Skip to contentSkip to conversation

5.15.86-rc1 review

PATCH In review Greg Kroah-Hartman 17 participants 756 messages Dec 28, 2022 – Jan 5, 2023 Page 3 of 6

Subsystems (11)
lore
Share & tools
lore
author maintainer
#270 lore raw

PATCH 267/731 wifi: cfg80211: Fix not unregister reg_pdev when load_builtin_regdb_keys() fails

In reply to Greg Kroah-Hartman

From: Chen Zhongjin <[email protected]>

[ Upstream commit 833a9fd28c9b7ccb39a334721379e992dc1c0c89 ]

In regulatory_init_db(), when it's going to return a error, reg_pdev should be unregistered. When load_builtin_regdb_keys() fails it doesn't do it and makes cfg80211 can't be reload with report:

sysfs: cannot create duplicate filename '/devices/platform/regulatory.0'
...

 <TASK>
 dump_stack_lvl+0x79/0x9b
 sysfs_warn_dup.cold+0x1c/0x29
 sysfs_create_dir_ns+0x22d/0x290
 kobject_add_internal+0x247/0x800
 kobject_add+0x135/0x1b0
 device_add+0x389/0x1be0
 platform_device_add+0x28f/0x790
 platform_device_register_full+0x376/0x4b0
 regulatory_init+0x9a/0x4b2 [cfg80211]
 cfg80211_init+0x84/0x113 [cfg80211]
 ...
Fixes 90a53e4432b1 ("cfg80211: implement regdb signature checking")
---
 net/wireless/reg.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
net/wireless/reg.c +3 -1
index 7b19a2087db9..d0fbe822e793 100644
@@ -4247,8 +4247,10 @@ static int __init regulatory_init_db(void)
 		return -EINVAL;
 
 	err = load_builtin_regdb_keys();
-	if (err)
+	if (err) {
+		platform_device_unregister(reg_pdev);
 		return err;
+	}
 
 	/* We always try to get an update for the static regdomain */
 	err = regulatory_hint_core(cfg80211_world_regdom->alpha2);
Signature · 1 line
2.35.1
author maintainer
#271 lore raw

PATCH 268/731 mt76: stop the radar detector after leaving dfs channel

In reply to Greg Kroah-Hartman

From: Xing Song <[email protected]>

[ Upstream commit 4e58ef4b6d727abdb071f7799aef763f8d6f2ad8 ]

The radar detctor is used for dfs channel. So it will start after switching to dfs channel and will stop after leaving. The TX will be blocked if radar detctor isn't stopped in non-dfs channel.

This patch resets the dfs state to indicate the radar detector needs to be stopped.

---
 drivers/net/wireless/mediatek/mt76/mt7615/main.c | 3 ++-
 drivers/net/wireless/mediatek/mt76/mt7915/main.c | 3 ++-
 2 files changed, 4 insertions(+), 2 deletions(-)
drivers/net/wireless/mediatek/mt76/mt7615/main.c +2 -1
index 8f1338dae211..96667b7d722d 100644
@@ -290,7 +290,8 @@ static void mt7615_init_dfs_state(struct mt7615_phy *phy)
 	if (hw->conf.flags & IEEE80211_CONF_OFFCHANNEL)
 		return;
 
-	if (!(chandef->chan->flags & IEEE80211_CHAN_RADAR))
+	if (!(chandef->chan->flags & IEEE80211_CHAN_RADAR) &&
+	    !(mphy->chandef.chan->flags & IEEE80211_CHAN_RADAR))
 		return;
 
 	if (mphy->chandef.chan->center_freq == chandef->chan->center_freq &&
drivers/net/wireless/mediatek/mt76/mt7915/main.c +2 -1
index 6aca470e2401..7a4f277a1622 100644
@@ -302,7 +302,8 @@ static void mt7915_init_dfs_state(struct mt7915_phy *phy)
 	if (hw->conf.flags & IEEE80211_CONF_OFFCHANNEL)
 		return;
 
-	if (!(chandef->chan->flags & IEEE80211_CHAN_RADAR))
+	if (!(chandef->chan->flags & IEEE80211_CHAN_RADAR) &&
+	    !(mphy->chandef.chan->flags & IEEE80211_CHAN_RADAR))
 		return;
 
 	if (mphy->chandef.chan->center_freq == chandef->chan->center_freq &&
Signature · 1 line
2.35.1
author maintainer
#272 lore raw

PATCH 269/731 wifi: mt76: mt7921: fix reporting of TX AGGR histogram

In reply to Greg Kroah-Hartman

From: Lorenzo Bianconi <[email protected]>

[ Upstream commit 028b4f22b37b88821fd87b56ce47b180583c774e ]

Similar to mt7915, fix stats clash between bins [4-7] in 802.11 tx aggregation histogram.

Fixes 163f4d22c118d ("mt76: mt7921: add MAC support")
---
 drivers/net/wireless/mediatek/mt76/mt7921/mac.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/wireless/mediatek/mt76/mt7921/mac.c +1 -1
index 426e7a32bdc8..6cf0c9b1b8b9 100644
@@ -1476,7 +1476,7 @@ mt7921_mac_update_mib_stats(struct mt7921_phy *phy)
 	mib->rts_retries_cnt += mt76_get_field(dev, MT_MIB_MB_BSDR1(0),
 					       MT_MIB_RTS_FAIL_COUNT_MASK);
 
-	for (i = 0, aggr1 = aggr0 + 4; i < 4; i++) {
+	for (i = 0, aggr1 = aggr0 + 8; i < 4; i++) {
 		u32 val, val2;
 
 		val = mt76_rr(dev, MT_TX_AGG_CNT(0, i));
Signature · 1 line
2.35.1
author maintainer
#273 lore raw

PATCH 270/731 wifi: mt76: fix coverity overrun-call in mt76_get_txpower()

In reply to Greg Kroah-Hartman

From: Deren Wu <[email protected]>

[ Upstream commit 03dd0d49de7db680a856fa566963bb8421f46368 ]

Make sure the nss is valid for nss_delta array. Return zero if the index is invalid.

Coverity message:
Event overrun-call: Overrunning callee's array of size 4 by passing argument "n_chains" (which evaluates to 15) in call to "mt76_tx_power_nss_delta".
int delta = mt76_tx_power_nss_delta(n_chains);

Fixes 07cda406308b ("mt76: fix rounding issues on converting per-chain and combined txpower")
---
 drivers/net/wireless/mediatek/mt76/mt76.h | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/net/wireless/mediatek/mt76/mt76.h +2 -1
index d1f00706d41e..4e4af6e17b50 100644
@@ -1021,8 +1021,9 @@ static inline bool mt76_is_skb_pktid(u8 pktid)
 static inline u8 mt76_tx_power_nss_delta(u8 nss)
 {
 	static const u8 nss_delta[4] = { 0, 6, 9, 12 };
+	u8 idx = nss - 1;
 
-	return nss_delta[nss - 1];
+	return (idx < ARRAY_SIZE(nss_delta)) ? nss_delta[idx] : 0;
 }
 
 static inline bool mt76_testmode_enabled(struct mt76_phy *phy)
Signature · 1 line
2.35.1
author maintainer
#274 lore raw

PATCH 271/731 regulator: core: fix module refcount leak in set_supply()

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit da46ee19cbd8344d6860816b4827a7ce95764867 ]

If create_regulator() fails in set_supply(), the module refcount needs be put to keep refcount balanced.

Fixes e2c09ae7a74d ("regulator: core: Increase refcount for regulator supply's module")
---
 drivers/regulator/core.c | 1 +
 1 file changed, 1 insertion(+)
drivers/regulator/core.c +1 -0
index 65900895a0b2..ab93debb817c 100644
@@ -1577,6 +1577,7 @@ static int set_supply(struct regulator_dev *rdev,
 
 	rdev->supply = create_regulator(supply_rdev, &rdev->dev, "SUPPLY");
 	if (rdev->supply == NULL) {
+		module_put(supply_rdev->owner);
 		err = -ENOMEM;
 		return err;
 	}
Signature · 1 line
2.35.1
author maintainer
#275 lore raw

PATCH 272/731 clk: qcom: lpass-sc7180: Fix pm_runtime usage

In reply to Greg Kroah-Hartman

From: Douglas Anderson <[email protected]>

[ Upstream commit ff1ccf59eaffd192efe21f7de9fb0c130faf1b1b ]

The sc7180 lpass clock controller's pm_runtime usage wasn't broken quite as spectacularly as the sc7280's pm_runtime usage, but it was still broken. Putting some printouts in at boot showed me this (with serial console enabled, which makes the prints slow and thus changes timing):

  [    3.109951] DOUG: my_pm_clk_resume, usage=1
  [    3.114767] DOUG: my_pm_clk_resume, usage=1
  [    3.664443] DOUG: my_pm_clk_suspend, usage=0
  [    3.897566] DOUG: my_pm_clk_suspend, usage=0
  [    3.910137] DOUG: my_pm_clk_resume, usage=1
  [    3.923217] DOUG: my_pm_clk_resume, usage=0
  [    4.440116] DOUG: my_pm_clk_suspend, usage=-1
  [    4.444982] DOUG: my_pm_clk_suspend, usage=0
  [   14.170501] DOUG: my_pm_clk_resume, usage=1
  [   14.176245] DOUG: my_pm_clk_resume, usage=0

...or this w/out serial console:

  [    0.556139] DOUG: my_pm_clk_resume, usage=1
  [    0.556279] DOUG: my_pm_clk_resume, usage=1
  [    1.058422] DOUG: my_pm_clk_suspend, usage=-1
  [    1.058464] DOUG: my_pm_clk_suspend, usage=0
  [    1.186250] DOUG: my_pm_clk_resume, usage=1
  [    1.186292] DOUG: my_pm_clk_resume, usage=0
  [    1.731536] DOUG: my_pm_clk_suspend, usage=-1
  [    1.731557] DOUG: my_pm_clk_suspend, usage=0
  [   10.288910] DOUG: my_pm_clk_resume, usage=1
  [   10.289496] DOUG: my_pm_clk_resume, usage=0

It seems to be doing roughly the right sequence of calls, but just like with sc7280 this is more by luck than anything. Having a usage of -1 is just not OK.

Let's fix this like we did with sc7280.

---
 drivers/clk/qcom/lpasscorecc-sc7180.c | 24 ++++++++++++++++--------
 1 file changed, 16 insertions(+), 8 deletions(-)
drivers/clk/qcom/lpasscorecc-sc7180.c +16 -8
index ac09b7b840ab..a5731994cbed 100644
@@ -356,7 +356,7 @@ static const struct qcom_cc_desc lpass_audio_hm_sc7180_desc = {
 	.num_gdscs = ARRAY_SIZE(lpass_audio_hm_sc7180_gdscs),
 };
 
-static int lpass_create_pm_clks(struct platform_device *pdev)
+static int lpass_setup_runtime_pm(struct platform_device *pdev)
 {
 	int ret;
 
@@ -375,7 +375,7 @@ static int lpass_create_pm_clks(struct platform_device *pdev)
 	if (ret < 0)
 		dev_err(&pdev->dev, "failed to acquire iface clock\n");
 
-	return ret;
+	return pm_runtime_resume_and_get(&pdev->dev);
 }
 
 static int lpass_core_cc_sc7180_probe(struct platform_device *pdev)
@@ -384,7 +384,7 @@ static int lpass_core_cc_sc7180_probe(struct platform_device *pdev)
 	struct regmap *regmap;
 	int ret;
 
-	ret = lpass_create_pm_clks(pdev);
+	ret = lpass_setup_runtime_pm(pdev);
 	if (ret)
 		return ret;
 
@@ -392,12 +392,14 @@ static int lpass_core_cc_sc7180_probe(struct platform_device *pdev)
 	desc = &lpass_audio_hm_sc7180_desc;
 	ret = qcom_cc_probe_by_index(pdev, 1, desc);
 	if (ret)
-		return ret;
+		goto exit;
 
 	lpass_core_cc_sc7180_regmap_config.name = "lpass_core_cc";
 	regmap = qcom_cc_map(pdev, &lpass_core_cc_sc7180_desc);
-	if (IS_ERR(regmap))
-		return PTR_ERR(regmap);
+	if (IS_ERR(regmap)) {
+		ret = PTR_ERR(regmap);
+		goto exit;
+	}
 
 	/*
 	 * Keep the CLK always-ON
@@ -415,6 +417,7 @@ static int lpass_core_cc_sc7180_probe(struct platform_device *pdev)
 	ret = qcom_cc_really_probe(pdev, &lpass_core_cc_sc7180_desc, regmap);
 
 	pm_runtime_mark_last_busy(&pdev->dev);
+exit:
 	pm_runtime_put_autosuspend(&pdev->dev);
 
 	return ret;
@@ -425,14 +428,19 @@ static int lpass_hm_core_probe(struct platform_device *pdev)
 	const struct qcom_cc_desc *desc;
 	int ret;
 
-	ret = lpass_create_pm_clks(pdev);
+	ret = lpass_setup_runtime_pm(pdev);
 	if (ret)
 		return ret;
 
 	lpass_core_cc_sc7180_regmap_config.name = "lpass_hm_core";
 	desc = &lpass_core_hm_sc7180_desc;
 
-	return qcom_cc_probe_by_index(pdev, 0, desc);
+	ret = qcom_cc_probe_by_index(pdev, 0, desc);
+
+	pm_runtime_mark_last_busy(&pdev->dev);
+	pm_runtime_put_autosuspend(&pdev->dev);
+
+	return ret;
 }
 
 static const struct of_device_id lpass_hm_sc7180_match_table[] = {
Signature · 1 line
2.35.1
author maintainer
#276 lore raw

PATCH 273/731 clk: qcom: clk-krait: fix wrong div2 functions

In reply to Greg Kroah-Hartman

From: Christian Marangi <[email protected]>

[ Upstream commit d676d3a3717cf726d3affedbe5ba98fc4ccad7b3 ]

Currently div2 value is applied to the wrong bits. This is caused by a bug in the code where the shift is done only for lpl, for anything else the mask is not shifted to the correct bits.

Fix this by correctly shift if lpl is not supported.

Fixes 4d7dc77babfe ("clk: qcom: Add support for Krait clocks")
---
 drivers/clk/qcom/clk-krait.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/clk/qcom/clk-krait.c +2 -0
index 90046428693c..e74fc81a14d0 100644
@@ -98,6 +98,8 @@ static int krait_div2_set_rate(struct clk_hw *hw, unsigned long rate,
 
 	if (d->lpl)
 		mask = mask << (d->shift + LPL_SHIFT) | mask << d->shift;
+	else
+		mask <<= d->shift;
 
 	spin_lock_irqsave(&krait_clock_reg_lock, flags);
 	val = krait_get_l2_indirect_reg(d->offset);
Signature · 1 line
2.35.1
author maintainer
#277 lore raw

PATCH 274/731 hsr: Add a rcu-read lock to hsr_forward_skb().

In reply to Greg Kroah-Hartman

From: Sebastian Andrzej Siewior <[email protected]>

[ Upstream commit 5aa2820177af650293b2f9f1873c1f6f8e4ad7a4 ]

hsr_forward_skb() a skb and keeps information in an on-stack hsr_frame_info. hsr_get_node() assigns hsr_frame_info::node_src which is from a RCU list. This pointer is used later in hsr_forward_do(). I don't see a reason why this pointer can't vanish midway since there is no guarantee that hsr_forward_skb() is invoked from an RCU read section.

Use rcu_read_lock() to protect hsr_frame_info::node_src from its assignment until it is no longer used.

Fixes f266a683a4804 ("net/hsr: Better frame dispatch")
---
 net/hsr/hsr_forward.c | 3 +++
 1 file changed, 3 insertions(+)
net/hsr/hsr_forward.c +3 -0
index 07892c4b6d0c..daf6abbadeb9 100644
@@ -572,11 +572,13 @@ void hsr_forward_skb(struct sk_buff *skb, struct hsr_port *port)
 {
 	struct hsr_frame_info frame;
 
+	rcu_read_lock();
 	if (fill_frame_info(&frame, skb, port) < 0)
 		goto out_drop;
 
 	hsr_register_frame_in(frame.node_src, port, frame.sequence_nr);
 	hsr_forward_do(&frame);
+	rcu_read_unlock();
 	/* Gets called for ingress frames as well as egress from master port.
 	 * So check and increment stats for master port only here.
 	 */
@@ -591,6 +593,7 @@ void hsr_forward_skb(struct sk_buff *skb, struct hsr_port *port)
 	return;
 
 out_drop:
+	rcu_read_unlock();
 	port->dev->stats.tx_dropped++;
 	kfree_skb(skb);
 }
Signature · 1 line
2.35.1
author maintainer
#278 lore raw

PATCH 275/731 hsr: Avoid double remove of a node.

In reply to Greg Kroah-Hartman

From: Sebastian Andrzej Siewior <[email protected]>

[ Upstream commit 0c74d9f79ec4299365bbe803baa736ae0068179e ]

Due to the hashed-MAC optimisation one problem become visible:
hsr_handle_sup_frame() walks over the list of available nodes and merges two node entries into one if based on the information in the supervision both MAC addresses belong to one node. The list-walk happens on a RCU protected list and delete operation happens under a lock.

If the supervision arrives on both slave interfaces at the same time then this delete operation can occur simultaneously on two CPUs. The result is the first-CPU deletes the from the list and the second CPUs BUGs while attempting to dereference a poisoned list-entry. This happens more likely with the optimisation because a new node for the mac_B entry is created once a packet has been received and removed (merged) once the supervision frame has been received.

Avoid removing/ cleaning up a hsr_node twice by adding a `removed' field which is set to true after the removal and checked before the removal.

Fixes f266a683a4804 ("net/hsr: Better frame dispatch")
---
 net/hsr/hsr_framereg.c | 16 +++++++++++-----
 net/hsr/hsr_framereg.h |  1 +
 2 files changed, 12 insertions(+), 5 deletions(-)
net/hsr/hsr_framereg.c +11 -5
index e31949479305..440788e5b3e2 100644
@@ -326,9 +326,12 @@ void hsr_handle_sup_frame(struct hsr_frame_info *frame)
 	node_real->addr_B_port = port_rcv->type;
 
 	spin_lock_bh(&hsr->list_lock);
-	list_del_rcu(&node_curr->mac_list);
+	if (!node_curr->removed) {
+		list_del_rcu(&node_curr->mac_list);
+		node_curr->removed = true;
+		kfree_rcu(node_curr, rcu_head);
+	}
 	spin_unlock_bh(&hsr->list_lock);
-	kfree_rcu(node_curr, rcu_head);
 
 done:
 	/* PRP uses v0 header */
@@ -502,9 +505,12 @@ void hsr_prune_nodes(struct timer_list *t)
 		if (time_is_before_jiffies(timestamp +
 				msecs_to_jiffies(HSR_NODE_FORGET_TIME))) {
 			hsr_nl_nodedown(hsr, node->macaddress_A);
-			list_del_rcu(&node->mac_list);
-			/* Note that we need to free this entry later: */
-			kfree_rcu(node, rcu_head);
+			if (!node->removed) {
+				list_del_rcu(&node->mac_list);
+				node->removed = true;
+				/* Note that we need to free this entry later: */
+				kfree_rcu(node, rcu_head);
+			}
 		}
 	}
 	spin_unlock_bh(&hsr->list_lock);
net/hsr/hsr_framereg.h +1 -0
index d9628e7a5f05..f9c83dc02ca5 100644
@@ -80,6 +80,7 @@ struct hsr_node {
 	bool			san_a;
 	bool			san_b;
 	u16			seq_out[HSR_PT_PORTS];
+	bool			removed;
 	struct rcu_head		rcu_head;
 };
 
Signature · 1 line
2.35.1
author maintainer
#279 lore raw

PATCH 276/731 hsr: Disable netpoll.

