attr_store+0x1c/0x30
   sysfs_kf_write+0xa8/0xd0
   kernfs_fop_write_iter+0x11c/0x1c8
   vfs_write+0x248/0x368
   ksys_write+0x7c/0xf8
   __arm64_sys_write+0x28/0x40
   invoke_syscall+0x4c/0xe8
   el0_svc_common+0x98/0xf0
   do_el0_svc+0x28/0x40
   el0_svc+0x54/0x1e0
   el0t_64_sync_handler+0x84/0x130
   el0t_64_sync+0x198/0x1a0

---
 drivers/net/ethernet/cadence/macb_main.c |   12 +++++++-----
 1 file changed, 7 insertions(+), 5 deletions(-)

@@ -5410,6 +5410,7 @@ static int __maybe_unused macb_suspend(s
 			/* write IP address into register */
 			tmp |= MACB_BFEXT(IP, be32_to_cpu(ifa->ifa_local));
 		}
+		spin_unlock_irqrestore(&bp->lock, flags);
 
 		/* Change interrupt handler and
 		 * Enable WoL IRQ on queue 0

@@ -5422,11 +5423,12 @@ static int __maybe_unused macb_suspend(s
 				dev_err(dev,
 					"Unable to request IRQ %d (error %d)\n",
 					bp->queues[0].irq, err);
-				spin_unlock_irqrestore(&bp->lock, flags);
 				return err;
 			}
+			spin_lock_irqsave(&bp->lock, flags);
 			queue_writel(bp->queues, IER, GEM_BIT(WOL));
 			gem_writel(bp, WOL, tmp);
+			spin_unlock_irqrestore(&bp->lock, flags);
 		} else {
 			err = devm_request_irq(dev, bp->queues[0].irq, macb_wol_interrupt,
 					       IRQF_SHARED, netdev->name, bp->queues);

@@ -5434,13 +5436,13 @@ static int __maybe_unused macb_suspend(s
 				dev_err(dev,
 					"Unable to request IRQ %d (error %d)\n",
 					bp->queues[0].irq, err);
-				spin_unlock_irqrestore(&bp->lock, flags);
 				return err;
 			}
+			spin_lock_irqsave(&bp->lock, flags);
 			queue_writel(bp->queues, IER, MACB_BIT(WOL));
 			macb_writel(bp, WOL, tmp);
+			spin_unlock_irqrestore(&bp->lock, flags);
 		}
-		spin_unlock_irqrestore(&bp->lock, flags);
 
 		enable_irq_wake(bp->queues[0].irq);
 	}

@@ -5507,6 +5509,8 @@ static int __maybe_unused macb_resume(st
 		queue_readl(bp->queues, ISR);
 		if (bp->caps & MACB_CAPS_ISR_CLEAR_ON_WRITE)
 			queue_writel(bp->queues, ISR, -1);
+		spin_unlock_irqrestore(&bp->lock, flags);
+
 		/* Replace interrupt handler on queue 0 */
 		devm_free_irq(dev, bp->queues[0].irq, bp->queues);
 		err = devm_request_irq(dev, bp->queues[0].irq, macb_interrupt,

@@ -5515,10 +5519,8 @@ static int __maybe_unused macb_resume(st
 			dev_err(dev,
 				"Unable to request IRQ %d (error %d)\n",
 				bp->queues[0].irq, err);
-			spin_unlock_irqrestore(&bp->lock, flags);
 			return err;
 		}
-		spin_unlock_irqrestore(&bp->lock, flags);
 
 		disable_irq_wake(bp->queues[0].irq);
 

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

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

From: Kevin Hao <[email protected]>

commit baa35a698cea26930679a20a7550bbb4c8319725 upstream.

Access to net_device::ip_ptr and its associated members must be protected by an RCU lock. Since we are modifying this piece of code, let's also move it to execute only when WAKE_ARP is enabled.

To minimize the duration of the RCU lock, a local variable is used to temporarily store the IP address. This change resolves the following
RCU check warning:
WARNING: suspicious RCU usage
7.0.0-rc3-next-20260310-yocto-standard+ #122 Not tainted
-----------------------------
drivers/net/ethernet/cadence/macb_main.c:5944 suspicious rcu_dereference_check() usage!

other info that might help us debug this:

rcu_scheduler_active = 2, debug_locks = 1
5 locks held by rtcwake/518:
#0: ffff000803ab1408 (sb_writers#5){.+.+}-{0:0}, at: vfs_write+0xf8/0x368 #1: ffff0008090bf088 (&of->mutex#2){+.+.}-{4:4}, at: kernfs_fop_write_iter+0xbc/0x1c8 #2: ffff00080098d588 (kn->active#70){.+.+}-{0:0}, at: kernfs_fop_write_iter+0xcc/0x1c8 #3: ffff800081c84888 (system_transition_mutex){+.+.}-{4:4}, at: pm_suspend+0x1ec/0x290 #4: ffff0008009ba0f8 (&dev->mutex){....}-{4:4}, at: device_suspend+0x118/0x4f0

  stack backtrace:
  CPU: 3 UID: 0 PID: 518 Comm: rtcwake Not tainted 7.0.0-rc3-next-20260310-yocto-standard+ #122 PREEMPT
  Hardware name: ZynqMP ZCU102 Rev1.1 (DT)
  Call trace:
   show_stack+0x24/0x38 (C)
   __dump_stack+0x28/0x38
   dump_stack_lvl+0x64/0x88
   dump_stack+0x18/0x24
   lockdep_rcu_suspicious+0x134/0x1d8
   macb_suspend+0xd8/0x4c0
   device_suspend+0x218/0x4f0
   dpm_suspend+0x244/0x3a0
   dpm_suspend_start+0x50/0x78
   suspend_devices_and_enter+0xec/0x560
   pm_suspend+0x194/0x290
   state_store+0x110/0x158
   kobj_attr_store+0x1c/0x30
   sysfs_kf_write+0xa8/0xd0
   kernfs_fop_write_iter+0x11c/0x1c8
   vfs_write+0x248/0x368
   ksys_write+0x7c/0xf8
   __arm64_sys_write+0x28/0x40
   invoke_syscall+0x4c/0xe8
   el0_svc_common+0x98/0xf0
   do_el0_svc+0x28/0x40
   el0_svc+0x54/0x1e0
   el0t_64_sync_handler+0x84/0x130
   el0t_64_sync+0x198/0x1a0

---
 drivers/net/ethernet/cadence/macb_main.c |   25 ++++++++++++++++---------
 1 file changed, 16 insertions(+), 9 deletions(-)

@@ -5352,9 +5352,9 @@ static int __maybe_unused macb_suspend(s
 	struct macb_queue *queue;
 	struct in_device *idev;
 	unsigned long flags;
+	u32 tmp, ifa_local;
 	unsigned int q;
 	int err;
-	u32 tmp;
 
 	if (!device_may_wakeup(&bp->dev->dev))
 		phy_exit(bp->sgmii_phy);

@@ -5363,14 +5363,21 @@ static int __maybe_unused macb_suspend(s
 		return 0;
 
 	if (bp->wol & MACB_WOL_ENABLED) {
-		/* Check for IP address in WOL ARP mode */
-		idev = __in_dev_get_rcu(bp->dev);
-		if (idev)
-			ifa = rcu_dereference(idev->ifa_list);
-		if ((bp->wolopts & WAKE_ARP) && !ifa) {
-			netdev_err(netdev, "IP address not assigned as required by WoL walk ARP\n");
-			return -EOPNOTSUPP;
+		if (bp->wolopts & WAKE_ARP) {
+			/* Check for IP address in WOL ARP mode */
+			rcu_read_lock();
+			idev = __in_dev_get_rcu(bp->dev);
+			if (idev)
+				ifa = rcu_dereference(idev->ifa_list);
+			if (!ifa) {
+				rcu_read_unlock();
+				netdev_err(netdev, "IP address not assigned as required by WoL walk ARP\n");
+				return -EOPNOTSUPP;
+			}
+			ifa_local = be32_to_cpu(ifa->ifa_local);
+			rcu_read_unlock();
 		}
+
 		spin_lock_irqsave(&bp->lock, flags);
 
 		/* Disable Tx and Rx engines before  disabling the queues,

@@ -5408,7 +5415,7 @@ static int __maybe_unused macb_suspend(s
 		if (bp->wolopts & WAKE_ARP) {
 			tmp |= MACB_BIT(ARP);
 			/* write IP address into register */
-			tmp |= MACB_BFEXT(IP, be32_to_cpu(ifa->ifa_local));
+			tmp |= MACB_BFEXT(IP, ifa_local);
 		}
 		spin_unlock_irqrestore(&bp->lock, flags);
 

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

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

From: Kevin Hao <[email protected]>

commit 647b8a2fe474474704110db6bd07f7a139e621eb upstream.

