5.4.148-rc1 review

From: Anirudh Rayabharam <[email protected]>

[ Upstream commit 258c81b341c8025d79073ce2d6ce19dcdc7d10d2 ]

In vhci_device_unlink_cleanup(), the URBs for unsent unlink requests are not given back. This sometimes causes usb_kill_urb to wait indefinitely for that urb to be given back. syzbot has reported a hung task issue [1] for this.

To fix this, give back the urbs corresponding to unsent unlink requests (unlink_tx list) similar to how urbs corresponding to unanswered unlink requests (unlink_rx list) are given back.

[1]: https://syzkaller.appspot.com/bug?id=08f12df95ae7da69814e64eb5515d5a85ed06b76

---
 drivers/usb/usbip/vhci_hcd.c | 24 ++++++++++++++++++++++++
 1 file changed, 24 insertions(+)

index 98636fbf7188..46a46cde2070 100644
@@ -952,8 +952,32 @@ static void vhci_device_unlink_cleanup(struct vhci_device *vdev)
 	spin_lock(&vdev->priv_lock);
 
 	list_for_each_entry_safe(unlink, tmp, &vdev->unlink_tx, list) {
+		struct urb *urb;
+
+		/* give back urb of unsent unlink request */
 		pr_info("unlink cleanup tx %lu\n", unlink->unlink_seqnum);
+
+		urb = pickup_urb_and_free_priv(vdev, unlink->unlink_seqnum);
+		if (!urb) {
+			list_del(&unlink->list);
+			kfree(unlink);
+			continue;
+		}
+
+		urb->status = -ENODEV;
+
+		usb_hcd_unlink_urb_from_ep(hcd, urb);
+
 		list_del(&unlink->list);
+
+		spin_unlock(&vdev->priv_lock);
+		spin_unlock_irqrestore(&vhci->lock, flags);
+
+		usb_hcd_giveback_urb(hcd, urb, urb->status);
+
+		spin_lock_irqsave(&vhci->lock, flags);
+		spin_lock(&vdev->priv_lock);
+
 		kfree(unlink);
 	}
 

2.30.2

From: Shuah Khan <[email protected]>

[ Upstream commit 66cce9e73ec61967ed1f97f30cee79bd9a2bb7ee ]

When a remote usb device is attached to the local Virtual USB Host Controller Root Hub port, the bound device driver may send a port reset command.

vhci_hcd accepts port resets only when the device doesn't have port address assigned to it. When reset happens device is in assigned/used state and vhci_hcd rejects it leaving the port in a stuck state.

This problem was found when a blue-tooth or xbox wireless dongle was passed through using usbip.

A few drivers reset the port during probe including mt76 driver specific to this bug report. Fix the problem with a change to honor reset requests when device is in used state (VDEV_ST_USED).

---
 drivers/usb/usbip/vhci_hcd.c | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

index 46a46cde2070..170abb06a8a4 100644
@@ -455,8 +455,14 @@ static int vhci_hub_control(struct usb_hcd *hcd, u16 typeReq, u16 wValue,
 			vhci_hcd->port_status[rhport] &= ~(1 << USB_PORT_FEAT_RESET);
 			vhci_hcd->re_timeout = 0;
 
+			/*
+			 * A few drivers do usb reset during probe when
+			 * the device could be in VDEV_ST_USED state
+			 */
 			if (vhci_hcd->vdev[rhport].ud.status ==
-			    VDEV_ST_NOTASSIGNED) {
+				VDEV_ST_NOTASSIGNED ||
+			    vhci_hcd->vdev[rhport].ud.status ==
+				VDEV_ST_USED) {
 				usbip_dbg_vhci_rh(
 					" enable rhport %d (status %u)\n",
 					rhport,

2.30.2

From: Sugar Zhang <[email protected]>

[ Upstream commit 53ca9b9777b95cdd689181d7c547e38dc79adad0 ]

API 'set_fmt' maybe called when PD is off, in the situation, any register access will hang the system. so, enable PD before r/w register.

---
 sound/soc/rockchip/rockchip_i2s.c | 19 ++++++++++++++-----
 1 file changed, 14 insertions(+), 5 deletions(-)

index 61c984f10d8e..f48b146cd96a 100644
@@ -186,7 +186,9 @@ static int rockchip_i2s_set_fmt(struct snd_soc_dai *cpu_dai,
 {
 	struct rk_i2s_dev *i2s = to_info(cpu_dai);
 	unsigned int mask = 0, val = 0;
+	int ret = 0;
 
+	pm_runtime_get_sync(cpu_dai->dev);
 	mask = I2S_CKR_MSS_MASK;
 	switch (fmt & SND_SOC_DAIFMT_MASTER_MASK) {
 	case SND_SOC_DAIFMT_CBS_CFS:

@@ -199,7 +201,8 @@ static int rockchip_i2s_set_fmt(struct snd_soc_dai *cpu_dai,
 		i2s->is_master_mode = false;
 		break;
 	default:
-		return -EINVAL;
+		ret = -EINVAL;
+		goto err_pm_put;
 	}
 
 	regmap_update_bits(i2s->regmap, I2S_CKR, mask, val);

@@ -213,7 +216,8 @@ static int rockchip_i2s_set_fmt(struct snd_soc_dai *cpu_dai,
 		val = I2S_CKR_CKP_POS;
 		break;
 	default:
-		return -EINVAL;
+		ret = -EINVAL;
+		goto err_pm_put;
 	}
 
 	regmap_update_bits(i2s->regmap, I2S_CKR, mask, val);

@@ -236,7 +240,8 @@ static int rockchip_i2s_set_fmt(struct snd_soc_dai *cpu_dai,
 		val = I2S_TXCR_TFS_PCM | I2S_TXCR_PBM_MODE(1);
 		break;
 	default:
-		return -EINVAL;
+		ret = -EINVAL;
+		goto err_pm_put;
 	}
 
 	regmap_update_bits(i2s->regmap, I2S_TXCR, mask, val);

@@ -259,12 +264,16 @@ static int rockchip_i2s_set_fmt(struct snd_soc_dai *cpu_dai,
 		val = I2S_RXCR_TFS_PCM | I2S_RXCR_PBM_MODE(1);
 		break;
 	default:
-		return -EINVAL;
+		ret = -EINVAL;
+		goto err_pm_put;
 	}
 
 	regmap_update_bits(i2s->regmap, I2S_RXCR, mask, val);
 
-	return 0;
+err_pm_put:
+	pm_runtime_put(cpu_dai->dev);
+
+	return ret;
 }
 
 static int rockchip_i2s_hw_params(struct snd_pcm_substream *substream,

2.30.2

From: Xiaotan Luo <[email protected]>

[ Upstream commit 1bf56843e664eef2525bdbfae6a561e98910f676 ]

- DSP_A: PCM delay 1 bit mode, L data MSB after FRM LRC
- DSP_B: PCM no delay mode, L data MSB during FRM LRC

---
 sound/soc/rockchip/rockchip_i2s.c | 16 ++++++++--------
 1 file changed, 8 insertions(+), 8 deletions(-)

index f48b146cd96a..086c90e09577 100644
@@ -233,12 +233,12 @@ static int rockchip_i2s_set_fmt(struct snd_soc_dai *cpu_dai,
 	case SND_SOC_DAIFMT_I2S:
 		val = I2S_TXCR_IBM_NORMAL;
 		break;
-	case SND_SOC_DAIFMT_DSP_A: /* PCM no delay mode */
-		val = I2S_TXCR_TFS_PCM;
-		break;
-	case SND_SOC_DAIFMT_DSP_B: /* PCM delay 1 mode */
+	case SND_SOC_DAIFMT_DSP_A: /* PCM delay 1 bit mode */
 		val = I2S_TXCR_TFS_PCM | I2S_TXCR_PBM_MODE(1);
 		break;
+	case SND_SOC_DAIFMT_DSP_B: /* PCM no delay mode */
+		val = I2S_TXCR_TFS_PCM;
+		break;
 	default:
 		ret = -EINVAL;
 		goto err_pm_put;

@@ -257,12 +257,12 @@ static int rockchip_i2s_set_fmt(struct snd_soc_dai *cpu_dai,
 	case SND_SOC_DAIFMT_I2S:
 		val = I2S_RXCR_IBM_NORMAL;
 		break;
-	case SND_SOC_DAIFMT_DSP_A: /* PCM no delay mode */
-		val = I2S_RXCR_TFS_PCM;
-		break;
-	case SND_SOC_DAIFMT_DSP_B: /* PCM delay 1 mode */
+	case SND_SOC_DAIFMT_DSP_A: /* PCM delay 1 bit mode */
 		val = I2S_RXCR_TFS_PCM | I2S_RXCR_PBM_MODE(1);
 		break;
+	case SND_SOC_DAIFMT_DSP_B: /* PCM no delay mode */
+		val = I2S_RXCR_TFS_PCM;
+		break;
 	default:
 		ret = -EINVAL;
 		goto err_pm_put;

2.30.2

From: Sean Keely <[email protected]>

[ Upstream commit 1ec06c2dee679e9f089e78ed20cb74ee90155f61 ]

On systems with multiple SH per SE compute_static_thread_mgmt_se# is split into independent masks, one for each SH, in the upper and lower 16 bits. We need to detect this and apply cu masking to each SH. The cu mask bits are assigned first to each SE, then to alternate SHs, then finally to higher CU id. This ensures that the maximum number of SPIs are engaged as early as possible while balancing CU assignment to each SH.

v2: Use max SH/SE rather than max SH in cu_per_sh.

v3: Fix comment blocks, ensure se_mask is initially zero filled,
and correctly assign se.sh.cu positions to unset bits in cu_mask.

---
 drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager.c | 84 +++++++++++++++-----
 drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager.h |  1 +
 2 files changed, 64 insertions(+), 21 deletions(-)

index 88813dad731f..c021519af810 100644
@@ -98,36 +98,78 @@ void mqd_symmetrically_map_cu_mask(struct mqd_manager *mm,
 		uint32_t *se_mask)
 {
 	struct kfd_cu_info cu_info;
-	uint32_t cu_per_se[KFD_MAX_NUM_SE] = {0};
-	int i, se, sh, cu = 0;
-
+	uint32_t cu_per_sh[KFD_MAX_NUM_SE][KFD_MAX_NUM_SH_PER_SE] = {0};
+	int i, se, sh, cu;
 	amdgpu_amdkfd_get_cu_info(mm->dev->kgd, &cu_info);
 
 	if (cu_mask_count > cu_info.cu_active_number)
 		cu_mask_count = cu_info.cu_active_number;
 
+	/* Exceeding these bounds corrupts the stack and indicates a coding error.
+	 * Returning with no CU's enabled will hang the queue, which should be
+	 * attention grabbing.
+	 */
+	if (cu_info.num_shader_engines > KFD_MAX_NUM_SE) {
+		pr_err("Exceeded KFD_MAX_NUM_SE, chip reports %d\n", cu_info.num_shader_engines);
+		return;
+	}
+	if (cu_info.num_shader_arrays_per_engine > KFD_MAX_NUM_SH_PER_SE) {
+		pr_err("Exceeded KFD_MAX_NUM_SH, chip reports %d\n",
+			cu_info.num_shader_arrays_per_engine * cu_info.num_shader_engines);
+		return;
+	}
+	/* Count active CUs per SH.
+	 *
+	 * Some CUs in an SH may be disabled.	HW expects disabled CUs to be
+	 * represented in the high bits of each SH's enable mask (the upper and lower
+	 * 16 bits of se_mask) and will take care of the actual distribution of
+	 * disabled CUs within each SH automatically.
+	 * Each half of se_mask must be filled only on bits 0-cu_per_sh[se][sh]-1.
+	 *
+	 * See note on Arcturus cu_bitmap layout in gfx_v9_0_get_cu_info.
+	 */
 	for (se = 0; se < cu_info.num_shader_engines; se++)
 		for (sh = 0; sh < cu_info.num_shader_arrays_per_engine; sh++)
-			cu_per_se[se] += hweight32(cu_info.cu_bitmap[se % 4][sh + (se / 4)]);
-
-	/* Symmetrically map cu_mask to all SEs:
-	 * cu_mask[0] bit0 -> se_mask[0] bit0;
-	 * cu_mask[0] bit1 -> se_mask[1] bit0;
-	 * ... (if # SE is 4)
-	 * cu_mask[0] bit4 -> se_mask[0] bit1;
+			cu_per_sh[se][sh] = hweight32(cu_info.cu_bitmap[se % 4][sh + (se / 4)]);
+
+	/* Symmetrically map cu_mask to all SEs & SHs:
+	 * se_mask programs up to 2 SH in the upper and lower 16 bits.
+	 *
+	 * Examples
+	 * Assuming 1 SH/SE, 4 SEs:
+	 * cu_mask[0] bit0 -> se_mask[0] bit0
+	 * cu_mask[0] bit1 -> se_mask[1] bit0
+	 * ...
+	 * cu_mask[0] bit4 -> se_mask[0] bit1
+	 * ...
+	 *
+	 * Assuming 2 SH/SE, 4 SEs
+	 * cu_mask[0] bit0 -> se_mask[0] bit0 (SE0,SH0,CU0)
+	 * cu_mask[0] bit1 -> se_mask[1] bit0 (SE1,SH0,CU0)
+	 * ...
+	 * cu_mask[0] bit4 -> se_mask[0] bit16 (SE0,SH1,CU0)
+	 * cu_mask[0] bit5 -> se_mask[1] bit16 (SE1,SH1,CU0)
+	 * ...
+	 * cu_mask[0] bit8 -> se_mask[0] bit1 (SE0,SH0,CU1)
 	 * ...
+	 *
+	 * First ensure all CUs are disabled, then enable user specified CUs.
 	 */
-	se = 0;
-	for (i = 0; i < cu_mask_count; i++) {
-		if (cu_mask[i / 32] & (1 << (i % 32)))
-			se_mask[se] |= 1 << cu;
-
-		do {
-			se++;
-			if (se == cu_info.num_shader_engines) {
-				se = 0;
-				cu++;
+	for (i = 0; i < cu_info.num_shader_engines; i++)
+		se_mask[i] = 0;
+
+	i = 0;
+	for (cu = 0; cu < 16; cu++) {
+		for (sh = 0; sh < cu_info.num_shader_arrays_per_engine; sh++) {
+			for (se = 0; se < cu_info.num_shader_engines; se++) {
+				if (cu_per_sh[se][sh] > cu) {
+					if (cu_mask[i / 32] & (1 << (i % 32)))
+						se_mask[se] |= 1 << (cu + sh * 16);
+					i++;
+					if (i == cu_mask_count)
+						return;
+				}
 			}
-		} while (cu >= cu_per_se[se] && cu < 32);
+		}
 	}
 }

index fbdb16418847..4edc012e3138 100644
@@ -27,6 +27,7 @@
 #include "kfd_priv.h"
 
 #define KFD_MAX_NUM_SE 8
+#define KFD_MAX_NUM_SH_PER_SE 2
 
 /**
  * struct mqd_manager

2.30.2

From: Zhang Qilong <[email protected]>

[ Upstream commit 0f5d44ac6e55551798dd3da0ff847c8df5990822 ]

If beacon_inject_active is true, we will return without freeing beacon. Fid that by freeing it before returning.

---
 drivers/net/wireless/intel/iwlwifi/mvm/mac-ctxt.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

index 9c417dd06291..7736621dca65 100644
@@ -1043,8 +1043,10 @@ int iwl_mvm_mac_ctxt_beacon_changed(struct iwl_mvm *mvm,
 		return -ENOMEM;
 
 #ifdef CONFIG_IWLWIFI_DEBUGFS
-	if (mvm->beacon_inject_active)
+	if (mvm->beacon_inject_active) {
+		dev_kfree_skb(beacon);
 		return -EBUSY;
+	}
 #endif
 
 	ret = iwl_mvm_mac_ctxt_send_beacon(mvm, vif, beacon);

2.30.2

From: Johannes Berg <[email protected]>

[ Upstream commit c6ce1c74ef2923b8ffd85f7f8b486f804f343b39 ]

When TVQM is enabled (iwl_mvm_has_new_tx_api() is true), then queue numbers are just sequentially assigned 0, 1, 2, ... Prior to TVQM, in DQA, there were some statically allocated queue numbers:
* IWL_MVM_DQA_AUX_QUEUE == 1,
* both IWL_MVM_DQA_INJECT_MONITOR_QUEUE and IWL_MVM_DQA_P2P_DEVICE_QUEUE == 2, and
* IWL_MVM_DQA_AP_PROBE_RESP_QUEUE == 9.

Now, these values are assigned to the members mvm->aux_queue, mvm->snif_queue, mvm->probe_queue and mvm->p2p_dev_queue by default. Normally, this doesn't really matter, and if TVQM is in fact available we override them to the real values after allocating a queue for use there.

However, this allocation doesn't always happen. For example, for mvm->p2p_dev_queue (== 2) it only happens when the P2P Device interface is started, if any. If it's not started, the value in mvm->p2p_dev_queue remains 2. This wouldn't really matter all that much if it weren't for iwl_mvm_is_static_queue() which checks a queue number against one of those four static numbers.

Now, if no P2P Device or monitor interface is added then queue 2 may be dynamically allocated, yet alias mvm->p2p_dev_queue or mvm->snif_queue, and thus iwl_mvm_is_static_queue() erroneously returns true for it. If it then gets full, all interface queues are stopped, instead of just backpressuring against the one TXQ that's really the only affected one.