In reply to Greg Kroah-Hartman

From: Sebastian Andrzej Siewior <[email protected]>

[ Upstream commit d5c7652eb16fa203d82546e0285136d7b321ffa9 ]

The hsr device is a software device. Its net_device_ops::ndo_start_xmit() routine will process the packet and then pass the resulting skb to dev_queue_xmit().
During processing, hsr acquires a lock with spin_lock_bh() (hsr_add_node()) which needs to be promoted to the _irq() suffix in order to avoid a potential deadlock.
Then there are the warnings in dev_queue_xmit() (due to local_bh_disable() with disabled interrupts) left.

Instead trying to address those (there is qdisc and…) for netpoll sake, just disable netpoll on hsr.

Disable netpoll on hsr and replace the _irqsave() locking with _bh().

Fixes f421436a591d3 ("net/hsr: Add support for the High-availability Seamless Redundancy protocol (HSRv0)")
---
 net/hsr/hsr_device.c  | 14 ++++++--------
 net/hsr/hsr_forward.c |  5 ++---
 2 files changed, 8 insertions(+), 11 deletions(-)
net/hsr/hsr_device.c +6 -8
index a1045c3d71b4..c44c6747a0bf 100644
@@ -278,7 +278,6 @@ static void send_hsr_supervision_frame(struct hsr_port *master,
 	__u8 type = HSR_TLV_LIFE_CHECK;
 	struct hsr_sup_payload *hsr_sp;
 	struct hsr_sup_tag *hsr_stag;
-	unsigned long irqflags;
 	struct sk_buff *skb;
 
 	*interval = msecs_to_jiffies(HSR_LIFE_CHECK_INTERVAL);
@@ -299,7 +298,7 @@ static void send_hsr_supervision_frame(struct hsr_port *master,
 	set_hsr_stag_HSR_ver(hsr_stag, hsr->prot_version);
 
 	/* From HSRv1 on we have separate supervision sequence numbers. */
-	spin_lock_irqsave(&master->hsr->seqnr_lock, irqflags);
+	spin_lock_bh(&hsr->seqnr_lock);
 	if (hsr->prot_version > 0) {
 		hsr_stag->sequence_nr = htons(hsr->sup_sequence_nr);
 		hsr->sup_sequence_nr++;
@@ -307,7 +306,7 @@ static void send_hsr_supervision_frame(struct hsr_port *master,
 		hsr_stag->sequence_nr = htons(hsr->sequence_nr);
 		hsr->sequence_nr++;
 	}
-	spin_unlock_irqrestore(&master->hsr->seqnr_lock, irqflags);
+	spin_unlock_bh(&hsr->seqnr_lock);
 
 	hsr_stag->HSR_TLV_type = type;
 	/* TODO: Why 12 in HSRv0? */
@@ -332,7 +331,6 @@ static void send_prp_supervision_frame(struct hsr_port *master,
 	struct hsr_priv *hsr = master->hsr;
 	struct hsr_sup_payload *hsr_sp;
 	struct hsr_sup_tag *hsr_stag;
-	unsigned long irqflags;
 	struct sk_buff *skb;
 
 	skb = hsr_init_skb(master);
@@ -347,7 +345,7 @@ static void send_prp_supervision_frame(struct hsr_port *master,
 	set_hsr_stag_HSR_ver(hsr_stag, (hsr->prot_version ? 1 : 0));
 
 	/* From HSRv1 on we have separate supervision sequence numbers. */
-	spin_lock_irqsave(&master->hsr->seqnr_lock, irqflags);
+	spin_lock_bh(&hsr->seqnr_lock);
 	hsr_stag->sequence_nr = htons(hsr->sup_sequence_nr);
 	hsr->sup_sequence_nr++;
 	hsr_stag->HSR_TLV_type = PRP_TLV_LIFE_CHECK_DD;
@@ -358,11 +356,11 @@ static void send_prp_supervision_frame(struct hsr_port *master,
 	ether_addr_copy(hsr_sp->macaddress_A, master->dev->dev_addr);
 
 	if (skb_put_padto(skb, ETH_ZLEN)) {
-		spin_unlock_irqrestore(&master->hsr->seqnr_lock, irqflags);
+		spin_unlock_bh(&hsr->seqnr_lock);
 		return;
 	}
 
-	spin_unlock_irqrestore(&master->hsr->seqnr_lock, irqflags);
+	spin_unlock_bh(&hsr->seqnr_lock);
 
 	hsr_forward_skb(skb, master);
 }
@@ -444,7 +442,7 @@ void hsr_dev_setup(struct net_device *dev)
 	dev->header_ops = &hsr_header_ops;
 	dev->netdev_ops = &hsr_device_ops;
 	SET_NETDEV_DEVTYPE(dev, &hsr_type);
-	dev->priv_flags |= IFF_NO_QUEUE;
+	dev->priv_flags |= IFF_NO_QUEUE | IFF_DISABLE_NETPOLL;
 
 	dev->needs_free_netdev = true;
 
net/hsr/hsr_forward.c +2 -3
index daf6abbadeb9..c8bcbd990f59 100644
@@ -458,7 +458,6 @@ static void handle_std_frame(struct sk_buff *skb,
 {
 	struct hsr_port *port = frame->port_rcv;
 	struct hsr_priv *hsr = port->hsr;
-	unsigned long irqflags;
 
 	frame->skb_hsr = NULL;
 	frame->skb_prp = NULL;
@@ -468,10 +467,10 @@ static void handle_std_frame(struct sk_buff *skb,
 		frame->is_from_san = true;
 	} else {
 		/* Sequence nr for the master node */
-		spin_lock_irqsave(&hsr->seqnr_lock, irqflags);
+		spin_lock_bh(&hsr->seqnr_lock);
 		frame->sequence_nr = hsr->sequence_nr;
 		hsr->sequence_nr++;
-		spin_unlock_irqrestore(&hsr->seqnr_lock, irqflags);
+		spin_unlock_bh(&hsr->seqnr_lock);
 	}
 }
 
Signature · 1 line
2.35.1
author maintainer
#280 lore raw

PATCH 277/731 hsr: Synchronize sending frames to have always incremented outgoing seq nr.

In reply to Greg Kroah-Hartman

From: Sebastian Andrzej Siewior <[email protected]>

[ Upstream commit 06afd2c31d338fa762548580c1bf088703dd1e03 ]

Sending frames via the hsr (master) device requires a sequence number which is tracked in hsr_priv::sequence_nr and protected by hsr_priv::seqnr_lock. Each time a new frame is sent, it will obtain a new id and then send it via the slave devices.
Each time a packet is sent (via hsr_forward_do()) the sequence number is checked via hsr_register_frame_out() to ensure that a frame is not handled twice. This make sense for the receiving side to ensure that the frame is not injected into the stack twice after it has been received from both slave ports.

There is no locking to cover the sending path which means the following scenario is possible:

CPU0 CPU1
hsr_dev_xmit(skb1) hsr_dev_xmit(skb2)
fill_frame_info() fill_frame_info()
hsr_fill_frame_info() hsr_fill_frame_info()
handle_std_frame() handle_std_frame()
skb1's sequence_nr = 1
skb2's sequence_nr = 2
hsr_forward_do() hsr_forward_do()

hsr_register_frame_out(, 2) // okay, send)

hsr_register_frame_out(, 1) // stop, lower seq duplicate

Both skbs (or their struct hsr_frame_info) received an unique id. However since skb2 was sent before skb1, the higher sequence number was recorded in hsr_register_frame_out() and the late arriving skb1 was dropped and never sent.

This scenario has been observed in a three node HSR setup, with node1 + node2 having ping and iperf running in parallel. From time to time ping reported a missing packet. Based on tracing that missing ping packet did not leave the system.

It might be possible (didn't check) to drop the sequence number check on the sending side. But if the higher sequence number leaves on wire before the lower does and the destination receives them in that order and it will drop the packet with the lower sequence number and never inject into the stack.
Therefore it seems the only way is to lock the whole path from obtaining the sequence number and sending via dev_queue_xmit() and assuming the packets leave on wire in the same order (and don't get reordered by the NIC).

Cover the whole path for the master interface from obtaining the ID until after it has been forwarded via hsr_forward_skb() to ensure the skbs are sent to the NIC in the order of the assigned sequence numbers.

Fixes f421436a591d3 ("net/hsr: Add support for the High-availability Seamless Redundancy protocol (HSRv0)")
---
 net/hsr/hsr_device.c  | 12 +++++++-----
 net/hsr/hsr_forward.c |  3 +--
 2 files changed, 8 insertions(+), 7 deletions(-)
net/hsr/hsr_device.c +7 -5
index c44c6747a0bf..7ce40b49c956 100644
@@ -219,7 +219,9 @@ static netdev_tx_t hsr_dev_xmit(struct sk_buff *skb, struct net_device *dev)
 		skb->dev = master->dev;
 		skb_reset_mac_header(skb);
 		skb_reset_mac_len(skb);
+		spin_lock_bh(&hsr->seqnr_lock);
 		hsr_forward_skb(skb, master);
+		spin_unlock_bh(&hsr->seqnr_lock);
 	} else {
 		atomic_long_inc(&dev->tx_dropped);
 		dev_kfree_skb_any(skb);
@@ -306,7 +308,6 @@ static void send_hsr_supervision_frame(struct hsr_port *master,
 		hsr_stag->sequence_nr = htons(hsr->sequence_nr);
 		hsr->sequence_nr++;
 	}
-	spin_unlock_bh(&hsr->seqnr_lock);
 
 	hsr_stag->HSR_TLV_type = type;
 	/* TODO: Why 12 in HSRv0? */
@@ -317,11 +318,13 @@ static void send_hsr_supervision_frame(struct hsr_port *master,
 	hsr_sp = skb_put(skb, sizeof(struct hsr_sup_payload));
 	ether_addr_copy(hsr_sp->macaddress_A, master->dev->dev_addr);
 
-	if (skb_put_padto(skb, ETH_ZLEN))
+	if (skb_put_padto(skb, ETH_ZLEN)) {
+		spin_unlock_bh(&hsr->seqnr_lock);
 		return;
+	}
 
 	hsr_forward_skb(skb, master);
-
+	spin_unlock_bh(&hsr->seqnr_lock);
 	return;
 }
 
@@ -360,9 +363,8 @@ static void send_prp_supervision_frame(struct hsr_port *master,
 		return;
 	}
 
-	spin_unlock_bh(&hsr->seqnr_lock);
-
 	hsr_forward_skb(skb, master);
+	spin_unlock_bh(&hsr->seqnr_lock);
 }
 
 /* Announce (supervision frame) timer function
net/hsr/hsr_forward.c +1 -2
index c8bcbd990f59..35382ed686d1 100644
@@ -467,10 +467,9 @@ static void handle_std_frame(struct sk_buff *skb,
 		frame->is_from_san = true;
 	} else {
 		/* Sequence nr for the master node */
-		spin_lock_bh(&hsr->seqnr_lock);
+		lockdep_assert_held(&hsr->seqnr_lock);
 		frame->sequence_nr = hsr->sequence_nr;
 		hsr->sequence_nr++;
-		spin_unlock_bh(&hsr->seqnr_lock);
 	}
 }
 
Signature · 1 line
2.35.1
author maintainer
#281 lore raw

PATCH 278/731 hsr: Synchronize sequence number updates.

In reply to Greg Kroah-Hartman

From: Sebastian Andrzej Siewior <[email protected]>

[ Upstream commit 5c7aa13210c3abdd34fd421f62347665ec6eb551 ]

hsr_register_frame_out() compares new sequence_nr vs the old one recorded in hsr_node::seq_out and if the new sequence_nr is higher then it will be written to hsr_node::seq_out as the new value.

This operation isn't locked so it is possible that two frames with the same sequence number arrive (via the two slave devices) and are fed to hsr_register_frame_out() at the same time. Both will pass the check and update the sequence counter later to the same value. As a result the content of the same packet is fed into the stack twice.

This was noticed by running ping and observing DUP being reported from time to time.

Instead of using the hsr_priv::seqnr_lock for the whole receive path (as it is for sending in the master node) add an additional lock that is only used for sequence number checks and updates.

Add a per-node lock that is used during sequence number reads and updates.

Fixes f421436a591d3 ("net/hsr: Add support for the High-availability Seamless Redundancy protocol (HSRv0)")
---
 net/hsr/hsr_framereg.c | 9 ++++++++-
 net/hsr/hsr_framereg.h | 2 ++
 2 files changed, 10 insertions(+), 1 deletion(-)
net/hsr/hsr_framereg.c +8 -1
index 440788e5b3e2..414bf4d3d3c9 100644
@@ -159,6 +159,7 @@ static struct hsr_node *hsr_add_node(struct hsr_priv *hsr,
 		return NULL;
 
 	ether_addr_copy(new_node->macaddress_A, addr);
+	spin_lock_init(&new_node->seq_out_lock);
 
 	/* We are only interested in time diffs here, so use current jiffies
 	 * as initialization. (0 could trigger an spurious ring error warning).
@@ -313,6 +314,7 @@ void hsr_handle_sup_frame(struct hsr_frame_info *frame)
 		goto done;
 
 	ether_addr_copy(node_real->macaddress_B, ethhdr->h_source);
+	spin_lock_bh(&node_real->seq_out_lock);
 	for (i = 0; i < HSR_PT_PORTS; i++) {
 		if (!node_curr->time_in_stale[i] &&
 		    time_after(node_curr->time_in[i], node_real->time_in[i])) {
@@ -323,6 +325,7 @@ void hsr_handle_sup_frame(struct hsr_frame_info *frame)
 		if (seq_nr_after(node_curr->seq_out[i], node_real->seq_out[i]))
 			node_real->seq_out[i] = node_curr->seq_out[i];
 	}
+	spin_unlock_bh(&node_real->seq_out_lock);
 	node_real->addr_B_port = port_rcv->type;
 
 	spin_lock_bh(&hsr->list_lock);
@@ -419,13 +422,17 @@ void hsr_register_frame_in(struct hsr_node *node, struct hsr_port *port,
 int hsr_register_frame_out(struct hsr_port *port, struct hsr_node *node,
 			   u16 sequence_nr)
 {
+	spin_lock_bh(&node->seq_out_lock);
 	if (seq_nr_before_or_eq(sequence_nr, node->seq_out[port->type]) &&
 	    time_is_after_jiffies(node->time_out[port->type] +
-	    msecs_to_jiffies(HSR_ENTRY_FORGET_TIME)))
+	    msecs_to_jiffies(HSR_ENTRY_FORGET_TIME))) {
+		spin_unlock_bh(&node->seq_out_lock);
 		return 1;
+	}
 
 	node->time_out[port->type] = jiffies;
 	node->seq_out[port->type] = sequence_nr;
+	spin_unlock_bh(&node->seq_out_lock);
 	return 0;
 }
 
net/hsr/hsr_framereg.h +2 -0
index f9c83dc02ca5..48990166e4c4 100644
@@ -69,6 +69,8 @@ void prp_update_san_info(struct hsr_node *node, bool is_sup);
 
 struct hsr_node {
 	struct list_head	mac_list;
+	/* Protect R/W access to seq_out */
+	spinlock_t		seq_out_lock;
 	unsigned char		macaddress_A[ETH_ALEN];
 	unsigned char		macaddress_B[ETH_ALEN];
 	/* Local slave through which AddrB frames are received from this node */
Signature · 1 line
2.35.1
author maintainer
#282 lore raw

PATCH 279/731 configfs: fix possible memory leak in configfs_create_dir()

In reply to Greg Kroah-Hartman

From: Chen Zhongjin <[email protected]>

[ Upstream commit c65234b283a65cfbfc94619655e820a5e55199eb ]

kmemleak reported memory leaks in configfs_create_dir():

unreferenced object 0xffff888009f6af00 (size 192):
comm "modprobe", pid 3777, jiffies 4295537735 (age 233.784s) backtrace:

    kmem_cache_alloc (mm/slub.c:3250 mm/slub.c:3256 mm/slub.c:3263 mm/slub.c:3273)
    new_fragment (./include/linux/slab.h:600 fs/configfs/dir.c:163)
    configfs_register_subsystem (fs/configfs/dir.c:1857)
    basic_write (drivers/hwtracing/stm/p_basic.c:14) stm_p_basic
    do_one_initcall (init/main.c:1296)
    do_init_module (kernel/module/main.c:2455)
    ...

unreferenced object 0xffff888003ba7180 (size 96):
comm "modprobe", pid 3777, jiffies 4295537735 (age 233.784s) backtrace:

    kmem_cache_alloc (mm/slub.c:3250 mm/slub.c:3256 mm/slub.c:3263 mm/slub.c:3273)
    configfs_new_dirent (./include/linux/slab.h:723 fs/configfs/dir.c:194)
    configfs_make_dirent (fs/configfs/dir.c:248)
    configfs_create_dir (fs/configfs/dir.c:296)
    configfs_attach_group.isra.28 (fs/configfs/dir.c:816 fs/configfs/dir.c:852)
    configfs_register_subsystem (fs/configfs/dir.c:1881)
    basic_write (drivers/hwtracing/stm/p_basic.c:14) stm_p_basic
    do_one_initcall (init/main.c:1296)
    do_init_module (kernel/module/main.c:2455)
    ...

This is because the refcount is not correct in configfs_make_dirent(). For normal stage, the refcount is changing as:

configfs_register_subsystem()
configfs_create_dir()
configfs_make_dirent()
configfs_new_dirent() # set s_count = 1 dentry->d_fsdata = configfs_get(sd); # s_count = 2
...
configfs_unregister_subsystem()
configfs_remove_dir()
remove_dir()
configfs_remove_dirent() # s_count = 1
dput() ...
*dentry_unlink_inode()*
configfs_d_iput() # s_count = 0, release

However, if we failed in configfs_create():

configfs_register_subsystem()
configfs_create_dir()
configfs_make_dirent() # s_count = 2
...
configfs_create() # fail
->out_remove:
configfs_remove_dirent(dentry)
configfs_put(sd) # s_count = 1
return PTR_ERR(inode);

There is no inode in the error path, so the configfs_d_iput() is lost and makes sd and fragment memory leaked.

To fix this, when we failed in configfs_create(), manually call configfs_put(sd) to keep the refcount correct.