The napi_consume_skb() function is not intended to be called in an IRQ disabled context. However, after commit 6bc8a5098bf4 ("net: macb: Fix tx_ptr_lock locking"), the freeing of TX SKBs is performed with IRQs disabled. To resolve the following call trace, use dev_consume_skb_any() for freeing TX SKBs:

   WARNING: kernel/softirq.c:430 at __local_bh_enable_ip+0x174/0x188, CPU#0: ksoftirqd/0/15
   Modules linked in:
   CPU: 0 UID: 0 PID: 15 Comm: ksoftirqd/0 Not tainted 7.0.0-rc4-next-20260319-yocto-standard-dirty #37 PREEMPT
   Hardware name: ZynqMP ZCU102 Rev1.1 (DT)
   pstate: 200000c5 (nzCv daIF -PAN -UAO -TCO -DIT -SSBS BTYPE=--)
   pc : __local_bh_enable_ip+0x174/0x188
   lr : local_bh_enable+0x24/0x38
   sp : ffff800082b3bb10
   x29: ffff800082b3bb10 x28: ffff0008031f3c00 x27: 000000000011ede0
   x26: ffff000800a7ff00 x25: ffff800083937ce8 x24: 0000000000017a80
   x23: ffff000803243a78 x22: 0000000000000040 x21: 0000000000000000
   x20: ffff000800394c80 x19: 0000000000000200 x18: 0000000000000001
   x17: 0000000000000001 x16: ffff000803240000 x15: 0000000000000000
   x14: ffffffffffffffff x13: 0000000000000028 x12: ffff000800395650
   x11: ffff8000821d1528 x10: ffff800081c2bc08 x9 : ffff800081c1e258
   x8 : 0000000100000301 x7 : ffff8000810426ec x6 : 0000000000000000
   x5 : 0000000000000001 x4 : 0000000000000001 x3 : 0000000000000000
   x2 : 0000000000000008 x1 : 0000000000000200 x0 : ffff8000810428dc
   Call trace:
    __local_bh_enable_ip+0x174/0x188 (P)
    local_bh_enable+0x24/0x38
    skb_attempt_defer_free+0x190/0x1d8
    napi_consume_skb+0x58/0x108
    macb_tx_poll+0x1a4/0x558
    __napi_poll+0x50/0x198
    net_rx_action+0x1f4/0x3d8
    handle_softirqs+0x16c/0x560
    run_ksoftirqd+0x44/0x80
    smpboot_thread_fn+0x1d8/0x338
    kthread+0x120/0x150
    ret_from_fork+0x10/0x20
   irq event stamp: 29751
   hardirqs last  enabled at (29750): [<ffff8000813be184>] _raw_spin_unlock_irqrestore+0x44/0x88
   hardirqs last disabled at (29751): [<ffff8000813bdf60>] _raw_spin_lock_irqsave+0x38/0x98
   softirqs last  enabled at (29150): [<ffff8000800f1aec>] handle_softirqs+0x504/0x560
   softirqs last disabled at (29153): [<ffff8000800f2fec>] run_ksoftirqd+0x44/0x80

---
 drivers/net/ethernet/cadence/macb_main.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -1129,7 +1129,7 @@ static void macb_tx_unmap(struct macb *b
 	}
 
 	if (tx_skb->skb) {
-		napi_consume_skb(tx_skb->skb, budget);
+		dev_consume_skb_any(tx_skb->skb);
 		tx_skb->skb = NULL;
 	}
 }

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

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

From: Sean Christopherson <[email protected]>

commit aad885e774966e97b675dfe928da164214a71605 upstream.

When installing an emulated MMIO SPTE, do so *after* dropping/zapping the existing SPTE (if it's shadow-present). While commit a54aa15c6bda3 was right about it being impossible to convert a shadow-present SPTE to an MMIO SPTE due to a _guest_ write, it failed to account for writes to guest memory that are outside the scope of KVM.

E.g. if host userspace modifies a shadowed gPTE to switch from a memslot to emulted MMIO and then the guest hits a relevant page fault, KVM will install the MMIO SPTE without first zapping the shadow-present SPTE.

  ------------[ cut here ]------------
  is_shadow_present_pte(*sptep)
  WARNING: arch/x86/kvm/mmu/mmu.c:484 at mark_mmio_spte+0xb2/0xc0 [kvm], CPU#0: vmx_ept_stale_r/4292
  Modules linked in: kvm_intel kvm irqbypass
  CPU: 0 UID: 1000 PID: 4292 Comm: vmx_ept_stale_r Not tainted 7.0.0-rc2-eafebd2d2ab0-sink-vm #319 PREEMPT
  Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 0.0.0 02/06/2015
  RIP: 0010:mark_mmio_spte+0xb2/0xc0 [kvm]
  Call Trace:
   <TASK>
   mmu_set_spte+0x237/0x440 [kvm]
   ept_page_fault+0x535/0x7f0 [kvm]
   kvm_mmu_do_page_fault+0xee/0x1f0 [kvm]
   kvm_mmu_page_fault+0x8d/0x620 [kvm]
   vmx_handle_exit+0x18c/0x5a0 [kvm_intel]
   kvm_arch_vcpu_ioctl_run+0xc55/0x1c20 [kvm]
   kvm_vcpu_ioctl+0x2d5/0x980 [kvm]
   __x64_sys_ioctl+0x8a/0xd0
   do_syscall_64+0xb5/0x730
   entry_SYSCALL_64_after_hwframe+0x4b/0x53
  RIP: 0033:0x47fa3f
   </TASK>
  ---[ end trace 0000000000000000 ]---

---
 arch/x86/kvm/mmu/mmu.c |   14 ++++++++------
 1 file changed, 8 insertions(+), 6 deletions(-)

@@ -2919,12 +2919,6 @@ static int mmu_set_spte(struct kvm_vcpu
 	bool prefetch = !fault || fault->prefetch;
 	bool write_fault = fault && fault->write;
 
-	if (unlikely(is_noslot_pfn(pfn))) {
-		vcpu->stat.pf_mmio_spte_created++;
-		mark_mmio_spte(vcpu, sptep, gfn, pte_access);
-		return RET_PF_EMULATE;
-	}
-
 	if (is_shadow_present_pte(*sptep)) {
 		/*
 		 * If we overwrite a PTE page pointer with a 2MB PMD, unlink

@@ -2944,6 +2938,14 @@ static int mmu_set_spte(struct kvm_vcpu
 			was_rmapped = 1;
 	}
 
+	if (unlikely(is_noslot_pfn(pfn))) {
+		vcpu->stat.pf_mmio_spte_created++;
+		mark_mmio_spte(vcpu, sptep, gfn, pte_access);
+		if (flush)
+			kvm_flush_remote_tlbs_gfn(vcpu->kvm, gfn, level);
+		return RET_PF_EMULATE;
+	}
+
 	wrprot = make_spte(vcpu, sp, slot, pte_access, gfn, pfn, *sptep, prefetch,
 			   true, host_writable, &spte);
 

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

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

From: Milos Nikic <[email protected]>

commit bac3190a8e79beff6ed221975e0c9b1b5f2a21da upstream.

This patch targets two internal state machine invariants in checkpoint.c residing inside functions that natively return integer error codes.

- In jbd2_cleanup_journal_tail(): A blocknr of 0 indicates a severely corrupted journal superblock. Replaced the J_ASSERT with a WARN_ON_ONCE and a graceful journal abort, returning -EFSCORRUPTED.

- In jbd2_log_do_checkpoint(): Replaced the J_ASSERT_BH checking for an unexpected buffer_jwrite state. If the warning triggers, we explicitly drop the just-taken get_bh() reference and call __flush_batch() to safely clean up any previously queued buffers in the j_chkpt_bhs array, preventing a memory leak before returning -EFSCORRUPTED.

---
 fs/jbd2/checkpoint.c |   15 +++++++++++++--
 1 file changed, 13 insertions(+), 2 deletions(-)

@@ -267,7 +267,15 @@ restart:
 			 */
 			BUFFER_TRACE(bh, "queue");
 			get_bh(bh);
-			J_ASSERT_BH(bh, !buffer_jwrite(bh));
+			if (WARN_ON_ONCE(buffer_jwrite(bh))) {
+				put_bh(bh); /* drop the ref we just took */
+				spin_unlock(&journal->j_list_lock);
+				/* Clean up any previously batched buffers */
+				if (batch_count)
+					__flush_batch(journal, &batch_count);
+				jbd2_journal_abort(journal, -EFSCORRUPTED);
+				return -EFSCORRUPTED;
+			}
 			journal->j_chkpt_bhs[batch_count++] = bh;
 			transaction->t_chp_stats.cs_written++;
 			transaction->t_checkpoint_list = jh->b_cpnext;

@@ -325,7 +333,10 @@ int jbd2_cleanup_journal_tail(journal_t
 
 	if (!jbd2_journal_get_log_tail(journal, &first_tid, &blocknr))
 		return 1;
-	J_ASSERT(blocknr != 0);
+	if (WARN_ON_ONCE(blocknr == 0)) {
+		jbd2_journal_abort(journal, -EFSCORRUPTED);
+		return -EFSCORRUPTED;
+	}
 
 	/*
 	 * We need to make sure that any blocks that were recently written out

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

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

From: Jassi Brar <[email protected]>

commit cfe02147e86307a17057ee4e3604f5f5919571d2 upstream.

The mbox_client for qcom-mpm sends NULL doorbell messages via mbox_send_message() but never signals TX completion.

Set knows_txdone=true and call mbox_client_txdone() after a successful send, matching the pattern used by other Qualcomm mailbox clients (smp2p, smsm, qcom_aoss etc).

---
 drivers/irqchip/irq-qcom-mpm.c |    3 +++
 1 file changed, 3 insertions(+)

@@ -306,6 +306,8 @@ static int mpm_pd_power_off(struct gener
 	if (ret < 0)
 		return ret;
 
+	mbox_client_txdone(priv->mbox_chan, 0);
+
 	return 0;
 }
 

@@ -434,6 +436,7 @@ static int qcom_mpm_init(struct device_n
 	}
 
 	priv->mbox_client.dev = dev;
+	priv->mbox_client.knows_txdone = true;
 	priv->mbox_chan = mbox_request_channel(&priv->mbox_client, 0);
 	if (IS_ERR(priv->mbox_chan)) {
 		ret = PTR_ERR(priv->mbox_chan);

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

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

From: Davidlohr Bueso <[email protected]>

commit 210d36d892de5195e6766c45519dfb1e65f3eb83 upstream.

Fuzzying/stressing futexes triggered:

    WARNING: kernel/futex/core.c:825 at wait_for_owner_exiting+0x7a/0x80, CPU#11: futex_lock_pi_s/524

When futex_lock_pi_atomic() sees the owner is exiting, it returns -EBUSY and stores a refcounted task pointer in 'exiting'.

After wait_for_owner_exiting() consumes that reference, the local pointer is never reset to nil. Upon a retry, if futex_lock_pi_atomic() returns a different error, the bogus pointer is passed to wait_for_owner_exiting().

CPU0 CPU1 CPU2
futex_lock_pi(uaddr)
// acquires the PI futex
exit()

    futex_cleanup_begin()
      futex_state = EXITING;
			     futex_lock_pi(uaddr)
			       futex_lock_pi_atomic()
				 attach_to_pi_owner()
				   // observes EXITING
				   *exiting = owner;  // takes ref
				   return -EBUSY
			       wait_for_owner_exiting(-EBUSY, owner)
				 put_task_struct();   // drops ref
			       // exiting still points to owner
			       goto retry;
			       futex_lock_pi_atomic()
				 lock_pi_update_atomic()
				   cmpxchg(uaddr)
					*uaddr ^= WAITERS // whatever
				   // value changed
				 return -EAGAIN;
			       wait_for_owner_exiting(-EAGAIN, exiting) // stale
				 WARN_ON_ONCE(exiting)

Fix this by resetting upon retry, essentially aligning it with requeue_pi.