This clearly can lead to issues, as everything is stopped even if just a single TXQ filled its corresponding HW queue, if it happens to have an appropriate number (2 or 9, AUX is always reassigned.) Due to a mac80211 bug, this also led to a situation in which the queues remained stopped across a deauthentication and then attempts to connect to a new AP started failing, but that's fixed separately.

Fix all of this by simply initializing the queue numbers to the invalid value until they're used, if TVQM is enabled, and also setting them back to that value when the queues are later freed again.

---
 drivers/net/wireless/intel/iwlwifi/mvm/ops.c | 24 +++++++++++++---
 drivers/net/wireless/intel/iwlwifi/mvm/sta.c | 30 ++++++++++++--------
 2 files changed, 38 insertions(+), 16 deletions(-)

index 8b0576cde797..a9aab6c690e8 100644
@@ -687,10 +687,26 @@ iwl_op_mode_mvm_start(struct iwl_trans *trans, const struct iwl_cfg *cfg,
 
 	mvm->fw_restart = iwlwifi_mod_params.fw_restart ? -1 : 0;
 
-	mvm->aux_queue = IWL_MVM_DQA_AUX_QUEUE;
-	mvm->snif_queue = IWL_MVM_DQA_INJECT_MONITOR_QUEUE;
-	mvm->probe_queue = IWL_MVM_DQA_AP_PROBE_RESP_QUEUE;
-	mvm->p2p_dev_queue = IWL_MVM_DQA_P2P_DEVICE_QUEUE;
+	if (iwl_mvm_has_new_tx_api(mvm)) {
+		/*
+		 * If we have the new TX/queue allocation API initialize them
+		 * all to invalid numbers. We'll rewrite the ones that we need
+		 * later, but that doesn't happen for all of them all of the
+		 * time (e.g. P2P Device is optional), and if a dynamic queue
+		 * ends up getting number 2 (IWL_MVM_DQA_P2P_DEVICE_QUEUE) then
+		 * iwl_mvm_is_static_queue() erroneously returns true, and we
+		 * might have things getting stuck.
+		 */
+		mvm->aux_queue = IWL_MVM_INVALID_QUEUE;
+		mvm->snif_queue = IWL_MVM_INVALID_QUEUE;
+		mvm->probe_queue = IWL_MVM_INVALID_QUEUE;
+		mvm->p2p_dev_queue = IWL_MVM_INVALID_QUEUE;
+	} else {
+		mvm->aux_queue = IWL_MVM_DQA_AUX_QUEUE;
+		mvm->snif_queue = IWL_MVM_DQA_INJECT_MONITOR_QUEUE;
+		mvm->probe_queue = IWL_MVM_DQA_AP_PROBE_RESP_QUEUE;
+		mvm->p2p_dev_queue = IWL_MVM_DQA_P2P_DEVICE_QUEUE;
+	}
 
 	mvm->sf_state = SF_UNINIT;
 	if (iwl_mvm_has_unified_ucode(mvm))

index 40cafcf40ccf..5df4bbb6c6de 100644
@@ -346,8 +346,9 @@ static int iwl_mvm_invalidate_sta_queue(struct iwl_mvm *mvm, int queue,
 }
 
 static int iwl_mvm_disable_txq(struct iwl_mvm *mvm, struct ieee80211_sta *sta,
-			       int queue, u8 tid, u8 flags)
+			       u16 *queueptr, u8 tid, u8 flags)
 {
+	int queue = *queueptr;
 	struct iwl_scd_txq_cfg_cmd cmd = {
 		.scd_queue = queue,
 		.action = SCD_CFG_DISABLE_QUEUE,

@@ -356,6 +357,7 @@ static int iwl_mvm_disable_txq(struct iwl_mvm *mvm, struct ieee80211_sta *sta,
 
 	if (iwl_mvm_has_new_tx_api(mvm)) {
 		iwl_trans_txq_free(mvm->trans, queue);
+		*queueptr = IWL_MVM_INVALID_QUEUE;
 
 		return 0;
 	}

@@ -517,6 +519,7 @@ static int iwl_mvm_free_inactive_queue(struct iwl_mvm *mvm, int queue,
 	u8 sta_id, tid;
 	unsigned long disable_agg_tids = 0;
 	bool same_sta;
+	u16 queue_tmp = queue;
 	int ret;
 
 	lockdep_assert_held(&mvm->mutex);

@@ -539,7 +542,7 @@ static int iwl_mvm_free_inactive_queue(struct iwl_mvm *mvm, int queue,
 		iwl_mvm_invalidate_sta_queue(mvm, queue,
 					     disable_agg_tids, false);
 
-	ret = iwl_mvm_disable_txq(mvm, old_sta, queue, tid, 0);
+	ret = iwl_mvm_disable_txq(mvm, old_sta, &queue_tmp, tid, 0);
 	if (ret) {
 		IWL_ERR(mvm,
 			"Failed to free inactive queue %d (ret=%d)\n",

@@ -1209,6 +1212,7 @@ static int iwl_mvm_sta_alloc_queue(struct iwl_mvm *mvm,
 	unsigned int wdg_timeout =
 		iwl_mvm_get_wd_timeout(mvm, mvmsta->vif, false, false);
 	int queue = -1;
+	u16 queue_tmp;
 	unsigned long disable_agg_tids = 0;
 	enum iwl_mvm_agg_state queue_state;
 	bool shared_queue = false, inc_ssn;

@@ -1357,7 +1361,8 @@ static int iwl_mvm_sta_alloc_queue(struct iwl_mvm *mvm,
 	return 0;
 
 out_err:
-	iwl_mvm_disable_txq(mvm, sta, queue, tid, 0);
+	queue_tmp = queue;
+	iwl_mvm_disable_txq(mvm, sta, &queue_tmp, tid, 0);
 
 	return ret;
 }

@@ -1795,7 +1800,7 @@ static void iwl_mvm_disable_sta_queues(struct iwl_mvm *mvm,
 		if (mvm_sta->tid_data[i].txq_id == IWL_MVM_INVALID_QUEUE)
 			continue;
 
-		iwl_mvm_disable_txq(mvm, sta, mvm_sta->tid_data[i].txq_id, i,
+		iwl_mvm_disable_txq(mvm, sta, &mvm_sta->tid_data[i].txq_id, i,
 				    0);
 		mvm_sta->tid_data[i].txq_id = IWL_MVM_INVALID_QUEUE;
 	}

@@ -2005,7 +2010,7 @@ static int iwl_mvm_add_int_sta_with_queue(struct iwl_mvm *mvm, int macidx,
 	ret = iwl_mvm_add_int_sta_common(mvm, sta, NULL, macidx, maccolor);
 	if (ret) {
 		if (!iwl_mvm_has_new_tx_api(mvm))
-			iwl_mvm_disable_txq(mvm, NULL, *queue,
+			iwl_mvm_disable_txq(mvm, NULL, queue,
 					    IWL_MAX_TID_COUNT, 0);
 		return ret;
 	}

@@ -2073,7 +2078,7 @@ int iwl_mvm_rm_snif_sta(struct iwl_mvm *mvm, struct ieee80211_vif *vif)
 	if (WARN_ON_ONCE(mvm->snif_sta.sta_id == IWL_MVM_INVALID_STA))
 		return -EINVAL;
 
-	iwl_mvm_disable_txq(mvm, NULL, mvm->snif_queue, IWL_MAX_TID_COUNT, 0);
+	iwl_mvm_disable_txq(mvm, NULL, &mvm->snif_queue, IWL_MAX_TID_COUNT, 0);
 	ret = iwl_mvm_rm_sta_common(mvm, mvm->snif_sta.sta_id);
 	if (ret)
 		IWL_WARN(mvm, "Failed sending remove station\n");

@@ -2090,7 +2095,7 @@ int iwl_mvm_rm_aux_sta(struct iwl_mvm *mvm)
 	if (WARN_ON_ONCE(mvm->aux_sta.sta_id == IWL_MVM_INVALID_STA))
 		return -EINVAL;
 
-	iwl_mvm_disable_txq(mvm, NULL, mvm->aux_queue, IWL_MAX_TID_COUNT, 0);
+	iwl_mvm_disable_txq(mvm, NULL, &mvm->aux_queue, IWL_MAX_TID_COUNT, 0);
 	ret = iwl_mvm_rm_sta_common(mvm, mvm->aux_sta.sta_id);
 	if (ret)
 		IWL_WARN(mvm, "Failed sending remove station\n");

@@ -2186,7 +2191,7 @@ static void iwl_mvm_free_bcast_sta_queues(struct iwl_mvm *mvm,
 					  struct ieee80211_vif *vif)
 {
 	struct iwl_mvm_vif *mvmvif = iwl_mvm_vif_from_mac80211(vif);
-	int queue;
+	u16 *queueptr, queue;
 
 	lockdep_assert_held(&mvm->mutex);
 

@@ -2195,10 +2200,10 @@ static void iwl_mvm_free_bcast_sta_queues(struct iwl_mvm *mvm,
 	switch (vif->type) {
 	case NL80211_IFTYPE_AP:
 	case NL80211_IFTYPE_ADHOC:
-		queue = mvm->probe_queue;
+		queueptr = &mvm->probe_queue;
 		break;
 	case NL80211_IFTYPE_P2P_DEVICE:
-		queue = mvm->p2p_dev_queue;
+		queueptr = &mvm->p2p_dev_queue;
 		break;
 	default:
 		WARN(1, "Can't free bcast queue on vif type %d\n",

@@ -2206,7 +2211,8 @@ static void iwl_mvm_free_bcast_sta_queues(struct iwl_mvm *mvm,
 		return;
 	}
 
-	iwl_mvm_disable_txq(mvm, NULL, queue, IWL_MAX_TID_COUNT, 0);
+	queue = *queueptr;
+	iwl_mvm_disable_txq(mvm, NULL, queueptr, IWL_MAX_TID_COUNT, 0);
 	if (iwl_mvm_has_new_tx_api(mvm))
 		return;
 

@@ -2441,7 +2447,7 @@ int iwl_mvm_rm_mcast_sta(struct iwl_mvm *mvm, struct ieee80211_vif *vif)
 
 	iwl_mvm_flush_sta(mvm, &mvmvif->mcast_sta, true, 0);
 
-	iwl_mvm_disable_txq(mvm, NULL, mvmvif->cab_queue, 0, 0);
+	iwl_mvm_disable_txq(mvm, NULL, &mvmvif->cab_queue, 0, 0);
 
 	ret = iwl_mvm_rm_sta_common(mvm, mvmvif->mcast_sta.sta_id);
 	if (ret)

2.30.2

From: Johannes Berg <[email protected]>

[ Upstream commit 6c608cd6962ebdf84fd3de6d42f88ed64d2f4e1b ]

BSS elements are protected using RCU, so we need to use RCU properly to access them, fix that.

---
 drivers/net/wireless/intel/iwlwifi/mvm/mac80211.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

index 09b1a6beee77..081cbc9ec736 100644
@@ -2970,16 +2970,20 @@ static void iwl_mvm_check_he_obss_narrow_bw_ru_iter(struct wiphy *wiphy,
 						    void *_data)
 {
 	struct iwl_mvm_he_obss_narrow_bw_ru_data *data = _data;
+	const struct cfg80211_bss_ies *ies;
 	const struct element *elem;
 
-	elem = cfg80211_find_elem(WLAN_EID_EXT_CAPABILITY, bss->ies->data,
-				  bss->ies->len);
+	rcu_read_lock();
+	ies = rcu_dereference(bss->ies);
+	elem = cfg80211_find_elem(WLAN_EID_EXT_CAPABILITY, ies->data,
+				  ies->len);
 
 	if (!elem || elem->datalen < 10 ||
 	    !(elem->data[10] &
 	      WLAN_EXT_CAPA10_OBSS_NARROW_BW_RU_TOLERANCE_SUPPORT)) {
 		data->tolerated = false;
 	}
+	rcu_read_unlock();
 }
 
 static void iwl_mvm_check_he_obss_narrow_bw_ru(struct ieee80211_hw *hw,

2.30.2

From: Yevgeny Kliteynik <[email protected]>

[ Upstream commit ec449ed8230cd30769de3cb70ee0fce293047372 ]

Under high stress, SW steering might get stuck on polling for completion that never comes.
For such cases QP needs to have protocol retransmission mechanism enabled. Currently the retransmission timeout is defined as 0 (unlimited). Fix this by defining a real timeout.

---
 drivers/net/ethernet/mellanox/mlx5/core/steering/dr_send.c | 1 +
 1 file changed, 1 insertion(+)

index f012aac83b10..401564b94eb1 100644
@@ -603,6 +603,7 @@ static int dr_cmd_modify_qp_rtr2rts(struct mlx5_core_dev *mdev,
 	MLX5_SET(qpc, qpc, log_ack_req_freq, 0);
 	MLX5_SET(qpc, qpc, retry_count, attr->retry_cnt);
 	MLX5_SET(qpc, qpc, rnr_retry, attr->rnr_retry);
+	MLX5_SET(qpc, qpc, primary_address_path.ack_timeout, 0x8); /* ~1ms */
 
 	return mlx5_core_qp_modify(mdev, MLX5_CMD_OP_RTR2RTS_QP, 0, qpc,
 				   &dr_qp->mqp);

2.30.2

From: Colin Ian King <[email protected]>

[ Upstream commit 0be883a0d795d9146f5325de582584147dd0dcdc ]

The check for count appears to be incorrect since a non-zero count check occurs a couple of statements earlier. Currently the check is always false and the dev->port->irq != PARPORT_IRQ_NONE part of the check is never tested and the if statement is dead-code. Fix this by removing the check on count.

Note that this code is pre-git history, so I can't find a sha for it.

Addresses-Coverity: ("Logically dead code")

---
 drivers/parport/ieee1284_ops.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

index 5d41dda6da4e..75daa16f38b7 100644
@@ -535,7 +535,7 @@ size_t parport_ieee1284_ecp_read_data (struct parport *port,
 				goto out;
 
 			/* Yield the port for a while. */
-			if (count && dev->port->irq != PARPORT_IRQ_NONE) {
+			if (dev->port->irq != PARPORT_IRQ_NONE) {
 				parport_release (dev);
 				schedule_timeout_interruptible(msecs_to_jiffies(40));
 				parport_claim_or_block (dev);

2.30.2

From: Zekun Shen <[email protected]>

[ Upstream commit 23151b9ae79e3bc4f6a0c4cd3a7f355f68dad128 ]

Bad header can have large length field which can cause OOB. cptr is the last bytes for read, and the eeprom is parsed from high to low address. The OOB, triggered by the condition length > cptr could cause memory error with a read on negative index.

There are some sanity check around length, but it is not compared with cptr (the remaining bytes). Here, the corrupted/bad EEPROM can cause panic.

I was able to reproduce the crash, but I cannot find the log and the reproducer now. After I applied the patch, the bug is no longer reproducible.

---
 drivers/net/wireless/ath/ath9k/ar9003_eeprom.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

index b4885a700296..b0a4ca3559fd 100644
@@ -3351,7 +3351,8 @@ static int ar9300_eeprom_restore_internal(struct ath_hw *ah,
 			"Found block at %x: code=%d ref=%d length=%d major=%d minor=%d\n",
 			cptr, code, reference, length, major, minor);
 		if ((!AR_SREV_9485(ah) && length >= 1024) ||
-		    (AR_SREV_9485(ah) && length > EEPROM_DATA_LEN_9485)) {
+		    (AR_SREV_9485(ah) && length > EEPROM_DATA_LEN_9485) ||
+		    (length > cptr)) {
 			ath_dbg(common, EEPROM, "Skipping bad header\n");
 			cptr -= COMP_HDR_LEN;
 			continue;

2.30.2

From: Miaoqing Pan <[email protected]>

[ Upstream commit 7c48662b9d56666219f526a71ace8c15e6e12f1f ]

The problem is that gpio_free() can sleep and the cfg_soc() can be called with spinlocks held. One problematic call tree is:

--> ath_reset_internal() takes &sc->sc_pcu_lock spin lock
--> ath9k_hw_reset()

      --> ath9k_hw_gpio_request_in()
         --> ath9k_hw_gpio_request()
            --> ath9k_hw_gpio_cfg_soc()

Remove gpio_free(), use error message instead, so we should make sure there is no GPIO conflict.

Also remove ath9k_hw_gpio_free() from ath9k_hw_apply_gpio_override(), as gpio_mask will never be set for SOC chips.