Fixes 7063fbf22611 ("[PATCH] configfs: User-driven configuration filesystem")
---
 fs/configfs/dir.c | 2 ++
 1 file changed, 2 insertions(+)
fs/configfs/dir.c +2 -0
index d1f9d2632202..ec6519e1ca3b 100644
@@ -316,6 +316,7 @@ static int configfs_create_dir(struct config_item *item, struct dentry *dentry,
 	return 0;
 
 out_remove:
+	configfs_put(dentry->d_fsdata);
 	configfs_remove_dirent(dentry);
 	return PTR_ERR(inode);
 }
@@ -382,6 +383,7 @@ int configfs_create_link(struct configfs_dirent *target, struct dentry *parent,
 	return 0;
 
 out_remove:
+	configfs_put(dentry->d_fsdata);
 	configfs_remove_dirent(dentry);
 	return PTR_ERR(inode);
 }
Signature · 1 line
2.35.1
author maintainer
#283 lore raw

PATCH 280/731 regulator: core: fix resource leak in regulator_register()

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit ba62319a42c50e6254e98b3f316464fac8e77968 ]

I got some resource leak reports while doing fault injection test:

OF: ERROR: memory leak, expected refcount 1 instead of 100, of_node_get()/of_node_put() unbalanced - destroy cset entry:
attach overlay node /i2c/pmic@64/regulators/buck1

unreferenced object 0xffff88810deea000 (size 512):
comm "490-i2c-rt5190a", pid 253, jiffies 4294859840 (age 5061.046s) hex dump (first 32 bytes):

    00 00 00 00 ad 4e ad de ff ff ff ff 00 00 00 00  .....N..........
    ff ff ff ff ff ff ff ff a0 1e 00 a1 ff ff ff ff  ................
  backtrace:
    [<00000000d78541e2>] kmalloc_trace+0x21/0x110
    [<00000000b343d153>] device_private_init+0x32/0xd0
    [<00000000be1f0c70>] device_add+0xb2d/0x1030
    [<00000000e3e6344d>] regulator_register+0xaf2/0x12a0
    [<00000000e2f5e754>] devm_regulator_register+0x57/0xb0
    [<000000008b898197>] rt5190a_probe+0x52a/0x861 [rt5190a_regulator]

unreferenced object 0xffff88810b617b80 (size 32):
comm "490-i2c-rt5190a", pid 253, jiffies 4294859904 (age 5060.983s) hex dump (first 32 bytes):

    72 65 67 75 6c 61 74 6f 72 2e 32 38 36 38 2d 53  regulator.2868-S
    55 50 50 4c 59 00 ff ff 29 00 00 00 2b 00 00 00  UPPLY...)...+...
  backtrace:
    [<000000009da9280d>] __kmalloc_node_track_caller+0x44/0x1b0
    [<0000000025c6a4e5>] kstrdup+0x3a/0x70
    [<00000000790efb69>] create_regulator+0xc0/0x4e0
    [<0000000005ed203a>] regulator_resolve_supply+0x2d4/0x440
    [<0000000045796214>] regulator_register+0x10b3/0x12a0
    [<00000000e2f5e754>] devm_regulator_register+0x57/0xb0
    [<000000008b898197>] rt5190a_probe+0x52a/0x861 [rt5190a_regulator]

After calling regulator_resolve_supply(), the 'rdev->supply' is set by set_supply(), after this set, in the error path, the resources need be released, so call regulator_put() to avoid the leaks.

Fixes aea6cb99703e ("regulator: resolve supply after creating regulator")
Fixes 8a866d527ac0 ("regulator: core: Resolve supply name earlier to prevent double-init")
---
 drivers/regulator/core.c | 1 +
 1 file changed, 1 insertion(+)
drivers/regulator/core.c +1 -0
index ab93debb817c..1a5aeb4868aa 100644
@@ -5548,6 +5548,7 @@ regulator_register(const struct regulator_desc *regulator_desc,
 	regulator_remove_coupling(rdev);
 	mutex_unlock(&regulator_list_mutex);
 wash:
+	regulator_put(rdev->supply);
 	kfree(rdev->coupling_desc.coupled_rdevs);
 	mutex_lock(&regulator_list_mutex);
 	regulator_ena_gpio_free(rdev);
Signature · 1 line
2.35.1
author maintainer
#284 lore raw

PATCH 281/731 hwmon: (jc42) Convert register access and caching to regmap/regcache

In reply to Greg Kroah-Hartman

From: Martin Blumenstingl <[email protected]>

[ Upstream commit 8f2fa4726faf01094d7a5be7bd0c120c565f54d9 ]

Switch the jc42 driver to use an I2C regmap to access the registers. Also move over to regmap's built-in caching instead of adding a custom caching implementation. This works for JC42_REG_TEMP_UPPER, JC42_REG_TEMP_LOWER and JC42_REG_TEMP_CRITICAL as these values never change except when explicitly written. The cache For JC42_REG_TEMP is dropped (regmap can't cache it because it's volatile, meaning it can change at any time) as well for simplicity and consistency with other drivers.

Stable-dep-of: 084ed144c448 ("hwmon: (jc42) Restore the min/max/critical temperatures on resume")

---
 drivers/hwmon/Kconfig |   1 +
 drivers/hwmon/jc42.c  | 233 ++++++++++++++++++++++++------------------
 2 files changed, 132 insertions(+), 102 deletions(-)
drivers/hwmon/Kconfig +1 -0
index 51f1caa10d11..17ba1d9ff075 100644
@@ -776,6 +776,7 @@ config SENSORS_IT87
 config SENSORS_JC42
 	tristate "JEDEC JC42.4 compliant memory module temperature sensors"
 	depends on I2C
+	select REGMAP_I2C
 	help
 	  If you say yes here, you get support for JEDEC JC42.4 compliant
 	  temperature sensors, which are used on many DDR3 memory modules for
drivers/hwmon/jc42.c +131 -102

364 lines not on this page · raw message

Signature · 1 line
2.35.1
author maintainer
#285 lore raw

PATCH 282/731 hwmon: (jc42) Restore the min/max/critical temperatures on resume

In reply to Greg Kroah-Hartman

From: Martin Blumenstingl <[email protected]>

[ Upstream commit 084ed144c448fd5bc8ed5a58247153fbbfd115c3 ]

The JC42 compatible thermal sensor on Kingston KSM32ES8/16ME DIMMs (using Micron E-Die) is an ST Microelectronics STTS2004 (manufacturer 0x104a, device 0x2201). It does not keep the previously programmed minimum, maximum and critical temperatures after system suspend and resume (which is a shutdown / startup cycle for the JC42 temperature sensor). This results in an alarm on system resume because the hardware default for these values is 0°C (so any environment temperature greater than 0°C will trigger the alarm).

Example before system suspend:
jc42-i2c-0-1a
Adapter: SMBus PIIX4 adapter port 0 at 0b00
temp1: +34.8°C (low = +0.0°C)

                         (high = +85.0°C, hyst = +85.0°C)
                         (crit = +95.0°C, hyst = +95.0°C)

Example after system resume (without this change):
jc42-i2c-0-1a
Adapter: SMBus PIIX4 adapter port 0 at 0b00
temp1: +34.8°C (low = +0.0°C) ALARM (HIGH, CRIT)

                         (high =  +0.0°C, hyst =  +0.0°C)
                         (crit =  +0.0°C, hyst =  +0.0°C)

Apply the cached values from the JC42_REG_TEMP_UPPER, JC42_REG_TEMP_LOWER, JC42_REG_TEMP_CRITICAL and JC42_REG_SMBUS (where the SMBUS register is not related to this issue but a side-effect of using regcache_sync() during system resume with the previously cached/programmed values. This fixes the alarm due to the hardware defaults of 0°C because the previously applied limits (set by userspace) are re-applied on system resume.

Fixes 175c490c9e7f ("hwmon: (jc42) Add support for STTS2004 and AT30TSE004")
---
 drivers/hwmon/jc42.c | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)
drivers/hwmon/jc42.c +9 -1
index 9a2a062eb7b8..5240bfdfcf2e 100644
@@ -567,6 +567,10 @@ static int jc42_suspend(struct device *dev)
 
 	data->config |= JC42_CFG_SHUTDOWN;
 	regmap_write(data->regmap, JC42_REG_CONFIG, data->config);
+
+	regcache_cache_only(data->regmap, true);
+	regcache_mark_dirty(data->regmap);
+
 	return 0;
 }
 
@@ -574,9 +578,13 @@ static int jc42_resume(struct device *dev)
 {
 	struct jc42_data *data = dev_get_drvdata(dev);
 
+	regcache_cache_only(data->regmap, false);
+
 	data->config &= ~JC42_CFG_SHUTDOWN;
 	regmap_write(data->regmap, JC42_REG_CONFIG, data->config);
-	return 0;
+
+	/* Restore cached register values to hardware */
+	return regcache_sync(data->regmap);
 }
 
 static const struct dev_pm_ops jc42_dev_pm_ops = {
Signature · 1 line
2.35.1
author maintainer
#286 lore raw

PATCH 283/731 bpf, sockmap: fix race in sock_map_free()

In reply to Greg Kroah-Hartman

From: Eric Dumazet <[email protected]>

[ Upstream commit 0a182f8d607464911756b4dbef5d6cad8de22469 ]

sock_map_free() calls release_sock(sk) without owning a reference on the socket. This can cause use-after-free as syzbot found [1]

Jakub Sitnicki already took care of a similar issue in sock_hash_free() in commit 75e68e5bf2c7 ("bpf, sockhash:
Synchronize delete from bucket list on map free")

[1]
refcount_t: decrement hit 0; leaking memory.

WARNING: CPU: 0 PID: 3785 at lib/refcount.c:31 refcount_warn_saturate+0x17c/0x1a0 lib/refcount.c:31
Modules linked in:
CPU: 0 PID: 3785 Comm: kworker/u4:6 Not tainted 6.1.0-rc7-syzkaller-00103-gef4d3ea40565 #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 10/26/2022
Workqueue: events_unbound bpf_map_free_deferred
RIP: 0010:refcount_warn_saturate+0x17c/0x1a0 lib/refcount.c:31
Code: 68 8b 31 c0 e8 75 71 15 fd 0f 0b e9 64 ff ff ff e8 d9 6e 4e fd c6 05 62 9c 3d 0a 01 48 c7 c7 80 bb 68 8b 31 c0 e8 54 71 15 fd <0f> 0b e9 43 ff ff ff 89 d9 80 e1 07 80 c1 03 38 c1 0f 8c a2 fe ff
RSP: 0018:ffffc9000456fb60 EFLAGS: 00010246
RAX: eae59bab72dcd700 RBX: 0000000000000004 RCX: ffff8880207057c0
RDX: 0000000000000000 RSI: 0000000000000201 RDI: 0000000000000000
RBP: 0000000000000004 R08: ffffffff816fdabd R09: fffff520008adee5
R10: fffff520008adee5 R11: 1ffff920008adee4 R12: 0000000000000004
R13: dffffc0000000000 R14: ffff88807b1c6c00 R15: 1ffff1100f638dcf
FS: 0000000000000000(0000) GS:ffff8880b9800000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000001b30c30000 CR3: 000000000d08e000 CR4: 00000000003506f0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Call Trace:
<TASK>
__refcount_dec include/linux/refcount.h:344 [inline]
refcount_dec include/linux/refcount.h:359 [inline]
__sock_put include/net/sock.h:779 [inline]
tcp_release_cb+0x2d0/0x360 net/ipv4/tcp_output.c:1092
release_sock+0xaf/0x1c0 net/core/sock.c:3468
sock_map_free+0x219/0x2c0 net/core/sock_map.c:356
process_one_work+0x81c/0xd10 kernel/workqueue.c:2289
worker_thread+0xb14/0x1330 kernel/workqueue.c:2436
kthread+0x266/0x300 kernel/kthread.c:376
ret_from_fork+0x1f/0x30 arch/x86/entry/entry_64.S:306
</TASK>
Fixes 7e81a3530206 ("bpf: Sockmap, ensure sock lock held during tear down")
Cc Jakub Sitnicki <[email protected]>
Cc John Fastabend <[email protected]>
Cc Alexei Starovoitov <[email protected]>
Cc Daniel Borkmann <[email protected]>
Cc Song Liu <[email protected]>
---
 net/core/sock_map.c | 2 ++
 1 file changed, 2 insertions(+)
net/core/sock_map.c +2 -0
index 4f4bc163a223..ae6013a8bce5 100644
@@ -349,11 +349,13 @@ static void sock_map_free(struct bpf_map *map)
 
 		sk = xchg(psk, NULL);
 		if (sk) {
+			sock_hold(sk);
 			lock_sock(sk);
 			rcu_read_lock();
 			sock_map_unref(sk, psk);
 			rcu_read_unlock();
 			release_sock(sk);
+			sock_put(sk);
 		}
 	}
 
Signature · 1 line
2.35.1
author maintainer
#287 lore raw

PATCH 284/731 ALSA: pcm: Set missing stop_operating flag at undoing trigger start

In reply to Greg Kroah-Hartman

From: Takashi Iwai <[email protected]>

[ Upstream commit 5c8cc93b06d1ff860327a273abf3ac006290d242 ]

When a PCM trigger-start fails at snd_pcm_do_start(), PCM core tries to undo the action at snd_pcm_undo_start() by issuing the trigger STOP manually. At that point, we forgot to set the stop_operating flag, hence the sync-stop won't be issued at the next prepare or other calls.

This patch adds the missing stop_operating flag at snd_pcm_undo_start().

---
 sound/core/pcm_native.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
sound/core/pcm_native.c +3 -1
index f38c2e5e9a29..44e06ef4ff0b 100644
@@ -1419,8 +1419,10 @@ static int snd_pcm_do_start(struct snd_pcm_substream *substream,
 static void snd_pcm_undo_start(struct snd_pcm_substream *substream,
 			       snd_pcm_state_t state)
 {
-	if (substream->runtime->trigger_master == substream)
+	if (substream->runtime->trigger_master == substream) {
 		substream->ops->trigger(substream, SNDRV_PCM_TRIGGER_STOP);
+		substream->runtime->stop_operating = true;
+	}
 }
 
 static void snd_pcm_post_start(struct snd_pcm_substream *substream,
Signature · 1 line
2.35.1
author maintainer
#288 lore raw

PATCH 285/731 media: saa7164: fix missing pci_disable_device()

In reply to Greg Kroah-Hartman

From: Liu Shixin <[email protected]>

[ Upstream commit 57fb35d7542384cac8f198cd1c927540ad38b61a ]

Add missing pci_disable_device() in the error path in saa7164_initdev().

Fixes 443c1228d505 ("V4L/DVB (12923): SAA7164: Add support for the NXP SAA7164 silicon")
---
 drivers/media/pci/saa7164/saa7164-core.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/media/pci/saa7164/saa7164-core.c +3 -1
index 7973ae42873a..c10997e2271d 100644
@@ -1259,7 +1259,7 @@ static int saa7164_initdev(struct pci_dev *pci_dev,
 
 	if (saa7164_dev_setup(dev) < 0) {
 		err = -EINVAL;
-		goto fail_free;
+		goto fail_dev;
 	}
 
 	/* print pci info */
@@ -1427,6 +1427,8 @@ static int saa7164_initdev(struct pci_dev *pci_dev,
 
 fail_irq:
 	saa7164_dev_unregister(dev);
+fail_dev:
+	pci_disable_device(pci_dev);
 fail_free:
 	v4l2_device_unregister(&dev->v4l2_dev);
 	kfree(dev);
Signature · 1 line
2.35.1
author maintainer
#289 lore raw

PATCH 286/731 ALSA: mts64: fix possible null-ptr-defer in snd_mts64_interrupt

In reply to Greg Kroah-Hartman

From: Gaosheng Cui <[email protected]>

[ Upstream commit cf2ea3c86ad90d63d1c572b43e1ca9276b0357ad ]

I got a null-ptr-defer error report when I do the following tests on the qemu platform:

make defconfig and CONFIG_PARPORT=m, CONFIG_PARPORT_PC=m, CONFIG_SND_MTS64=m

Then making test scripts:
cat>test_mod1.sh<<EOF
modprobe snd-mts64
modprobe snd-mts64
EOF

Executing the script, perhaps several times, we will get a null-ptr-defer report, as follow:

syzkaller:~# ./test_mod.sh
snd_mts64: probe of snd_mts64.0 failed with error -5
modprobe: ERROR: could not insert 'snd_mts64': No such device
 BUG: kernel NULL pointer dereference, address: 0000000000000000
 #PF: supervisor write access in kernel mode
 #PF: error_code(0x0002) - not-present page
 PGD 0 P4D 0
 Oops: 0002 [#1] PREEMPT SMP PTI
 CPU: 0 PID: 205 Comm: modprobe Not tainted 6.1.0-rc8-00588-g76dcd734eca2 #6
 Call Trace:
  <IRQ>
  snd_mts64_interrupt+0x24/0xa0 [snd_mts64]
  parport_irq_handler+0x37/0x50 [parport]
  __handle_irq_event_percpu+0x39/0x190
  handle_irq_event_percpu+0xa/0x30
  handle_irq_event+0x2f/0x50
  handle_edge_irq+0x99/0x1b0
  __common_interrupt+0x5d/0x100
  common_interrupt+0xa0/0xc0
  </IRQ>
  <TASK>
  asm_common_interrupt+0x22/0x40
 RIP: 0010:_raw_write_unlock_irqrestore+0x11/0x30
  parport_claim+0xbd/0x230 [parport]
  snd_mts64_probe+0x14a/0x465 [snd_mts64]
  platform_probe+0x3f/0xa0
  really_probe+0x129/0x2c0
  __driver_probe_device+0x6d/0xc0
  driver_probe_device+0x1a/0xa0
  __device_attach_driver+0x7a/0xb0
  bus_for_each_drv+0x62/0xb0
  __device_attach+0xe4/0x180
  bus_probe_device+0x82/0xa0
  device_add+0x550/0x920
  platform_device_add+0x106/0x220
  snd_mts64_attach+0x2e/0x80 [snd_mts64]
  port_check+0x14/0x20 [parport]
  bus_for_each_dev+0x6e/0xc0
  __parport_register_driver+0x7c/0xb0 [parport]
  snd_mts64_module_init+0x31/0x1000 [snd_mts64]
  do_one_initcall+0x3c/0x1f0
  do_init_module+0x46/0x1c6
  load_module+0x1d8d/0x1e10
  __do_sys_finit_module+0xa2/0xf0
  do_syscall_64+0x37/0x90
  entry_SYSCALL_64_after_hwframe+0x63/0xcd
  </TASK>
 Kernel panic - not syncing: Fatal exception in interrupt
 Rebooting in 1 seconds..

The mts wa not initialized during interrupt, we add check for mts to fix this bug.

Fixes 68ab801e32bb ("[ALSA] Add snd-mts64 driver for ESI Miditerminal 4140")
---
 sound/drivers/mts64.c | 3 +++
 1 file changed, 3 insertions(+)
sound/drivers/mts64.c +3 -0
index d3bc9e8c407d..f0d34cf70c3e 100644
@@ -815,6 +815,9 @@ static void snd_mts64_interrupt(void *private)
 	u8 status, data;
 	struct snd_rawmidi_substream *substream;
 
+	if (!mts)
+		return;
+
 	spin_lock(&mts->lock);
 	ret = mts64_read(mts->pardev->port);
 	data = ret & 0x00ff;
Signature · 1 line
2.35.1
author maintainer
#290 lore raw

PATCH 287/731 xprtrdma: Fix regbuf data not freed in rpcrdma_req_create()

In reply to Greg Kroah-Hartman

From: Zhang Xiaoxu <[email protected]>

[ Upstream commit 9181f40fb2952fd59ecb75e7158620c9c669eee3 ]

If rdma receive buffer allocate failed, should call rpcrdma_regbuf_free() to free the send buffer, otherwise, the buffer data will be leaked.