---
 kernel/futex/pi.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

@@ -918,7 +918,7 @@ int fixup_pi_owner(u32 __user *uaddr, st
 int futex_lock_pi(u32 __user *uaddr, unsigned int flags, ktime_t *time, int trylock)
 {
 	struct hrtimer_sleeper timeout, *to;
-	struct task_struct *exiting = NULL;
+	struct task_struct *exiting;
 	struct rt_mutex_waiter rt_waiter;
 	struct futex_hash_bucket *hb;
 	struct futex_q q = futex_q_init;

@@ -933,6 +933,7 @@ int futex_lock_pi(u32 __user *uaddr, uns
 	to = futex_setup_timer(time, &timeout, flags, 0);
 
 retry:
+	exiting = NULL;
 	ret = get_futex_key(uaddr, flags, &q.key, FUTEX_WRITE);
 	if (unlikely(ret != 0))
 		goto out;

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

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

From: Joy Zou <[email protected]>

commit 2e7b5cf72e51c9cf9c8b75190189c757df31ddd9 upstream.

Configure only the requested channel when a fixed channel is specified to avoid modifying other channels unintentionally.

Fix parameter configuration when a fixed DMA channel is requested on i.MX9 AON domain and i.MX8QM/QXP/DXL platforms. When a client requests a fixed channel (e.g., channel 6), the driver traverses channels 0-5 and may unintentionally modify their configuration if they are unused.

This leads to issues such as setting the is_multi_fifo flag unexpectedly, causing memcpy tests to fail when using the dmatest tool.

Only affect edma memcpy test when the channel is fixed.

---
 drivers/dma/fsl-edma-main.c |   26 +++++++++++---------------
 1 file changed, 11 insertions(+), 15 deletions(-)

@@ -185,10 +185,8 @@ static struct dma_chan *fsl_edma3_xlate(
 			return NULL;
 		i = fsl_chan - fsl_edma->chans;
 
-		fsl_chan->priority = dma_spec->args[1];
-		fsl_chan->is_rxchan = dma_spec->args[2] & FSL_EDMA_RX;
-		fsl_chan->is_remote = dma_spec->args[2] & FSL_EDMA_REMOTE;
-		fsl_chan->is_multi_fifo = dma_spec->args[2] & FSL_EDMA_MULTI_FIFO;
+		if (!b_chmux && i != dma_spec->args[0])
+			continue;
 
 		if ((dma_spec->args[2] & FSL_EDMA_EVEN_CH) && (i & 0x1))
 			continue;

@@ -196,17 +194,15 @@ static struct dma_chan *fsl_edma3_xlate(
 		if ((dma_spec->args[2] & FSL_EDMA_ODD_CH) && !(i & 0x1))
 			continue;
 
-		if (!b_chmux && i == dma_spec->args[0]) {
-			chan = dma_get_slave_channel(chan);
-			chan->device->privatecnt++;
-			return chan;
-		} else if (b_chmux && !fsl_chan->srcid) {
-			/* if controller support channel mux, choose a free channel */
-			chan = dma_get_slave_channel(chan);
-			chan->device->privatecnt++;
-			fsl_chan->srcid = dma_spec->args[0];
-			return chan;
-		}
+		fsl_chan->srcid = dma_spec->args[0];
+		fsl_chan->priority = dma_spec->args[1];
+		fsl_chan->is_rxchan = dma_spec->args[2] & FSL_EDMA_RX;
+		fsl_chan->is_remote = dma_spec->args[2] & FSL_EDMA_REMOTE;
+		fsl_chan->is_multi_fifo = dma_spec->args[2] & FSL_EDMA_MULTI_FIFO;
+
+		chan = dma_get_slave_channel(chan);
+		chan->device->privatecnt++;
+		return chan;
 	}
 	return NULL;
 }

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

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

From: Claudiu Beznea <[email protected]>

commit abb863e6213dc41a58ef8bb3289b7e77460dabf3 upstream.

The driver lists (ld_free, ld_queue) are used in rz_dmac_free_chan_resources(), rz_dmac_terminate_all(), rz_dmac_issue_pending(), and rz_dmac_irq_handler_thread(), all under the virtual channel lock. Take the same lock in rz_dmac_prep_slave_sg() and rz_dmac_prep_dma_memcpy() as well to avoid concurrency issues, since these functions also check whether the lists are empty and update or remove list entries.

---
 drivers/dma/sh/rz-dmac.c |   63 ++++++++++++++++++++++++++---------------------
 1 file changed, 35 insertions(+), 28 deletions(-)

@@ -10,6 +10,7 @@
  */
 
 #include <linux/bitfield.h>
+#include <linux/cleanup.h>
 #include <linux/dma-mapping.h>
 #include <linux/dmaengine.h>
 #include <linux/interrupt.h>

@@ -424,6 +425,7 @@ static int rz_dmac_alloc_chan_resources(
 		if (!desc)
 			break;
 
+		/* No need to lock. This is called only for the 1st client. */
 		list_add_tail(&desc->node, &channel->ld_free);
 		channel->descs_allocated++;
 	}

@@ -479,18 +481,21 @@ rz_dmac_prep_dma_memcpy(struct dma_chan
 	dev_dbg(dmac->dev, "%s channel: %d src=0x%pad dst=0x%pad len=%zu\n",
 		__func__, channel->index, &src, &dest, len);
 
-	if (list_empty(&channel->ld_free))
-		return NULL;
+	scoped_guard(spinlock_irqsave, &channel->vc.lock) {
+		if (list_empty(&channel->ld_free))
+			return NULL;
+
+		desc = list_first_entry(&channel->ld_free, struct rz_dmac_desc, node);
+
+		desc->type = RZ_DMAC_DESC_MEMCPY;
+		desc->src = src;
+		desc->dest = dest;
+		desc->len = len;
+		desc->direction = DMA_MEM_TO_MEM;
 
-	desc = list_first_entry(&channel->ld_free, struct rz_dmac_desc, node);
-
-	desc->type = RZ_DMAC_DESC_MEMCPY;
-	desc->src = src;
-	desc->dest = dest;
-	desc->len = len;
-	desc->direction = DMA_MEM_TO_MEM;
+		list_move_tail(channel->ld_free.next, &channel->ld_queue);
+	}
 
-	list_move_tail(channel->ld_free.next, &channel->ld_queue);
 	return vchan_tx_prep(&channel->vc, &desc->vd, flags);
 }
 

@@ -506,27 +511,29 @@ rz_dmac_prep_slave_sg(struct dma_chan *c
 	int dma_length = 0;
 	int i = 0;
 
-	if (list_empty(&channel->ld_free))
-		return NULL;
-
-	desc = list_first_entry(&channel->ld_free, struct rz_dmac_desc, node);
+	scoped_guard(spinlock_irqsave, &channel->vc.lock) {
+		if (list_empty(&channel->ld_free))
+			return NULL;
+
+		desc = list_first_entry(&channel->ld_free, struct rz_dmac_desc, node);
+
+		for_each_sg(sgl, sg, sg_len, i)
+			dma_length += sg_dma_len(sg);
+
+		desc->type = RZ_DMAC_DESC_SLAVE_SG;
+		desc->sg = sgl;
+		desc->sgcount = sg_len;
+		desc->len = dma_length;
+		desc->direction = direction;
+
+		if (direction == DMA_DEV_TO_MEM)
+			desc->src = channel->src_per_address;
+		else
+			desc->dest = channel->dst_per_address;
 
-	for_each_sg(sgl, sg, sg_len, i) {
-		dma_length += sg_dma_len(sg);
+		list_move_tail(channel->ld_free.next, &channel->ld_queue);
 	}
 
-	desc->type = RZ_DMAC_DESC_SLAVE_SG;
-	desc->sg = sgl;
-	desc->sgcount = sg_len;
-	desc->len = dma_length;
-	desc->direction = direction;
-
-	if (direction == DMA_DEV_TO_MEM)
-		desc->src = channel->src_per_address;
-	else
-		desc->dest = channel->dst_per_address;
-
-	list_move_tail(channel->ld_free.next, &channel->ld_queue);
 	return vchan_tx_prep(&channel->vc, &desc->vd, flags);
 }
 

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

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

From: Claudiu Beznea <[email protected]>

commit 89a8567d84bde88cb7cdbbac2ab2299c4f991490 upstream.

Both rz_dmac_disable_hw() and rz_dmac_irq_handle_channel() update the CHCTRL register. To avoid concurrency issues when configuring functionalities exposed by this registers, take the virtual channel lock. All other CHCTRL updates were already protected by the same lock.

Previously, rz_dmac_disable_hw() disabled and re-enabled local IRQs, before accessing CHCTRL registers but this does not ensure race-free access. Remove the local IRQ disable/enable code as well.

---
 drivers/dma/sh/rz-dmac.c |    9 ++++-----
 1 file changed, 4 insertions(+), 5 deletions(-)

@@ -286,13 +286,10 @@ static void rz_dmac_disable_hw(struct rz
 {
 	struct dma_chan *chan = &channel->vc.chan;
 	struct rz_dmac *dmac = to_rz_dmac(chan->device);
-	unsigned long flags;
 
 	dev_dbg(dmac->dev, "%s channel %d\n", __func__, channel->index);
 
-	local_irq_save(flags);
 	rz_dmac_ch_writel(channel, CHCTRL_DEFAULT, CHCTRL, 1);
-	local_irq_restore(flags);
 }
 
 static void rz_dmac_set_dmars_register(struct rz_dmac *dmac, int nr, u32 dmars)

@@ -545,8 +542,8 @@ static int rz_dmac_terminate_all(struct
 	unsigned int i;
 	LIST_HEAD(head);
 
-	rz_dmac_disable_hw(channel);
 	spin_lock_irqsave(&channel->vc.lock, flags);
+	rz_dmac_disable_hw(channel);
 	for (i = 0; i < DMAC_NR_LMDESC; i++)
 		lmdesc[i].header = 0;
 

@@ -677,7 +674,9 @@ static void rz_dmac_irq_handle_channel(s
 	if (chstat & CHSTAT_ER) {
 		dev_err(dmac->dev, "DMAC err CHSTAT_%d = %08X\n",
 			channel->index, chstat);
-		rz_dmac_ch_writel(channel, CHCTRL_DEFAULT, CHCTRL, 1);
+
+		scoped_guard(spinlock_irqsave, &channel->vc.lock)
+			rz_dmac_ch_writel(channel, CHCTRL_DEFAULT, CHCTRL, 1);
 		goto done;
 	}
 

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

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

From: Eric Huang <[email protected]>

commit 14b81abe7bdc25f8097906fc2f91276ffedb2d26 upstream.