---
 drivers/net/wireless/ath/ath9k/hw.c | 12 +++++++-----
 1 file changed, 7 insertions(+), 5 deletions(-)

index 9fd8e64288ff..7e2e22b6bbbc 100644
@@ -1622,7 +1622,6 @@ static void ath9k_hw_apply_gpio_override(struct ath_hw *ah)
 		ath9k_hw_gpio_request_out(ah, i, NULL,
 					  AR_GPIO_OUTPUT_MUX_AS_OUTPUT);
 		ath9k_hw_set_gpio(ah, i, !!(ah->gpio_val & BIT(i)));
-		ath9k_hw_gpio_free(ah, i);
 	}
 }
 

@@ -2730,14 +2729,17 @@ static void ath9k_hw_gpio_cfg_output_mux(struct ath_hw *ah, u32 gpio, u32 type)
 static void ath9k_hw_gpio_cfg_soc(struct ath_hw *ah, u32 gpio, bool out,
 				  const char *label)
 {
+	int err;
+
 	if (ah->caps.gpio_requested & BIT(gpio))
 		return;
 
-	/* may be requested by BSP, free anyway */
-	gpio_free(gpio);
-
-	if (gpio_request_one(gpio, out ? GPIOF_OUT_INIT_LOW : GPIOF_IN, label))
+	err = gpio_request_one(gpio, out ? GPIOF_OUT_INIT_LOW : GPIOF_IN, label);
+	if (err) {
+		ath_err(ath9k_hw_common(ah), "request GPIO%d failed:%d\n",
+			gpio, err);
 		return;
+	}
 
 	ah->caps.gpio_requested |= BIT(gpio);
 }

2.30.2

From: 王贇 <[email protected]>

[ Upstream commit 733c99ee8be9a1410287cdbb943887365e83b2d6 ]

In netlbl_cipsov4_add_std() when 'doi_def->map.std' alloc failed, we sometime observe panic:

  BUG: kernel NULL pointer dereference, address:
  ...
  RIP: 0010:cipso_v4_doi_free+0x3a/0x80
  ...
  Call Trace:
   netlbl_cipsov4_add_std+0xf4/0x8c0
   netlbl_cipsov4_add+0x13f/0x1b0
   genl_family_rcv_msg_doit.isra.15+0x132/0x170
   genl_rcv_msg+0x125/0x240

This is because in cipso_v4_doi_free() there is no check on 'doi_def->map.std' when 'doi_def->type' equal 1, which is possibe, since netlbl_cipsov4_add_std() haven't initialize it before alloc 'doi_def->map.std'.

This patch just add the check to prevent panic happen for similar cases.

---
 net/netlabel/netlabel_cipso_v4.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

index 8cd3daf0e3db..1778e4e8ce24 100644
@@ -144,8 +144,8 @@ static int netlbl_cipsov4_add_std(struct genl_info *info,
 		return -ENOMEM;
 	doi_def->map.std = kzalloc(sizeof(*doi_def->map.std), GFP_KERNEL);
 	if (doi_def->map.std == NULL) {
-		ret_val = -ENOMEM;
-		goto add_std_failure;
+		kfree(doi_def);
+		return -ENOMEM;
 	}
 	doi_def->type = CIPSO_V4_MAP_TRANS;
 

2.30.2

From: Haimin Zhang <[email protected]>

[ Upstream commit efe487fce3061d94222c6501d7be3aa549b3dc78 ]

syzbot report an array-index-out-of-bounds in taprio_change index 16 is out of range for type '__u16 [16]'
that's because mqprio->num_tc is lager than TC_MAX_QUEUE,so we check the return value of netdev_set_num_tc.

---
 net/sched/sch_taprio.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

index a4de4853c79d..da9ed0613eb7 100644
@@ -1503,7 +1503,9 @@ static int taprio_change(struct Qdisc *sch, struct nlattr *opt,
 	taprio_set_picos_per_byte(dev, q);
 
 	if (mqprio) {
-		netdev_set_num_tc(dev, mqprio->num_tc);
+		err = netdev_set_num_tc(dev, mqprio->num_tc);
+		if (err)
+			goto free_sched;
 		for (i = 0; i < mqprio->num_tc; i++)
 			netdev_set_tc_queue(dev, i,
 					    mqprio->count[i],

2.30.2

From: Yang Yingliang <[email protected]>

[ Upstream commit a39ff4a47f3e1da3b036817ef436b1a9be10783a ]

It will cause null-ptr-deref if platform_get_resource() returns NULL, we need check the return value.

---
 drivers/net/ethernet/wiznet/w5100.c | 2 ++
 1 file changed, 2 insertions(+)

index bede1ff289c5..a65b7291e12a 100644
@@ -1052,6 +1052,8 @@ static int w5100_mmio_probe(struct platform_device *pdev)
 		mac_addr = data->mac_addr;
 
 	mem = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	if (!mem)
+		return -EINVAL;
 	if (resource_size(mem) < W5100_BUS_DIRECT_SIZE)
 		ops = &w5100_mmio_indirect_ops;
 	else

2.30.2

From: Mikulas Patocka <[email protected]>

commit 030f653078316a9cc9ca6bd1b0234dcf858be35d upstream.

I was debugging some crashes on parisc and I found out that there is a crash possibility if a function using alloca is interrupted by a signal. The reason for the crash is that the gcc alloca implementation leaves garbage in the upper 32 bits of the sp register. This normally doesn't matter (the upper bits are ignored because the PSW W-bit is clear), however the signal delivery routine in the kernel uses full 64 bits of sp and it fails with -EFAULT if the upper 32 bits are not zero.

I created this program that demonstrates the problem:

#include <stdlib.h>
#include <unistd.h>
#include <signal.h>
#include <alloca.h>

static __attribute__((noinline,noclone)) void aa(int *size)

{
	void * volatile p = alloca(-*size);
	while (1) ;
}

static void handler(int sig)

{
	write(1, "signal delivered\n", 17);
	_exit(0);
}

int main(void)

{
	int size = -0x100;
	signal(SIGALRM, handler);
	alarm(1);
	aa(&size);
}

If you compile it with optimizations, it will crash. The "aa" function has this disassembly:

000106a0 <aa>:

   106a0:       08 03 02 41     copy r3,r1
   106a4:       08 1e 02 43     copy sp,r3
   106a8:       6f c1 00 80     stw,ma r1,40(sp)
   106ac:       37 dc 3f c1     ldo -20(sp),ret0
   106b0:       0c 7c 12 90     stw ret0,8(r3)
   106b4:       0f 40 10 9c     ldw 0(r26),ret0		; ret0 = 0x00000000FFFFFF00
   106b8:       97 9c 00 7e     subi 3f,ret0,ret0	; ret0 = 0xFFFFFFFF0000013F
   106bc:       d7 80 1c 1a     depwi 0,31,6,ret0	; ret0 = 0xFFFFFFFF00000100
   106c0:       0b 9e 0a 1e     add,l sp,ret0,sp	;   sp = 0xFFFFFFFFxxxxxxxx
   106c4:       e8 1f 1f f7     b,l,n 106c4 <aa+0x24>,r0

This patch fixes the bug by truncating the "usp" variable to 32 bits.

---
 arch/parisc/kernel/signal.c |    6 ++++++
 1 file changed, 6 insertions(+)

@@ -238,6 +238,12 @@ setup_rt_frame(struct ksignal *ksig, sig
 #endif
 	
 	usp = (regs->gr[30] & ~(0x01UL));
+#ifdef CONFIG_64BIT
+	if (is_compat_task()) {
+		/* The gcc alloca implementation leaves garbage in the upper 32 bits of sp */
+		usp = (compat_uint_t)usp;
+	}
+#endif
 	/*FIXME: frame_size parameter is unused, remove it. */
 	frame = get_sigframe(&ksig->ka, usp, sizeof(*frame));
 

From: chenying <[email protected]>

commit 52d5a0c6bd8a89f460243ed937856354f8f253a3 upstream.

If function ovl_instantiate() returns an error, ovl_cleanup will be called and try to remove newdentry from wdir, but the newdentry has been moved to udir at this time. This will causes BUG_ON(victim->d_parent->d_inode != dir) in fs/namei.c:may_delete.

---
 fs/overlayfs/dir.c |    6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

@@ -513,8 +513,10 @@ static int ovl_create_over_whiteout(stru
 			goto out_cleanup;
 	}
 	err = ovl_instantiate(dentry, inode, newdentry, hardlink);
-	if (err)
-		goto out_cleanup;
+	if (err) {
+		ovl_cleanup(udir, newdentry);
+		dput(newdentry);
+	}
 out_dput:
 	dput(upper);
 out_unlock:

From: Maciej W. Rozycki <[email protected]>

commit 44d01fc86d952f5a8b8b32bdb4841504d5833d95 upstream.

Update BusLogic driver's messaging system to use pr_cont() for continuation lines, bringing messy output:

pci 0000:00:13.0: PCI->APIC IRQ transform: INT A -> IRQ 17

scsi: ***** BusLogic SCSI Driver Version 2.1.17 of 12 September 2013 *****
scsi: Copyright 1995-1998 by Leonard N. Zubkoff <[email protected]>
scsi0: Configuring BusLogic Model BT-958 PCI Wide Ultra SCSI Host Adapter
scsi0:   Firmware Version: 5.07B, I/O Address: 0x7000, IRQ Channel: 17/Level
scsi0:   PCI Bus: 0, Device: 19, Address:
0xE0012000,
Host Adapter SCSI ID: 7
scsi0:   Parity Checking: Enabled, Extended Translation: Enabled
scsi0:   Synchronous Negotiation: Ultra, Wide Negotiation: Enabled
scsi0:   Disconnect/Reconnect: Enabled, Tagged Queuing: Enabled
scsi0:   Scatter/Gather Limit: 128 of 8192 segments, Mailboxes: 211
scsi0:   Driver Queue Depth: 211, Host Adapter Queue Depth: 192
scsi0:   Tagged Queue Depth:

Automatic
, Untagged Queue Depth: 3
scsi0: SCSI Bus Termination: Both Enabled
, SCAM: Disabled

scsi0: *** BusLogic BT-958 Initialized Successfully *** scsi host0: BusLogic BT-958

back to order:

pci 0000:00:13.0: PCI->APIC IRQ transform: INT A -> IRQ 17

scsi: ***** BusLogic SCSI Driver Version 2.1.17 of 12 September 2013 *****
scsi: Copyright 1995-1998 by Leonard N. Zubkoff <[email protected]>
scsi0: Configuring BusLogic Model BT-958 PCI Wide Ultra SCSI Host Adapter
scsi0:   Firmware Version: 5.07B, I/O Address: 0x7000, IRQ Channel: 17/Level
scsi0:   PCI Bus: 0, Device: 19, Address: 0xE0012000, Host Adapter SCSI ID: 7
scsi0:   Parity Checking: Enabled, Extended Translation: Enabled
scsi0:   Synchronous Negotiation: Ultra, Wide Negotiation: Enabled
scsi0:   Disconnect/Reconnect: Enabled, Tagged Queuing: Enabled
scsi0:   Scatter/Gather Limit: 128 of 8192 segments, Mailboxes: 211
scsi0:   Driver Queue Depth: 211, Host Adapter Queue Depth: 192
scsi0:   Tagged Queue Depth: Automatic, Untagged Queue Depth: 3
scsi0:   SCSI Bus Termination: Both Enabled, SCAM: Disabled
scsi0: *** BusLogic BT-958 Initialized Successfully ***
scsi host0: BusLogic BT-958

Also diagnostic output such as with the BusLogic=TraceConfiguration parameter is affected and becomes vertical and therefore hard to read. This has now been corrected, e.g.:

pci 0000:00:13.0: PCI->APIC IRQ transform: INT A -> IRQ 17

blogic_cmd(86) Status = 30:  4 ==>  4: FF 05 93 00
blogic_cmd(95) Status = 28: (Modify I/O Address)
blogic_cmd(91) Status = 30:  1 ==>  1: 01
blogic_cmd(04) Status = 30:  4 ==>  4: 41 41 35 30
blogic_cmd(8D) Status = 30: 14 ==> 14: 45 DC 00 20 00 00 00 00 00 40 30 37 42 1D
scsi: ***** BusLogic SCSI Driver Version 2.1.17 of 12 September 2013 *****
scsi: Copyright 1995-1998 by Leonard N. Zubkoff <[email protected]>
blogic_cmd(04) Status = 30:  4 ==>  4: 41 41 35 30
blogic_cmd(0B) Status = 30:  3 ==>  3: 00 08 07
blogic_cmd(0D) Status = 30: 34 ==> 34: 03 01 07 04 00 00 00 00 00 00 00 00 00 00 00 00 FF 42 44 46 FF 00 00 00 00 00 00 00 00 00 FF 00 FF 00
blogic_cmd(8D) Status = 30: 14 ==> 14: 45 DC 00 20 00 00 00 00 00 40 30 37 42 1D
blogic_cmd(84) Status = 30:  1 ==>  1: 37
blogic_cmd(8B) Status = 30:  5 ==>  5: 39 35 38 20 20
blogic_cmd(85) Status = 30:  1 ==>  1: 42
blogic_cmd(86) Status = 30:  4 ==>  4: FF 05 93 00
blogic_cmd(91) Status = 30: 64 ==> 64: 41 46 3E 20 39 35 38 20 20 00 C4 00 04 01 07 2F 07 04 35 FF FF FF FF FF FF FF FF FF FF 01 00 FE FF 08 FF FF 00 00 00 00 00 00 00 01 00 01 00 00 FF FF 00 00 00 00 00 00 00 00 00 00 00 00 00 FC
scsi0: Configuring BusLogic Model BT-958 PCI Wide Ultra SCSI Host Adapter

etc.

---
 drivers/scsi/BusLogic.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

@@ -3601,7 +3601,7 @@ static void blogic_msg(enum blogic_msgle
 			if (buf[0] != '\n' || len > 1)
 				printk("%sscsi%d: %s", blogic_msglevelmap[msglevel], adapter->host_no, buf);
 		} else
-			printk("%s", buf);
+			pr_cont("%s", buf);
 	} else {
 		if (begin) {
 			if (adapter != NULL && adapter->adapter_initd)

@@ -3609,7 +3609,7 @@ static void blogic_msg(enum blogic_msgle
 			else
 				printk("%s%s", blogic_msglevelmap[msglevel], buf);
 		} else
-			printk("%s", buf);
+			pr_cont("%s", buf);
 	}
 	begin = (buf[len - 1] == '\n');
 }

From: Saurav Kashyap <[email protected]>

commit 62e0dec59c1e139dab55aff5aa442adc97804271 upstream.

Avoid allocating firmware dump and only allocate a single queue for a kexec kernel.

---
 drivers/scsi/qla2xxx/qla_os.c |    6 ++++++
 1 file changed, 6 insertions(+)

@@ -15,6 +15,7 @@
 #include <linux/slab.h>
 #include <linux/blk-mq-pci.h>
 #include <linux/refcount.h>
+#include <linux/crash_dump.h>
 
 #include <scsi/scsi_tcq.h>
 #include <scsi/scsicam.h>

@@ -2799,6 +2800,11 @@ qla2x00_probe_one(struct pci_dev *pdev,
 			return ret;
 	}
 
+	if (is_kdump_kernel()) {
+		ql2xmqsupport = 0;
+		ql2xallocfwdump = 0;
+	}
+
 	/* This may fail but that's ok */
 	pci_enable_pcie_error_reporting(pdev);
 

From: Saurav Kashyap <[email protected]>

commit c8fadf019964d0eb1da410ba8b629494d3339db9 upstream.

The first invocation of function find_first_zero_bit will return 0 and queue_id gets set to 0.

An index of queue_pair_map also gets set to 0.

	qpair_id = find_first_zero_bit(ha->qpair_qid_map, ha->max_qpairs);

        set_bit(qpair_id, ha->qpair_qid_map);
        ha->queue_pair_map[qpair_id] = qpair;

In the alloc_queue callback driver checks the map, if queue is already allocated:

	ha->queue_pair_map[qidx]

This works fine as long as max_qpairs is greater than nvme_max_hw_queues(8) since the size of the queue_pair_map is equal to max_qpair. In case nr_cpus is less than 8, max_qpairs is less than 8. This creates wrong value returned as qpair.

[ 1572.353669] qla2xxx [0000:24:00.3]-2121:6: Returning existing qpair of 4e00000000000000 for idx=2
[ 1572.354458] general protection fault: 0000 [#1] SMP PTI
[ 1572.354461] CPU: 1 PID: 44 Comm: kworker/1:1H Kdump: loaded Tainted: G          IOE    --------- -  - 4.18.0-304.el8.x86_64 #1
[ 1572.354462] Hardware name: HP ProLiant DL380p Gen8, BIOS P70 03/01/2013
[ 1572.354467] Workqueue: kblockd blk_mq_run_work_fn
[ 1572.354485] RIP: 0010:qla_nvme_post_cmd+0x92/0x760 [qla2xxx]
[ 1572.354486] Code: 84 24 5c 01 00 00 00 00 b8 0a 74 1e 66 83 79 48 00 0f 85 a8 03 00 00 48 8b 44 24 08 48 89 ee 4c 89 e7 8b 50 24 e8 5e 8e 00 00 <f0> 41 ff 47 04 0f ae f0 41 f6 47 24 04 74 19 f0 41 ff 4f 04 b8 f0
[ 1572.354487] RSP: 0018:ffff9c81c645fc90 EFLAGS: 00010246
[ 1572.354489] RAX: 0000000000000001 RBX: ffff8ea3e5070138 RCX: 0000000000000001
[ 1572.354490] RDX: 0000000000000001 RSI: 0000000000000001 RDI: ffff8ea4c866b800
[ 1572.354491] RBP: ffff8ea4c866b800 R08: 0000000000005010 R09: ffff8ea4c866b800
[ 1572.354492] R10: 0000000000000001 R11: 000000069d1ca3ff R12: ffff8ea4bc460000
[ 1572.354493] R13: ffff8ea3e50702b0 R14: ffff8ea4c4c16a58 R15: 4e00000000000000
[ 1572.354494] FS:  0000000000000000(0000) GS:ffff8ea4dfd00000(0000) knlGS:0000000000000000
[ 1572.354495] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 1572.354496] CR2: 000055884504fa58 CR3: 00000005a1410001 CR4: 00000000000606e0
[ 1572.354497] Call Trace:
[ 1572.354503]  ? check_preempt_curr+0x62/0x90
[ 1572.354506]  ? dma_direct_map_sg+0x72/0x1f0
[ 1572.354509]  ? nvme_fc_start_fcp_op.part.32+0x175/0x460 [nvme_fc]
[ 1572.354511]  ? blk_mq_dispatch_rq_list+0x11c/0x730
[ 1572.354515]  ? __switch_to_asm+0x35/0x70
[ 1572.354516]  ? __switch_to_asm+0x41/0x70
[ 1572.354518]  ? __switch_to_asm+0x35/0x70
[ 1572.354519]  ? __switch_to_asm+0x41/0x70
[ 1572.354521]  ? __switch_to_asm+0x35/0x70
[ 1572.354522]  ? __switch_to_asm+0x41/0x70
[ 1572.354523]  ? __switch_to_asm+0x35/0x70
[ 1572.354525]  ? entry_SYSCALL_64_after_hwframe+0xb9/0xca
[ 1572.354527]  ? __switch_to_asm+0x41/0x70
[ 1572.354529]  ? __blk_mq_sched_dispatch_requests+0xc6/0x170
[ 1572.354531]  ? blk_mq_sched_dispatch_requests+0x30/0x60
[ 1572.354532]  ? __blk_mq_run_hw_queue+0x51/0xd0
[ 1572.354535]  ? process_one_work+0x1a7/0x360
[ 1572.354537]  ? create_worker+0x1a0/0x1a0
[ 1572.354538]  ? worker_thread+0x30/0x390
[ 1572.354540]  ? create_worker+0x1a0/0x1a0
[ 1572.354541]  ? kthread+0x116/0x130
[ 1572.354543]  ? kthread_flush_work_fn+0x10/0x10
[ 1572.354545]  ? ret_from_fork+0x35/0x40

Fix is to use index 0 for admin and first IO queue.