Fixes bb93a1ae2bf4 ("xprtrdma: Allocate req's regbufs at xprt create time")
---
 net/sunrpc/xprtrdma/verbs.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
net/sunrpc/xprtrdma/verbs.c +1 -1
index 1295f9ab839f..507ba8b79992 100644
@@ -858,7 +858,7 @@ struct rpcrdma_req *rpcrdma_req_create(struct rpcrdma_xprt *r_xprt, size_t size,
 	return req;
 
 out3:
-	kfree(req->rl_sendbuf);
+	rpcrdma_regbuf_free(req->rl_sendbuf);
 out2:
 	kfree(req);
 out1:
Signature · 1 line
2.35.1
author maintainer
#291 lore raw

PATCH 288/731 SUNRPC: Fix missing release socket in rpc_sockname()

In reply to Greg Kroah-Hartman

From: Wang ShaoBo <[email protected]>

[ Upstream commit 50fa355bc0d75911fe9d5072a5ba52cdb803aff7 ]

socket dynamically created is not released when getting an unintended address family type in rpc_sockname(), direct to out_release for calling sock_release().

Fixes 2e738fdce22f ("SUNRPC: Add API to acquire source address")
---
 net/sunrpc/clnt.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
net/sunrpc/clnt.c +1 -1
index ca2a494d727b..bbeb80e1133d 100644
@@ -1375,7 +1375,7 @@ static int rpc_sockname(struct net *net, struct sockaddr *sap, size_t salen,
 		break;
 	default:
 		err = -EAFNOSUPPORT;
-		goto out;
+		goto out_release;
 	}
 	if (err < 0) {
 		dprintk("RPC:       can't bind UDP socket (%d)\n", err);
Signature · 1 line
2.35.1
author maintainer
#292 lore raw

PATCH 289/731 NFSv4.x: Fail client initialisation if state manager thread cant run

In reply to Greg Kroah-Hartman

From: Trond Myklebust <[email protected]>

[ Upstream commit b4e4f66901658fae0614dea5bf91062a5387eda7 ]

If the state manager thread fails to start, then we should just mark the client initialisation as failed so that other processes or threads don't get stuck in nfs_wait_client_init_complete().

Fixes 4697bd5e9419 ("NFSv4: Fix a race in the net namespace mount notification")
---
 fs/nfs/nfs4state.c | 2 ++
 1 file changed, 2 insertions(+)
fs/nfs/nfs4state.c +2 -0
index ecac56be6cb7..0cd803b4d90c 100644
@@ -1227,6 +1227,8 @@ void nfs4_schedule_state_manager(struct nfs_client *clp)
 	if (IS_ERR(task)) {
 		printk(KERN_ERR "%s: kthread_run: %ld\n",
 			__func__, PTR_ERR(task));
+		if (!nfs_client_init_is_complete(clp))
+			nfs_mark_client_ready(clp, PTR_ERR(task));
 		nfs4_clear_state_manager_bit(clp);
 		nfs_put_client(clp);
 		module_put(THIS_MODULE);
Signature · 1 line
2.35.1
author maintainer
#293 lore raw

PATCH 290/731 riscv, bpf: Emit fixed-length instructions for BPF_PSEUDO_FUNC

In reply to Greg Kroah-Hartman

From: Pu Lehui <[email protected]>

[ Upstream commit b54b6003612a376e7be32cbc5c1af3754bbbbb3d ]

For BPF_PSEUDO_FUNC instruction, verifier will refill imm with correct addresses of bpf_calls and then run last pass of JIT. Since the emit_imm of RV64 is variable-length, which will emit appropriate length instructions accorroding to the imm, it may broke ctx->offset, and lead to unpredictable problem, such as inaccurate jump. So let's fix it with fixed-length instructions.

---
 arch/riscv/net/bpf_jit_comp64.c | 29 ++++++++++++++++++++++++++++-
 1 file changed, 28 insertions(+), 1 deletion(-)
arch/riscv/net/bpf_jit_comp64.c +28 -1
index 3af4131c22c7..2e3f1a626a3a 100644
@@ -120,6 +120,25 @@ static bool in_auipc_jalr_range(s64 val)
 		val < ((1L << 31) - (1L << 11));
 }
 
+/* Emit fixed-length instructions for address */
+static int emit_addr(u8 rd, u64 addr, bool extra_pass, struct rv_jit_context *ctx)
+{
+	u64 ip = (u64)(ctx->insns + ctx->ninsns);
+	s64 off = addr - ip;
+	s64 upper = (off + (1 << 11)) >> 12;
+	s64 lower = off & 0xfff;
+
+	if (extra_pass && !in_auipc_jalr_range(off)) {
+		pr_err("bpf-jit: target offset 0x%llx is out of range\n", off);
+		return -ERANGE;
+	}
+
+	emit(rv_auipc(rd, upper), ctx);
+	emit(rv_addi(rd, rd, lower), ctx);
+	return 0;
+}
+
+/* Emit variable-length instructions for 32-bit and 64-bit imm */
 static void emit_imm(u8 rd, s64 val, struct rv_jit_context *ctx)
 {
 	/* Note that the immediate from the add is sign-extended,
@@ -887,7 +906,15 @@ int bpf_jit_emit_insn(const struct bpf_insn *insn, struct rv_jit_context *ctx,
 		u64 imm64;
 
 		imm64 = (u64)insn1.imm << 32 | (u32)imm;
-		emit_imm(rd, imm64, ctx);
+		if (bpf_pseudo_func(insn)) {
+			/* fixed-length insns for extra jit pass */
+			ret = emit_addr(rd, imm64, extra_pass, ctx);
+			if (ret)
+				return ret;
+		} else {
+			emit_imm(rd, imm64, ctx);
+		}
+
 		return 1;
 	}
 
Signature · 1 line
2.35.1
author maintainer
#294 lore raw

PATCH 291/731 mmc: alcor: fix return value check of mmc_add_host()

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit e93d1468f429475a753d6baa79b853b7ee5ef8c0 ]

mmc_add_host() may return error, if we ignore its return value, the memory that allocated in mmc_alloc_host() will be leaked and it will lead a kernel crash because of deleting not added device in the remove path.

So fix this by checking the return value and calling mmc_free_host() in the error path.

Fixes c5413ad815a6 ("mmc: add new Alcor Micro Cardreader SD/MMC driver")
---
 drivers/mmc/host/alcor.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
drivers/mmc/host/alcor.c +4 -1
index bfb8efeb7eb8..d01df01d4b4d 100644
@@ -1114,7 +1114,10 @@ static int alcor_pci_sdmmc_drv_probe(struct platform_device *pdev)
 	alcor_hw_init(host);
 
 	dev_set_drvdata(&pdev->dev, host);
-	mmc_add_host(mmc);
+	ret = mmc_add_host(mmc);
+	if (ret)
+		goto free_host;
+
 	return 0;
 
 free_host:
Signature · 1 line
2.35.1
author maintainer
#295 lore raw

PATCH 292/731 mmc: moxart: fix return value check of mmc_add_host()

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit 0ca18d09c744fb030ae9bc5836c3e357e0237dea ]

mmc_add_host() may return error, if we ignore its return value, the memory that allocated in mmc_alloc_host() will be leaked and it will lead a kernel crash because of deleting not added device in the remove path.

So fix this by checking the return value and goto error path which will call mmc_free_host().

Fixes 1b66e94e6b99 ("mmc: moxart: Add MOXA ART SD/MMC driver")
---
 drivers/mmc/host/moxart-mmc.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/mmc/host/moxart-mmc.c +3 -1
index dfc3ffd5b1f8..52ed30f2d9f4 100644
@@ -665,7 +665,9 @@ static int moxart_probe(struct platform_device *pdev)
 		goto out;
 
 	dev_set_drvdata(dev, mmc);
-	mmc_add_host(mmc);
+	ret = mmc_add_host(mmc);
+	if (ret)
+		goto out;
 
 	dev_dbg(dev, "IRQ=%d, FIFO is %d bytes\n", irq, host->fifo_width);
 
Signature · 1 line
2.35.1
author maintainer
#296 lore raw

PATCH 293/731 mmc: mxcmmc: fix return value check of mmc_add_host()

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit cde600af7b413c9fe03e85c58c4279df90e91d13 ]

mmc_add_host() may return error, if we ignore its return value, the memory that allocated in mmc_alloc_host() will be leaked and it will lead a kernel crash because of deleting not added device in the remove path.

So fix this by checking the return value and goto error path which will call mmc_free_host().

Fixes d96be879ff46 ("mmc: Add a MX2/MX3 specific SDHC driver")
---
 drivers/mmc/host/mxcmmc.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/mmc/host/mxcmmc.c +3 -1
index 9bf95ba217fa..97227ad71715 100644
@@ -1143,7 +1143,9 @@ static int mxcmci_probe(struct platform_device *pdev)
 
 	timer_setup(&host->watchdog, mxcmci_watchdog, 0);
 
-	mmc_add_host(mmc);
+	ret = mmc_add_host(mmc);
+	if (ret)
+		goto out_free_dma;
 
 	return 0;
 
Signature · 1 line
2.35.1
author maintainer
#297 lore raw

PATCH 294/731 mmc: pxamci: fix return value check of mmc_add_host()

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit 80e1ef3afb8bfbe768380b70ffe1b6cab87d1a3b ]

mmc_add_host() may return error, if we ignore its return value, the memory that allocated in mmc_alloc_host() will be leaked and it will lead a kernel crash because of deleting not added device in the remove path.

So fix this by checking the return value and goto error path which will call mmc_free_host(), besides, ->exit() need be called to uninit the pdata.

---
 drivers/mmc/host/pxamci.c | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)
drivers/mmc/host/pxamci.c +6 -1
index 55868b6b8658..e25e9bb34eb3 100644
@@ -763,7 +763,12 @@ static int pxamci_probe(struct platform_device *pdev)
 			dev_warn(dev, "gpio_ro and get_ro() both defined\n");
 	}
 
-	mmc_add_host(mmc);
+	ret = mmc_add_host(mmc);
+	if (ret) {
+		if (host->pdata && host->pdata->exit)
+			host->pdata->exit(dev, mmc);
+		goto out;
+	}
 
 	return 0;
 
Signature · 1 line
2.35.1
author maintainer
#298 lore raw

PATCH 295/731 mmc: rtsx_pci: fix return value check of mmc_add_host()

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit 0c87db77423a282b3b38b8a6daf057b822680516 ]

mmc_add_host() may return error, if we ignore its return value, the memory that allocated in mmc_alloc_host() will be leaked and it will lead a kernel crash because of deleting not added device in the remove path.

So fix this by checking the return value and calling mmc_free_host() in the error path, beside, runtime PM also needs be disabled.

Fixes ff984e57d36e ("mmc: Add realtek pcie sdmmc host driver")
---
 drivers/mmc/host/rtsx_pci_sdmmc.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)
drivers/mmc/host/rtsx_pci_sdmmc.c +8 -1
index e1580f78c6b2..8098726dcc0b 100644
@@ -1474,6 +1474,7 @@ static int rtsx_pci_sdmmc_drv_probe(struct platform_device *pdev)
 	struct realtek_pci_sdmmc *host;
 	struct rtsx_pcr *pcr;
 	struct pcr_handle *handle = pdev->dev.platform_data;
+	int ret;
 
 	if (!handle)
 		return -ENXIO;
@@ -1511,7 +1512,13 @@ static int rtsx_pci_sdmmc_drv_probe(struct platform_device *pdev)
 	pm_runtime_mark_last_busy(&pdev->dev);
 	pm_runtime_use_autosuspend(&pdev->dev);
 
-	mmc_add_host(mmc);
+	ret = mmc_add_host(mmc);
+	if (ret) {
+		pm_runtime_dont_use_autosuspend(&pdev->dev);
+		pm_runtime_disable(&pdev->dev);
+		mmc_free_host(mmc);
+		return ret;
+	}
 
 	return 0;
 }
Signature · 1 line
2.35.1
author maintainer
#299 lore raw

PATCH 296/731 mmc: rtsx_usb_sdmmc: fix return value check of mmc_add_host()

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit fc38a5a10e9e5a75eb9189854abeb8405b214cc9 ]

mmc_add_host() may return error, if we ignore its return value, the memory that allocated in mmc_alloc_host() will be leaked and it will lead a kernel crash because of deleting not added device in the remove path.

So fix this by checking the return value and calling mmc_free_host() in the error path, besides, led_classdev_unregister() and pm_runtime_disable() also need be called.

Fixes c7f6558d84af ("mmc: Add realtek USB sdmmc host driver")
---
 drivers/mmc/host/rtsx_usb_sdmmc.c | 11 ++++++++++-
 1 file changed, 10 insertions(+), 1 deletion(-)
drivers/mmc/host/rtsx_usb_sdmmc.c +10 -1
index 5fe4528e296e..1be3a355f10d 100644
@@ -1332,6 +1332,7 @@ static int rtsx_usb_sdmmc_drv_probe(struct platform_device *pdev)
 #ifdef RTSX_USB_USE_LEDS_CLASS
 	int err;
 #endif
+	int ret;
 
 	ucr = usb_get_intfdata(to_usb_interface(pdev->dev.parent));
 	if (!ucr)
@@ -1368,7 +1369,15 @@ static int rtsx_usb_sdmmc_drv_probe(struct platform_device *pdev)
 	INIT_WORK(&host->led_work, rtsx_usb_update_led);
 
 #endif
-	mmc_add_host(mmc);
+	ret = mmc_add_host(mmc);
+	if (ret) {
+#ifdef RTSX_USB_USE_LEDS_CLASS
+		led_classdev_unregister(&host->led);
+#endif
+		mmc_free_host(mmc);
+		pm_runtime_disable(&pdev->dev);
+		return ret;
+	}
 
 	return 0;
 }
Signature · 1 line
2.35.1
author maintainer
#300 lore raw

PATCH 297/731 mmc: toshsd: fix return value check of mmc_add_host()

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit f670744a316ea983113a65313dcd387b5a992444 ]

mmc_add_host() may return error, if we ignore its return value, the memory that allocated in mmc_alloc_host() will be leaked and it will lead a kernel crash because of deleting not added device in the remove path.

So fix this by checking the return value and goto error path which will call mmc_free_host(), besides, free_irq() also needs be called.

Fixes a5eb8bbd66cc ("mmc: add Toshiba PCI SD controller driver")
---
 drivers/mmc/host/toshsd.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
drivers/mmc/host/toshsd.c +5 -1
index 8d037c2071ab..497791ffada6 100644
@@ -651,7 +651,9 @@ static int toshsd_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 	if (ret)
 		goto unmap;
 
-	mmc_add_host(mmc);
+	ret = mmc_add_host(mmc);
+	if (ret)
+		goto free_irq;
 
 	base = pci_resource_start(pdev, 0);
 	dev_dbg(&pdev->dev, "MMIO %pa, IRQ %d\n", &base, pdev->irq);
@@ -660,6 +662,8 @@ static int toshsd_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 
 	return 0;
 
+free_irq:
+	free_irq(pdev->irq, host);
 unmap:
 	pci_iounmap(pdev, host->ioaddr);
 release:
Signature · 1 line
2.35.1
author maintainer
#301 lore raw

PATCH 298/731 mmc: vub300: fix return value check of mmc_add_host()

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit 0613ad2401f88bdeae5594c30afe318e93b14676 ]

mmc_add_host() may return error, if we ignore its return value, the memory that allocated in mmc_alloc_host() will be leaked and it will lead a kernel crash because of deleting not added device in the remove path.

So fix this by checking the return value and goto error path which will call mmc_free_host(), besides, the timer added before mmc_add_host() needs be del.

And this patch fixes another missing call mmc_free_host() if usb_control_msg() fails.

Fixes 88095e7b473a ("mmc: Add new VUB300 USB-to-SD/SDIO/MMC driver")
---
 drivers/mmc/host/vub300.c | 11 ++++++++---
 1 file changed, 8 insertions(+), 3 deletions(-)
drivers/mmc/host/vub300.c +8 -3
index 97beece62fec..ab36ec479747 100644
@@ -2299,14 +2299,14 @@ static int vub300_probe(struct usb_interface *interface,
 				0x0000, 0x0000, &vub300->system_port_status,
 				sizeof(vub300->system_port_status), 1000);
 	if (retval < 0) {
-		goto error4;
+		goto error5;
 	} else if (sizeof(vub300->system_port_status) == retval) {
 		vub300->card_present =
 			(0x0001 & vub300->system_port_status.port_flags) ? 1 : 0;
 		vub300->read_only =
 			(0x0010 & vub300->system_port_status.port_flags) ? 1 : 0;
 	} else {
-		goto error4;
+		goto error5;
 	}
 	usb_set_intfdata(interface, vub300);
 	INIT_DELAYED_WORK(&vub300->pollwork, vub300_pollwork_thread);
@@ -2329,8 +2329,13 @@ static int vub300_probe(struct usb_interface *interface,
 			 "USB vub300 remote SDIO host controller[%d]"
 			 "connected with no SD/SDIO card inserted\n",
 			 interface_to_InterfaceNumber(interface));
-	mmc_add_host(mmc);
+	retval = mmc_add_host(mmc);
+	if (retval)
+		goto error6;
+
 	return 0;
+error6:
+	del_timer_sync(&vub300->inactivity_timer);
 error5:
 	mmc_free_host(mmc);
 	/*
Signature · 1 line
2.35.1
author maintainer
#302 lore raw

PATCH 299/731 mmc: wmt-sdmmc: fix return value check of mmc_add_host()

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit 29276d56f6ed138db0f38cd31aedc0b725c8c76c ]

mmc_add_host() may return error, if we ignore its return value, the memory that allocated in mmc_alloc_host() will be leaked and it will lead a kernel crash because of deleting not added device in the remove path.

So fix this by checking the return value and goto error path which will call mmc_free_host(), besides, clk_disable_unprepare() also needs be called.

Fixes 3a96dff0f828 ("mmc: SD/MMC Host Controller for Wondermedia WM8505/WM8650")
---
 drivers/mmc/host/wmt-sdmmc.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
drivers/mmc/host/wmt-sdmmc.c +5 -1
index 8df722ec57ed..393319548857 100644
@@ -859,11 +859,15 @@ static int wmt_mci_probe(struct platform_device *pdev)
 	/* configure the controller to a known 'ready' state */
 	wmt_reset_hardware(mmc);
 
-	mmc_add_host(mmc);
+	ret = mmc_add_host(mmc);
+	if (ret)
+		goto fail7;
 
 	dev_info(&pdev->dev, "WMT SDHC Controller initialized\n");
 
 	return 0;
+fail7:
+	clk_disable_unprepare(priv->clk_sdmmc);
 fail6:
 	clk_put(priv->clk_sdmmc);
 fail5_and_a_half:
Signature · 1 line
2.35.1
author maintainer
#303 lore raw

PATCH 300/731 mmc: atmel-mci: fix return value check of mmc_add_host()

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit 9e6e8c43726673ca2abcaac87640b9215fd72f4c ]

mmc_add_host() may return error, if we ignore its return value, it will lead two issues:
1. The memory that allocated in mmc_alloc_host() is leaked.
2. In the remove() path, mmc_remove_host() will be called to delete device, but it's not added yet, it will lead a kernel crash because of null-ptr-deref in device_del().