PASID resue could cause interrupt issue when process immediately runs into hw state left by previous process exited with the same PASID, it's possible that page faults are still pending in the IH ring buffer when the process exits and frees up its PASID. To prevent the case, it uses idr cyclic allocator same as kernel pid's.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c |   45 ++++++++++++++++++++++----------
 drivers/gpu/drm/amd/amdgpu/amdgpu_ids.h |    1 
 drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c  |    1 
 3 files changed, 34 insertions(+), 13 deletions(-)

@@ -35,10 +35,13 @@
  * PASIDs are global address space identifiers that can be shared
  * between the GPU, an IOMMU and the driver. VMs on different devices
  * may use the same PASID if they share the same address
- * space. Therefore PASIDs are allocated using a global IDA. VMs are
- * looked up from the PASID per amdgpu_device.
+ * space. Therefore PASIDs are allocated using IDR cyclic allocator
+ * (similar to kernel PID allocation) which naturally delays reuse.
+ * VMs are looked up from the PASID per amdgpu_device.
  */
-static DEFINE_IDA(amdgpu_pasid_ida);
+
+static DEFINE_IDR(amdgpu_pasid_idr);
+static DEFINE_SPINLOCK(amdgpu_pasid_idr_lock);
 
 /* Helper to free pasid from a fence callback */
 struct amdgpu_pasid_cb {

@@ -50,8 +53,8 @@ struct amdgpu_pasid_cb {
  * amdgpu_pasid_alloc - Allocate a PASID
  * @bits: Maximum width of the PASID in bits, must be at least 1
  *
- * Allocates a PASID of the given width while keeping smaller PASIDs
- * available if possible.
+ * Uses kernel's IDR cyclic allocator (same as PID allocation).
+ * Allocates sequentially with automatic wrap-around.
  *
  * Returns a positive integer on success. Returns %-EINVAL if bits==0.
  * Returns %-ENOSPC if no PASID was available. Returns %-ENOMEM on

@@ -59,14 +62,15 @@ struct amdgpu_pasid_cb {
  */
 int amdgpu_pasid_alloc(unsigned int bits)
 {
-	int pasid = -EINVAL;
+	int pasid;
 
-	for (bits = min(bits, 31U); bits > 0; bits--) {
-		pasid = ida_alloc_range(&amdgpu_pasid_ida, 1U << (bits - 1),
-					(1U << bits) - 1, GFP_KERNEL);
-		if (pasid != -ENOSPC)
-			break;
-	}
+	if (bits == 0)
+		return -EINVAL;
+
+	spin_lock(&amdgpu_pasid_idr_lock);
+	pasid = idr_alloc_cyclic(&amdgpu_pasid_idr, NULL, 1,
+				 1U << bits, GFP_KERNEL);
+	spin_unlock(&amdgpu_pasid_idr_lock);
 
 	if (pasid >= 0)
 		trace_amdgpu_pasid_allocated(pasid);

@@ -81,7 +85,10 @@ int amdgpu_pasid_alloc(unsigned int bits
 void amdgpu_pasid_free(u32 pasid)
 {
 	trace_amdgpu_pasid_freed(pasid);
-	ida_free(&amdgpu_pasid_ida, pasid);
+
+	spin_lock(&amdgpu_pasid_idr_lock);
+	idr_remove(&amdgpu_pasid_idr, pasid);
+	spin_unlock(&amdgpu_pasid_idr_lock);
 }
 
 static void amdgpu_pasid_free_cb(struct dma_fence *fence,

@@ -635,3 +642,15 @@ void amdgpu_vmid_mgr_fini(struct amdgpu_
 		}
 	}
 }
+
+/**
+ * amdgpu_pasid_mgr_cleanup - cleanup PASID manager
+ *
+ * Cleanup the IDR allocator.
+ */
+void amdgpu_pasid_mgr_cleanup(void)
+{
+	spin_lock(&amdgpu_pasid_idr_lock);
+	idr_destroy(&amdgpu_pasid_idr);
+	spin_unlock(&amdgpu_pasid_idr_lock);
+}

@@ -75,6 +75,7 @@ int amdgpu_pasid_alloc(unsigned int bits
 void amdgpu_pasid_free(u32 pasid);
 void amdgpu_pasid_free_delayed(struct dma_resv *resv,
 			       u32 pasid);
+void amdgpu_pasid_mgr_cleanup(void);
 
 bool amdgpu_vmid_had_gpu_reset(struct amdgpu_device *adev,
 			       struct amdgpu_vmid *id);

@@ -2721,6 +2721,7 @@ void amdgpu_vm_manager_fini(struct amdgp
 	xa_destroy(&adev->vm_manager.pasids);
 
 	amdgpu_vmid_mgr_fini(adev);
+	amdgpu_pasid_mgr_cleanup();
 }
 
 /**

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

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

From: Imre Deak <[email protected]>

commit 77fcf58df15edcf3f5b5421f24814fb72796def9 upstream.

Clearing the DP tunnel stream BW in the atomic state involves getting the tunnel group state, which can fail. Handle the error accordingly.

This fixes at least one issue where drm_dp_tunnel_atomic_set_stream_bw() failed to get the tunnel group state returning -EDEADLK, which wasn't handled. This lead to the ctx->contended warn later in modeset_lock() while taking a WW mutex for another object in the same atomic state, and thus within the same already contended WW context.

Moving intel_crtc_state_alloc() later would avoid freeing saved_state on the error path; this stable patch leaves that simplification for a follow-up.

---
 drivers/gpu/drm/i915/display/intel_display.c   |    8 +++++++-
 drivers/gpu/drm/i915/display/intel_dp_tunnel.c |   20 ++++++++++++++------
 drivers/gpu/drm/i915/display/intel_dp_tunnel.h |   11 +++++++----
 3 files changed, 28 insertions(+), 11 deletions(-)

@@ -4588,6 +4588,7 @@ intel_crtc_prepare_cleared_state(struct
 		intel_atomic_get_new_crtc_state(state, crtc);
 	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
 	struct intel_crtc_state *saved_state;
+	int err;
 
 	saved_state = intel_crtc_state_alloc(crtc);
 	if (!saved_state)

@@ -4596,7 +4597,12 @@ intel_crtc_prepare_cleared_state(struct
 	/* free the old crtc_state->hw members */
 	intel_crtc_free_hw_state(crtc_state);
 
-	intel_dp_tunnel_atomic_clear_stream_bw(state, crtc_state);
+	err = intel_dp_tunnel_atomic_clear_stream_bw(state, crtc_state);
+	if (err) {
+		kfree(saved_state);
+
+		return err;
+	}
 
 	/* FIXME: before the switch to atomic started, a new pipe_config was
 	 * kzalloc'd. Code that depends on any field being zero should be

@@ -622,19 +622,27 @@ int intel_dp_tunnel_atomic_compute_strea
  *
  * Clear any DP tunnel stream BW requirement set by
  * intel_dp_tunnel_atomic_compute_stream_bw().
+ *
+ * Returns 0 in case of success, a negative error code otherwise.
  */
-void intel_dp_tunnel_atomic_clear_stream_bw(struct intel_atomic_state *state,
-					    struct intel_crtc_state *crtc_state)
+int intel_dp_tunnel_atomic_clear_stream_bw(struct intel_atomic_state *state,
+					   struct intel_crtc_state *crtc_state)
 {
 	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	int err;
 
 	if (!crtc_state->dp_tunnel_ref.tunnel)
-		return;
+		return 0;
+
+	err = drm_dp_tunnel_atomic_set_stream_bw(&state->base,
+						 crtc_state->dp_tunnel_ref.tunnel,
+						 crtc->pipe, 0);
+	if (err)
+		return err;
 
-	drm_dp_tunnel_atomic_set_stream_bw(&state->base,
-					   crtc_state->dp_tunnel_ref.tunnel,
-					   crtc->pipe, 0);
 	drm_dp_tunnel_ref_put(&crtc_state->dp_tunnel_ref);
+
+	return 0;
 }
 
 /**

@@ -39,8 +39,8 @@ int intel_dp_tunnel_atomic_compute_strea
 					     struct intel_dp *intel_dp,
 					     const struct intel_connector *connector,
 					     struct intel_crtc_state *crtc_state);
-void intel_dp_tunnel_atomic_clear_stream_bw(struct intel_atomic_state *state,
-					    struct intel_crtc_state *crtc_state);
+int intel_dp_tunnel_atomic_clear_stream_bw(struct intel_atomic_state *state,
+					   struct intel_crtc_state *crtc_state);
 
 int intel_dp_tunnel_atomic_add_state_for_crtc(struct intel_atomic_state *state,
 					      struct intel_crtc *crtc);

@@ -87,9 +87,12 @@ intel_dp_tunnel_atomic_compute_stream_bw
 	return 0;
 }
 
-static inline void
+static inline int
 intel_dp_tunnel_atomic_clear_stream_bw(struct intel_atomic_state *state,
-				       struct intel_crtc_state *crtc_state) {}
+				       struct intel_crtc_state *crtc_state)
+{
+	return 0;
+}
 
 static inline int
 intel_dp_tunnel_atomic_add_state_for_crtc(struct intel_atomic_state *state,

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

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

From: Li Jun <[email protected]>

commit 3a28daa9b7d7c2ddf2c722e9e95d7e0928bf0cd1 upstream.