---
 drivers/scsi/qla2xxx/qla_nvme.c |    5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

@@ -84,8 +84,9 @@ static int qla_nvme_alloc_queue(struct n
 	struct qla_hw_data *ha;
 	struct qla_qpair *qpair;
 
-	if (!qidx)
-		qidx++;
+	/* Map admin queue and 1st IO queue to index 0 */
+	if (qidx)
+		qidx--;
 
 	vha = (struct scsi_qla_host *)lport->private;
 	ha = vha->hw;

From: Pratik R. Sampat <[email protected]>

commit f34ee9cb2c5ac5af426fee6fa4591a34d187e696 upstream.

In the numa=off kernel command-line configuration init_chip_info() loops around the number of chips and attempts to copy the cpumask of that node which is NULL for all iterations after the first chip.

Hence, store the cpu mask for each chip instead of derving cpumask from node while populating the "chips" struct array and copy that to the chips[i].mask

---
 drivers/cpufreq/powernv-cpufreq.c |   16 ++++++++++++++--
 1 file changed, 14 insertions(+), 2 deletions(-)

@@ -36,6 +36,7 @@
 #define MAX_PSTATE_SHIFT	32
 #define LPSTATE_SHIFT		48
 #define GPSTATE_SHIFT		56
+#define MAX_NR_CHIPS		32
 
 #define MAX_RAMP_DOWN_TIME				5120
 /*

@@ -1050,12 +1051,20 @@ static int init_chip_info(void)
 	unsigned int *chip;
 	unsigned int cpu, i;
 	unsigned int prev_chip_id = UINT_MAX;
+	cpumask_t *chip_cpu_mask;
 	int ret = 0;
 
 	chip = kcalloc(num_possible_cpus(), sizeof(*chip), GFP_KERNEL);
 	if (!chip)
 		return -ENOMEM;
 
+	/* Allocate a chip cpu mask large enough to fit mask for all chips */
+	chip_cpu_mask = kcalloc(MAX_NR_CHIPS, sizeof(cpumask_t), GFP_KERNEL);
+	if (!chip_cpu_mask) {
+		ret = -ENOMEM;
+		goto free_and_return;
+	}
+
 	for_each_possible_cpu(cpu) {
 		unsigned int id = cpu_to_chip_id(cpu);
 

@@ -1063,22 +1072,25 @@ static int init_chip_info(void)
 			prev_chip_id = id;
 			chip[nr_chips++] = id;
 		}
+		cpumask_set_cpu(cpu, &chip_cpu_mask[nr_chips-1]);
 	}
 
 	chips = kcalloc(nr_chips, sizeof(struct chip), GFP_KERNEL);
 	if (!chips) {
 		ret = -ENOMEM;
-		goto free_and_return;
+		goto out_free_chip_cpu_mask;
 	}
 
 	for (i = 0; i < nr_chips; i++) {
 		chips[i].id = chip[i];
-		cpumask_copy(&chips[i].mask, cpumask_of_node(chip[i]));
+		cpumask_copy(&chips[i].mask, &chip_cpu_mask[i]);
 		INIT_WORK(&chips[i].throttle, powernv_cpufreq_work_fn);
 		for_each_cpu(cpu, &chips[i].mask)
 			per_cpu(chip_info, cpu) =  &chips[i];
 	}
 
+out_free_chip_cpu_mask:
+	kfree(chip_cpu_mask);
 free_and_return:
 	kfree(chip);
 	return ret;

From: Halil Pasic <[email protected]>

commit 93ebb6828723b8aef114415c4dc3518342f7dcad upstream.

Since commit 903cd0f315fe ("swiotlb: Use is_swiotlb_force_bounce for swiotlb data bouncing") if code sets swiotlb_force it needs to do so before the swiotlb is initialised. Otherwise io_tlb_default_mem->force_bounce will not get set to true, and devices that use (the default) swiotlb will not bounce despite switolb_force having the value of SWIOTLB_FORCE.

Let us restore swiotlb functionality for PV by fulfilling this new requirement.

This change addresses what turned out to be a fragility in commit 64e1f0c531d1 ("s390/mm: force swiotlb for protected virtualization"), which ain't exactly broken in its original context, but could give us some more headache if people backport the broken change and forget this fix.

---
 arch/s390/mm/init.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -168,9 +168,9 @@ static void pv_init(void)
 		return;
 
 	/* make sure bounce buffers are shared */
+	swiotlb_force = SWIOTLB_FORCE;
 	swiotlb_init(1);
 	swiotlb_update_mem_attributes();
-	swiotlb_force = SWIOTLB_FORCE;
 }
 
 void __init mem_init(void)

From: Liu Zixian <[email protected]>

commit 13db8c50477d83ad3e3b9b0ae247e5cd833a7ae4 upstream.

After fork, the child process will get incorrect (2x) hugetlb_usage. If a process uses 5 2MB hugetlb pages in an anonymous mapping,

	HugetlbPages:	   10240 kB

and then forks, the child will show,

	HugetlbPages:	   20480 kB

The reason for double the amount is because hugetlb_usage will be copied from the parent and then increased when we copy page tables from parent to child. Child will have 2x actual usage.

Fix this by adding hugetlb_count_init in mm_init.

---
 include/linux/hugetlb.h |    9 +++++++++
 kernel/fork.c           |    1 +
 2 files changed, 10 insertions(+)

@@ -542,6 +542,11 @@ static inline spinlock_t *huge_pte_lockp
 
 void hugetlb_report_usage(struct seq_file *m, struct mm_struct *mm);
 
+static inline void hugetlb_count_init(struct mm_struct *mm)
+{
+	atomic_long_set(&mm->hugetlb_usage, 0);
+}
+
 static inline void hugetlb_count_add(long l, struct mm_struct *mm)
 {
 	atomic_long_add(l, &mm->hugetlb_usage);

@@ -711,6 +716,10 @@ static inline spinlock_t *huge_pte_lockp
 	return &mm->page_table_lock;
 }
 
+static inline void hugetlb_count_init(struct mm_struct *mm)
+{
+}
+
 static inline void hugetlb_report_usage(struct seq_file *f, struct mm_struct *m)
 {
 }

@@ -1028,6 +1028,7 @@ static struct mm_struct *mm_init(struct
 	mm->pmd_huge_pte = NULL;
 #endif
 	mm_init_uprobes_state(mm);
+	hugetlb_count_init(mm);
 
 	if (current->mm) {
 		mm->flags = current->mm->flags & MMF_INIT_MASK;

From: Rik van Riel <[email protected]>

commit 32d4f4b782bb8f0ceb78c6b5dc46eb577ae25bf7 upstream.

Commit f56ce412a59d ("mm: memcontrol: fix occasional OOMs due to proportional memory.low reclaim") introduced a divide by zero corner case when oomd is being used in combination with cgroup memory.low protection.

When oomd decides to kill a cgroup, it will force the cgroup memory to be reclaimed after killing the tasks, by writing to the memory.max file for that cgroup, forcing the remaining page cache and reclaimable slab to be reclaimed down to zero.

Previously, on cgroups with some memory.low protection that would result in the memory being reclaimed down to the memory.low limit, or likely not at all, having the page cache reclaimed asynchronously later.

With f56ce412a59d the oomd write to memory.max tries to reclaim all the way down to zero, which may race with another reclaimer, to the point of ending up with the divide by zero below.

This patch implements the obvious fix.

---
 mm/vmscan.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -2513,7 +2513,7 @@ out:
 			cgroup_size = max(cgroup_size, protection);
 
 			scan = lruvec_size - lruvec_size * protection /
-				cgroup_size;
+				(cgroup_size + 1);
 
 			/*
 			 * Minimally target SWAP_CLUSTER_MAX pages to keep

From: Vasily Averin <[email protected]>

commit fab827dbee8c2e06ca4ba000fa6c48bcf9054aba upstream.

Commit 5d097056c9a0 ("kmemcg: account certain kmem allocations to memcg") enabled memcg accounting for pids allocated from init_pid_ns.pid_cachep, but forgot to adjust the setting for nested pid namespaces. As a result, pid memory is not accounted exactly where it is really needed, inside memcg-limited containers with their own pid namespaces.

Pid was one the first kernel objects enabled for memcg accounting. init_pid_ns.pid_cachep marked by SLAB_ACCOUNT and we can expect that any new pids in the system are memcg-accounted.

Though recently I've noticed that it is wrong. nested pid namespaces creates own slab caches for pid objects, nested pids have increased size because contain id both for all parent and for own pid namespaces. The problem is that these slab caches are _NOT_ marked by SLAB_ACCOUNT, as a result any pids allocated in nested pid namespaces are not memcg-accounted.

Pid struct in nested pid namespace consumes up to 500 bytes memory, 100000 such objects gives us up to ~50Mb unaccounted memory, this allow container to exceed assigned memcg limits.

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

@@ -53,7 +53,8 @@ static struct kmem_cache *create_pid_cac
 	mutex_lock(&pid_caches_mutex);
 	/* Name collision forces to do allocation under mutex. */
 	if (!*pkc)
-		*pkc = kmem_cache_create(name, len, 0, SLAB_HWCACHE_ALIGN, 0);
+		*pkc = kmem_cache_create(name, len, 0,
+					 SLAB_HWCACHE_ALIGN | SLAB_ACCOUNT, 0);
 	mutex_unlock(&pid_caches_mutex);
 	/* current can fail, but someone else can succeed. */
 	return READ_ONCE(*pkc);

From: Patryk Duda <[email protected]>

commit 3abc16af57c9939724df92fcbda296b25cc95168 upstream.

Sometimes kernel is trying to probe Fingerprint MCU (FPMCU) when it hasn't initialized SPI yet. This can happen because FPMCU is restarted during system boot and kernel can send message in short window eg. between sysjump to RW and SPI initialization.

---
 drivers/platform/chrome/cros_ec_proto.c |    9 +++++++++
 1 file changed, 9 insertions(+)

@@ -213,6 +213,15 @@ static int cros_ec_host_command_proto_qu
 	msg->insize = sizeof(struct ec_response_get_protocol_info);
 
 	ret = send_command(ec_dev, msg);
+	/*
+	 * Send command once again when timeout occurred.
+	 * Fingerprint MCU (FPMCU) is restarted during system boot which
+	 * introduces small window in which FPMCU won't respond for any
+	 * messages sent by kernel. There is no need to wait before next
+	 * attempt because we waited at least EC_MSG_DEADLINE_MS.
+	 */
+	if (ret == -ETIMEDOUT)
+		ret = send_command(ec_dev, msg);
 
 	if (ret < 0) {
 		dev_dbg(ec_dev->dev,

From: Kees Cook <[email protected]>

commit f9398f15605a50110bf570aaa361163a85113dd1 upstream.

The static initializer test got accidentally converted to a dynamic initializer. Fix this and retain the giant padding hole without using an aligned struct member.

---
 lib/test_stackinit.c |   20 +++++++-------------
 1 file changed, 7 insertions(+), 13 deletions(-)

@@ -67,10 +67,10 @@ static bool range_contains(char *haystac
 #define INIT_STRUCT_none		/**/
 #define INIT_STRUCT_zero		= { }
 #define INIT_STRUCT_static_partial	= { .two = 0, }
-#define INIT_STRUCT_static_all		= { .one = arg->one,		\
-					    .two = arg->two,		\
-					    .three = arg->three,	\
-					    .four = arg->four,		\
+#define INIT_STRUCT_static_all		= { .one = 0,			\
+					    .two = 0,			\
+					    .three = 0,			\
+					    .four = 0,			\
 					}
 #define INIT_STRUCT_dynamic_partial	= { .two = arg->two, }
 #define INIT_STRUCT_dynamic_all		= { .one = arg->one,		\

@@ -84,8 +84,7 @@ static bool range_contains(char *haystac
 					var.one = 0;			\
 					var.two = 0;			\
 					var.three = 0;			\
-					memset(&var.four, 0,		\
-					       sizeof(var.four))
+					var.four = 0
 
 /*
  * @name: unique string name for the test

@@ -208,18 +207,13 @@ struct test_small_hole {
 	unsigned long four;
 };
 
-/* Try to trigger unhandled padding in a structure. */
-struct test_aligned {
-	u32 internal1;
-	u64 internal2;
-} __aligned(64);
-
+/* Trigger unhandled padding in a structure. */
 struct test_big_hole {
 	u8 one;
 	u8 two;
 	u8 three;
 	/* 61 byte padding hole here. */
-	struct test_aligned four;
+	u8 four __aligned(64);
 } __aligned(64);
 
 struct test_trailing_hole {

From: Jan Hoffmann <[email protected]>

commit 552799f8b3b0074d2617f53a63a088f9514a66e3 upstream.

Currently, outgoing packets larger than 1496 bytes are dropped when tagged VLAN is used on a switch port.

Add the frame check sequence length to the value of the register GSWIP_MAC_FLEN to fix this. This matches the lantiq_ppa vendor driver, which uses a value consisting of 1518 bytes for the MAC frame, plus the lengths of special tag and VLAN tags.

---
 drivers/net/dsa/lantiq_gswip.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

@@ -837,7 +837,8 @@ static int gswip_setup(struct dsa_switch
 
 	gswip_switch_mask(priv, 0, GSWIP_MAC_CTRL_2_MLEN,
 			  GSWIP_MAC_CTRL_2p(cpu_port));
-	gswip_switch_w(priv, VLAN_ETH_FRAME_LEN + 8, GSWIP_MAC_FLEN);
+	gswip_switch_w(priv, VLAN_ETH_FRAME_LEN + 8 + ETH_FCS_LEN,
+		       GSWIP_MAC_FLEN);
 	gswip_switch_mask(priv, 0, GSWIP_BM_QUEUE_GCTRL_GL_MOD,
 			  GSWIP_BM_QUEUE_GCTRL);
 

From: David Heidelberg <[email protected]>

commit cb0927ab80d224c9074f53d1a55b087d12ec5a85 upstream.

Without this fix boot throws NULL ptr exception at msm_dsi_manager_setup_encoder on devices like Nexus 7 2013 (MDP4 v4.4).

---
 drivers/gpu/drm/msm/disp/mdp4/mdp4_kms.c |    4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

@@ -405,6 +405,7 @@ struct msm_kms *mdp4_kms_init(struct drm
 {
 	struct platform_device *pdev = to_platform_device(dev->dev);
 	struct mdp4_platform_config *config = mdp4_get_config(pdev);
+	struct msm_drm_private *priv = dev->dev_private;
 	struct mdp4_kms *mdp4_kms;
 	struct msm_kms *kms = NULL;
 	struct msm_gem_address_space *aspace;

@@ -419,7 +420,8 @@ struct msm_kms *mdp4_kms_init(struct drm
 
 	mdp_kms_init(&mdp4_kms->base, &kms_funcs);
 
-	kms = &mdp4_kms->base.base;
+	priv->kms = &mdp4_kms->base.base;
+	kms = priv->kms;
 
 	mdp4_kms->dev = dev;
 

From: Andrey Grodzovsky <[email protected]>

commit ea7acd7c5967542353430947f3faf699e70602e5 upstream.

With added CPU domain to placement you can have
now 3 placemnts at once.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_object.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -200,7 +200,7 @@ void amdgpu_bo_placement_from_domain(str
 		c++;
 	}
 
-	BUG_ON(c >= AMDGPU_BO_MAX_PLACEMENTS);
+	BUG_ON(c > AMDGPU_BO_MAX_PLACEMENTS);
 
 	placement->num_placement = c;
 	placement->placement = places;

From: Alyssa Rosenzweig <[email protected]>

commit b5fab345654c603c07525100d744498f28786929 upstream.

In lock_region, simplify the calculation of the region_width parameter. This field is the size, but encoded as ceil(log2(size)) - 1. ceil(log2(size)) may be computed directly as fls(size - 1). However, we want to use the 64-bit versions as the amount to lock can exceed 32-bits.