So fix this by checking the return value and calling mmc_free_host() in the error path.

Fixes 7d2be0749a59 ("atmel-mci: Driver for Atmel on-chip MMC controllers")
---
 drivers/mmc/host/atmel-mci.c | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)
drivers/mmc/host/atmel-mci.c +6 -3
index 807177c953f3..6f971a3e7e49 100644
@@ -2223,6 +2223,7 @@ static int atmci_init_slot(struct atmel_mci *host,
 {
 	struct mmc_host			*mmc;
 	struct atmel_mci_slot		*slot;
+	int ret;
 
 	mmc = mmc_alloc_host(sizeof(struct atmel_mci_slot), &host->pdev->dev);
 	if (!mmc)
@@ -2306,11 +2307,13 @@ static int atmci_init_slot(struct atmel_mci *host,
 
 	host->slot[id] = slot;
 	mmc_regulator_get_supply(mmc);
-	mmc_add_host(mmc);
+	ret = mmc_add_host(mmc);
+	if (ret) {
+		mmc_free_host(mmc);
+		return ret;
+	}
 
 	if (gpio_is_valid(slot->detect_pin)) {
-		int ret;
-
 		timer_setup(&slot->detect_timer, atmci_detect_change, 0);
 
 		ret = request_irq(gpio_to_irq(slot->detect_pin),
Signature · 1 line
2.35.1
author maintainer
#304 lore raw

PATCH 301/731 mmc: omap_hsmmc: fix return value check of mmc_add_host()

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit a525cad241c339ca00bf7ebf03c5180f2a9b767c ]

mmc_add_host() may return error, if we ignore its return value, it will lead two issues:
1. The memory that allocated in mmc_alloc_host() is leaked.
2. In the remove() path, mmc_remove_host() will be called to delete device, but it's not added yet, it will lead a kernel crash because of null-ptr-deref in device_del().

Fix this by checking the return value and goto error path wihch will call mmc_free_host().

Fixes a45c6cb81647 ("[ARM] 5369/1: omap mmc: Add new omap hsmmc controller for 2430 and 34xx, v3")
---
 drivers/mmc/host/omap_hsmmc.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/mmc/host/omap_hsmmc.c +3 -1
index 2f8038d69f67..eb0bd46b7e81 100644
@@ -1987,7 +1987,9 @@ static int omap_hsmmc_probe(struct platform_device *pdev)
 	if (!ret)
 		mmc->caps |= MMC_CAP_SDIO_IRQ;
 
-	mmc_add_host(mmc);
+	ret = mmc_add_host(mmc);
+	if (ret)
+		goto err_irq;
 
 	if (mmc_pdata(host)->name != NULL) {
 		ret = device_create_file(&mmc->class_dev, &dev_attr_slot_name);
Signature · 1 line
2.35.1
author maintainer
#305 lore raw

PATCH 302/731 mmc: meson-gx: fix return value check of mmc_add_host()

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit 90935f16f2650ab7416fa2ffbe5c28cb39cf3f1e ]

mmc_add_host() may return error, if we ignore its return value, it will lead two issues:
1. The memory that allocated in mmc_alloc_host() is leaked.
2. In the remove() path, mmc_remove_host() will be called to delete device, but it's not added yet, it will lead a kernel crash because of null-ptr-deref in device_del().

Fix this by checking the return value and goto error path which will call mmc_free_host().

Fixes 51c5d8447bd7 ("MMC: meson: initial support for GX platforms")
---
 drivers/mmc/host/meson-gx-mmc.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/mmc/host/meson-gx-mmc.c +3 -1
index 9b2e2548bd18..753f9ea254d4 100644
@@ -1291,7 +1291,9 @@ static int meson_mmc_probe(struct platform_device *pdev)
 	}
 
 	mmc->ops = &meson_mmc_ops;
-	mmc_add_host(mmc);
+	ret = mmc_add_host(mmc);
+	if (ret)
+		goto err_free_irq;
 
 	return 0;
 
Signature · 1 line
2.35.1
author maintainer
#306 lore raw

PATCH 303/731 mmc: via-sdmmc: fix return value check of mmc_add_host()

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit e4e46fb61e3bb4628170810d3f2b996b709b90d9 ]

mmc_add_host() may return error, if we ignore its return value, it will lead two issues:
1. The memory that allocated in mmc_alloc_host() is leaked.
2. In the remove() path, mmc_remove_host() will be called to delete device, but it's not added yet, it will lead a kernel crash because of null-ptr-deref in device_del().

Fix this by checking the return value and goto error path which will call mmc_free_host().

Fixes f0bf7f61b840 ("mmc: Add new via-sdmmc host controller driver")
---
 drivers/mmc/host/via-sdmmc.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/mmc/host/via-sdmmc.c +3 -1
index 88662a90ed96..a2b0d9461665 100644
@@ -1151,7 +1151,9 @@ static int via_sd_probe(struct pci_dev *pcidev,
 	    pcidev->subsystem_device == 0x3891)
 		sdhost->quirks = VIA_CRDR_QUIRK_300MS_PWRDELAY;
 
-	mmc_add_host(mmc);
+	ret = mmc_add_host(mmc);
+	if (ret)
+		goto unmap;
 
 	return 0;
 
Signature · 1 line
2.35.1
author maintainer
#307 lore raw

PATCH 304/731 mmc: wbsd: fix return value check of mmc_add_host()

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit dc5b9b50fc9d1334407e316e6e29a5097ef833bd ]

mmc_add_host() may return error, if we ignore its return value, it will lead two issues:
1. The memory that allocated in mmc_alloc_host() is leaked.
2. In the remove() path, mmc_remove_host() will be called to delete device, but it's not added yet, it will lead a kernel crash because of null-ptr-deref in device_del().

So fix this by checking the return value and goto error path which will call mmc_free_host(), besides, other resources also need be released.

---
 drivers/mmc/host/wbsd.c | 12 +++++++++++-
 1 file changed, 11 insertions(+), 1 deletion(-)
drivers/mmc/host/wbsd.c +11 -1
index 67ecd342fe5f..7c7ec8d10232 100644
@@ -1698,7 +1698,17 @@ static int wbsd_init(struct device *dev, int base, int irq, int dma,
 	 */
 	wbsd_init_device(host);
 
-	mmc_add_host(mmc);
+	ret = mmc_add_host(mmc);
+	if (ret) {
+		if (!pnp)
+			wbsd_chip_poweroff(host);
+
+		wbsd_release_resources(host);
+		wbsd_free_mmc(dev);
+
+		mmc_free_host(mmc);
+		return ret;
+	}
 
 	pr_info("%s: W83L51xD", mmc_hostname(mmc));
 	if (host->chip_id != 0)
Signature · 1 line
2.35.1
author maintainer
#308 lore raw

PATCH 305/731 mmc: mmci: fix return value check of mmc_add_host()

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit b38a20f29a49ae04d23750d104b25400b792b98c ]

mmc_add_host() may return error, if we ignore its return value, it will lead two issues:
1. The memory that allocated in mmc_alloc_host() is leaked.
2. In the remove() path, mmc_remove_host() will be called to delete device, but it's not added yet, it will lead a kernel crash because of null-ptr-deref in device_del().

So fix this by checking the return value and goto error path which will call mmc_free_host().

---
 drivers/mmc/host/mmci.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/mmc/host/mmci.c +3 -1
index 3765e2f4ad98..2c4eda83ca18 100644
@@ -2254,7 +2254,9 @@ static int mmci_probe(struct amba_device *dev,
 	pm_runtime_set_autosuspend_delay(&dev->dev, 50);
 	pm_runtime_use_autosuspend(&dev->dev);
 
-	mmc_add_host(mmc);
+	ret = mmc_add_host(mmc);
+	if (ret)
+		goto clk_disable;
 
 	pm_runtime_put(&dev->dev);
 	return 0;
Signature · 1 line
2.35.1
author maintainer
#309 lore raw

PATCH 306/731 mmc: renesas_sdhi: alway populate SCC pointer

In reply to Greg Kroah-Hartman

From: Wolfram Sang <[email protected]>

[ Upstream commit 3d4f9898c1c74323dd61d6a8a0efca9401232ad4 ]

We need the SCC pointer to reset the device, so populate it even when we don't need it for tuning.

---
 drivers/mmc/host/renesas_sdhi_core.c | 12 +++++++-----
 1 file changed, 7 insertions(+), 5 deletions(-)
drivers/mmc/host/renesas_sdhi_core.c +7 -5
index 387f2a4f693a..f443debbcb98 100644
@@ -1037,11 +1037,14 @@ int renesas_sdhi_probe(struct platform_device *pdev,
 	if (ver >= SDHI_VER_GEN3_SD)
 		host->get_timeout_cycles = renesas_sdhi_gen3_get_cycles;
 
+	/* Check for SCC so we can reset it if needed */
+	if (of_data && of_data->scc_offset && ver >= SDHI_VER_GEN2_SDR104)
+		priv->scc_ctl = host->ctl + of_data->scc_offset;
+
 	/* Enable tuning iff we have an SCC and a supported mode */
-	if (of_data && of_data->scc_offset &&
-	    (host->mmc->caps & MMC_CAP_UHS_SDR104 ||
-	     host->mmc->caps2 & (MMC_CAP2_HS200_1_8V_SDR |
-				 MMC_CAP2_HS400_1_8V))) {
+	if (priv->scc_ctl && (host->mmc->caps & MMC_CAP_UHS_SDR104 ||
+	    host->mmc->caps2 & (MMC_CAP2_HS200_1_8V_SDR |
+				MMC_CAP2_HS400_1_8V))) {
 		const struct renesas_sdhi_scc *taps = of_data->taps;
 		bool use_4tap = priv->quirks && priv->quirks->hs400_4taps;
 		bool hit = false;
@@ -1061,7 +1064,6 @@ int renesas_sdhi_probe(struct platform_device *pdev,
 		if (!hit)
 			dev_warn(&host->pdev->dev, "Unknown clock rate for tuning\n");
 
-		priv->scc_ctl = host->ctl + of_data->scc_offset;
 		host->check_retune = renesas_sdhi_check_scc_error;
 		host->ops.execute_tuning = renesas_sdhi_execute_tuning;
 		host->ops.prepare_hs400_tuning = renesas_sdhi_prepare_hs400_tuning;
Signature · 1 line
2.35.1
author maintainer
#310 lore raw

PATCH 307/731 memstick: ms_block: Add error handling support for add_disk()

In reply to Greg Kroah-Hartman

From: Luis Chamberlain <[email protected]>

[ Upstream commit 2304c55fd506fcd5e1a59ae21a306ee82507340b ]

We never checked for errors on add_disk() as this function returned void. Now that this is fixed, use the shiny new error handling.

Contrary to the typical removal which delays the put_disk() until later, since we are failing on a probe we immediately put the disk on failure from add_disk by using blk_cleanup_disk().

Stable-dep-of: 4f431a047a5c ("memstick/ms_block: Add check for alloc_ordered_workqueue")

---
 drivers/memstick/core/ms_block.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
drivers/memstick/core/ms_block.c +5 -1
index f854822f84d6..29a69243cbd0 100644
@@ -2157,10 +2157,14 @@ static int msb_init_disk(struct memstick_dev *card)
 		set_disk_ro(msb->disk, 1);
 
 	msb_start(card);
-	device_add_disk(&card->dev, msb->disk, NULL);
+	rc = device_add_disk(&card->dev, msb->disk, NULL);
+	if (rc)
+		goto out_cleanup_disk;
 	dbg("Disk added");
 	return 0;
 
+out_cleanup_disk:
+	blk_cleanup_disk(msb->disk);
 out_free_tag_set:
 	blk_mq_free_tag_set(&msb->tag_set);
 out_release_id:
Signature · 1 line
2.35.1
author maintainer
#311 lore raw

PATCH 308/731 memstick/ms_block: Add check for alloc_ordered_workqueue

In reply to Greg Kroah-Hartman

From: Jiasheng Jiang <[email protected]>

[ Upstream commit 4f431a047a5c8698ed4b67e2760cfbeb5fffb69d ]

As the alloc_ordered_workqueue may return NULL pointer, it should be better to add check for the return value. Moreover, the msb->io_queue should be freed if error occurs later.

Fixes 0ab30494bc4f ("memstick: add support for legacy memorysticks")
---
 drivers/memstick/core/ms_block.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)
drivers/memstick/core/ms_block.c +8 -1
index 29a69243cbd0..7619c30b4ee1 100644
@@ -2150,6 +2150,11 @@ static int msb_init_disk(struct memstick_dev *card)
 
 	msb->usage_count = 1;
 	msb->io_queue = alloc_ordered_workqueue("ms_block", WQ_MEM_RECLAIM);
+	if (!msb->io_queue) {
+		rc = -ENOMEM;
+		goto out_cleanup_disk;
+	}
+
 	INIT_WORK(&msb->io_work, msb_io_work);
 	sg_init_table(msb->prealloc_sg, MS_BLOCK_MAX_SEGS+1);
 
@@ -2159,10 +2164,12 @@ static int msb_init_disk(struct memstick_dev *card)
 	msb_start(card);
 	rc = device_add_disk(&card->dev, msb->disk, NULL);
 	if (rc)
-		goto out_cleanup_disk;
+		goto out_destroy_workqueue;
 	dbg("Disk added");
 	return 0;
 
+out_destroy_workqueue:
+	destroy_workqueue(msb->io_queue);
 out_cleanup_disk:
 	blk_cleanup_disk(msb->disk);
 out_free_tag_set:
Signature · 1 line
2.35.1
author maintainer
#312 lore raw

PATCH 309/731 mmc: core: Normalize the error handling branch in sd_read_ext_regs()

In reply to Greg Kroah-Hartman

From: Zhen Lei <[email protected]>

[ Upstream commit fc02e2b52389c8fde02852b2f959c0b45f042bbd ]

Let's use pr_err() to output the error messages and let's extend a comment to clarify why returning 0 (success) in one case make sense.

Fixes c784f92769ae ("mmc: core: Read the SD function extension registers for power management")
[Ulf: Clarified the comment and the commit-msg]
---
 drivers/mmc/core/sd.c | 11 ++++++++---
 1 file changed, 8 insertions(+), 3 deletions(-)
drivers/mmc/core/sd.c +8 -3
index 86a8a1f56583..592166e53dce 100644
@@ -1252,7 +1252,7 @@ static int sd_read_ext_regs(struct mmc_card *card)
 	 */
 	err = sd_read_ext_reg(card, 0, 0, 0, 512, gen_info_buf);
 	if (err) {
-		pr_warn("%s: error %d reading general info of SD ext reg\n",
+		pr_err("%s: error %d reading general info of SD ext reg\n",
 			mmc_hostname(card->host), err);
 		goto out;
 	}
@@ -1266,7 +1266,12 @@ static int sd_read_ext_regs(struct mmc_card *card)
 	/* Number of extensions to be find. */
 	num_ext = gen_info_buf[4];
 
-	/* We support revision 0, but limit it to 512 bytes for simplicity. */
+	/*
+	 * We only support revision 0 and limit it to 512 bytes for simplicity.
+	 * No matter what, let's return zero to allow us to continue using the
+	 * card, even if we can't support the features from the SD function
+	 * extensions registers.
+	 */
 	if (rev != 0 || len > 512) {
 		pr_warn("%s: non-supported SD ext reg layout\n",
 			mmc_hostname(card->host));
@@ -1281,7 +1286,7 @@ static int sd_read_ext_regs(struct mmc_card *card)
 	for (i = 0; i < num_ext; i++) {
 		err = sd_parse_ext_reg(card, gen_info_buf, &next_ext_addr);
 		if (err) {
-			pr_warn("%s: error %d parsing SD ext reg\n",
+			pr_err("%s: error %d parsing SD ext reg\n",
 				mmc_hostname(card->host), err);
 			goto out;
 		}
Signature · 1 line
2.35.1
author maintainer
#313 lore raw

PATCH 310/731 regulator: qcom-labibb: Fix missing of_node_put() in qcom_labibb_regulator_probe()

In reply to Greg Kroah-Hartman

From: Yuan Can <[email protected]>

[ Upstream commit cf34ac6aa2b12fb0c3aacfdcae8acd7904b949ec ]

The reg_node needs to be released through of_node_put() in the error handling path when of_irq_get_byname() failed.

Fixes 390af53e0411 ("regulator: qcom-labibb: Implement short-circuit and over-current IRQs")
---
 drivers/regulator/qcom-labibb-regulator.c | 1 +
 1 file changed, 1 insertion(+)
drivers/regulator/qcom-labibb-regulator.c +1 -0
index 639b71eb41ff..bcf7140f3bc9 100644
@@ -822,6 +822,7 @@ static int qcom_labibb_regulator_probe(struct platform_device *pdev)
 			if (irq == 0)
 				irq = -EINVAL;
 
+			of_node_put(reg_node);
 			return dev_err_probe(vreg->dev, irq,
 					     "Short-circuit irq not found.\n");
 		}
Signature · 1 line
2.35.1
author maintainer
#314 lore raw

PATCH 311/731 media: c8sectpfe: Add of_node_put() when breaking out of loop

In reply to Greg Kroah-Hartman

From: Liang He <[email protected]>

[ Upstream commit 63ff05a1ad242a5a0f897921c87b70d601bda59c ]

In configure_channels(), we should call of_node_put() when breaking out of for_each_child_of_node() which will automatically increase and decrease the refcount.

Fixes c5f5d0f99794 ("[media] c8sectpfe: STiH407/10 Linux DVB demux support")
---
 drivers/media/platform/sti/c8sectpfe/c8sectpfe-core.c | 1 +
 1 file changed, 1 insertion(+)
drivers/media/platform/sti/c8sectpfe/c8sectpfe-core.c +1 -0
index 338b205ae3a7..88d0188397e7 100644
@@ -940,6 +940,7 @@ static int configure_channels(struct c8sectpfei *fei)
 		if (ret) {
 			dev_err(fei->dev,
 				"configure_memdma_and_inputblock failed\n");
+			of_node_put(child);
 			goto err_unmap;
 		}
 		index++;
Signature · 1 line
2.35.1
author maintainer
#315 lore raw

PATCH 312/731 media: coda: Add check for dcoda_iram_alloc

In reply to Greg Kroah-Hartman

From: Jiasheng Jiang <[email protected]>

[ Upstream commit 6b8082238fb8bb20f67e46388123e67a5bbc558d ]

As the coda_iram_alloc may return NULL pointer, it should be better to check the return value
in order to avoid NULL poineter dereference, same as the others.