1. Replace "of_find_node_by_path("/")" with "of_root" to avoid multiple calls to "of_node_put()".

2. Fix a potential kernel oops during early boot when memory allocation fails while parsing CPU model from device tree.

---
 arch/loongarch/kernel/env.c |    7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)

@@ -42,16 +42,15 @@ static int __init init_cpu_fullname(void
 	int cpu, ret;
 	char *cpuname;
 	const char *model;
-	struct device_node *root;
 
 	/* Parsing cpuname from DTS model property */
-	root = of_find_node_by_path("/");
-	ret = of_property_read_string(root, "model", &model);
+	ret = of_property_read_string(of_root, "model", &model);
 	if (ret == 0) {
 		cpuname = kstrdup(model, GFP_KERNEL);
+		if (!cpuname)
+			return -ENOMEM;
 		loongson_sysconf.cpuname = strsep(&cpuname, " ");
 	}
-	of_node_put(root);
 
 	if (loongson_sysconf.cpuname && !strncmp(loongson_sysconf.cpuname, "Loongson", 8)) {
 		for (cpu = 0; cpu < NR_CPUS; cpu++)

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

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

From: Huacai Chen <[email protected]>

commit 95db0c9f526d583634cddb2e5914718570fbac87 upstream.

1. Hardware limitation: GPU, DC and VPU are typically PCI device 06.0, 06.1 and 06.2. They share some hardware resources, so when configure the PCI 06.0 device BAR1, DMA memory access cannot be performed through this BAR, otherwise it will cause hardware abnormalities.

2. In typical scenarios of reboot or S3/S4, DC access to memory through BAR is not prohibited, resulting in GPU DMA hangs.

3. Workaround method: When configuring the 06.0 device BAR1, turn off the memory access of DC, GPU and VPU (via DC's CRTC registers).

---
 arch/loongarch/pci/pci.c |   80 +++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 80 insertions(+)

@@ -6,9 +6,11 @@
 #include <linux/export.h>
 #include <linux/init.h>
 #include <linux/acpi.h>
+#include <linux/delay.h>
 #include <linux/types.h>
 #include <linux/pci.h>
 #include <linux/vgaarb.h>
+#include <linux/io-64-nonatomic-lo-hi.h>
 #include <asm/cacheflush.h>
 #include <asm/loongson.h>
 

@@ -16,6 +18,9 @@
 #define PCI_DEVICE_ID_LOONGSON_DC1      0x7a06
 #define PCI_DEVICE_ID_LOONGSON_DC2      0x7a36
 #define PCI_DEVICE_ID_LOONGSON_DC3      0x7a46
+#define PCI_DEVICE_ID_LOONGSON_GPU1     0x7a15
+#define PCI_DEVICE_ID_LOONGSON_GPU2     0x7a25
+#define PCI_DEVICE_ID_LOONGSON_GPU3     0x7a35
 
 int raw_pci_read(unsigned int domain, unsigned int bus, unsigned int devfn,
 						int reg, int len, u32 *val)

@@ -100,3 +105,78 @@ static void pci_fixup_vgadev(struct pci_
 DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_LOONGSON, PCI_DEVICE_ID_LOONGSON_DC1, pci_fixup_vgadev);
 DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_LOONGSON, PCI_DEVICE_ID_LOONGSON_DC2, pci_fixup_vgadev);
 DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_LOONGSON, PCI_DEVICE_ID_LOONGSON_DC3, pci_fixup_vgadev);
+
+#define CRTC_NUM_MAX		2
+#define CRTC_OUTPUT_ENABLE	0x100
+
+static void loongson_gpu_fixup_dma_hang(struct pci_dev *pdev, bool on)
+{
+	u32 i, val, count, crtc_offset, device;
+	void __iomem *crtc_reg, *base, *regbase;
+	static u32 crtc_status[CRTC_NUM_MAX] = { 0 };
+
+	base = pdev->bus->ops->map_bus(pdev->bus, pdev->devfn + 1, 0);
+	device = readw(base + PCI_DEVICE_ID);
+
+	regbase = ioremap(readq(base + PCI_BASE_ADDRESS_0) & ~0xffull, SZ_64K);
+	if (!regbase) {
+		pci_err(pdev, "Failed to ioremap()\n");
+		return;
+	}
+
+	switch (device) {
+	case PCI_DEVICE_ID_LOONGSON_DC2:
+		crtc_reg = regbase + 0x1240;
+		crtc_offset = 0x10;
+		break;
+	case PCI_DEVICE_ID_LOONGSON_DC3:
+		crtc_reg = regbase;
+		crtc_offset = 0x400;
+		break;
+	}
+
+	for (i = 0; i < CRTC_NUM_MAX; i++, crtc_reg += crtc_offset) {
+		val = readl(crtc_reg);
+
+		if (!on)
+			crtc_status[i] = val;
+
+		/* No need to fixup if the status is off at startup. */
+		if (!(crtc_status[i] & CRTC_OUTPUT_ENABLE))
+			continue;
+
+		if (on)
+			val |= CRTC_OUTPUT_ENABLE;
+		else
+			val &= ~CRTC_OUTPUT_ENABLE;
+
+		mb();
+		writel(val, crtc_reg);
+
+		for (count = 0; count < 40; count++) {
+			val = readl(crtc_reg) & CRTC_OUTPUT_ENABLE;
+			if ((on && val) || (!on && !val))
+				break;
+			udelay(1000);
+		}
+
+		pci_info(pdev, "DMA hang fixup at reg[0x%lx]: 0x%x\n",
+				(unsigned long)crtc_reg & 0xffff, readl(crtc_reg));
+	}
+
+	iounmap(regbase);
+}
+
+static void pci_fixup_dma_hang_early(struct pci_dev *pdev)
+{
+	loongson_gpu_fixup_dma_hang(pdev, false);
+}
+DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_LOONGSON, PCI_DEVICE_ID_LOONGSON_GPU2, pci_fixup_dma_hang_early);
+DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_LOONGSON, PCI_DEVICE_ID_LOONGSON_GPU3, pci_fixup_dma_hang_early);
+
+static void pci_fixup_dma_hang_final(struct pci_dev *pdev)
+{
+	loongson_gpu_fixup_dma_hang(pdev, true);
+}
+DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_LOONGSON, PCI_DEVICE_ID_LOONGSON_GPU2, pci_fixup_dma_hang_final);
+DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_LOONGSON, PCI_DEVICE_ID_LOONGSON_GPU3, pci_fixup_dma_hang_final);

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

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

From: Huacai Chen <[email protected]>

commit 2db06c15d8c7a0ccb6108524e16cd9163753f354 upstream.

kvm_get_vcpu_by_cpuid() takes a cpuid parameter whose type is int, so cpuid can be negative. Let kvm_get_vcpu_by_cpuid() return NULL for this case so as to make it more robust.

This fix an out-of-bounds access to kvm_arch::phyid_map::phys_map[].

---
 arch/loongarch/kvm/vcpu.c |    3 +++
 1 file changed, 3 insertions(+)

@@ -531,6 +531,9 @@ struct kvm_vcpu *kvm_get_vcpu_by_cpuid(s
 {
 	struct kvm_phyid_map *map;
 
+	if (cpuid < 0)
+		return NULL;
+
 	if (cpuid >= KVM_MAX_PHYID)
 		return NULL;
 

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

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

From: Yuto Ohnuki <[email protected]>

commit 4f24a767e3d64a5f58c595b5c29b6063a201f1e3 upstream.

The unmount sequence in xfs_unmount_flush_inodes() pushed the AIL while background reclaim and inodegc are still running. This is broken independently of any use-after-free issues - background reclaim and inodegc should not be running while the AIL is being pushed during unmount, as inodegc can dirty and insert inodes into the AIL during the flush, and background reclaim can race to abort and free dirty inodes.

Reorder xfs_unmount_flush_inodes() to stop inodegc and cancel background reclaim before pushing the AIL. Stop inodegc before cancelling m_reclaim_work because the inodegc worker can re-queue m_reclaim_work via xfs_inodegc_set_reclaimable.

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

@@ -588,8 +588,9 @@ xfs_unmount_check(
  * have been retrying in the background.  This will prevent never-ending
  * retries in AIL pushing from hanging the unmount.
  *
- * Finally, we can push the AIL to clean all the remaining dirty objects, then
- * reclaim the remaining inodes that are still in memory at this point in time.
+ * Stop inodegc and background reclaim before pushing the AIL so that they
+ * are not running while the AIL is being flushed. Then push the AIL to
+ * clean all the remaining dirty objects and reclaim the remaining inodes.
  */
 static void
 xfs_unmount_flush_inodes(

@@ -601,9 +602,9 @@ xfs_unmount_flush_inodes(
 
 	xfs_set_unmounting(mp);
 
-	xfs_ail_push_all_sync(mp->m_ail);
 	xfs_inodegc_stop(mp);
 	cancel_delayed_work_sync(&mp->m_reclaim_work);
+	xfs_ail_push_all_sync(mp->m_ail);
 	xfs_reclaim_inodes(mp);
 	xfs_health_unmount(mp);
 }

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

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

From: Yuto Ohnuki <[email protected]>

commit 394d70b86fae9fe865e7e6d9540b7696f73aa9b6 upstream.

In xfs_inode_item_push() and xfs_qm_dquot_logitem_push(), the AIL lock is dropped to perform buffer IO. Once the cluster buffer no longer protects the log item from reclaim, the log item may be freed by background reclaim or the dquot shrinker. The subsequent spin_lock() call dereferences lip->li_ailp, which is a use-after-free.

Fix this by saving the ailp pointer in a local variable while the AIL lock is held and the log item is guaranteed to be valid.