This avoids undefined (and completely wrong) behaviour when locking all memory (size ~0). In this case, the old code would "round up" ~0 to the nearest page, overflowing to 0. Since fls(0) == 0, this would calculate a region width of 10 + 0 = 10. But then the code would shift by (region_width - 11) = -1. As shifting by a negative number is undefined, UBSAN flags the bug. Of course, even if it were defined the behaviour is wrong, instead of locking all memory almost none would get locked.

The new form of the calculation corrects this special case and avoids the undefined behaviour.

---
 drivers/gpu/drm/panfrost/panfrost_mmu.c |   19 +++++--------------
 1 file changed, 5 insertions(+), 14 deletions(-)

@@ -56,21 +56,12 @@ static void lock_region(struct panfrost_
 {
 	u8 region_width;
 	u64 region = iova & PAGE_MASK;
-	/*
-	 * fls returns:
-	 * 1 .. 32
-	 *
-	 * 10 + fls(num_pages)
-	 * results in the range (11 .. 42)
-	 */
-
-	size = round_up(size, PAGE_SIZE);
 
-	region_width = 10 + fls(size >> PAGE_SHIFT);
-	if ((size >> PAGE_SHIFT) != (1ul << (region_width - 11))) {
-		/* not pow2, so must go up to the next pow2 */
-		region_width += 1;
-	}
+	/* The size is encoded as ceil(log2) minus(1), which may be calculated
+	 * with fls. The size must be clamped to hardware bounds.
+	 */
+	size = max_t(u64, size, PAGE_SIZE);
+	region_width = fls64(size - 1) - 1;
 	region |= region_width;
 
 	/* Lock the region that needs to be updated */

From: Alyssa Rosenzweig <[email protected]>

commit a77b58825d7221d4a45c47881c35a47ba003aa73 upstream.

Mali virtual addresses are 48-bit. Use a u64 instead of size_t to ensure we can express the "lock everything" condition as ~0ULL without overflow. This code was silently broken on any platform where a size_t is less than 48-bits; in particular, it was broken on 32-bit armv7 platforms which remain in use with panfrost. (Mainly RK3288)

---
 drivers/gpu/drm/panfrost/panfrost_mmu.c |   12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

@@ -52,7 +52,7 @@ static int write_cmd(struct panfrost_dev
 }
 
 static void lock_region(struct panfrost_device *pfdev, u32 as_nr,
-			u64 iova, size_t size)
+			u64 iova, u64 size)
 {
 	u8 region_width;
 	u64 region = iova & PAGE_MASK;

@@ -72,7 +72,7 @@ static void lock_region(struct panfrost_
 
 
 static int mmu_hw_do_operation_locked(struct panfrost_device *pfdev, int as_nr,
-				      u64 iova, size_t size, u32 op)
+				      u64 iova, u64 size, u32 op)
 {
 	if (as_nr < 0)
 		return 0;

@@ -89,7 +89,7 @@ static int mmu_hw_do_operation_locked(st
 
 static int mmu_hw_do_operation(struct panfrost_device *pfdev,
 			       struct panfrost_mmu *mmu,
-			       u64 iova, size_t size, u32 op)
+			       u64 iova, u64 size, u32 op)
 {
 	int ret;
 

@@ -106,7 +106,7 @@ static void panfrost_mmu_enable(struct p
 	u64 transtab = cfg->arm_mali_lpae_cfg.transtab;
 	u64 memattr = cfg->arm_mali_lpae_cfg.memattr;
 
-	mmu_hw_do_operation_locked(pfdev, as_nr, 0, ~0UL, AS_COMMAND_FLUSH_MEM);
+	mmu_hw_do_operation_locked(pfdev, as_nr, 0, ~0ULL, AS_COMMAND_FLUSH_MEM);
 
 	mmu_write(pfdev, AS_TRANSTAB_LO(as_nr), transtab & 0xffffffffUL);
 	mmu_write(pfdev, AS_TRANSTAB_HI(as_nr), transtab >> 32);

@@ -122,7 +122,7 @@ static void panfrost_mmu_enable(struct p
 
 static void panfrost_mmu_disable(struct panfrost_device *pfdev, u32 as_nr)
 {
-	mmu_hw_do_operation_locked(pfdev, as_nr, 0, ~0UL, AS_COMMAND_FLUSH_MEM);
+	mmu_hw_do_operation_locked(pfdev, as_nr, 0, ~0ULL, AS_COMMAND_FLUSH_MEM);
 
 	mmu_write(pfdev, AS_TRANSTAB_LO(as_nr), 0);
 	mmu_write(pfdev, AS_TRANSTAB_HI(as_nr), 0);

@@ -222,7 +222,7 @@ static size_t get_pgsize(u64 addr, size_
 
 static void panfrost_mmu_flush_range(struct panfrost_device *pfdev,
 				     struct panfrost_mmu *mmu,
-				     u64 iova, size_t size)
+				     u64 iova, u64 size)
 {
 	if (mmu->as < 0)
 		return;

From: Alyssa Rosenzweig <[email protected]>

commit bd7ffbc3ca12629aeb66fb9e28cf42b7f37e3e3b upstream.

When locking a region, we currently clamp to a PAGE_SIZE as the minimum lock region. While this is valid for Midgard, it is invalid for Bifrost, where the minimum locking size is 8x larger than the 4k page size. Add a hardware definition for the minimum lock region size (corresponding to KBASE_LOCK_REGION_MIN_SIZE_LOG2 in kbase) and respect it.

---
 drivers/gpu/drm/panfrost/panfrost_mmu.c  |    2 +-
 drivers/gpu/drm/panfrost/panfrost_regs.h |    2 ++
 2 files changed, 3 insertions(+), 1 deletion(-)

@@ -60,7 +60,7 @@ static void lock_region(struct panfrost_
 	/* The size is encoded as ceil(log2) minus(1), which may be calculated
 	 * with fls. The size must be clamped to hardware bounds.
 	 */
-	size = max_t(u64, size, PAGE_SIZE);
+	size = max_t(u64, size, AS_LOCK_REGION_MIN_SIZE);
 	region_width = fls64(size - 1) - 1;
 	region |= region_width;
 

@@ -318,6 +318,8 @@
 #define AS_FAULTSTATUS_ACCESS_TYPE_READ		(0x2 << 8)
 #define AS_FAULTSTATUS_ACCESS_TYPE_WRITE	(0x3 << 8)
 
+#define AS_LOCK_REGION_MIN_SIZE                 (1ULL << 15)
+
 #define gpu_write(dev, reg, data) writel(data, dev->iomem + reg)
 #define gpu_read(dev, reg) readl(dev->iomem + reg)
 

From: Anand Jain <[email protected]>

commit 6f93e834fa7c5faa0372e46828b4b2a966ac61d7 upstream.

The mount option max_inline ranges from 0 to the sectorsize (which is now equal to page size). But we parse the mount options too early and before the actual sectorsize is read from the superblock. So the upper limit of max_inline is unaware of the actual sectorsize and is limited by the temporary sectorsize 4096, even on a system where the default sectorsize is 64K.

Fix this by reading the superblock sectorsize before the mount option parse.

---
 fs/btrfs/disk-io.c |   45 +++++++++++++++++++++++----------------------
 1 file changed, 23 insertions(+), 22 deletions(-)

@@ -2894,6 +2894,29 @@ int open_ctree(struct super_block *sb,
 	 */
 	fs_info->compress_type = BTRFS_COMPRESS_ZLIB;
 
+	/*
+	 * Flag our filesystem as having big metadata blocks if they are bigger
+	 * than the page size
+	 */
+	if (btrfs_super_nodesize(disk_super) > PAGE_SIZE) {
+		if (!(features & BTRFS_FEATURE_INCOMPAT_BIG_METADATA))
+			btrfs_info(fs_info,
+				"flagging fs with big metadata feature");
+		features |= BTRFS_FEATURE_INCOMPAT_BIG_METADATA;
+	}
+
+	/* Set up fs_info before parsing mount options */
+	nodesize = btrfs_super_nodesize(disk_super);
+	sectorsize = btrfs_super_sectorsize(disk_super);
+	stripesize = sectorsize;
+	fs_info->dirty_metadata_batch = nodesize * (1 + ilog2(nr_cpu_ids));
+	fs_info->delalloc_batch = sectorsize * 512 * (1 + ilog2(nr_cpu_ids));
+
+	/* Cache block sizes */
+	fs_info->nodesize = nodesize;
+	fs_info->sectorsize = sectorsize;
+	fs_info->stripesize = stripesize;
+
 	ret = btrfs_parse_options(fs_info, options, sb->s_flags);
 	if (ret) {
 		err = ret;

@@ -2921,28 +2944,6 @@ int open_ctree(struct super_block *sb,
 		btrfs_info(fs_info, "has skinny extents");
 
 	/*
-	 * flag our filesystem as having big metadata blocks if
-	 * they are bigger than the page size
-	 */
-	if (btrfs_super_nodesize(disk_super) > PAGE_SIZE) {
-		if (!(features & BTRFS_FEATURE_INCOMPAT_BIG_METADATA))
-			btrfs_info(fs_info,
-				"flagging fs with big metadata feature");
-		features |= BTRFS_FEATURE_INCOMPAT_BIG_METADATA;
-	}
-
-	nodesize = btrfs_super_nodesize(disk_super);
-	sectorsize = btrfs_super_sectorsize(disk_super);
-	stripesize = sectorsize;
-	fs_info->dirty_metadata_batch = nodesize * (1 + ilog2(nr_cpu_ids));
-	fs_info->delalloc_batch = sectorsize * 512 * (1 + ilog2(nr_cpu_ids));
-
-	/* Cache block sizes */
-	fs_info->nodesize = nodesize;
-	fs_info->sectorsize = sectorsize;
-	fs_info->stripesize = stripesize;
-
-	/*
 	 * mixed block groups end up with duplicate but slightly offset
 	 * extent buffers for the same range.  It leads to corruptions
 	 */

From: Juergen Gross <[email protected]>

commit f68aa100d815b5b4467fd1c3abbe3b99d65fd028 upstream.

A Xen PV guest doesn't have a legacy RTC device, so reset the legacy RTC flag. Otherwise the following WARN splat will occur at boot:

[    1.333404] WARNING: CPU: 1 PID: 1 at /home/gross/linux/head/drivers/rtc/rtc-mc146818-lib.c:25 mc146818_get_time+0x1be/0x210
[    1.333404] Modules linked in:
[    1.333404] CPU: 1 PID: 1 Comm: swapper/0 Tainted: G        W         5.14.0-rc7-default+ #282
[    1.333404] RIP: e030:mc146818_get_time+0x1be/0x210
[    1.333404] Code: c0 64 01 c5 83 fd 45 89 6b 14 7f 06 83 c5 64 89 6b 14 41 83 ec 01 b8 02 00 00 00 44 89 63 10 5b 5d 41 5c 41 5d 41 5e 41 5f c3 <0f> 0b 48 c7 c7 30 0e ef 82 4c 89 e6 e8 71 2a 24 00 48 c7 c0 ff ff
[    1.333404] RSP: e02b:ffffc90040093df8 EFLAGS: 00010002
[    1.333404] RAX: 00000000000000ff RBX: ffffc90040093e34 RCX: 0000000000000000
[    1.333404] RDX: 0000000000000001 RSI: 0000000000000000 RDI: 000000000000000d
[    1.333404] RBP: ffffffff82ef0e30 R08: ffff888005013e60 R09: 0000000000000000
[    1.333404] R10: ffffffff82373e9b R11: 0000000000033080 R12: 0000000000000200
[    1.333404] R13: 0000000000000000 R14: 0000000000000002 R15: ffffffff82cdc6d4
[    1.333404] FS:  0000000000000000(0000) GS:ffff88807d440000(0000) knlGS:0000000000000000
[    1.333404] CS:  10000e030 DS: 0000 ES: 0000 CR0: 0000000080050033
[    1.333404] CR2: 0000000000000000 CR3: 000000000260a000 CR4: 0000000000050660
[    1.333404] Call Trace:
[    1.333404]  ? wakeup_sources_sysfs_init+0x30/0x30
[    1.333404]  ? rdinit_setup+0x2b/0x2b
[    1.333404]  early_resume_init+0x23/0xa4
[    1.333404]  ? cn_proc_init+0x36/0x36
[    1.333404]  do_one_initcall+0x3e/0x200
[    1.333404]  kernel_init_freeable+0x232/0x28e
[    1.333404]  ? rest_init+0xd0/0xd0
[    1.333404]  kernel_init+0x16/0x120
[    1.333404]  ret_from_fork+0x1f/0x30

---
 arch/x86/xen/enlighten_pv.c |    7 +++++++
 1 file changed, 7 insertions(+)

@@ -1183,6 +1183,11 @@ static void __init xen_dom0_set_legacy_f
 	x86_platform.legacy.rtc = 1;
 }
 
+static void __init xen_domu_set_legacy_features(void)
+{
+	x86_platform.legacy.rtc = 0;
+}
+
 /* First C function to be called on Xen boot */
 asmlinkage __visible void __init xen_start_kernel(void)
 {

@@ -1353,6 +1358,8 @@ asmlinkage __visible void __init xen_sta
 		add_preferred_console("xenboot", 0, NULL);
 		if (pci_xen)
 			x86_init.pci.arch_init = pci_xen_init;
+		x86_platform.set_legacy_features =
+				xen_domu_set_legacy_features;
 	} else {
 		const struct dom0_vga_console_info *info =
 			(void *)((char *)xen_start_info +

From: Adrian Bunk <[email protected]>

commit 52ce14c134a003fee03d8fc57442c05a55b53715 upstream.

This function is called to enable SR-IOV when available, not enabling interfaces without VFs was a regression.

---
 drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -1246,7 +1246,7 @@ int bnx2x_iov_init_one(struct bnx2x *bp,
 
 	/* SR-IOV capability was enabled but there are no VFs*/
 	if (iov->total == 0) {
-		err = -EINVAL;
+		err = 0;
 		goto failed;
 	}
 

From: Mark Brown <[email protected]>

commit e35ac9d0b56e9efefaeeb84b635ea26c2839ea86 upstream.

When we need a buffer for SVE register state we call sve_alloc() to make sure that one is there. In order to avoid repeated allocations and frees we keep the buffer around unless we change vector length and just memset() it to ensure a clean register state. The function that deals with this takes the task to operate on as an argument, however in the case where we do a memset() we initialise using the SVE state size for the current task rather than the task passed as an argument.

This is only an issue in the case where we are setting the register state for a task via ptrace and the task being configured has a different vector length to the task tracing it. In the case where the buffer is larger in the traced process we will leak old state from the traced process to itself, in the case where the buffer is smaller in the traced process we will overflow the buffer and corrupt memory.

---
 arch/arm64/kernel/fpsimd.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -498,7 +498,7 @@ size_t sve_state_size(struct task_struct
 void sve_alloc(struct task_struct *task)
 {
 	if (task->thread.sve_state) {
-		memset(task->thread.sve_state, 0, sve_state_size(current));
+		memset(task->thread.sve_state, 0, sve_state_size(task));
 		return;
 	}
 

From: Juergen Gross <[email protected]>

commit 0560204b360a332c321124dbc5cdfd3364533a74 upstream.

If there is no legacy RTC device, don't try to use it for storing trace data across suspend/resume.

---
 drivers/base/power/trace.c |   10 ++++++++++
 1 file changed, 10 insertions(+)

@@ -13,6 +13,7 @@
 #include <linux/export.h>
 #include <linux/rtc.h>
 #include <linux/suspend.h>
+#include <linux/init.h>
 
 #include <linux/mc146818rtc.h>
 

@@ -165,6 +166,9 @@ void generate_pm_trace(const void *trace
 	const char *file = *(const char **)(tracedata + 2);
 	unsigned int user_hash_value, file_hash_value;
 
+	if (!x86_platform.legacy.rtc)
+		return;
+
 	user_hash_value = user % USERHASH;
 	file_hash_value = hash_string(lineno, file, FILEHASH);
 	set_magic_time(user_hash_value, file_hash_value, dev_hash_value);

@@ -267,6 +271,9 @@ static struct notifier_block pm_trace_nb
 
 static int early_resume_init(void)
 {
+	if (!x86_platform.legacy.rtc)
+		return 0;
+
 	hash_value_early_read = read_magic_time();
 	register_pm_notifier(&pm_trace_nb);
 	return 0;

@@ -277,6 +284,9 @@ static int late_resume_init(void)
 	unsigned int val = hash_value_early_read;
 	unsigned int user, file, dev;
 
+	if (!x86_platform.legacy.rtc)
+		return 0;
+
 	user = val % USERHASH;
 	val = val / USERHASH;
 	file = val % FILEHASH;

From: Evan Quan <[email protected]>

commit 60b78ed088ebe1a872ee1320b6c5ad6ee2c4bd9a upstream.

Some AMD GPUs have built-in USB xHCI and USB Type-C UCSI controllers with power dependencies between the GPU and the other functions as in 6d2e369f0d4c ("PCI: Add NVIDIA GPU multi-function power dependencies").

Add device link support for the AMD integrated USB xHCI and USB Type-C UCSI controllers.

Without this, runtime power management, including GPU resume and temp and fan sensors don't work correctly.