Fixes b313bcc9a467 ("[media] coda: simplify IRAM setup")
---
 drivers/media/platform/coda/coda-bit.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/media/platform/coda/coda-bit.c +2 -2
index c484c008ab02..284608845b85 100644
@@ -852,7 +852,7 @@ static void coda_setup_iram(struct coda_ctx *ctx)
 		/* Only H.264BP and H.263P3 are considered */
 		iram_info->buf_dbk_y_use = coda_iram_alloc(iram_info, w64);
 		iram_info->buf_dbk_c_use = coda_iram_alloc(iram_info, w64);
-		if (!iram_info->buf_dbk_c_use)
+		if (!iram_info->buf_dbk_y_use || !iram_info->buf_dbk_c_use)
 			goto out;
 		iram_info->axi_sram_use |= dbk_bits;
 
@@ -876,7 +876,7 @@ static void coda_setup_iram(struct coda_ctx *ctx)
 
 		iram_info->buf_dbk_y_use = coda_iram_alloc(iram_info, w128);
 		iram_info->buf_dbk_c_use = coda_iram_alloc(iram_info, w128);
-		if (!iram_info->buf_dbk_c_use)
+		if (!iram_info->buf_dbk_y_use || !iram_info->buf_dbk_c_use)
 			goto out;
 		iram_info->axi_sram_use |= dbk_bits;
 
Signature · 1 line
2.35.1
author maintainer
#316 lore raw

PATCH 313/731 media: coda: Add check for kmalloc

In reply to Greg Kroah-Hartman

From: Jiasheng Jiang <[email protected]>

[ Upstream commit 6e5e5defdb8b0186312c2f855ace175aee6daf9b ]

As the kmalloc may return NULL pointer,
it should be better to check the return value
in order to avoid NULL poineter dereference,
same as the others.

Fixes cb1d3a336371 ("[media] coda: add CODA7541 JPEG support")
---
 drivers/media/platform/coda/coda-bit.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)
drivers/media/platform/coda/coda-bit.c +8 -2
index 284608845b85..582a6c581f3c 100644
@@ -1082,10 +1082,16 @@ static int coda_start_encoding(struct coda_ctx *ctx)
 	}
 
 	if (dst_fourcc == V4L2_PIX_FMT_JPEG) {
-		if (!ctx->params.jpeg_qmat_tab[0])
+		if (!ctx->params.jpeg_qmat_tab[0]) {
 			ctx->params.jpeg_qmat_tab[0] = kmalloc(64, GFP_KERNEL);
-		if (!ctx->params.jpeg_qmat_tab[1])
+			if (!ctx->params.jpeg_qmat_tab[0])
+				return -ENOMEM;
+		}
+		if (!ctx->params.jpeg_qmat_tab[1]) {
 			ctx->params.jpeg_qmat_tab[1] = kmalloc(64, GFP_KERNEL);
+			if (!ctx->params.jpeg_qmat_tab[1])
+				return -ENOMEM;
+		}
 		coda_set_jpeg_compression_quality(ctx, ctx->params.jpeg_quality);
 	}
 
Signature · 1 line
2.35.1
author maintainer
#317 lore raw

PATCH 314/731 clk: samsung: Fix memory leak in _samsung_clk_register_pll()

In reply to Greg Kroah-Hartman

From: Xiu Jianfeng <[email protected]>

[ Upstream commit 5174e5b0d1b669a489524192b6adcbb3c54ebc72 ]

If clk_register() fails, @pll->rate_table may have allocated memory by kmemdup(), so it needs to be freed, otherwise will cause memory leak issue, this patch fixes it.

Fixes 3ff6e0d8d64d ("clk: samsung: Add support to register rate_table for samsung plls")
---
 drivers/clk/samsung/clk-pll.c | 1 +
 1 file changed, 1 insertion(+)
drivers/clk/samsung/clk-pll.c +1 -0
index 5873a9354b50..4909e940f0ab 100644
@@ -1385,6 +1385,7 @@ static void __init _samsung_clk_register_pll(struct samsung_clk_provider *ctx,
 	if (ret) {
 		pr_err("%s: failed to register pll clock %s : %d\n",
 			__func__, pll_clk->name, ret);
+		kfree(pll->rate_table);
 		kfree(pll);
 		return;
 	}
Signature · 1 line
2.35.1
author maintainer
#318 lore raw

PATCH 315/731 spi: spi-gpio: Dont set MOSI as an input if not 3WIRE mode

In reply to Greg Kroah-Hartman

From: Kris Bahnsen <[email protected]>

[ Upstream commit 3a6f994f848a69deb2bf3cd9d130dd0c09730e55 ]

The addition of 3WIRE support would affect MOSI direction even when still in standard (4 wire) mode. This can lead to MOSI being at an invalid logic level when a device driver sets an SPI message with a NULL tx_buf.

spi.h states that if tx_buf is NULL then "zeros will be shifted out ... " If MOSI is tristated then the data shifted out is subject to pull resistors, keepers, or in the absence of those, noise.

This issue came to light when using spi-gpio connected to an ADS7843 touchscreen controller. MOSI pulled high when clocking MISO data in caused the SPI device to interpret this as a command which would put the device in an unexpected and non-functional state.

Fixes 4b859db2c606 ("spi: spi-gpio: add SPI_3WIRE support")
Fixes 5132b3d28371 ("spi: gpio: Support 3WIRE high-impedance turn-around")
---
 drivers/spi/spi-gpio.c | 16 +++++++++++++---
 1 file changed, 13 insertions(+), 3 deletions(-)
drivers/spi/spi-gpio.c +13 -3
index 0584f4d2fde2..3ffdab6caac2 100644
@@ -244,9 +244,19 @@ static int spi_gpio_set_direction(struct spi_device *spi, bool output)
 	if (output)
 		return gpiod_direction_output(spi_gpio->mosi, 1);
 
-	ret = gpiod_direction_input(spi_gpio->mosi);
-	if (ret)
-		return ret;
+	/*
+	 * Only change MOSI to an input if using 3WIRE mode.
+	 * Otherwise, MOSI could be left floating if there is
+	 * no pull resistor connected to the I/O pin, or could
+	 * be left logic high if there is a pull-up. Transmitting
+	 * logic high when only clocking MISO data in can put some
+	 * SPI devices in to a bad state.
+	 */
+	if (spi->mode & SPI_3WIRE) {
+		ret = gpiod_direction_input(spi_gpio->mosi);
+		if (ret)
+			return ret;
+	}
 	/*
 	 * Send a turnaround high impedance cycle when switching
 	 * from output to input. Theoretically there should be
Signature · 1 line
2.35.1
author maintainer
#319 lore raw

PATCH 316/731 wifi: rtl8xxxu: Add __packed to struct rtl8723bu_c2h

In reply to Greg Kroah-Hartman

From: Bitterblue Smith <[email protected]>

[ Upstream commit dd469a754afdb782ba3033cee102147493dc39f4 ]

This struct is used to access a sequence of bytes received from the wifi chip. It must not have any padding bytes between the members.

This doesn't change anything on my system, possibly because currently none of the members need more than byte alignment.

Fixes b2b43b7837ba ("rtl8xxxu: Initial functionality to handle C2H events for 8723bu")
---
 drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.h +1 -1
index 7ddce3c3f0c4..cd3ff9847ced 100644
@@ -1190,7 +1190,7 @@ struct rtl8723bu_c2h {
 			u8 bw;
 		} __packed ra_report;
 	};
-};
+} __packed;
 
 struct rtl8xxxu_fileops;
 
Signature · 1 line
2.35.1
author maintainer
#320 lore raw

PATCH 317/731 wifi: rtl8xxxu: Fix the channel width reporting

In reply to Greg Kroah-Hartman

From: Bitterblue Smith <[email protected]>

[ Upstream commit 76c16af2cb10282274596e21add2c9f0b95c941b ]

The gen 2 chips RTL8192EU and RTL8188FU periodically send the driver reports about the TX rate, and the driver passes these reports to sta_statistics. The reports from RTL8192EU may or may not include the channel width. The reports from RTL8188FU do not include it.

Only access the c2h->ra_report.bw field if the report (skb) is big enough.

The other problem fixed here is that the code was actually never changing the channel width initially reported by rtl8xxxu_bss_info_changed because the value of RATE_INFO_BW_20 is 0.

Fixes 0985d3a410ac ("rtl8xxxu: Feed current txrate information for mac80211")
---
 drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu_core.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu_core.c +7 -3
index 2a4ecbb0fce6..7370d92a3bda 100644
@@ -5517,7 +5517,6 @@ static void rtl8xxxu_c2hcmd_callback(struct work_struct *work)
 			rarpt->txrate.flags = 0;
 			rate = c2h->ra_report.rate;
 			sgi = c2h->ra_report.sgi;
-			bw = c2h->ra_report.bw;
 
 			if (rate < DESC_RATE_MCS0) {
 				rarpt->txrate.legacy =
@@ -5534,8 +5533,13 @@ static void rtl8xxxu_c2hcmd_callback(struct work_struct *work)
 						RATE_INFO_FLAGS_SHORT_GI;
 				}
 
-				if (bw == RATE_INFO_BW_20)
-					rarpt->txrate.bw |= RATE_INFO_BW_20;
+				if (skb->len >= offsetofend(typeof(*c2h), ra_report.bw)) {
+					if (c2h->ra_report.bw == RTL8XXXU_CHANNEL_WIDTH_40)
+						bw = RATE_INFO_BW_40;
+					else
+						bw = RATE_INFO_BW_20;
+					rarpt->txrate.bw = bw;
+				}
 			}
 			bit_rate = cfg80211_calculate_bitrate(&rarpt->txrate);
 			rarpt->bit_rate = bit_rate;
Signature · 1 line
2.35.1
author maintainer
#321 lore raw

PATCH 318/731 wifi: brcmfmac: Fix error return code in brcmf_sdio_download_firmware()

In reply to Greg Kroah-Hartman

From: Wang Yufen <[email protected]>

[ Upstream commit c2f2924bc7f9ea75ef8d95863e710168f8196256 ]

Fix to return a negative error code instead of 0 when brcmf_chip_set_active() fails. In addition, change the return value for brcmf_pcie_exit_download_state() to keep consistent.

Fixes d380ebc9b6fb ("brcmfmac: rename chip download functions")
---
 drivers/net/wireless/broadcom/brcm80211/brcmfmac/pcie.c | 2 +-
 drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c | 1 +
 2 files changed, 2 insertions(+), 1 deletion(-)
drivers/net/wireless/broadcom/brcm80211/brcmfmac/pcie.c +1 -1
index 3ff4997e1c97..358021a33b8a 100644
@@ -626,7 +626,7 @@ static int brcmf_pcie_exit_download_state(struct brcmf_pciedev_info *devinfo,
 	}
 
 	if (!brcmf_chip_set_active(devinfo->ci, resetintr))
-		return -EINVAL;
+		return -EIO;
 	return 0;
 }
 
drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c +1 -0
index f7961b22e051..5006aa831751 100644
@@ -3411,6 +3411,7 @@ static int brcmf_sdio_download_firmware(struct brcmf_sdio *bus,
 	/* Take arm out of reset */
 	if (!brcmf_chip_set_active(bus->ci, rstvec)) {
 		brcmf_err("error getting out of ARM core reset\n");
+		bcmerror = -EIO;
 		goto err;
 	}
 
Signature · 1 line
2.35.1
author maintainer
#322 lore raw

PATCH 319/731 blktrace: Fix output non-blktrace event when blk_classic option enabled

In reply to Greg Kroah-Hartman

From: Yang Jihong <[email protected]>

[ Upstream commit f596da3efaf4130ff61cd029558845808df9bf99 ]

When the blk_classic option is enabled, non-blktrace events must be filtered out. Otherwise, events of other types are output in the blktrace classic format, which is unexpected.

The problem can be triggered in the following ways:

  # echo 1 > /sys/kernel/debug/tracing/options/blk_classic
  # echo 1 > /sys/kernel/debug/tracing/events/enable
  # echo blk > /sys/kernel/debug/tracing/current_tracer
  # cat /sys/kernel/debug/tracing/trace_pipe
---
 kernel/trace/blktrace.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
kernel/trace/blktrace.c +2 -1
index eaa98e2b468f..16b0d3fa56e0 100644
@@ -1547,7 +1547,8 @@ blk_trace_event_print_binary(struct trace_iterator *iter, int flags,
 
 static enum print_line_t blk_tracer_print_line(struct trace_iterator *iter)
 {
-	if (!(blk_tracer_flags.val & TRACE_BLK_OPT_CLASSIC))
+	if ((iter->ent->type != TRACE_BLK) ||
+	    !(blk_tracer_flags.val & TRACE_BLK_OPT_CLASSIC))
 		return TRACE_TYPE_UNHANDLED;
 
 	return print_one_line(iter, true);
Signature · 1 line
2.35.1
author maintainer
#323 lore raw

PATCH 320/731 bpf: Do not zero-extend kfunc return values

In reply to Greg Kroah-Hartman

From: Björn Töpel <[email protected]>

[ Upstream commit d35af0a7feb077c43ff0233bba5a8c6e75b73e35 ]

In BPF all global functions, and BPF helpers return a 64-bit value. For kfunc calls, this is not the case, and they can return e.g. 32-bit values.

The return register R0 for kfuncs calls can therefore be marked as subreg_def != DEF_NOT_SUBREG. In general, if a register is marked with subreg_def != DEF_NOT_SUBREG, some archs (where bpf_jit_needs_zext() returns true) require the verifier to insert explicit zero-extension instructions.

For kfuncs calls, however, the caller should do sign/zero extension for return values. In other words, the compiler is responsible to insert proper instructions, not the verifier.

An example, provided by Yonghong Song:

$ cat t.c
extern unsigned foo(void);
unsigned bar1(void) {
     return foo();
}
unsigned bar2(void) {
     if (foo()) return 10; else return 20;
}
$ clang -target bpf -mcpu=v3 -O2 -c t.c && llvm-objdump -d t.o
t.o:    file format elf64-bpf

Disassembly of section .text:

0000000000000000 <bar1>:

	0:       85 10 00 00 ff ff ff ff call -0x1
	1:       95 00 00 00 00 00 00 00 exit

0000000000000010 <bar2>:

	2:       85 10 00 00 ff ff ff ff call -0x1
	3:       bc 01 00 00 00 00 00 00 w1 = w0
	4:       b4 00 00 00 14 00 00 00 w0 = 0x14
	5:       16 01 01 00 00 00 00 00 if w1 == 0x0 goto +0x1 <LBB1_2>
	6:       b4 00 00 00 0a 00 00 00 w0 = 0xa

0000000000000038 <LBB1_2>:

	7:       95 00 00 00 00 00 00 00 exit

If the return value of 'foo()' is used in the BPF program, the proper zero-extension will be done.

Currently, the verifier correctly marks, say, a 32-bit return value as subreg_def != DEF_NOT_SUBREG, but will fail performing the actual zero-extension, due to a verifier bug in opt_subreg_zext_lo32_rnd_hi32(). load_reg is not properly set to R0, and the following path will be taken:

		if (WARN_ON(load_reg == -1)) {
			verbose(env, "verifier bug. zext_dst is set, but no reg is defined\n");
			return -EFAULT;
		}

A longer discussion from v1 can be found in the link below.

Correct the verifier by avoiding doing explicit zero-extension of R0 for kfunc calls. Note that R0 will still be marked as a sub-register for return values smaller than 64-bit.

---
 kernel/bpf/verifier.c | 4 ++++
 1 file changed, 4 insertions(+)
kernel/bpf/verifier.c +4 -0
index b1ca4dbdeecf..488225bb42f6 100644
@@ -12132,6 +12132,10 @@ static int opt_subreg_zext_lo32_rnd_hi32(struct bpf_verifier_env *env,
 		if (!bpf_jit_needs_zext() && !is_cmpxchg_insn(&insn))
 			continue;
 
+		/* Zero-extension is done by the caller. */
+		if (bpf_pseudo_kfunc_call(&insn))
+			continue;
+
 		if (WARN_ON(load_reg == -1)) {
 			verbose(env, "verifier bug. zext_dst is set, but no reg is defined\n");
 			return -EFAULT;
Signature · 1 line
2.35.1
author maintainer
#324 lore raw

PATCH 321/731 clk: socfpga: Fix memory leak in socfpga_gate_init()

In reply to Greg Kroah-Hartman

From: Xiu Jianfeng <[email protected]>

[ Upstream commit 0b8ba891ad4d1ef6bfa4c72efc83f9f9f855f68b ]

Free @socfpga_clk and @ops on the error path to avoid memory leak issue.

Fixes a30a67be7b6e ("clk: socfpga: Don't have get_parent for single parent ops")
---
 drivers/clk/socfpga/clk-gate.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
drivers/clk/socfpga/clk-gate.c +4 -1
index 1ec9678d8cd3..ee2a2d284113 100644
@@ -188,8 +188,10 @@ void __init socfpga_gate_init(struct device_node *node)
 		return;
 
 	ops = kmemdup(&gateclk_ops, sizeof(gateclk_ops), GFP_KERNEL);
-	if (WARN_ON(!ops))
+	if (WARN_ON(!ops)) {
+		kfree(socfpga_clk);
 		return;
+	}
 
 	rc = of_property_read_u32_array(node, "clk-gate", clk_gate, 2);
 	if (rc)
@@ -243,6 +245,7 @@ void __init socfpga_gate_init(struct device_node *node)
 
 	err = clk_hw_register(NULL, hw_clk);
 	if (err) {
+		kfree(ops);
 		kfree(socfpga_clk);
 		return;
 	}
Signature · 1 line
2.35.1
author maintainer
#325 lore raw

PATCH 322/731 net: vmw_vsock: vmci: Check memcpy_from_msg()

In reply to Greg Kroah-Hartman

From: Artem Chernyshev <[email protected]>

[ Upstream commit 44aa5a6dba8283bfda28b1517af4de711c5652a4 ]

vmci_transport_dgram_enqueue() does not check the return value of memcpy_from_msg(). If memcpy_from_msg() fails, it is possible that uninitialized memory contents are sent unintentionally instead of user's message in the datagram to the destination. Return with an error if memcpy_from_msg() fails.

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

Fixes 0f7db23a07af ("vmci_transport: switch ->enqeue_dgram, ->enqueue_stream and ->dequeue_stream to msghdr")
---
 net/vmw_vsock/vmci_transport.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
net/vmw_vsock/vmci_transport.c +5 -1
index b17dc9745188..94c1112f1c8c 100644
@@ -1711,7 +1711,11 @@ static int vmci_transport_dgram_enqueue(
 	if (!dg)
 		return -ENOMEM;
 
-	memcpy_from_msg(VMCI_DG_PAYLOAD(dg), msg, len);
+	err = memcpy_from_msg(VMCI_DG_PAYLOAD(dg), msg, len);
+	if (err) {
+		kfree(dg);
+		return err;
+	}
 
 	dg->dst = vmci_make_handle(remote_addr->svm_cid,
 				   remote_addr->svm_port);
Signature · 1 line
2.35.1
author maintainer
#326 lore raw

PATCH 323/731 net: defxx: Fix missing err handling in dfx_init()

In reply to Greg Kroah-Hartman

From: Yongqiang Liu <[email protected]>

[ Upstream commit ae18dcdff0f8d7e84cd3fd9f496518b5e72d185d ]

When eisa_driver_register() or tc_register_driver() failed, the modprobe defxx would fail with some err log as follows:

Error: Driver 'defxx' is already registered, aborting...

Fix this issue by adding err hanling in dfx_init().