---
 fs/xfs/xfs_dquot_item.c |    9 +++++++--
 fs/xfs/xfs_inode_item.c |    9 +++++++--
 2 files changed, 14 insertions(+), 4 deletions(-)

@@ -126,6 +126,7 @@ xfs_qm_dquot_logitem_push(
 	struct xfs_dq_logitem	*qlip = DQUOT_ITEM(lip);
 	struct xfs_dquot	*dqp = qlip->qli_dquot;
 	struct xfs_buf		*bp;
+	struct xfs_ail		*ailp = lip->li_ailp;
 	uint			rval = XFS_ITEM_SUCCESS;
 	int			error;
 

@@ -154,7 +155,7 @@ xfs_qm_dquot_logitem_push(
 		goto out_unlock;
 	}
 
-	spin_unlock(&lip->li_ailp->ail_lock);
+	spin_unlock(&ailp->ail_lock);
 
 	error = xfs_dquot_use_attached_buf(dqp, &bp);
 	if (error == -EAGAIN) {

@@ -173,9 +174,13 @@ xfs_qm_dquot_logitem_push(
 			rval = XFS_ITEM_FLUSHING;
 	}
 	xfs_buf_relse(bp);
+	/*
+	 * The buffer no longer protects the log item from reclaim, so
+	 * do not reference lip after this point.
+	 */
 
 out_relock_ail:
-	spin_lock(&lip->li_ailp->ail_lock);
+	spin_lock(&ailp->ail_lock);
 out_unlock:
 	xfs_dqunlock(dqp);
 	return rval;

@@ -726,6 +726,7 @@ xfs_inode_item_push(
 	struct xfs_inode_log_item *iip = INODE_ITEM(lip);
 	struct xfs_inode	*ip = iip->ili_inode;
 	struct xfs_buf		*bp = lip->li_buf;
+	struct xfs_ail		*ailp = lip->li_ailp;
 	uint			rval = XFS_ITEM_SUCCESS;
 	int			error;
 

@@ -748,7 +749,7 @@ xfs_inode_item_push(
 	if (!xfs_buf_trylock(bp))
 		return XFS_ITEM_LOCKED;
 
-	spin_unlock(&lip->li_ailp->ail_lock);
+	spin_unlock(&ailp->ail_lock);
 
 	/*
 	 * We need to hold a reference for flushing the cluster buffer as it may

@@ -772,7 +773,11 @@ xfs_inode_item_push(
 		rval = XFS_ITEM_LOCKED;
 	}
 
-	spin_lock(&lip->li_ailp->ail_lock);
+	/*
+	 * The buffer no longer protects the log item from reclaim, so
+	 * do not reference lip after this point.
+	 */
+	spin_lock(&ailp->ail_lock);
 	return rval;
 }
 

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

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

From: hongao <[email protected]>

commit 268378b6ad20569af0d1957992de1c8b16c6e900 upstream.

xchk_quota_item can return early after calling xchk_fblock_process_error. When that helper returns false, the function returned immediately without dropping dq->q_qlock, which can leave the dquot lock held and risk lock leaks or deadlocks in later quota operations.

Fix this by unlocking dq->q_qlock before the early return.

---
 fs/xfs/scrub/quota.c |    4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

@@ -174,8 +174,10 @@ xchk_quota_item(
 
 	error = xchk_quota_item_bmap(sc, dq, offset);
 	xchk_iunlock(sc, XFS_ILOCK_SHARED);
-	if (!xchk_fblock_process_error(sc, XFS_DATA_FORK, offset, &error))
+	if (!xchk_fblock_process_error(sc, XFS_DATA_FORK, offset, &error)) {
+		mutex_unlock(&dq->q_qlock);
 		return error;
+	}
 
 	/*
 	 * Warn if the hard limits are larger than the fs.

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

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

From: Long Li <[email protected]>

commit d72f2084e30966097c8eae762e31986a33c3c0ae upstream.

The ri_total checks for SET/REPLACE operations are hardcoded to 3, but xfs_attri_item_size() only emits a value iovec when value_len > 0, so ri_total is 2 when value_len == 0.

For PPTR_SET/PPTR_REMOVE/PPTR_REPLACE, value_len is validated by xfs_attri_validate() to be exactly sizeof(struct xfs_parent_rec) and is never zero, so their hardcoded checks remain correct.

This problem may cause log recovery failures. The following script can be used to reproduce the problem:

#!/bin/bash
mkfs.xfs -f /dev/sda
mount /dev/sda /mnt/test/
touch /mnt/test/file
for i in {1..200}; do

         attr -s "user.attr_$i" -V "value_$i" /mnt/test/file > /dev/null

done
echo 1 > /sys/fs/xfs/debug/larp
echo 1 > /sys/fs/xfs/sda/errortag/larp
attr -s "user.zero" -V "" /mnt/test/file
echo 0 > /sys/fs/xfs/sda/errortag/larp
umount /mnt/test
mount /dev/sda /mnt/test/ # mount failed

Fix this by deriving the expected count dynamically as "2 + !!value_len" for SET/REPLACE operations.

---
 fs/xfs/xfs_attr_item.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

@@ -1052,8 +1052,8 @@ xlog_recover_attri_commit_pass2(
 		break;
 	case XFS_ATTRI_OP_FLAGS_SET:
 	case XFS_ATTRI_OP_FLAGS_REPLACE:
-		/* Log item, attr name, attr value */
-		if (item->ri_total != 3) {
+		/* Log item, attr name, optional attr value */
+		if (item->ri_total != 2 + !!attri_formatp->alfi_value_len) {
 			XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp,
 					     attri_formatp, len);
 			return -EFSCORRUPTED;

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

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

From: Darrick J. Wong <[email protected]>

commit 70685c291ef82269180758130394ecdc4496b52c upstream.

xlog_recovery_iget* never set @ip to a valid pointer if they return an error, so this irele will walk off a dangling pointer. Fix that.

---
 fs/xfs/xfs_attr_item.c |    1 -
 1 file changed, 1 deletion(-)

@@ -658,7 +658,6 @@ xfs_attri_recover_work(
 		break;
 	}
 	if (error) {
-		xfs_irele(ip);
 		XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, attrp,
 				sizeof(*attrp));
 		return ERR_PTR(-EFSCORRUPTED);

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

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

From: Darrick J. Wong <[email protected]>

commit e31c53a8060e134111ed095783fee0aa0c43b080 upstream.

The xfile/xmbuf shmem file descriptions are no longer as detailed as they were when online fsck was first merged, because moving to static strings in commit 60382993a2e180 ("xfs: get rid of the xchk_xfile_*_descr calls") removed a memory allocation and hence a source of failure.

However this makes encoding the description in the tracepoints sort of a waste of memory. David Laight also points out that file_path doesn't zero the whole buffer which causes exposure of stale trace bytes, and Steven Rostedt wonders why we're not using a dynamic array for the file path.

I don't think this is worth fixing, so let's just rip it out.

---
 fs/xfs/scrub/trace.h |   12 ++----------
 fs/xfs/xfs_trace.h   |   11 ++---------
 2 files changed, 4 insertions(+), 19 deletions(-)

@@ -959,20 +959,12 @@ TRACE_EVENT(xfile_create,
 	TP_STRUCT__entry(
 		__field(dev_t, dev)
 		__field(unsigned long, ino)
-		__array(char, pathname, MAXNAMELEN)
 	),
 	TP_fast_assign(
-		char		*path;
-
 		__entry->ino = file_inode(xf->file)->i_ino;
-		path = file_path(xf->file, __entry->pathname, MAXNAMELEN);
-		if (IS_ERR(path))
-			strncpy(__entry->pathname, "(unknown)",
-					sizeof(__entry->pathname));
 	),
-	TP_printk("xfino 0x%lx path '%s'",
-		  __entry->ino,
-		  __entry->pathname)
+	TP_printk("xfino 0x%lx",
+		  __entry->ino)
 );
 
 TRACE_EVENT(xfile_destroy,

@@ -4710,23 +4710,16 @@ TRACE_EVENT(xmbuf_create,
 	TP_STRUCT__entry(
 		__field(dev_t, dev)
 		__field(unsigned long, ino)
-		__array(char, pathname, MAXNAMELEN)
 	),
 	TP_fast_assign(
-		char		*path;
 		struct file	*file = btp->bt_file;
 
 		__entry->dev = btp->bt_mount->m_super->s_dev;
 		__entry->ino = file_inode(file)->i_ino;
-		path = file_path(file, __entry->pathname, MAXNAMELEN);
-		if (IS_ERR(path))
-			strncpy(__entry->pathname, "(unknown)",
-					sizeof(__entry->pathname));
 	),
-	TP_printk("dev %d:%d xmino 0x%lx path '%s'",
+	TP_printk("dev %d:%d xmino 0x%lx",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
-		  __entry->ino,
-		  __entry->pathname)
+		  __entry->ino)
 );
 
 TRACE_EVENT(xmbuf_free,

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

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

From: Simon Weber <[email protected]>

commit b1d682f1990c19fb1d5b97d13266210457092bcd upstream.

Fix an issue arising when ext4 features has_journal, ea_inode, and encrypt are activated simultaneously, leading to ENOSPC when creating an encrypted file.

Fix by passing XATTR_CREATE flag to xattr_set_handle function if a handle is specified, i.e., when the function is called in the control flow of creating a new inode. This aligns the number of jbd2 credits set_handle checks for with the number allocated for creating a new inode.

ext4_set_context must not be called with a non-null handle (fs_data) if fscrypt context xattr is not guaranteed to not exist yet. The only other usage of this function currently is when handling the ioctl FS_IOC_SET_ENCRYPTION_POLICY, which calls it with fs_data=NULL.

---
 fs/ext4/crypto.c |    9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

@@ -163,10 +163,17 @@ static int ext4_set_context(struct inode
 	 */
 
 	if (handle) {
+		/*
+		 * Since the inode is new it is ok to pass the
+		 * XATTR_CREATE flag. This is necessary to match the
+		 * remaining journal credits check in the set_handle
+		 * function with the credits allocated for the new
+		 * inode.
+		 */
 		res = ext4_xattr_set_handle(handle, inode,
 					    EXT4_XATTR_INDEX_ENCRYPTION,
 					    EXT4_XATTR_NAME_ENCRYPTION_CONTEXT,
-					    ctx, len, 0);
+					    ctx, len, XATTR_CREATE);
 		if (!res) {
 			ext4_set_inode_flag(inode, EXT4_INODE_ENCRYPT);
 			ext4_clear_inode_state(inode,

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

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

From: Deepanshu Kartikey <[email protected]>

commit ed9356a30e59c7cc3198e7fc46cfedf3767b9b17 upstream.

Add a check in ext4_setattr() to convert files from inline data storage to extent-based storage when truncate() grows the file size beyond the inline capacity. This prevents the filesystem from entering an inconsistent state where the inline data flag is set but the file size exceeds what can be stored inline.