Reported-at: https://gitlab.freedesktop.org/drm/amd/-/issues/1704

---
 drivers/pci/quirks.c |    9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)

@@ -5394,7 +5394,7 @@ DECLARE_PCI_FIXUP_CLASS_FINAL(PCI_VENDOR
 			      PCI_CLASS_MULTIMEDIA_HD_AUDIO, 8, quirk_gpu_hda);
 
 /*
- * Create device link for NVIDIA GPU with integrated USB xHCI Host
+ * Create device link for GPUs with integrated USB xHCI Host
  * controller to VGA.
  */
 static void quirk_gpu_usb(struct pci_dev *usb)

@@ -5403,9 +5403,11 @@ static void quirk_gpu_usb(struct pci_dev
 }
 DECLARE_PCI_FIXUP_CLASS_FINAL(PCI_VENDOR_ID_NVIDIA, PCI_ANY_ID,
 			      PCI_CLASS_SERIAL_USB, 8, quirk_gpu_usb);
+DECLARE_PCI_FIXUP_CLASS_FINAL(PCI_VENDOR_ID_ATI, PCI_ANY_ID,
+			      PCI_CLASS_SERIAL_USB, 8, quirk_gpu_usb);
 
 /*
- * Create device link for NVIDIA GPU with integrated Type-C UCSI controller
+ * Create device link for GPUs with integrated Type-C UCSI controller
  * to VGA. Currently there is no class code defined for UCSI device over PCI
  * so using UNKNOWN class for now and it will be updated when UCSI
  * over PCI gets a class code.

@@ -5418,6 +5420,9 @@ static void quirk_gpu_usb_typec_ucsi(str
 DECLARE_PCI_FIXUP_CLASS_FINAL(PCI_VENDOR_ID_NVIDIA, PCI_ANY_ID,
 			      PCI_CLASS_SERIAL_UNKNOWN, 8,
 			      quirk_gpu_usb_typec_ucsi);
+DECLARE_PCI_FIXUP_CLASS_FINAL(PCI_VENDOR_ID_ATI, PCI_ANY_ID,
+			      PCI_CLASS_SERIAL_UNKNOWN, 8,
+			      quirk_gpu_usb_typec_ucsi);
 
 /*
  * Enable the NVIDIA GPU integrated HDA controller if the BIOS left it

From: Ernst Sjöstrand <[email protected]>

commit 67a44e659888569a133a8f858c8230e9d7aad1d5 upstream.

Seems like newer cards can have even more instances now. Found by UBSAN: array-index-out-of-bounds in drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c:318:29 index 8 is out of range for type 'uint32_t *[8]'

---
 drivers/gpu/drm/amd/amdgpu/amdgpu.h |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -762,7 +762,7 @@ enum amd_hw_ip_block_type {
 	MAX_HWIP
 };
 
-#define HWIP_MAX_INSTANCE	8
+#define HWIP_MAX_INSTANCE	10
 
 struct amd_powerplay {
 	void *pp_handle;

From: Lucas Stach <[email protected]>

commit 78edefc05e41352099ffb8f06f8d9b2d091e29cd upstream.

Being able to have the refcount manipulation in an assignment makes it much easier to parse the code.

---
 drivers/gpu/drm/etnaviv/etnaviv_buffer.c     |    3 +--
 drivers/gpu/drm/etnaviv/etnaviv_gem.c        |    3 +--
 drivers/gpu/drm/etnaviv/etnaviv_gem_submit.c |    3 +--
 drivers/gpu/drm/etnaviv/etnaviv_gpu.c        |    6 ++----
 drivers/gpu/drm/etnaviv/etnaviv_mmu.h        |    4 +++-
 5 files changed, 8 insertions(+), 11 deletions(-)

@@ -397,8 +397,7 @@ void etnaviv_buffer_queue(struct etnaviv
 		if (switch_mmu_context) {
 			struct etnaviv_iommu_context *old_context = gpu->mmu_context;
 
-			etnaviv_iommu_context_get(mmu_context);
-			gpu->mmu_context = mmu_context;
+			gpu->mmu_context = etnaviv_iommu_context_get(mmu_context);
 			etnaviv_iommu_context_put(old_context);
 		}
 

@@ -304,8 +304,7 @@ struct etnaviv_vram_mapping *etnaviv_gem
 		list_del(&mapping->obj_node);
 	}
 
-	etnaviv_iommu_context_get(mmu_context);
-	mapping->context = mmu_context;
+	mapping->context = etnaviv_iommu_context_get(mmu_context);
 	mapping->use = 1;
 
 	ret = etnaviv_iommu_map_gem(mmu_context, etnaviv_obj,

@@ -534,8 +534,7 @@ int etnaviv_ioctl_gem_submit(struct drm_
 		goto err_submit_objects;
 
 	submit->ctx = file->driver_priv;
-	etnaviv_iommu_context_get(submit->ctx->mmu);
-	submit->mmu_context = submit->ctx->mmu;
+	submit->mmu_context = etnaviv_iommu_context_get(submit->ctx->mmu);
 	submit->exec_state = args->exec_state;
 	submit->flags = args->flags;
 

@@ -1307,12 +1307,10 @@ struct dma_fence *etnaviv_gpu_submit(str
 	}
 
 	if (!gpu->mmu_context) {
-		etnaviv_iommu_context_get(submit->mmu_context);
-		gpu->mmu_context = submit->mmu_context;
+		gpu->mmu_context = etnaviv_iommu_context_get(submit->mmu_context);
 		etnaviv_gpu_start_fe_idleloop(gpu);
 	} else {
-		etnaviv_iommu_context_get(gpu->mmu_context);
-		submit->prev_mmu_context = gpu->mmu_context;
+		submit->prev_mmu_context = etnaviv_iommu_context_get(gpu->mmu_context);
 	}
 
 	if (submit->nr_pmrs) {

@@ -105,9 +105,11 @@ void etnaviv_iommu_dump(struct etnaviv_i
 struct etnaviv_iommu_context *
 etnaviv_iommu_context_init(struct etnaviv_iommu_global *global,
 			   struct etnaviv_cmdbuf_suballoc *suballoc);
-static inline void etnaviv_iommu_context_get(struct etnaviv_iommu_context *ctx)
+static inline struct etnaviv_iommu_context *
+etnaviv_iommu_context_get(struct etnaviv_iommu_context *ctx)
 {
 	kref_get(&ctx->refcount);
+	return ctx;
 }
 void etnaviv_iommu_context_put(struct etnaviv_iommu_context *ctx);
 void etnaviv_iommu_restore(struct etnaviv_gpu *gpu,

From: Lucas Stach <[email protected]>

commit cda7532916f7bc860b36a1806cb8352e6f63dacb upstream.

The prev context is the MMU context at the time of the job queueing in hardware. As a job might be queued multiple times due to recovery after a GPU hang, we need to make sure to put the stale prev MMU context from a prior queuing, to avoid the reference and thus the MMU context leaking.

---
 drivers/gpu/drm/etnaviv/etnaviv_gpu.c |    2 ++
 1 file changed, 2 insertions(+)

@@ -1310,6 +1310,8 @@ struct dma_fence *etnaviv_gpu_submit(str
 		gpu->mmu_context = etnaviv_iommu_context_get(submit->mmu_context);
 		etnaviv_gpu_start_fe_idleloop(gpu);
 	} else {
+		if (submit->prev_mmu_context)
+			etnaviv_iommu_context_put(submit->prev_mmu_context);
 		submit->prev_mmu_context = etnaviv_iommu_context_get(gpu->mmu_context);
 	}
 

From: Lucas Stach <[email protected]>

commit 23e0f5a57d0ecec86e1fc82194acd94aede21a46 upstream.

While the DMA frontend can only be active when the MMU context is set, the reverse isn't necessarily true, as the frontend can be stopped while the MMU state is kept. Stop treating mmu_context being set as a indication that the frontend is running and instead add a explicit property.

---
 drivers/gpu/drm/etnaviv/etnaviv_gpu.c |   10 ++++++++--
 drivers/gpu/drm/etnaviv/etnaviv_gpu.h |    1 +
 2 files changed, 9 insertions(+), 2 deletions(-)

@@ -545,6 +545,8 @@ static int etnaviv_hw_reset(struct etnav
 	/* We rely on the GPU running, so program the clock */
 	etnaviv_gpu_update_clock(gpu);
 
+	gpu->fe_running = false;
+
 	return 0;
 }
 

@@ -607,6 +609,8 @@ void etnaviv_gpu_start_fe(struct etnaviv
 			  VIVS_MMUv2_SEC_COMMAND_CONTROL_ENABLE |
 			  VIVS_MMUv2_SEC_COMMAND_CONTROL_PREFETCH(prefetch));
 	}
+
+	gpu->fe_running = true;
 }
 
 static void etnaviv_gpu_start_fe_idleloop(struct etnaviv_gpu *gpu)

@@ -1306,7 +1310,7 @@ struct dma_fence *etnaviv_gpu_submit(str
 		goto out_unlock;
 	}
 
-	if (!gpu->mmu_context) {
+	if (!gpu->fe_running) {
 		gpu->mmu_context = etnaviv_iommu_context_get(submit->mmu_context);
 		etnaviv_gpu_start_fe_idleloop(gpu);
 	} else {

@@ -1530,7 +1534,7 @@ int etnaviv_gpu_wait_idle(struct etnaviv
 
 static int etnaviv_gpu_hw_suspend(struct etnaviv_gpu *gpu)
 {
-	if (gpu->initialized && gpu->mmu_context) {
+	if (gpu->initialized && gpu->fe_running) {
 		/* Replace the last WAIT with END */
 		mutex_lock(&gpu->lock);
 		etnaviv_buffer_end(gpu);

@@ -1545,6 +1549,8 @@ static int etnaviv_gpu_hw_suspend(struct
 
 		etnaviv_iommu_context_put(gpu->mmu_context);
 		gpu->mmu_context = NULL;
+
+		gpu->fe_running = false;
 	}
 
 	gpu->exec_state = -1;

@@ -101,6 +101,7 @@ struct etnaviv_gpu {
 	struct workqueue_struct *wq;
 	struct drm_gpu_scheduler sched;
 	bool initialized;
+	bool fe_running;
 
 	/* 'ring'-buffer: */
 	struct etnaviv_cmdbuf buffer;

From: Lucas Stach <[email protected]>

commit 8f3eea9d01d7b0f95b0fe04187c0059019ada85b upstream.

The MMU state may be kept across a runtime suspend/resume cycle, as we avoid a full hardware reset to keep the latency of the runtime PM small.

Don't pretend that the MMU state is lost in driver state. The MMU context is pushed out when new HW jobs with a different context are coming in. The only exception to this is when the GPU is unbound, in which case we need to make sure to also free the last active context.

---
 drivers/gpu/drm/etnaviv/etnaviv_gpu.c |    6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

@@ -1547,9 +1547,6 @@ static int etnaviv_gpu_hw_suspend(struct
 		 */
 		etnaviv_gpu_wait_idle(gpu, 100);
 
-		etnaviv_iommu_context_put(gpu->mmu_context);
-		gpu->mmu_context = NULL;
-
 		gpu->fe_running = false;
 	}
 

@@ -1698,6 +1695,9 @@ static void etnaviv_gpu_unbind(struct de
 	etnaviv_gpu_hw_suspend(gpu);
 #endif
 
+	if (gpu->mmu_context)
+		etnaviv_iommu_context_put(gpu->mmu_context);
+
 	if (gpu->initialized) {
 		etnaviv_cmdbuf_free(&gpu->buffer);
 		etnaviv_iommu_global_fini(gpu);

From: Lucas Stach <[email protected]>

commit 725cbc7884c37f3b4f1777bc1aea6432cded8ca5 upstream.

When the GPU is reset both the current exec state, as well as all MMU state is lost. Move the driver side state tracking into the reset function to keep hardware and software state from diverging.

---
 drivers/gpu/drm/etnaviv/etnaviv_gpu.c |    5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

@@ -546,6 +546,8 @@ static int etnaviv_hw_reset(struct etnav
 	etnaviv_gpu_update_clock(gpu);
 
 	gpu->fe_running = false;
+	gpu->exec_state = -1;
+	gpu->mmu_context = NULL;
 
 	return 0;
 }

@@ -794,7 +796,6 @@ int etnaviv_gpu_init(struct etnaviv_gpu
 	/* Now program the hardware */
 	mutex_lock(&gpu->lock);
 	etnaviv_gpu_hw_init(gpu);
-	gpu->exec_state = -1;
 	mutex_unlock(&gpu->lock);
 
 	pm_runtime_mark_last_busy(gpu->dev);

@@ -998,8 +999,6 @@ void etnaviv_gpu_recover_hang(struct etn
 	spin_unlock(&gpu->event_spinlock);
 
 	etnaviv_gpu_hw_init(gpu);
-	gpu->exec_state = -1;
-	gpu->mmu_context = NULL;
 
 	mutex_unlock(&gpu->lock);
 	pm_runtime_mark_last_busy(gpu->dev);

From: Lucas Stach <[email protected]>

commit f978a5302f5566480c58ffae64a16d34456801bd upstream.

After a reset the GPU is no longer using the MMU context and may be restarted with a different context. While the mmu_state proeprly was cleared, the context wasn't unreferenced, leading to a memory leak.

---
 drivers/gpu/drm/etnaviv/etnaviv_gpu.c |    2 ++
 1 file changed, 2 insertions(+)

@@ -547,6 +547,8 @@ static int etnaviv_hw_reset(struct etnav
 
 	gpu->fe_running = false;
 	gpu->exec_state = -1;
+	if (gpu->mmu_context)
+		etnaviv_iommu_context_put(gpu->mmu_context);
 	gpu->mmu_context = NULL;
 
 	return 0;

From: Lucas Stach <[email protected]>

commit d6408538f091fb22d47f792d4efa58143d56c3fb upstream.

Move the refcount manipulation of the MMU context to the point where the hardware state is programmed. At that point it is also known if a previous MMU state is still there, or the state needs to be reprogrammed with a potentially different context.

---
 drivers/gpu/drm/etnaviv/etnaviv_gpu.c      |   24 ++++++++++++------------
 drivers/gpu/drm/etnaviv/etnaviv_iommu.c    |    4 ++++
 drivers/gpu/drm/etnaviv/etnaviv_iommu_v2.c |    8 ++++++++
 3 files changed, 24 insertions(+), 12 deletions(-)

@@ -617,17 +617,19 @@ void etnaviv_gpu_start_fe(struct etnaviv
 	gpu->fe_running = true;
 }
 
-static void etnaviv_gpu_start_fe_idleloop(struct etnaviv_gpu *gpu)
+static void etnaviv_gpu_start_fe_idleloop(struct etnaviv_gpu *gpu,
+					  struct etnaviv_iommu_context *context)
 {
-	u32 address = etnaviv_cmdbuf_get_va(&gpu->buffer,
-				&gpu->mmu_context->cmdbuf_mapping);
 	u16 prefetch;
+	u32 address;
 
 	/* setup the MMU */
-	etnaviv_iommu_restore(gpu, gpu->mmu_context);
+	etnaviv_iommu_restore(gpu, context);
 
 	/* Start command processor */
 	prefetch = etnaviv_buffer_init(gpu);
+	address = etnaviv_cmdbuf_get_va(&gpu->buffer,
+					&gpu->mmu_context->cmdbuf_mapping);
 
 	etnaviv_gpu_start_fe(gpu, address, prefetch);
 }

@@ -1311,14 +1313,12 @@ struct dma_fence *etnaviv_gpu_submit(str
 		goto out_unlock;
 	}
 
-	if (!gpu->fe_running) {
-		gpu->mmu_context = etnaviv_iommu_context_get(submit->mmu_context);
-		etnaviv_gpu_start_fe_idleloop(gpu);
-	} else {
-		if (submit->prev_mmu_context)
-			etnaviv_iommu_context_put(submit->prev_mmu_context);
-		submit->prev_mmu_context = etnaviv_iommu_context_get(gpu->mmu_context);
-	}
+	if (!gpu->fe_running)
+		etnaviv_gpu_start_fe_idleloop(gpu, submit->mmu_context);
+
+	if (submit->prev_mmu_context)
+		etnaviv_iommu_context_put(submit->prev_mmu_context);
+	submit->prev_mmu_context = etnaviv_iommu_context_get(gpu->mmu_context);
 
 	if (submit->nr_pmrs) {
 		gpu->event[event[1]].sync_point = &sync_point_perfmon_sample_pre;

@@ -92,6 +92,10 @@ static void etnaviv_iommuv1_restore(stru
 	struct etnaviv_iommuv1_context *v1_context = to_v1_context(context);
 	u32 pgtable;
 
+	if (gpu->mmu_context)
+		etnaviv_iommu_context_put(gpu->mmu_context);
+	gpu->mmu_context = etnaviv_iommu_context_get(context);
+
 	/* set base addresses */
 	gpu_write(gpu, VIVS_MC_MEMORY_BASE_ADDR_RA, context->global->memory_base);
 	gpu_write(gpu, VIVS_MC_MEMORY_BASE_ADDR_FE, context->global->memory_base);

@@ -172,6 +172,10 @@ static void etnaviv_iommuv2_restore_nons
 	if (gpu_read(gpu, VIVS_MMUv2_CONTROL) & VIVS_MMUv2_CONTROL_ENABLE)
 		return;
 
+	if (gpu->mmu_context)
+		etnaviv_iommu_context_put(gpu->mmu_context);
+	gpu->mmu_context = etnaviv_iommu_context_get(context);
+
 	prefetch = etnaviv_buffer_config_mmuv2(gpu,
 				(u32)v2_context->mtlb_dma,
 				(u32)context->global->bad_page_dma);

@@ -192,6 +196,10 @@ static void etnaviv_iommuv2_restore_sec(
 	if (gpu_read(gpu, VIVS_MMUv2_SEC_CONTROL) & VIVS_MMUv2_SEC_CONTROL_ENABLE)
 		return;
 
+	if (gpu->mmu_context)
+		etnaviv_iommu_context_put(gpu->mmu_context);
+	gpu->mmu_context = etnaviv_iommu_context_get(context);
+
 	gpu_write(gpu, VIVS_MMUv2_PTA_ADDRESS_LOW,
 		  lower_32_bits(context->global->v2.pta_dma));
 	gpu_write(gpu, VIVS_MMUv2_PTA_ADDRESS_HIGH,

From: Lucas Stach <[email protected]>

commit f2faea8b64125852fa9acc6771c07fc0311a039b upstream.