Fixes e89a2cfb7d7b5 ("[TC] defxx: TURBOchannel support")
---
 drivers/net/fddi/defxx.c | 22 ++++++++++++++++++----
 1 file changed, 18 insertions(+), 4 deletions(-)
drivers/net/fddi/defxx.c +18 -4
index 6d1e3f49a3d3..ebf502290e5f 100644
@@ -3831,10 +3831,24 @@ static int dfx_init(void)
 	int status;
 
 	status = pci_register_driver(&dfx_pci_driver);
-	if (!status)
-		status = eisa_driver_register(&dfx_eisa_driver);
-	if (!status)
-		status = tc_register_driver(&dfx_tc_driver);
+	if (status)
+		goto err_pci_register;
+
+	status = eisa_driver_register(&dfx_eisa_driver);
+	if (status)
+		goto err_eisa_register;
+
+	status = tc_register_driver(&dfx_tc_driver);
+	if (status)
+		goto err_tc_register;
+
+	return 0;
+
+err_tc_register:
+	eisa_driver_unregister(&dfx_eisa_driver);
+err_eisa_register:
+	pci_unregister_driver(&dfx_pci_driver);
+err_pci_register:
 	return status;
 }
 
Signature · 1 line
2.35.1
author maintainer
#327 lore raw

PATCH 324/731 net: stmmac: selftests: fix potential memleak in stmmac_test_arpoffload()

In reply to Greg Kroah-Hartman

From: Zhang Changzhong <[email protected]>

[ Upstream commit f150b63f3fa5fdd81e0dd6151e8850268e29438c ]

The skb allocated by stmmac_test_get_arp_skb() hasn't been released in some error handling case, which will lead to a memory leak. Fix this up by adding kfree_skb() to release skb.

Compile tested only.

Fixes 5e3fb0a6e2b3 ("net: stmmac: selftests: Implement the ARP Offload test")
---
 drivers/net/ethernet/stmicro/stmmac/stmmac_selftests.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)
drivers/net/ethernet/stmicro/stmmac/stmmac_selftests.c +6 -2
index dd5c4ef92ef3..ea7200b7b647 100644
@@ -1654,12 +1654,16 @@ static int stmmac_test_arpoffload(struct stmmac_priv *priv)
 	}
 
 	ret = stmmac_set_arp_offload(priv, priv->hw, true, ip_addr);
-	if (ret)
+	if (ret) {
+		kfree_skb(skb);
 		goto cleanup;
+	}
 
 	ret = dev_set_promiscuity(priv->dev, 1);
-	if (ret)
+	if (ret) {
+		kfree_skb(skb);
 		goto cleanup;
+	}
 
 	ret = dev_direct_xmit(skb, 0);
 	if (ret)
Signature · 1 line
2.35.1
author maintainer
#328 lore raw

PATCH 325/731 net: stmmac: fix possible memory leak in stmmac_dvr_probe()

In reply to Greg Kroah-Hartman

From: Gaosheng Cui <[email protected]>

[ Upstream commit a137f3f27f9290933fe7e40e6dc8a445781c31a2 ]

The bitmap_free() should be called to free priv->af_xdp_zc_qps when create_singlethread_workqueue() fails, otherwise there will be a memory leak, so we add the err path error_wq_init to fix it.

Fixes bba2556efad6 ("net: stmmac: Enable RX via AF_XDP zero-copy")
---
 drivers/net/ethernet/stmicro/stmmac/stmmac_main.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/net/ethernet/stmicro/stmmac/stmmac_main.c +2 -1
index eba97adaf1fb..d987ca727a21 100644
@@ -7046,7 +7046,7 @@ int stmmac_dvr_probe(struct device *device,
 	priv->wq = create_singlethread_workqueue("stmmac_wq");
 	if (!priv->wq) {
 		dev_err(priv->device, "failed to create workqueue\n");
-		return -ENOMEM;
+		goto error_wq_init;
 	}
 
 	INIT_WORK(&priv->service_task, stmmac_service_task);
@@ -7273,6 +7273,7 @@ int stmmac_dvr_probe(struct device *device,
 	stmmac_napi_del(ndev);
 error_hw_init:
 	destroy_workqueue(priv->wq);
+error_wq_init:
 	bitmap_free(priv->af_xdp_zc_qps);
 
 	return ret;
Signature · 1 line
2.35.1
author maintainer
#329 lore raw

PATCH 326/731 drivers: net: qlcnic: Fix potential memory leak in qlcnic_sriov_init()

In reply to Greg Kroah-Hartman

From: Yuan Can <[email protected]>

[ Upstream commit 01de1123322e4fe1bbd0fcdf0982511b55519c03 ]

If vp alloc failed in qlcnic_sriov_init(), all previously allocated vp needs to be freed.

Fixes f197a7aa6288 ("qlcnic: VF-PF communication channel implementation")
---
 drivers/net/ethernet/qlogic/qlcnic/qlcnic_sriov_common.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/net/ethernet/qlogic/qlcnic/qlcnic_sriov_common.c +2 -0
index 42a44c97572a..df9b84f6600f 100644
@@ -221,6 +221,8 @@ int qlcnic_sriov_init(struct qlcnic_adapter *adapter, int num_vfs)
 	return 0;
 
 qlcnic_destroy_async_wq:
+	while (i--)
+		kfree(sriov->vf_info[i].vp);
 	destroy_workqueue(bc->bc_async_wq);
 
 qlcnic_destroy_trans_wq:
Signature · 1 line
2.35.1
author maintainer
#330 lore raw

PATCH 327/731 of: overlay: fix null pointer dereferencing in find_dup_cset_node_entry() and find_dup_cset_prop()

In reply to Greg Kroah-Hartman

From: ruanjinjie <[email protected]>

[ Upstream commit ee9d7a0e754568180a2f8ebc4aad226278a9116f ]

When kmalloc() fail to allocate memory in kasprintf(), fn_1 or fn_2 will be NULL, and strcmp() will cause null pointer dereference.

Fixes 2fe0e8769df9 ("of: overlay: check prevents multiple fragments touching same property")
---
 drivers/of/overlay.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/of/overlay.c +2 -2
index d1187123c4fc..424682372417 100644
@@ -547,7 +547,7 @@ static int find_dup_cset_node_entry(struct overlay_changeset *ovcs,
 
 		fn_1 = kasprintf(GFP_KERNEL, "%pOF", ce_1->np);
 		fn_2 = kasprintf(GFP_KERNEL, "%pOF", ce_2->np);
-		node_path_match = !strcmp(fn_1, fn_2);
+		node_path_match = !fn_1 || !fn_2 || !strcmp(fn_1, fn_2);
 		kfree(fn_1);
 		kfree(fn_2);
 		if (node_path_match) {
@@ -582,7 +582,7 @@ static int find_dup_cset_prop(struct overlay_changeset *ovcs,
 
 		fn_1 = kasprintf(GFP_KERNEL, "%pOF", ce_1->np);
 		fn_2 = kasprintf(GFP_KERNEL, "%pOF", ce_2->np);
-		node_path_match = !strcmp(fn_1, fn_2);
+		node_path_match = !fn_1 || !fn_2 || !strcmp(fn_1, fn_2);
 		kfree(fn_1);
 		kfree(fn_2);
 		if (node_path_match &&
Signature · 1 line
2.35.1
author maintainer
#331 lore raw

PATCH 328/731 ethernet: s2io: dont call dev_kfree_skb() under spin_lock_irqsave()

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit 6cee96e09df54ae17784c0f38a49e0ed8229b825 ]

It is not allowed to call kfree_skb() or consume_skb() from hardware interrupt context or with hardware interrupts being disabled.

It should use dev_kfree_skb_irq() or dev_consume_skb_irq() instead. The difference between them is free reason, dev_kfree_skb_irq() means the SKB is dropped in error and dev_consume_skb_irq() means the SKB is consumed in normal.

In this case, dev_kfree_skb() is called in free_tx_buffers() to drop the SKBs in tx buffers, when the card is down, so replace it with dev_kfree_skb_irq() here.

Fixes 1da177e4c3f4 ("Linux-2.6.12-rc2")
---
 drivers/net/ethernet/neterion/s2io.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/ethernet/neterion/s2io.c +1 -1
index f4703f53bcdc..65ccdbe665e5 100644
@@ -2386,7 +2386,7 @@ static void free_tx_buffers(struct s2io_nic *nic)
 			skb = s2io_txdl_getskb(&mac_control->fifos[i], txdp, j);
 			if (skb) {
 				swstats->mem_freed += skb->truesize;
-				dev_kfree_skb(skb);
+				dev_kfree_skb_irq(skb);
 				cnt++;
 			}
 		}
Signature · 1 line
2.35.1
author maintainer
#332 lore raw

PATCH 329/731 net: farsync: Fix kmemleak when rmmods farsync

In reply to Greg Kroah-Hartman

From: Li Zetao <[email protected]>

[ Upstream commit 2f623aaf9f31de968dea6169849706a2f9be444c ]

There are two memory leaks reported by kmemleak:

unreferenced object 0xffff888114b20200 (size 128):

    comm "modprobe", pid 4846, jiffies 4295146524 (age 401.345s)
    hex dump (first 32 bytes):
      e0 62 57 09 81 88 ff ff e0 62 57 09 81 88 ff ff  .bW......bW.....
      01 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
    backtrace:
      [<ffffffff815bcd82>] kmalloc_trace+0x22/0x60
      [<ffffffff83d35c78>] __hw_addr_add_ex+0x198/0x6c0
      [<ffffffff83d3989d>] dev_addr_init+0x13d/0x230
      [<ffffffff83d1063d>] alloc_netdev_mqs+0x10d/0xe50
      [<ffffffff82b4a06e>] alloc_hdlcdev+0x2e/0x80
      [<ffffffffa016a741>] fst_add_one+0x601/0x10e0 [farsync]
      ...

unreferenced object 0xffff88810b85b000 (size 1024):

    comm "modprobe", pid 4846, jiffies 4295146523 (age 401.346s)
    hex dump (first 32 bytes):
      00 00 b0 02 00 c9 ff ff 00 70 0a 00 00 c9 ff ff  .........p......
      00 00 00 f2 00 00 00 f3 0a 00 00 00 02 00 00 00  ................
    backtrace:
      [<ffffffff815bcd82>] kmalloc_trace+0x22/0x60
      [<ffffffffa016a294>] fst_add_one+0x154/0x10e0 [farsync]
      [<ffffffff82060e83>] local_pci_probe+0xd3/0x170
      ...

The root cause is traced to the netdev and fst_card_info are not freed when removes one fst in fst_remove_one(), which may trigger oom if repeated insmod and rmmod module.

Fix it by adding free_netdev() and kfree() in fst_remove_one(), just as the operations on the error handling path in fst_add_one().

Fixes 1da177e4c3f4 ("Linux-2.6.12-rc2")
---
 drivers/net/wan/farsync.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/net/wan/farsync.c +2 -0
index 6a212c085435..5b01642ca44e 100644
@@ -2545,6 +2545,7 @@ fst_remove_one(struct pci_dev *pdev)
 		struct net_device *dev = port_to_dev(&card->ports[i]);
 
 		unregister_hdlc_device(dev);
+		free_netdev(dev);
 	}
 
 	fst_disable_intr(card);
@@ -2564,6 +2565,7 @@ fst_remove_one(struct pci_dev *pdev)
 				  card->tx_dma_handle_card);
 	}
 	fst_card_array[card->card_no] = NULL;
+	kfree(card);
 }
 
 static struct pci_driver fst_driver = {
Signature · 1 line
2.35.1
author maintainer
#333 lore raw

PATCH 330/731 net/tunnel: wait until all sk_user_data reader finish before releasing the sock

In reply to Greg Kroah-Hartman

From: Hangbin Liu <[email protected]>

[ Upstream commit 3cf7203ca620682165706f70a1b12b5194607dce ]

There is a race condition in vxlan that when deleting a vxlan device during receiving packets, there is a possibility that the sock is released after getting vxlan_sock vs from sk_user_data. Then in later vxlan_ecn_decapsulate(), vxlan_get_sk_family() we will got NULL pointer dereference. e.g.

#0 [ffffa25ec6978a38] machine_kexec at ffffffff8c669757 #1 [ffffa25ec6978a90] __crash_kexec at ffffffff8c7c0a4d #2 [ffffa25ec6978b58] crash_kexec at ffffffff8c7c1c48 #3 [ffffa25ec6978b60] oops_end at ffffffff8c627f2b #4 [ffffa25ec6978b80] page_fault_oops at ffffffff8c678fcb #5 [ffffa25ec6978bd8] exc_page_fault at ffffffff8d109542 #6 [ffffa25ec6978c00] asm_exc_page_fault at ffffffff8d200b62

      [exception RIP: vxlan_ecn_decapsulate+0x3b]
      RIP: ffffffffc1014e7b  RSP: ffffa25ec6978cb0  RFLAGS: 00010246
      RAX: 0000000000000008  RBX: ffff8aa000888000  RCX: 0000000000000000
      RDX: 000000000000000e  RSI: ffff8a9fc7ab803e  RDI: ffff8a9fd1168700
      RBP: ffff8a9fc7ab803e   R8: 0000000000700000   R9: 00000000000010ae
      R10: ffff8a9fcb748980  R11: 0000000000000000  R12: ffff8a9fd1168700
      R13: ffff8aa000888000  R14: 00000000002a0000  R15: 00000000000010ae
      ORIG_RAX: ffffffffffffffff  CS: 0010  SS: 0018

#7 [ffffa25ec6978ce8] vxlan_rcv at ffffffffc10189cd [vxlan] #8 [ffffa25ec6978d90] udp_queue_rcv_one_skb at ffffffff8cfb6507 #9 [ffffa25ec6978dc0] udp_unicast_rcv_skb at ffffffff8cfb6e45
#10 [ffffa25ec6978dc8] __udp4_lib_rcv at ffffffff8cfb8807 #11 [ffffa25ec6978e20] ip_protocol_deliver_rcu at ffffffff8cf76951 #12 [ffffa25ec6978e48] ip_local_deliver at ffffffff8cf76bde #13 [ffffa25ec6978ea0] __netif_receive_skb_one_core at ffffffff8cecde9b #14 [ffffa25ec6978ec8] process_backlog at ffffffff8cece139 #15 [ffffa25ec6978f00] __napi_poll at ffffffff8ceced1a
#16 [ffffa25ec6978f28] net_rx_action at ffffffff8cecf1f3 #17 [ffffa25ec6978fa0] __softirqentry_text_start at ffffffff8d4000ca #18 [ffffa25ec6978ff0] do_softirq at ffffffff8c6fbdc3

Reproducer: https://github.com/Mellanox/ovs-tests/blob/master/test-ovs-vxlan-remove-tunnel-during-traffic.sh

Fix this by waiting for all sk_user_data reader to finish before releasing the sock.

Fixes 6a93cc905274 ("udp-tunnel: Add a few more UDP tunnel APIs")
---
 net/ipv4/udp_tunnel_core.c | 1 +
 1 file changed, 1 insertion(+)
net/ipv4/udp_tunnel_core.c +1 -0
index 46101fd67a47..1ff5b8e30bb9 100644
@@ -179,6 +179,7 @@ EXPORT_SYMBOL_GPL(udp_tunnel_xmit_skb);
 void udp_tunnel_sock_release(struct socket *sock)
 {
 	rcu_assign_sk_user_data(sock->sk, NULL);
+	synchronize_rcu();
 	kernel_sock_shutdown(sock, SHUT_RDWR);
 	sock_release(sock);
 }
Signature · 1 line
2.35.1
author maintainer
#334 lore raw

PATCH 331/731 net: apple: mace: dont call dev_kfree_skb() under spin_lock_irqsave()

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit 3dfe3486c1cd4f82b466b7d307f23777137b8acc ]

It is not allowed to call kfree_skb() or consume_skb() from hardware interrupt context or with hardware interrupts being disabled.

It should use dev_kfree_skb_irq() or dev_consume_skb_irq() instead. The difference between them is free reason, dev_kfree_skb_irq() means the SKB is dropped in error and dev_consume_skb_irq() means the SKB is consumed in normal.

In this case, dev_kfree_skb() is called in mace_tx_timeout() to drop the SKB, when tx timeout, so replace it with dev_kfree_skb_irq().

Fixes 1da177e4c3f4 ("Linux-2.6.12-rc2")
---
 drivers/net/ethernet/apple/mace.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/ethernet/apple/mace.c +1 -1
index 4b80e3a52a19..44037e9e197f 100644
@@ -841,7 +841,7 @@ static void mace_tx_timeout(struct timer_list *t)
     if (mp->tx_bad_runt) {
 	mp->tx_bad_runt = 0;
     } else if (i != mp->tx_fill) {
-	dev_kfree_skb(mp->tx_bufs[i]);
+	dev_kfree_skb_irq(mp->tx_bufs[i]);
 	if (++i >= N_TX_RING)
 	    i = 0;
 	mp->tx_empty = i;
Signature · 1 line
2.35.1
author maintainer
#335 lore raw

PATCH 332/731 net: apple: bmac: dont call dev_kfree_skb() under spin_lock_irqsave()

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit 5fe02e046e6422c4adfdbc50206ec7186077da24 ]

It is not allowed to call kfree_skb() or consume_skb() from hardware interrupt context or with hardware interrupts being disabled.

It should use dev_kfree_skb_irq() or dev_consume_skb_irq() instead. The difference between them is free reason, dev_kfree_skb_irq() means the SKB is dropped in error and dev_consume_skb_irq() means the SKB is consumed in normal.

In this case, dev_kfree_skb() is called in bmac_tx_timeout() to drop the SKB, when tx timeout, so replace it with dev_kfree_skb_irq().

Fixes 1da177e4c3f4 ("Linux-2.6.12-rc2")
---
 drivers/net/ethernet/apple/bmac.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/ethernet/apple/bmac.c +1 -1
index a989d2df59ad..7a966361d83f 100644
@@ -1511,7 +1511,7 @@ static void bmac_tx_timeout(struct timer_list *t)
 	i = bp->tx_empty;
 	++dev->stats.tx_errors;
 	if (i != bp->tx_fill) {
-		dev_kfree_skb(bp->tx_bufs[i]);
+		dev_kfree_skb_irq(bp->tx_bufs[i]);
 		bp->tx_bufs[i] = NULL;
 		if (++i >= N_TX_RING) i = 0;
 		bp->tx_empty = i;
Signature · 1 line
2.35.1
author maintainer
#336 lore raw

PATCH 333/731 net: emaclite: dont call dev_kfree_skb() under spin_lock_irqsave()

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit d1678bf45f21fa5ae4a456f821858679556ea5f8 ]

It is not allowed to call kfree_skb() or consume_skb() from hardware interrupt context or with hardware interrupts being disabled.

It should use dev_kfree_skb_irq() or dev_consume_skb_irq() instead. The difference between them is free reason, dev_kfree_skb_irq() means the SKB is dropped in error and dev_consume_skb_irq() means the SKB is consumed in normal.

In this case, dev_kfree_skb() is called in xemaclite_tx_timeout() to drop the SKB, when tx timeout, so replace it with dev_kfree_skb_irq().