Without this fix, the following sequence causes a kernel BUG_ON():

1. Mount filesystem with inode that has inline flag set and small size
2. truncate(file, 50MB) - grows size but inline flag remains set
3. sendfile() attempts to write data
4. ext4_write_inline_data() hits BUG_ON(write_size > inline_capacity)

The crash occurs because ext4_write_inline_data() expects inline storage to accommodate the write, but the actual inline capacity (~60 bytes for i_block + ~96 bytes for xattrs) is far smaller than the file size and write request.

The fix checks if the new size from setattr exceeds the inode's actual inline capacity (EXT4_I(inode)->i_inline_size) and converts the file to extent-based storage before proceeding with the size change.

This addresses the root cause by ensuring the inline data flag and file size remain consistent during truncate operations.

---
 fs/ext4/inode.c |   12 ++++++++++++
 1 file changed, 12 insertions(+)

@@ -5490,6 +5490,18 @@ int ext4_setattr(struct mnt_idmap *idmap
 		if (attr->ia_size == inode->i_size)
 			inc_ivers = false;
 
+		/*
+		 * If file has inline data but new size exceeds inline capacity,
+		 * convert to extent-based storage first to prevent inconsistent
+		 * state (inline flag set but size exceeds inline capacity).
+		 */
+		if (ext4_has_inline_data(inode) &&
+		    attr->ia_size > EXT4_I(inode)->i_inline_size) {
+			error = ext4_convert_inline_data(inode);
+			if (error)
+				goto err_out;
+		}
+
 		if (shrink) {
 			if (ext4_should_order_data(inode)) {
 				error = ext4_begin_ordered_truncate(inode,

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

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

From: Jan Kara <[email protected]>

commit f4a2b42e78914ff15630e71289adc589c3a8eb45 upstream.

There are cases where ext4_bio_write_page() gets called for a page which has no buffers to submit. This happens e.g. when the part of the file is actually a hole, when we cannot allocate blocks due to being called from jbd2, or in data=journal mode when checkpointing writes the buffers earlier. In these cases we just return from ext4_bio_write_page() however if the page didn't need redirtying, we will leave stale DIRTY and/or TOWRITE tags in xarray because those get cleared only in __folio_start_writeback(). As a result we can leave these tags set in mappings even after a final sync on filesystem that's getting remounted read-only or that's being frozen. Various assertions can then get upset when writeback is started on such filesystems (Gerald reported assertion in ext4_journal_check_start() firing).

Fix the problem by cycling the page through writeback state even if we decide nothing needs to be written for it so that xarray tags get properly updated. This is slightly silly (we could update the xarray tags directly) but I don't think a special helper messing with xarray tags is really worth it in this relatively rare corner case.

---
 fs/ext4/page-io.c |   10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)

@@ -506,9 +506,15 @@ int ext4_bio_write_folio(struct ext4_io_
 		nr_to_submit++;
 	} while ((bh = bh->b_this_page) != head);
 
-	/* Nothing to submit? Just unlock the folio... */
-	if (!nr_to_submit)
+	if (!nr_to_submit) {
+		/*
+		 * We have nothing to submit. Just cycle the folio through
+		 * writeback state to properly update xarray tags.
+		 */
+		__folio_start_writeback(folio, keep_towrite);
+		folio_end_writeback(folio);
 		return 0;
+	}
 
 	bh = head = folio_buffers(folio);
 

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

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

From: Jan Kara <[email protected]>

commit 1308255bbf8452762f89f44f7447ce137ecdbcff upstream.

When inode metadata is changed, we sometimes just call ext4_mark_inode_dirty() to track modified metadata. This copies inode metadata into block buffer which is enough when we are journalling metadata. However when we are running in nojournal mode we currently fail to write the dirtied inode buffer during fsync(2) because the inode is not marked as dirty. Use explicit ext4_write_inode() call to make sure the inode table buffer is written to the disk. This is a band aid solution but proper solution requires a much larger rewrite including changes in metadata bh tracking infrastructure.

---
 fs/ext4/fsync.c |   16 ++++++++++++++--
 1 file changed, 14 insertions(+), 2 deletions(-)

@@ -83,11 +83,23 @@ static int ext4_fsync_nojournal(struct f
 				int datasync, bool *needs_barrier)
 {
 	struct inode *inode = file->f_inode;
+	struct writeback_control wbc = {
+		.sync_mode = WB_SYNC_ALL,
+		.nr_to_write = 0,
+	};
 	int ret;
 
 	ret = generic_buffers_fsync_noflush(file, start, end, datasync);
-	if (!ret)
-		ret = ext4_sync_parent(inode);
+	if (ret)
+		return ret;
+
+	/* Force writeout of inode table buffer to disk */
+	ret = ext4_write_inode(inode, &wbc);
+	if (ret)
+		return ret;
+
+	ret = ext4_sync_parent(inode);
+
 	if (test_opt(inode->i_sb, BARRIER))
 		*needs_barrier = true;
 

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

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

From: Jan Kara <[email protected]>

commit bd060afa7cc3e0ad30afa9ecc544a78638498555 upstream.

recently_deleted() checks whether inode has been used in the near past. However this can give false positive result when inode table is not initialized yet and we are in fact comparing to random garbage (or stale itable block of a filesystem before mkfs). Ultimately this results in uninitialized inodes being skipped during inode allocation and possibly they are never initialized and thus e2fsck complains. Verify if the inode has been initialized before checking for dtime.

---
 fs/ext4/ialloc.c |    6 ++++++
 1 file changed, 6 insertions(+)

@@ -686,6 +686,12 @@ static int recently_deleted(struct super
 	if (unlikely(!gdp))
 		return 0;
 
+	/* Inode was never used in this filesystem? */
+	if (ext4_has_group_desc_csum(sb) &&
+	    (gdp->bg_flags & cpu_to_le16(EXT4_BG_INODE_UNINIT) ||
+	     ino >= EXT4_INODES_PER_GROUP(sb) - ext4_itable_unused_count(sb, gdp)))
+		return 0;
+
 	bh = sb_find_get_block(sb, ext4_inode_table(sb, gdp) +
 		       (ino / inodes_per_block));
 	if (!bh || !buffer_uptodate(bh))

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

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

From: Yuto Ohnuki <[email protected]>

commit 356227096eb66e41b23caf7045e6304877322edf upstream.

Replace BUG_ON() with proper error handling when inline data size exceeds PAGE_SIZE. This prevents kernel panic and allows the system to continue running while properly reporting the filesystem corruption.

The error is logged via ext4_error_inode(), the buffer head is released to prevent memory leak, and -EFSCORRUPTED is returned to indicate filesystem corruption.

---
 fs/ext4/inline.c |   10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

@@ -517,7 +517,15 @@ static int ext4_read_inline_folio(struct
 		goto out;
 
 	len = min_t(size_t, ext4_get_inline_size(inode), i_size_read(inode));
-	BUG_ON(len > PAGE_SIZE);
+
+	if (len > PAGE_SIZE) {
+		ext4_error_inode(inode, __func__, __LINE__, 0,
+				 "inline size %zu exceeds PAGE_SIZE", len);
+		ret = -EFSCORRUPTED;
+		brelse(iloc.bh);
+		goto out;
+	}
+
 	kaddr = kmap_local_folio(folio, 0);
 	ret = ext4_read_inline_data(inode, kaddr, len, &iloc);
 	kaddr = folio_zero_tail(folio, len, kaddr + len);

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

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

From: Tejas Bharambe <[email protected]>

commit 2acb5c12ebd860f30e4faf67e6cc8c44ddfe5fe8 upstream.

ext4_ext_correct_indexes() walks up the extent tree correcting index entries when the first extent in a leaf is modified. Before accessing path[k].p_idx->ei_block, there is no validation that p_idx falls within the valid range of index entries for that level.

If the on-disk extent header contains a corrupted or crafted eh_entries value, p_idx can point past the end of the allocated buffer, causing a slab-out-of-bounds read.

Fix this by validating path[k].p_idx against EXT_LAST_INDEX() at both access sites: before the while loop and inside it. Return -EFSCORRUPTED if the index pointer is out of range, consistent with how other bounds violations are handled in the ext4 extent tree code.

---
 fs/ext4/extents.c |   15 +++++++++++++++
 1 file changed, 15 insertions(+)

@@ -1739,6 +1739,13 @@ static int ext4_ext_correct_indexes(hand
 	err = ext4_ext_get_access(handle, inode, path + k);
 	if (err)
 		return err;
+	if (unlikely(path[k].p_idx > EXT_LAST_INDEX(path[k].p_hdr))) {
+		EXT4_ERROR_INODE(inode,
+				 "path[%d].p_idx %p > EXT_LAST_INDEX %p",
+				 k, path[k].p_idx,
+				 EXT_LAST_INDEX(path[k].p_hdr));
+		return -EFSCORRUPTED;
+	}
 	path[k].p_idx->ei_block = border;
 	err = ext4_ext_dirty(handle, inode, path + k);
 	if (err)

@@ -1751,6 +1758,14 @@ static int ext4_ext_correct_indexes(hand
 		err = ext4_ext_get_access(handle, inode, path + k);
 		if (err)
 			goto clean;
+		if (unlikely(path[k].p_idx > EXT_LAST_INDEX(path[k].p_hdr))) {
+			EXT4_ERROR_INODE(inode,
+					 "path[%d].p_idx %p > EXT_LAST_INDEX %p",
+					 k, path[k].p_idx,
+					 EXT_LAST_INDEX(path[k].p_hdr));
+			err = -EFSCORRUPTED;
+			goto clean;
+		}
 		path[k].p_idx->ei_block = border;
 		err = ext4_ext_dirty(handle, inode, path + k);
 		if (err)

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

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

From: Edward Adam Davis <[email protected]>

commit 5422fe71d26d42af6c454ca9527faaad4e677d6c upstream.

On the mkdir/mknod path, when mapping logical blocks to physical blocks, if inserting a new extent into the extent tree fails (in this example, because the file system disabled the huge file feature when marking the inode as dirty), ext4_ext_map_blocks() only calls ext4_free_blocks() to reclaim the physical block without deleting the corresponding data in the extent tree. This causes subsequent mkdir operations to reference the previously reclaimed physical block number again, even though this physical block is already being used by the xattr block. Therefore, a situation arises where both the directory and xattr are using the same buffer head block in memory simultaneously.