When we forcefully evict a mapping from the the address space and thus the MMU context, the MMU context is leaked, as the mapping no longer points to it, so it doesn't get freed when the GEM object is destroyed. Add the mssing context put to fix the leak.

---
 drivers/gpu/drm/etnaviv/etnaviv_mmu.c |    1 +
 1 file changed, 1 insertion(+)

@@ -204,6 +204,7 @@ static int etnaviv_iommu_find_iova(struc
 		 */
 		list_for_each_entry_safe(m, n, &list, scan_node) {
 			etnaviv_iommu_remove_mapping(context, m);
+			etnaviv_iommu_context_put(m->context);
 			m->context = NULL;
 			list_del_init(&m->mmu_node);
 			list_del_init(&m->scan_node);

From: Alexander Egorenkov <[email protected]>

commit d76b14f3971a0638b6cd0da289f8b48acee287d0 upstream.

Prevent out-of-range access if the returned SCLP SCCB response is smaller in size than the address of the Secure-IPL flag.

---
 drivers/s390/char/sclp_early.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

@@ -40,13 +40,14 @@ static void __init sclp_early_facilities
 	sclp.has_gisaf = !!(sccb->fac118 & 0x08);
 	sclp.has_hvs = !!(sccb->fac119 & 0x80);
 	sclp.has_kss = !!(sccb->fac98 & 0x01);
-	sclp.has_sipl = !!(sccb->cbl & 0x4000);
 	if (sccb->fac85 & 0x02)
 		S390_lowcore.machine_flags |= MACHINE_FLAG_ESOP;
 	if (sccb->fac91 & 0x40)
 		S390_lowcore.machine_flags |= MACHINE_FLAG_TLB_GUEST;
 	if (sccb->cpuoff > 134)
 		sclp.has_diag318 = !!(sccb->byte_134 & 0x80);
+	if (sccb->cpuoff > 137)
+		sclp.has_sipl = !!(sccb->cbl & 0x4000);
 	sclp.rnmax = sccb->rnmax ? sccb->rnmax : sccb->rnmax2;
 	sclp.rzm = sccb->rnsize ? sccb->rnsize : sccb->rnsize2;
 	sclp.rzm <<= 20;

From: Mike Rapoport <[email protected]>

commit 34b1999da935a33be6239226bfa6cd4f704c5c88 upstream.

Jiri Olsa reported a fault when running:

  # cat /proc/kallsyms | grep ksys_read
  ffffffff8136d580 T ksys_read
  # objdump -d --start-address=0xffffffff8136d580 --stop-address=0xffffffff8136d590 /proc/kcore

/proc/kcore: file format elf64-x86-64

Segmentation fault

  general protection fault, probably for non-canonical address 0xf887ffcbff000: 0000 [#1] SMP PTI
  CPU: 12 PID: 1079 Comm: objdump Not tainted 5.14.0-rc5qemu+ #508
  Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.14.0-4.fc34 04/01/2014
  RIP: 0010:kern_addr_valid
  Call Trace:
   read_kcore
   ? rcu_read_lock_sched_held
   ? rcu_read_lock_sched_held
   ? rcu_read_lock_sched_held
   ? trace_hardirqs_on
   ? rcu_read_lock_sched_held
   ? lock_acquire
   ? lock_acquire
   ? rcu_read_lock_sched_held
   ? lock_acquire
   ? rcu_read_lock_sched_held
   ? rcu_read_lock_sched_held
   ? rcu_read_lock_sched_held
   ? lock_release
   ? _raw_spin_unlock
   ? __handle_mm_fault
   ? rcu_read_lock_sched_held
   ? lock_acquire
   ? rcu_read_lock_sched_held
   ? lock_release
   proc_reg_read
   ? vfs_read
   vfs_read
   ksys_read
   do_syscall_64
   entry_SYSCALL_64_after_hwframe

The fault happens because kern_addr_valid() dereferences existent but not present PMD in the high kernel mappings.

Such PMDs are created when free_kernel_image_pages() frees regions larger than 2Mb. In this case, a part of the freed memory is mapped with PMDs and the set_memory_np_noalias() -> ... -> __change_page_attr() sequence will mark the PMD as not present rather than wipe it completely.

Have kern_addr_valid() check whether higher level page table entries are present before trying to dereference them to fix this issue and to avoid similar issues in the future.

Stable backporting note:
------------------------

Note that the stable marking is for all active stable branches because there could be cases where pagetable entries exist but are not valid - see 9a14aefc1d28 ("x86: cpa, fix lookup_address"), for example. So make sure to be on the safe side here and use pXY_present() accessors rather than pXY_none() which could #GP when accessing pages in the direct map.

Also see:

c40a56a7818c ("x86/mm/init: Remove freed kernel image areas from alias mapping")

for more info.

---
 arch/x86/mm/init_64.c |    6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

@@ -1355,18 +1355,18 @@ int kern_addr_valid(unsigned long addr)
 		return 0;
 
 	p4d = p4d_offset(pgd, addr);
-	if (p4d_none(*p4d))
+	if (!p4d_present(*p4d))
 		return 0;
 
 	pud = pud_offset(p4d, addr);
-	if (pud_none(*pud))
+	if (!pud_present(*pud))
 		return 0;
 
 	if (pud_large(*pud))
 		return pfn_valid(pud_pfn(*pud));
 
 	pmd = pmd_offset(pud, addr);
-	if (pmd_none(*pmd))
+	if (!pmd_present(*pmd))
 		return 0;
 
 	if (pmd_large(*pmd))

From: Xin Long <[email protected]>

commit cc19862ffe454a5b632ca202e5a51bfec9f89fd2 upstream.

syzbot reported an use-after-free crash:

  BUG: KASAN: use-after-free in tipc_recvmsg+0xf77/0xf90 net/tipc/socket.c:1979
  Call Trace:
   tipc_recvmsg+0xf77/0xf90 net/tipc/socket.c:1979
   sock_recvmsg_nosec net/socket.c:943 [inline]
   sock_recvmsg net/socket.c:961 [inline]
   sock_recvmsg+0xca/0x110 net/socket.c:957
   tipc_conn_rcv_from_sock+0x162/0x2f0 net/tipc/topsrv.c:398
   tipc_conn_recv_work+0xeb/0x190 net/tipc/topsrv.c:421
   process_one_work+0x98d/0x1630 kernel/workqueue.c:2276
   worker_thread+0x658/0x11f0 kernel/workqueue.c:2422

As Hoang pointed out, it was caused by skb_cb->bytes_read still accessed after calling tsk_advance_rx_queue() to free the skb in tipc_recvmsg().

This patch is to fix it by accessing skb_cb->bytes_read earlier than calling tsk_advance_rx_queue().

---
 net/tipc/socket.c |    8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

@@ -1849,10 +1849,12 @@ static int tipc_recvmsg(struct socket *s
 		tipc_node_distr_xmit(sock_net(sk), &xmitq);
 	}
 
-	if (!skb_cb->bytes_read)
-		tsk_advance_rx_queue(sk);
+	if (skb_cb->bytes_read)
+		goto exit;
+
+	tsk_advance_rx_queue(sk);
 
-	if (likely(!connected) || skb_cb->bytes_read)
+	if (likely(!connected))
 		goto exit;
 
 	/* Send connection flow control advertisement when applicable */

From: Eric Dumazet <[email protected]>

commit 550ac9c1aaaaf51fd42e20d461f0b1cdbd55b3d2 upstream.

syszbot triggers this warning, which looks something we can easily prevent.

If we initialize priv->list_field in chnl_net_init(), then always use list_del_init(), we can remove robust_list_del() completely.

WARNING: CPU: 0 PID: 3233 at net/caif/chnl_net.c:67 robust_list_del net/caif/chnl_net.c:67 [inline]
WARNING: CPU: 0 PID: 3233 at net/caif/chnl_net.c:67 chnl_net_uninit+0xc9/0x2e0 net/caif/chnl_net.c:375
Modules linked in:
CPU: 0 PID: 3233 Comm: syz-executor.3 Not tainted 5.14.0-syzkaller #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
RIP: 0010:robust_list_del net/caif/chnl_net.c:67 [inline]
RIP: 0010:chnl_net_uninit+0xc9/0x2e0 net/caif/chnl_net.c:375
Code: 89 eb e8 3a a3 ba f8 48 89 d8 48 c1 e8 03 42 80 3c 28 00 0f 85 bf 01 00 00 48 81 fb 00 14 4e 8d 48 8b 2b 75 d0 e8 17 a3 ba f8 <0f> 0b 5b 5d 41 5c 41 5d e9 0a a3 ba f8 4c 89 e3 e8 02 a3 ba f8 4c
RSP: 0018:ffffc90009067248 EFLAGS: 00010202
RAX: 0000000000008780 RBX: ffffffff8d4e1400 RCX: ffffc9000fd34000
RDX: 0000000000040000 RSI: ffffffff88bb6e49 RDI: 0000000000000003
RBP: ffff88802cd9ee08 R08: 0000000000000000 R09: ffffffff8d0e6647
R10: ffffffff88bb6dc2 R11: 0000000000000000 R12: ffff88803791ae08
R13: dffffc0000000000 R14: 00000000e600ffce R15: ffff888073ed3480
FS:  00007fed10fa0700(0000) GS:ffff8880b9d00000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000001b2c322000 CR3: 00000000164a6000 CR4: 00000000001506e0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Call Trace:
 register_netdevice+0xadf/0x1500 net/core/dev.c:10347
 ipcaif_newlink+0x4c/0x260 net/caif/chnl_net.c:468
 __rtnl_newlink+0x106d/0x1750 net/core/rtnetlink.c:3458
 rtnl_newlink+0x64/0xa0 net/core/rtnetlink.c:3506
 rtnetlink_rcv_msg+0x413/0xb80 net/core/rtnetlink.c:5572
 netlink_rcv_skb+0x153/0x420 net/netlink/af_netlink.c:2504
 netlink_unicast_kernel net/netlink/af_netlink.c:1314 [inline]
 netlink_unicast+0x533/0x7d0 net/netlink/af_netlink.c:1340
 netlink_sendmsg+0x86d/0xdb0 net/netlink/af_netlink.c:1929
 sock_sendmsg_nosec net/socket.c:704 [inline]
 sock_sendmsg+0xcf/0x120 net/socket.c:724
 __sys_sendto+0x21c/0x320 net/socket.c:2036
 __do_sys_sendto net/socket.c:2048 [inline]
 __se_sys_sendto net/socket.c:2044 [inline]
 __x64_sys_sendto+0xdd/0x1b0 net/socket.c:2044
 do_syscall_x64 arch/x86/entry/common.c:50 [inline]
 do_syscall_64+0x35/0xb0 arch/x86/entry/common.c:80
 entry_SYSCALL_64_after_hwframe+0x44/0xae

---
 net/caif/chnl_net.c |   19 +++----------------
 1 file changed, 3 insertions(+), 16 deletions(-)

@@ -53,20 +53,6 @@ struct chnl_net {
 	enum caif_states state;
 };
 
-static void robust_list_del(struct list_head *delete_node)
-{
-	struct list_head *list_node;
-	struct list_head *n;
-	ASSERT_RTNL();
-	list_for_each_safe(list_node, n, &chnl_net_list) {
-		if (list_node == delete_node) {
-			list_del(list_node);
-			return;
-		}
-	}
-	WARN_ON(1);
-}
-
 static int chnl_recv_cb(struct cflayer *layr, struct cfpkt *pkt)
 {
 	struct sk_buff *skb;

@@ -368,6 +354,7 @@ static int chnl_net_init(struct net_devi
 	ASSERT_RTNL();
 	priv = netdev_priv(dev);
 	strncpy(priv->name, dev->name, sizeof(priv->name));
+	INIT_LIST_HEAD(&priv->list_field);
 	return 0;
 }
 

@@ -376,7 +363,7 @@ static void chnl_net_uninit(struct net_d
 	struct chnl_net *priv;
 	ASSERT_RTNL();
 	priv = netdev_priv(dev);
-	robust_list_del(&priv->list_field);
+	list_del_init(&priv->list_field);
 }
 
 static const struct net_device_ops netdev_ops = {

@@ -541,7 +528,7 @@ static void __exit chnl_exit_module(void
 	rtnl_lock();
 	list_for_each_safe(list_node, _tmp, &chnl_net_list) {
 		dev = list_entry(list_node, struct chnl_net, list_field);
-		list_del(list_node);
+		list_del_init(list_node);
 		delete_device(dev);
 	}
 	rtnl_unlock();

From: Randy Dunlap <[email protected]>

commit 7366c23ff492ad260776a3ee1aaabba9fc773a8b upstream.

Building dp83640.c on arch/parisc/ produces a build warning for PAGE0 being redefined. Since the macro is not used in the dp83640 driver, just make it a comment for documentation purposes.

In file included from ../drivers/net/phy/dp83640.c:23:
../drivers/net/phy/dp83640_reg.h:8: warning: "PAGE0" redefined

    8 | #define PAGE0                     0x0000
                 from ../drivers/net/phy/dp83640.c:11:

../arch/parisc/include/asm/page.h:187: note: this is the location of the previous definition
187 | #define PAGE0 ((struct zeropage *)__PAGE_OFFSET)

---
 drivers/net/phy/dp83640_reg.h |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -5,7 +5,7 @@
 #ifndef HAVE_DP83640_REGISTERS
 #define HAVE_DP83640_REGISTERS
 
-#define PAGE0                     0x0000
+/* #define PAGE0                  0x0000 */
 #define PHYCR2                    0x001c /* PHY Control Register 2 */
 
 #define PAGE4                     0x0004

From: Lin, Zhenpeng <[email protected]>

commit d9ea761fdd197351890418acd462c51f241014a7 upstream.

Commit 2677d2067731 ("dccp: don't free ccid2_hc_tx_sock ...") fixed a UAF but reintroduced CVE-2017-6074.

When the sock is cloned, two dccps_hc_tx_ccid will reference to the same ccid. So one can free the ccid object twice from two socks after cloning.

This issue was found by "Hadar Manor" as well and assigned with CVE-2020-16119, which was fixed in Ubuntu's kernel. So here I port the patch from Ubuntu to fix it.

The patch prevents cloned socks from referencing the same ccid.

---
 net/dccp/minisocks.c |    2 ++
 1 file changed, 2 insertions(+)

@@ -94,6 +94,8 @@ struct sock *dccp_create_openreq_child(c
 		newdp->dccps_role	    = DCCP_ROLE_SERVER;
 		newdp->dccps_hc_rx_ackvec   = NULL;
 		newdp->dccps_service_list   = NULL;
+		newdp->dccps_hc_rx_ccid     = NULL;
+		newdp->dccps_hc_tx_ccid     = NULL;
 		newdp->dccps_service	    = dreq->dreq_service;
 		newdp->dccps_timestamp_echo = dreq->dreq_timestamp_echo;
 		newdp->dccps_timestamp_time = dreq->dreq_timestamp_time;

From: Xiyu Yang <[email protected]>

commit 9b6ff7eb666415e1558f1ba8a742f5db6a9954de upstream.

The reference count leak issue may take place in an error handling path. If both conditions of tunnel->version == L2TP_HDR_VER_3 and the return value of l2tp_v3_ensure_opt_in_linear is nonzero, the function would directly jump to label invalid, without decrementing the reference count of the l2tp_session object session increased earlier by l2tp_tunnel_get_session(). This may result in refcount leaks.

Fix this issue by decrease the reference count before jumping to the label invalid.

---
 net/l2tp/l2tp_core.c |    4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

@@ -886,8 +886,10 @@ static int l2tp_udp_recv_core(struct l2t
 	}
 
 	if (tunnel->version == L2TP_HDR_VER_3 &&
-	    l2tp_v3_ensure_opt_in_linear(session, skb, &ptr, &optr))
+	    l2tp_v3_ensure_opt_in_linear(session, skb, &ptr, &optr)) {
+		l2tp_session_dec_refcount(session);
 		goto error;
+	}
 
 	l2tp_recv_common(session, skb, ptr, optr, hdrflags, length);
 	l2tp_session_dec_refcount(session);

From: Florian Fainelli <[email protected]>

commit e3f0cc1a945fcefec0c7c9d9dfd028a51daa1846 upstream.

A number of users have reported that they were not able to get the PHY to successfully link up, especially after commit c36757eb9dee ("net:
phy: consider AN_RESTART status when reading link status") where we stopped reading just BMSR, but we also read BMCR to determine the link status.

Andrius at NetBSD did a wonderful job at debugging the problem and found out that the MDIO bus clock frequency would be incorrectly set back to its default value which would prevent the MDIO bus controller from reading PHY registers properly. Back when we only read BMSR, if we read all 1s, we could falsely indicate a link status, though in general there is a cable plugged in, so this went unnoticed. After a second read of BMCR was added, a wrong read will lead to the inability to determine a link UP condition which is when it started to be visibly broken, even if it was long before that.

The fix consists in restoring the value of the MD_CSR register that was set prior to the MAC reset.

---
 drivers/net/ethernet/rdc/r6040.c |    9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

@@ -119,6 +119,8 @@
 #define PHY_ST		0x8A	/* PHY status register */
 #define MAC_SM		0xAC	/* MAC status machine */
 #define  MAC_SM_RST	0x0002	/* MAC status machine reset */
+#define MD_CSC		0xb6	/* MDC speed control register */
+#define  MD_CSC_DEFAULT	0x0030
 #define MAC_ID		0xBE	/* Identifier register */
 
 #define TX_DCNT		0x80	/* TX descriptor count */

@@ -354,8 +356,9 @@ static void r6040_reset_mac(struct r6040
 {
 	void __iomem *ioaddr = lp->base;
 	int limit = MAC_DEF_TIMEOUT;
-	u16 cmd;
+	u16 cmd, md_csc;
 
+	md_csc = ioread16(ioaddr + MD_CSC);
 	iowrite16(MAC_RST, ioaddr + MCR1);
 	while (limit--) {
 		cmd = ioread16(ioaddr + MCR1);

@@ -367,6 +370,10 @@ static void r6040_reset_mac(struct r6040
 	iowrite16(MAC_SM_RST, ioaddr + MAC_SM);
 	iowrite16(0, ioaddr + MAC_SM);
 	mdelay(5);
+
+	/* Restore MDIO clock frequency */
+	if (md_csc != MD_CSC_DEFAULT)
+		iowrite16(md_csc, ioaddr + MD_CSC);
 }
 
 static void r6040_init_mac_regs(struct net_device *dev)

From: Hoang Le <[email protected]>

commit f4bb62e64c88c93060c051195d3bbba804e56945 upstream.

In tipc_sk_enqueue() we use hardcoded 2 jiffies to extract socket buffer from generic queue to particular socket. The 2 jiffies is too short in case there are other high priority tasks get CPU cycles for multiple jiffies update. As result, no buffer could be enqueued to particular socket.

To solve this, we switch to use constant timeout 20msecs. Then, the function will be expired between 2 jiffies (CONFIG_100HZ) and 20 jiffies (CONFIG_1000HZ).

---
 net/tipc/socket.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -2275,7 +2275,7 @@ static int tipc_sk_backlog_rcv(struct so
 static void tipc_sk_enqueue(struct sk_buff_head *inputq, struct sock *sk,
 			    u32 dport, struct sk_buff_head *xmitq)
 {
-	unsigned long time_limit = jiffies + 2;
+	unsigned long time_limit = jiffies + usecs_to_jiffies(20000);
 	struct sk_buff *skb;
 	unsigned int lim;
 	atomic_t *dcnt;

From: Michael Petlan <[email protected]>

commit 57f0ff059e3daa4e70a811cb1d31a49968262d20 upstream.

It's later supposed to be either a correct address or NULL. Without the initialization, it may contain an undefined value which results in the following segmentation fault:

  # perf top --sort comm -g --ignore-callees=do_idle

terminates with:

  #0  0x00007ffff56b7685 in __strlen_avx2 () from /lib64/libc.so.6
  #1  0x00007ffff55e3802 in strdup () from /lib64/libc.so.6
  #2  0x00005555558cb139 in hist_entry__init (callchain_size=<optimized out>, sample_self=true, template=0x7fffde7fb110, he=0x7fffd801c250) at util/hist.c:489
  #3  hist_entry__new (template=template@entry=0x7fffde7fb110, sample_self=sample_self@entry=true) at util/hist.c:564
  #4  0x00005555558cb4ba in hists__findnew_entry (hists=hists@entry=0x5555561d9e38, entry=entry@entry=0x7fffde7fb110, al=al@entry=0x7fffde7fb420,
      sample_self=sample_self@entry=true) at util/hist.c:657
  #5  0x00005555558cba1b in __hists__add_entry (hists=hists@entry=0x5555561d9e38, al=0x7fffde7fb420, sym_parent=<optimized out>, bi=bi@entry=0x0, mi=mi@entry=0x0,
      sample=sample@entry=0x7fffde7fb4b0, sample_self=true, ops=0x0, block_info=0x0) at util/hist.c:288
  #6  0x00005555558cbb70 in hists__add_entry (sample_self=true, sample=0x7fffde7fb4b0, mi=0x0, bi=0x0, sym_parent=<optimized out>, al=<optimized out>, hists=0x5555561d9e38)
      at util/hist.c:1056
  #7  iter_add_single_cumulative_entry (iter=0x7fffde7fb460, al=<optimized out>) at util/hist.c:1056
  #8  0x00005555558cc8a4 in hist_entry_iter__add (iter=iter@entry=0x7fffde7fb460, al=al@entry=0x7fffde7fb420, max_stack_depth=<optimized out>, arg=arg@entry=0x7fffffff7db0)
      at util/hist.c:1231
  #9  0x00005555557cdc9a in perf_event__process_sample (machine=<optimized out>, sample=0x7fffde7fb4b0, evsel=<optimized out>, event=<optimized out>, tool=0x7fffffff7db0)
      at builtin-top.c:842
  #10 deliver_event (qe=<optimized out>, qevent=<optimized out>) at builtin-top.c:1202
  #11 0x00005555558a9318 in do_flush (show_progress=false, oe=0x7fffffff80e0) at util/ordered-events.c:244
  #12 __ordered_events__flush (oe=oe@entry=0x7fffffff80e0, how=how@entry=OE_FLUSH__TOP, timestamp=timestamp@entry=0) at util/ordered-events.c:323
  #13 0x00005555558a9789 in __ordered_events__flush (timestamp=<optimized out>, how=<optimized out>, oe=<optimized out>) at util/ordered-events.c:339
  #14 ordered_events__flush (how=OE_FLUSH__TOP, oe=0x7fffffff80e0) at util/ordered-events.c:341
  #15 ordered_events__flush (oe=oe@entry=0x7fffffff80e0, how=how@entry=OE_FLUSH__TOP) at util/ordered-events.c:339
  #16 0x00005555557cd631 in process_thread (arg=0x7fffffff7db0) at builtin-top.c:1114
  #17 0x00007ffff7bb817a in start_thread () from /lib64/libpthread.so.0
  #18 0x00007ffff5656dc3 in clone () from /lib64/libc.so.6

If you look at the frame #2, the code is:

488 if (he->srcline) {
489 he->srcline = strdup(he->srcline);
490 if (he->srcline == NULL)
491 goto err_rawdata;
492 }

If he->srcline is not NULL (it is not NULL if it is uninitialized rubbish), it gets strdupped and strdupping a rubbish random string causes the problem.

Also, if you look at the commit 1fb7d06a509e, it adds the srcline property into the struct, but not initializing it everywhere needed.

Committer notes:

Now I see, when using --ignore-callees=do_idle we end up here at line 2189 in add_callchain_ip():

2181 if (al.sym != NULL) {
2182 if (perf_hpp_list.parent && !*parent &&
2183 symbol__match_regex(al.sym, &parent_regex))
2184 *parent = al.sym;
2185 else if (have_ignore_callees && root_al &&
2186 symbol__match_regex(al.sym, &ignore_callees_regex)) {
2187 /* Treat this symbol as the root,
2188 forgetting its callees. */
2189 *root_al = al;
2190 callchain_cursor_reset(cursor);
2191 }
2192 }

And the al that doesn't have the ->srcline field initialized will be copied to the root_al, so then, back to:

1211 int hist_entry_iter__add(struct hist_entry_iter *iter, struct addr_location *al,
1212 int max_stack_depth, void *arg)
1213 {
1214 int err, err2;
1215 struct map *alm = NULL;
1216
1217 if (al)
1218 alm = map__get(al->map);
1219
1220 err = sample__resolve_callchain(iter->sample, &callchain_cursor, &iter->parent,
1221 iter->evsel, al, max_stack_depth);
1222 if (err) {
1223 map__put(alm);
1224 return err;
1225 }
1226
1227 err = iter->ops->prepare_entry(iter, al);
1228 if (err)
1229 goto out;
1230
1231 err = iter->ops->add_single_entry(iter, al);
1232 if (err)
1233 goto out;
1234

That al at line 1221 is what hist_entry_iter__add() (called from sample__resolve_callchain()) saw as 'root_al', and then:

        iter->ops->add_single_entry(iter, al);

will go on with al->srcline with a bogus value, I'll add the above sequence to the cset and apply, thanks!

---
 tools/perf/util/machine.c |    1 +
 1 file changed, 1 insertion(+)

@@ -2020,6 +2020,7 @@ static int add_callchain_ip(struct threa
 
 	al.filtered = 0;
 	al.sym = NULL;
+	al.srcline = NULL;
 	if (!cpumode) {
 		thread__find_cpumode_addr_location(thread, ip, &al);
 	} else {

From: Saeed Mahameed <[email protected]>

commit dfe6fd72b5f1878b16aa2c8603e031bbcd66b96d upstream.

Handle error flow on mlx5_core_alloc_pd() failure, read_fw_strings_work must be canceled.

---
 drivers/net/ethernet/mellanox/mlx5/core/diag/fw_tracer.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

@@ -1005,7 +1005,7 @@ int mlx5_fw_tracer_init(struct mlx5_fw_t
 	err = mlx5_core_alloc_pd(dev, &tracer->buff.pdn);
 	if (err) {
 		mlx5_core_warn(dev, "FWTracer: Failed to allocate PD %d\n", err);
-		return err;
+		goto err_cancel_work;
 	}
 
 	err = mlx5_fw_tracer_create_mkey(tracer);

@@ -1029,6 +1029,7 @@ err_notifier_unregister:
 	mlx5_core_destroy_mkey(dev, &tracer->buff.mkey);
 err_dealloc_pd:
 	mlx5_core_dealloc_pd(dev, tracer->buff.pdn);
+err_cancel_work:
 	cancel_work_sync(&tracer->read_fw_strings_work);
 	return err;
 }

From: Maor Gottlieb <[email protected]>

commit ee27e330a953595903979ffdb84926843595a9fe upstream.

Fixes the below flow of sleeping in atomic context by releasing the RCU lock before calling to free_match_list.

build_match_list() <- disables preempt
-> free_match_list()
-> tree_put_node()

      -> down_write_ref_node() <- take write lock

---
 drivers/net/ethernet/mellanox/mlx5/core/fs_core.c |    5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

@@ -1606,9 +1606,9 @@ static int build_match_list(struct match
 
 		curr_match = kmalloc(sizeof(*curr_match), GFP_ATOMIC);
 		if (!curr_match) {
+			rcu_read_unlock();
 			free_match_list(match_head, ft_locked);
-			err = -ENOMEM;
-			goto out;
+			return -ENOMEM;
 		}
 		if (!tree_get_node(&g->node)) {
 			kfree(curr_match);

@@ -1617,7 +1617,6 @@ static int build_match_list(struct match
 		curr_match->g = g;
 		list_add_tail(&curr_match->list, &match_head->list);
 	}
-out:
 	rcu_read_unlock();
 	return err;
 }

From: Baptiste Lepers <[email protected]>

commit b89a05b21f46150ac10a962aa50109250b56b03b upstream.

In perf_event_addr_filters_apply, the task associated with the event (event->ctx->task) is read using READ_ONCE at the beginning of the function, checked, and then re-read from event->ctx->task, voiding all guarantees of the checks. Reuse the value that was read by READ_ONCE to ensure the consistency of the task struct throughout the function.

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

@@ -9259,7 +9259,7 @@ static void perf_event_addr_filters_appl
 		return;
 
 	if (ifh->nr_file_filters) {
-		mm = get_task_mm(event->ctx->task);
+		mm = get_task_mm(task);
 		if (!mm)
 			goto restart;
 

From: Paolo Abeni <[email protected]>

commit 3c4cea8fa7f71f00c5279547043a84bc2a4d8b8c upstream.

If the sendmsg() call in vhost_tx_batch() fails, both the 'batched_xdp' and 'done_idx' indexes are left unchanged. If such failure happens when batched_xdp == VHOST_NET_BATCH, the next call to vhost_net_build_xdp() will access and write memory outside the xdp buffers area.

Since sendmsg() can only error with EBADFD, this change addresses the issue explicitly freeing the XDP buffers batch on error.

---
 drivers/vhost/net.c |   11 ++++++++++-
 1 file changed, 10 insertions(+), 1 deletion(-)

@@ -466,7 +466,7 @@ static void vhost_tx_batch(struct vhost_
 		.num = nvq->batched_xdp,
 		.ptr = nvq->xdp,
 	};
-	int err;
+	int i, err;
 
 	if (nvq->batched_xdp == 0)
 		goto signal_used;

@@ -475,6 +475,15 @@ static void vhost_tx_batch(struct vhost_
 	err = sock->ops->sendmsg(sock, msghdr, 0);
 	if (unlikely(err < 0)) {
 		vq_err(&nvq->vq, "Fail to batch sending packets\n");
+
+		/* free pages owned by XDP; since this is an unlikely error path,
+		 * keep it simple and avoid more complex bulk update for the
+		 * used pages
+		 */
+		for (i = 0; i < nvq->batched_xdp; ++i)
+			put_page(virt_to_head_page(nvq->xdp[i].data));
+		nvq->batched_xdp = 0;
+		nvq->done_idx = 0;
 		return;
 	}
 

From: Eric Dumazet <[email protected]>

commit 04f08eb44b5011493d77b602fdec29ff0f5c6cd5 upstream.

syzbot reported another data-race in af_unix [1]

Lets change __skb_insert() to use WRITE_ONCE() when changing skb head qlen.

Also, change unix_dgram_poll() to use lockless version of unix_recvq_full()

It is verry possible we can switch all/most unix_recvq_full() to the lockless version, this will be done in a future kernel version.

[1] HEAD commit: 8596e589b787732c8346f0482919e83cc9362db1

BUG: KCSAN: data-race in skb_queue_tail / unix_dgram_poll

write to 0xffff88814eeb24e0 of 4 bytes by task 25815 on cpu 0:

 __skb_insert include/linux/skbuff.h:1938 [inline]
 __skb_queue_before include/linux/skbuff.h:2043 [inline]
 __skb_queue_tail include/linux/skbuff.h:2076 [inline]
 skb_queue_tail+0x80/0xa0 net/core/skbuff.c:3264
 unix_dgram_sendmsg+0xff2/0x1600 net/unix/af_unix.c:1850
 sock_sendmsg_nosec net/socket.c:703 [inline]
 sock_sendmsg net/socket.c:723 [inline]
 ____sys_sendmsg+0x360/0x4d0 net/socket.c:2392
 ___sys_sendmsg net/socket.c:2446 [inline]
 __sys_sendmmsg+0x315/0x4b0 net/socket.c:2532
 __do_sys_sendmmsg net/socket.c:2561 [inline]
 __se_sys_sendmmsg net/socket.c:2558 [inline]
 __x64_sys_sendmmsg+0x53/0x60 net/socket.c:2558
 do_syscall_x64 arch/x86/entry/common.c:50 [inline]
 do_syscall_64+0x3d/0x90 arch/x86/entry/common.c:80
 entry_SYSCALL_64_after_hwframe+0x44/0xae

read to 0xffff88814eeb24e0 of 4 bytes by task 25834 on cpu 1:

 skb_queue_len include/linux/skbuff.h:1869 [inline]
 unix_recvq_full net/unix/af_unix.c:194 [inline]
 unix_dgram_poll+0x2bc/0x3e0 net/unix/af_unix.c:2777
 sock_poll+0x23e/0x260 net/socket.c:1288
 vfs_poll include/linux/poll.h:90 [inline]
 ep_item_poll fs/eventpoll.c:846 [inline]
 ep_send_events fs/eventpoll.c:1683 [inline]
 ep_poll fs/eventpoll.c:1798 [inline]
 do_epoll_wait+0x6ad/0xf00 fs/eventpoll.c:2226
 __do_sys_epoll_wait fs/eventpoll.c:2238 [inline]
 __se_sys_epoll_wait fs/eventpoll.c:2233 [inline]
 __x64_sys_epoll_wait+0xf6/0x120 fs/eventpoll.c:2233
 do_syscall_x64 arch/x86/entry/common.c:50 [inline]
 do_syscall_64+0x3d/0x90 arch/x86/entry/common.c:80
 entry_SYSCALL_64_after_hwframe+0x44/0xae

value changed: 0x0000001b -> 0x00000001

Reported by Kernel Concurrency Sanitizer on:

CPU: 1 PID: 25834 Comm: syz-executor.1 Tainted: G        W         5.14.0-syzkaller #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011

---
 include/linux/skbuff.h |    2 +-
 net/unix/af_unix.c     |    2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

@@ -1887,7 +1887,7 @@ static inline void __skb_insert(struct s
 	WRITE_ONCE(newsk->prev, prev);
 	WRITE_ONCE(next->prev, newsk);
 	WRITE_ONCE(prev->next, newsk);
-	list->qlen++;
+	WRITE_ONCE(list->qlen, list->qlen + 1);
 }
 
 static inline void __skb_queue_splice(const struct sk_buff_head *list,

@@ -2734,7 +2734,7 @@ static __poll_t unix_dgram_poll(struct f
 
 		other = unix_peer(sk);
 		if (other && unix_peer(other) != sk &&
-		    unix_recvq_full(other) &&
+		    unix_recvq_full_lockless(other) &&
 		    unix_dgram_peer_wake_me(sk, other))
 			writable = 0;