The above causes ext4_xattr_block_set() to enter an infinite loop about "inserted" and cannot release the inode lock, ultimately leading to the 143s blocking problem mentioned in [1].

If the metadata is corrupted, then trying to remove some extent space can do even more harm. Also in case EXT4_GET_BLOCKS_DELALLOC_RESERVE was passed, remove space wrongly update quota information. Jan Kara suggests distinguishing between two cases:

1) The error is ENOSPC or EDQUOT - in this case the filesystem is fully consistent and we must maintain its consistency including all the accounting. However these errors can happen only early before we've inserted the extent into the extent tree. So current code works correctly for this case.

2) Some other error - this means metadata is corrupted. We should strive to do as few modifications as possible to limit damage. So I'd just skip freeing of allocated blocks.

[1]

INFO: task syz.0.17:5995 blocked for more than 143 seconds.
Call Trace:
 inode_lock_nested include/linux/fs.h:1073 [inline]
 __start_dirop fs/namei.c:2923 [inline]
 start_dirop fs/namei.c:2934 [inline]

---
 fs/ext4/extents.c |    8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

@@ -4458,9 +4458,13 @@ got_allocated_blocks:
 	path = ext4_ext_insert_extent(handle, inode, path, &newex, flags);
 	if (IS_ERR(path)) {
 		err = PTR_ERR(path);
-		if (allocated_clusters) {
+		/*
+		 * Gracefully handle out of space conditions. If the filesystem
+		 * is inconsistent, we'll just leak allocated blocks to avoid
+		 * causing even more damage.
+		 */
+		if (allocated_clusters && (err == -EDQUOT || err == -ENOSPC)) {
 			int fb_flags = 0;
-
 			/*
 			 * free data blocks we just allocated.
 			 * not a good idea to call discard here directly,

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

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

From: Ye Bin <[email protected]>

commit 46066e3a06647c5b186cc6334409722622d05c44 upstream.

There's issue as follows:
...
EXT4-fs (mmcblk0p1): Delayed block allocation failed for inode 206 at logical offset 0 with max blocks 1 with error 117 EXT4-fs (mmcblk0p1): This should not happen!! Data will be lost

EXT4-fs (mmcblk0p1): Delayed block allocation failed for inode 206 at logical offset 0 with max blocks 1 with error 117 EXT4-fs (mmcblk0p1): This should not happen!! Data will be lost

EXT4-fs (mmcblk0p1): Delayed block allocation failed for inode 206 at logical offset 0 with max blocks 1 with error 117 EXT4-fs (mmcblk0p1): This should not happen!! Data will be lost

EXT4-fs (mmcblk0p1): Delayed block allocation failed for inode 206 at logical offset 0 with max blocks 1 with error 117 EXT4-fs (mmcblk0p1): This should not happen!! Data will be lost

EXT4-fs (mmcblk0p1): Delayed block allocation failed for inode 2243 at logical offset 0 with max blocks 1 with error 117 EXT4-fs (mmcblk0p1): This should not happen!! Data will be lost

EXT4-fs (mmcblk0p1): Delayed block allocation failed for inode 2239 at logical offset 0 with max blocks 1 with error 117 EXT4-fs (mmcblk0p1): This should not happen!! Data will be lost

EXT4-fs (mmcblk0p1): error count since last fsck: 1
EXT4-fs (mmcblk0p1): initial error at time 1765597433: ext4_mb_generate_buddy:760 EXT4-fs (mmcblk0p1): last error at time 1765597433: ext4_mb_generate_buddy:760
...

According to the log analysis, blocks are always requested from the corrupted block group. This may happen as follows:
ext4_mb_find_by_goal
ext4_mb_load_buddy
ext4_mb_load_buddy_gfp
ext4_mb_init_cache
ext4_read_block_bitmap_nowait
ext4_wait_block_bitmap
ext4_validate_block_bitmap
if (!grp || EXT4_MB_GRP_BBITMAP_CORRUPT(grp))
return -EFSCORRUPTED; // There's no logs.
if (err)
return err; // Will return error
ext4_lock_group(ac->ac_sb, group);
if (unlikely(EXT4_MB_GRP_BBITMAP_CORRUPT(e4b->bd_info))) // Unreachable
goto out;

After commit 9008a58e5dce ("ext4: make the bitmap read routines return real error codes") merged, Commit 163a203ddb36 ("ext4: mark block group as corrupt on block bitmap error") is no real solution for allocating blocks from corrupted block groups. This is because if 'EXT4_MB_GRP_BBITMAP_CORRUPT(e4b->bd_info)' is true, then 'ext4_mb_load_buddy()' may return an error. This means that the block allocation will fail.
Therefore, check block group if corrupted when ext4_mb_load_buddy() returns error.

---
 fs/ext4/mballoc.c |    6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

@@ -2440,8 +2440,12 @@ int ext4_mb_find_by_goal(struct ext4_all
 		return 0;
 
 	err = ext4_mb_load_buddy(ac->ac_sb, group, e4b);
-	if (err)
+	if (err) {
+		if (EXT4_MB_GRP_BBITMAP_CORRUPT(e4b->bd_info) &&
+		    !(ac->ac_flags & EXT4_MB_HINT_GOAL_ONLY))
+			return 0;
 		return err;
+	}
 
 	ext4_lock_group(ac->ac_sb, group);
 	if (unlikely(EXT4_MB_GRP_BBITMAP_CORRUPT(e4b->bd_info)))

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

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

From: Helen Koike <[email protected]>

commit 3822743dc20386d9897e999dbb990befa3a5b3f8 upstream.

bigalloc with s_first_data_block != 0 is not supported, reject mounting it.

---
 fs/ext4/super.c |    7 +++++++
 1 file changed, 7 insertions(+)

@@ -3631,6 +3631,13 @@ int ext4_feature_set_ok(struct super_blo
 			 "extents feature\n");
 		return 0;
 	}
+	if (ext4_has_feature_bigalloc(sb) &&
+	    le32_to_cpu(EXT4_SB(sb)->s_es->s_first_data_block)) {
+		ext4_msg(sb, KERN_WARNING,
+			 "bad geometry: bigalloc file system with non-zero "
+			 "first_data_block\n");
+		return 0;
+	}
 
 #if !IS_ENABLED(CONFIG_QUOTA) || !IS_ENABLED(CONFIG_QFMT_V2)
 	if (!readonly && (ext4_has_feature_quota(sb) ||

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

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

From: Jiayuan Chen <[email protected]>

commit d15e4b0a418537aafa56b2cb80d44add83e83697 upstream.

Commit b98535d09179 ("ext4: fix bug_on in start_this_handle during umount filesystem") moved ext4_unregister_sysfs() before flushing s_sb_upd_work to prevent new error work from being queued via /proc/fs/ext4/xx/mb_groups reads during unmount. However, this introduced a use-after-free because update_super_work calls ext4_notify_error_sysfs() -> sysfs_notify() which accesses the kobject's kernfs_node after it has been freed by kobject_del() in ext4_unregister_sysfs():

update_super_work ext4_put_super
----------------- --------------

                                   ext4_unregister_sysfs(sb)
                                     kobject_del(&sbi->s_kobj)
                                       __kobject_del()
                                         sysfs_remove_dir()
                                           kobj->sd = NULL
                                         sysfs_put(sd)
                                           kernfs_put()  // RCU free
  ext4_notify_error_sysfs(sbi)
    sysfs_notify(&sbi->s_kobj)
      kn = kobj->sd              // stale pointer
      kernfs_get(kn)             // UAF on freed kernfs_node
                                   ext4_journal_destroy()
                                     flush_work(&sbi->s_sb_upd_work)

Instead of reordering the teardown sequence, fix this by making ext4_notify_error_sysfs() detect that sysfs has already been torn down by checking s_kobj.state_in_sysfs, and skipping the sysfs_notify() call in that case. A dedicated mutex (s_error_notify_mutex) serializes ext4_notify_error_sysfs() against kobject_del() in ext4_unregister_sysfs() to prevent TOCTOU races where the kobject could be deleted between the state_in_sysfs check and the sysfs_notify() call.

---
 fs/ext4/ext4.h  |    1 +
 fs/ext4/super.c |    1 +
 fs/ext4/sysfs.c |   10 +++++++++-
 3 files changed, 11 insertions(+), 1 deletion(-)

@@ -1534,6 +1534,7 @@ struct ext4_sb_info {
 	struct proc_dir_entry *s_proc;
 	struct kobject s_kobj;
 	struct completion s_kobj_unregister;
+	struct mutex s_error_notify_mutex; /* protects sysfs_notify vs kobject_del */
 	struct super_block *s_sb;
 	struct buffer_head *s_mmp_bh;
 

@@ -5313,6 +5313,7 @@ static int __ext4_fill_super(struct fs_c
 
 	timer_setup(&sbi->s_err_report, print_daily_error_info, 0);
 	spin_lock_init(&sbi->s_error_lock);
+	mutex_init(&sbi->s_error_notify_mutex);
 	INIT_WORK(&sbi->s_sb_upd_work, update_super_work);
 
 	err = ext4_group_desc_init(sb, es, logical_sb_block, &first_not_zeroed);

@@ -551,7 +551,10 @@ static const struct kobj_type ext4_feat_
 
 void ext4_notify_error_sysfs(struct ext4_sb_info *sbi)
 {
-	sysfs_notify(&sbi->s_kobj, NULL, "errors_count");
+	mutex_lock(&sbi->s_error_notify_mutex);
+	if (sbi->s_kobj.state_in_sysfs)
+		sysfs_notify(&sbi->s_kobj, NULL, "errors_count");
+	mutex_unlock(&sbi->s_error_notify_mutex);
 }
 
 static struct kobject *ext4_root;

@@ -564,8 +567,10 @@ int ext4_register_sysfs(struct super_blo
 	int err;
 
 	init_completion(&sbi->s_kobj_unregister);
+	mutex_lock(&sbi->s_error_notify_mutex);
 	err = kobject_init_and_add(&sbi->s_kobj, &ext4_sb_ktype, ext4_root,
 				   "%s", sb->s_id);
+	mutex_unlock(&sbi->s_error_notify_mutex);
 	if (err) {
 		kobject_put(&sbi->s_kobj);
 		wait_for_completion(&sbi->s_kobj_unregister);