Fixes bb81b2ddfa19 ("net: add Xilinx emac lite device driver")
---
 drivers/net/ethernet/xilinx/xilinx_emaclite.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/ethernet/xilinx/xilinx_emaclite.c +1 -1
index 97c1d1ecba34..b1971c4d5313 100644
@@ -543,7 +543,7 @@ static void xemaclite_tx_timeout(struct net_device *dev, unsigned int txqueue)
 	xemaclite_enable_interrupts(lp);
 
 	if (lp->deferred_skb) {
-		dev_kfree_skb(lp->deferred_skb);
+		dev_kfree_skb_irq(lp->deferred_skb);
 		lp->deferred_skb = NULL;
 		dev->stats.tx_errors++;
 	}
Signature · 1 line
2.35.1
author maintainer
#337 lore raw

PATCH 334/731 net: ethernet: dnet: dont call dev_kfree_skb() under spin_lock_irqsave()

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit f07fadcbee2a5e84caa67c7c445424200bffb60b ]

It is not allowed to call kfree_skb() or consume_skb() from hardware interrupt context or with hardware interrupts being disabled.

In this case, the lock is used to protected 'bp', so we can move dev_kfree_skb() after the spin_unlock_irqrestore().

Fixes 4796417417a6 ("dnet: Dave DNET ethernet controller driver (updated)")
---
 drivers/net/ethernet/dnet.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/net/ethernet/dnet.c +2 -2
index 6c51cf991dad..14dc2e13bf03 100644
@@ -550,11 +550,11 @@ static netdev_tx_t dnet_start_xmit(struct sk_buff *skb, struct net_device *dev)
 
 	skb_tx_timestamp(skb);
 
+	spin_unlock_irqrestore(&bp->lock, flags);
+
 	/* free the buffer */
 	dev_kfree_skb(skb);
 
-	spin_unlock_irqrestore(&bp->lock, flags);
-
 	return NETDEV_TX_OK;
 }
 
Signature · 1 line
2.35.1
author maintainer
#338 lore raw

PATCH 335/731 hamradio: dont call dev_kfree_skb() under spin_lock_irqsave()

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit 3727f742915f04f6fc550b80cf406999bd4e90d0 ]

It is not allowed to call kfree_skb() or consume_skb() from hardware interrupt context or with hardware interrupts being disabled.

It should use dev_kfree_skb_irq() or dev_consume_skb_irq() instead. The difference between them is free reason, dev_kfree_skb_irq() means the SKB is dropped in error and dev_consume_skb_irq() means the SKB is consumed in normal.

In scc_discard_buffers(), dev_kfree_skb() is called to discard the SKBs, so replace it with dev_kfree_skb_irq().

In scc_net_tx(), dev_kfree_skb() is called to drop the SKB that exceed queue length, so replace it with dev_kfree_skb_irq().

Fixes 1da177e4c3f4 ("Linux-2.6.12-rc2")
---
 drivers/net/hamradio/scc.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
drivers/net/hamradio/scc.c +3 -3
index e0bb131a33d7..39db3cae4dd1 100644
@@ -301,12 +301,12 @@ static inline void scc_discard_buffers(struct scc_channel *scc)
 	spin_lock_irqsave(&scc->lock, flags);	
 	if (scc->tx_buff != NULL)
 	{
-		dev_kfree_skb(scc->tx_buff);
+		dev_kfree_skb_irq(scc->tx_buff);
 		scc->tx_buff = NULL;
 	}
 	
 	while (!skb_queue_empty(&scc->tx_queue))
-		dev_kfree_skb(skb_dequeue(&scc->tx_queue));
+		dev_kfree_skb_irq(skb_dequeue(&scc->tx_queue));
 
 	spin_unlock_irqrestore(&scc->lock, flags);
 }
@@ -1668,7 +1668,7 @@ static netdev_tx_t scc_net_tx(struct sk_buff *skb, struct net_device *dev)
 	if (skb_queue_len(&scc->tx_queue) > scc->dev->tx_queue_len) {
 		struct sk_buff *skb_del;
 		skb_del = skb_dequeue(&scc->tx_queue);
-		dev_kfree_skb(skb_del);
+		dev_kfree_skb_irq(skb_del);
 	}
 	skb_queue_tail(&scc->tx_queue, skb);
 	netif_trans_update(dev);
Signature · 1 line
2.35.1
author maintainer
#339 lore raw

PATCH 336/731 net: amd: lance: dont call dev_kfree_skb() under spin_lock_irqsave()

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit 6151d105dfce8c23edf30eed35e97f3d9b96a35c ]

It is not allowed to call kfree_skb() or consume_skb() from hardware interrupt context or with hardware interrupts being disabled.

It should use dev_kfree_skb_irq() or dev_consume_skb_irq() instead. The difference between them is free reason, dev_kfree_skb_irq() means the SKB is dropped in error and dev_consume_skb_irq() means the SKB is consumed in normal.

In these two cases, dev_kfree_skb() is called consume the xmited SKB, so replace it with dev_consume_skb_irq().

Fixes 1da177e4c3f4 ("Linux-2.6.12-rc2")
---
 drivers/net/ethernet/amd/atarilance.c | 2 +-
 drivers/net/ethernet/amd/lance.c      | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)
drivers/net/ethernet/amd/atarilance.c +1 -1
index 9d2f49fd945e..a0971ed00453 100644
@@ -821,7 +821,7 @@ lance_start_xmit(struct sk_buff *skb, struct net_device *dev)
 	lp->memcpy_f( PKTBUF_ADDR(head), (void *)skb->data, skb->len );
 	head->flag = TMD1_OWN_CHIP | TMD1_ENP | TMD1_STP;
 	dev->stats.tx_bytes += skb->len;
-	dev_kfree_skb( skb );
+	dev_consume_skb_irq(skb);
 	lp->cur_tx++;
 	while( lp->cur_tx >= TX_RING_SIZE && lp->dirty_tx >= TX_RING_SIZE ) {
 		lp->cur_tx -= TX_RING_SIZE;
drivers/net/ethernet/amd/lance.c +1 -1
index 945bf1d87507..6c2d72024e21 100644
@@ -999,7 +999,7 @@ static netdev_tx_t lance_start_xmit(struct sk_buff *skb,
 		skb_copy_from_linear_data(skb, &lp->tx_bounce_buffs[entry], skb->len);
 		lp->tx_ring[entry].base =
 			((u32)isa_virt_to_bus((lp->tx_bounce_buffs + entry)) & 0xffffff) | 0x83000000;
-		dev_kfree_skb(skb);
+		dev_consume_skb_irq(skb);
 	} else {
 		lp->tx_skbuff[entry] = skb;
 		lp->tx_ring[entry].base = ((u32)isa_virt_to_bus(skb->data) & 0xffffff) | 0x83000000;
Signature · 1 line
2.35.1
author maintainer
#340 lore raw

PATCH 337/731 af_unix: call proto_unregister() in the error path in af_unix_init()

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit 73e341e0281a35274629e9be27eae2f9b1b492bf ]

If register unix_stream_proto returns error, unix_dgram_proto needs be unregistered.

Fixes 94531cfcbe79 ("af_unix: Add unix_stream_proto for sockmap")
---
 net/unix/af_unix.c | 1 +
 1 file changed, 1 insertion(+)
net/unix/af_unix.c +1 -0
index a579e28bd213..545823c1d5ed 100644
@@ -3401,6 +3401,7 @@ static int __init af_unix_init(void)
 	rc = proto_register(&unix_stream_proto, 1);
 	if (rc != 0) {
 		pr_crit("%s: Cannot create unix_sock SLAB cache!\n", __func__);
+		proto_unregister(&unix_dgram_proto);
 		goto out;
 	}
 
Signature · 1 line
2.35.1
author maintainer
#341 lore raw

PATCH 338/731 net: amd-xgbe: Fix logic around active and passive cables

In reply to Greg Kroah-Hartman

From: Tom Lendacky <[email protected]>

[ Upstream commit 4998006c73afe44e2f639d55bd331c6c26eb039f ]

SFP+ active and passive cables are copper cables with fixed SFP+ end connectors. Due to a misinterpretation of this, SFP+ active cables could end up not being recognized, causing the driver to fail to establish a connection.

Introduce a new enum in SFP+ cable types, XGBE_SFP_CABLE_FIBER, that is the default cable type, and handle active and passive cables when they are specifically detected.

Fixes abf0a1c2b26a ("amd-xgbe: Add support for SFP+ modules")
---
 drivers/net/ethernet/amd/xgbe/xgbe-phy-v2.c | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)
drivers/net/ethernet/amd/xgbe/xgbe-phy-v2.c +6 -3
index a7166cd1179f..a9a734454973 100644
@@ -189,6 +189,7 @@ enum xgbe_sfp_cable {
 	XGBE_SFP_CABLE_UNKNOWN = 0,
 	XGBE_SFP_CABLE_ACTIVE,
 	XGBE_SFP_CABLE_PASSIVE,
+	XGBE_SFP_CABLE_FIBER,
 };
 
 enum xgbe_sfp_base {
@@ -1149,16 +1150,18 @@ static void xgbe_phy_sfp_parse_eeprom(struct xgbe_prv_data *pdata)
 	phy_data->sfp_tx_fault = xgbe_phy_check_sfp_tx_fault(phy_data);
 	phy_data->sfp_rx_los = xgbe_phy_check_sfp_rx_los(phy_data);
 
-	/* Assume ACTIVE cable unless told it is PASSIVE */
+	/* Assume FIBER cable unless told otherwise */
 	if (sfp_base[XGBE_SFP_BASE_CABLE] & XGBE_SFP_BASE_CABLE_PASSIVE) {
 		phy_data->sfp_cable = XGBE_SFP_CABLE_PASSIVE;
 		phy_data->sfp_cable_len = sfp_base[XGBE_SFP_BASE_CU_CABLE_LEN];
-	} else {
+	} else if (sfp_base[XGBE_SFP_BASE_CABLE] & XGBE_SFP_BASE_CABLE_ACTIVE) {
 		phy_data->sfp_cable = XGBE_SFP_CABLE_ACTIVE;
+	} else {
+		phy_data->sfp_cable = XGBE_SFP_CABLE_FIBER;
 	}
 
 	/* Determine the type of SFP */
-	if (phy_data->sfp_cable == XGBE_SFP_CABLE_PASSIVE &&
+	if (phy_data->sfp_cable != XGBE_SFP_CABLE_FIBER &&
 	    xgbe_phy_sfp_bit_rate(sfp_eeprom, XGBE_SFP_SPEED_10000))
 		phy_data->sfp_base = XGBE_SFP_BASE_10000_CR;
 	else if (sfp_base[XGBE_SFP_BASE_10GBE_CC] & XGBE_SFP_BASE_10GBE_CC_SR)
Signature · 1 line
2.35.1
author maintainer
#342 lore raw

PATCH 339/731 net: amd-xgbe: Check only the minimum speed for active/passive cables

In reply to Greg Kroah-Hartman

From: Tom Lendacky <[email protected]>

[ Upstream commit f8ab263d4d48e6dab752029bf562f20a2ee630ed ]

There are cables that exist that can support speeds in excess of 10GbE. The driver, however, restricts the EEPROM advertised nominal bitrate to a specific range, which can prevent usage of cables that can support, for example, up to 25GbE.

Rather than checking that an active or passive cable supports a specific range, only check for a minimum supported speed.

Fixes abf0a1c2b26a ("amd-xgbe: Add support for SFP+ modules")
---
 drivers/net/ethernet/amd/xgbe/xgbe-phy-v2.c | 14 ++------------
 1 file changed, 2 insertions(+), 12 deletions(-)
drivers/net/ethernet/amd/xgbe/xgbe-phy-v2.c +2 -12
index a9a734454973..97e32c0490f8 100644
@@ -237,10 +237,7 @@ enum xgbe_sfp_speed {
 
 #define XGBE_SFP_BASE_BR			12
 #define XGBE_SFP_BASE_BR_1GBE_MIN		0x0a
-#define XGBE_SFP_BASE_BR_1GBE_MAX		0x0d
 #define XGBE_SFP_BASE_BR_10GBE_MIN		0x64
-#define XGBE_SFP_BASE_BR_10GBE_MAX		0x68
-#define XGBE_MOLEX_SFP_BASE_BR_10GBE_MAX	0x78
 
 #define XGBE_SFP_BASE_CU_CABLE_LEN		18
 
@@ -827,29 +824,22 @@ static void xgbe_phy_sfp_phy_settings(struct xgbe_prv_data *pdata)
 static bool xgbe_phy_sfp_bit_rate(struct xgbe_sfp_eeprom *sfp_eeprom,
 				  enum xgbe_sfp_speed sfp_speed)
 {
-	u8 *sfp_base, min, max;
+	u8 *sfp_base, min;
 
 	sfp_base = sfp_eeprom->base;
 
 	switch (sfp_speed) {
 	case XGBE_SFP_SPEED_1000:
 		min = XGBE_SFP_BASE_BR_1GBE_MIN;
-		max = XGBE_SFP_BASE_BR_1GBE_MAX;
 		break;
 	case XGBE_SFP_SPEED_10000:
 		min = XGBE_SFP_BASE_BR_10GBE_MIN;
-		if (memcmp(&sfp_eeprom->base[XGBE_SFP_BASE_VENDOR_NAME],
-			   XGBE_MOLEX_VENDOR, XGBE_SFP_BASE_VENDOR_NAME_LEN) == 0)
-			max = XGBE_MOLEX_SFP_BASE_BR_10GBE_MAX;
-		else
-			max = XGBE_SFP_BASE_BR_10GBE_MAX;
 		break;
 	default:
 		return false;
 	}
 
-	return ((sfp_base[XGBE_SFP_BASE_BR] >= min) &&
-		(sfp_base[XGBE_SFP_BASE_BR] <= max));
+	return sfp_base[XGBE_SFP_BASE_BR] >= min;
 }
 
 static void xgbe_phy_free_phy_device(struct xgbe_prv_data *pdata)
Signature · 1 line
2.35.1
author maintainer
#343 lore raw

PATCH 340/731 can: tcan4x5x: Remove invalid write in clear_interrupts

In reply to Greg Kroah-Hartman

From: Markus Schneider-Pargmann <[email protected]>

[ Upstream commit 40c9e4f676abbe194541d88e796341c92d5a13c0 ]

Register 0x824 TCAN4X5X_MCAN_INT_REG is a read-only register. Any writes to this register do not have any effect.

Remove this write. The m_can driver aldready clears the interrupts in m_can_isr() by writing to M_CAN_IR which is translated to register 0x1050 which is a writable version of this register.

Fixes 5443c226ba91 ("can: tcan4x5x: Add tcan4x5x driver to the kernel")
---
 drivers/net/can/m_can/tcan4x5x-core.c | 5 -----
 1 file changed, 5 deletions(-)
drivers/net/can/m_can/tcan4x5x-core.c +0 -5
index 04687b15b250..9b735eccec8a 100644
@@ -204,11 +204,6 @@ static int tcan4x5x_clear_interrupts(struct m_can_classdev *cdev)
 	if (ret)
 		return ret;
 
-	ret = tcan4x5x_write_tcan_reg(cdev, TCAN4X5X_MCAN_INT_REG,
-				      TCAN4X5X_ENABLE_MCAN_INT);
-	if (ret)
-		return ret;
-
 	ret = tcan4x5x_write_tcan_reg(cdev, TCAN4X5X_INT_FLAGS,
 				      TCAN4X5X_CLEAR_ALL_INT);
 	if (ret)
Signature · 1 line
2.35.1
author maintainer
#344 lore raw

PATCH 341/731 can: m_can: Call the RAM init directly from m_can_chip_config

In reply to Greg Kroah-Hartman

From: Vivek Yadav <[email protected]>

[ Upstream commit eaacfeaca7ad0804b9a6eff7afeba93a87db7638 ]

When we try to access the mcan message ram addresses during the probe, hclk is gated by any other drivers or disabled, because of that probe gets failed.

Move the mram init functionality to mcan chip config called by m_can_start from mcan open function, by that time clocks are enabled.

Stable-dep-of: 67727a17a6b3 ("can: tcan4x5x: Fix use of register error status mask")

---
 drivers/net/can/m_can/m_can.c          | 32 +++++++++++++++++++++-----
 drivers/net/can/m_can/m_can_platform.c |  4 ----
 drivers/net/can/m_can/tcan4x5x-core.c  |  5 ----
 3 files changed, 26 insertions(+), 15 deletions(-)
drivers/net/can/m_can/m_can.c +26 -6
index 46ab6155795c..e027229c1955 100644
@@ -1248,10 +1248,17 @@ static int m_can_set_bittiming(struct net_device *dev)
  * - setup bittiming
  * - configure timestamp generation
  */
-static void m_can_chip_config(struct net_device *dev)
+static int m_can_chip_config(struct net_device *dev)
 {
 	struct m_can_classdev *cdev = netdev_priv(dev);
 	u32 cccr, test;
+	int err;
+
+	err = m_can_init_ram(cdev);
+	if (err) {
+		dev_err(cdev->dev, "Message RAM configuration failed\n");
+		return err;
+	}
 
 	m_can_config_endisable(cdev, true);
 
@@ -1375,18 +1382,25 @@ static void m_can_chip_config(struct net_device *dev)
 
 	if (cdev->ops->init)
 		cdev->ops->init(cdev);
+
+	return 0;
 }
 
-static void m_can_start(struct net_device *dev)
+static int m_can_start(struct net_device *dev)
 {
 	struct m_can_classdev *cdev = netdev_priv(dev);
+	int ret;
 
 	/* basic m_can configuration */
-	m_can_chip_config(dev);
+	ret = m_can_chip_config(dev);
+	if (ret)
+		return ret;
 
 	cdev->can.state = CAN_STATE_ERROR_ACTIVE;
 
 	m_can_enable_all_interrupts(cdev);
+
+	return 0;
 }
 
 static int m_can_set_mode(struct net_device *dev, enum can_mode mode)
@@ -1824,7 +1838,9 @@ static int m_can_open(struct net_device *dev)
 	}
 
 	/* start the m_can controller */
-	m_can_start(dev);
+	err = m_can_start(dev);
+	if (err)
+		goto exit_irq_fail;
 
 	can_led_event(dev, CAN_LED_EVENT_OPEN);
 
@@ -2082,9 +2098,13 @@ int m_can_class_resume(struct device *dev)
 		ret = m_can_clk_start(cdev);
 		if (ret)
 			return ret;
+		ret  = m_can_start(ndev);
+		if (ret) {
+			m_can_clk_stop(cdev);
+
+			return ret;
+		}
 
-		m_can_init_ram(cdev);
-		m_can_start(ndev);
 		netif_device_attach(ndev);
 		netif_start_queue(ndev);
 	}
drivers/net/can/m_can/m_can_platform.c +0 -4
index eee47bad0592..de6d8e01bf2e 100644
@@ -140,10 +140,6 @@ static int m_can_plat_probe(struct platform_device *pdev)
 
 	platform_set_drvdata(pdev, mcan_class);
 
-	ret = m_can_init_ram(mcan_class);
-	if (ret)
-		goto probe_fail;
-
 	pm_runtime_enable(mcan_class->dev);
 	ret = m_can_class_register(mcan_class);
 	if (ret)
drivers/net/can/m_can/tcan4x5x-core.c +0 -5
index 9b735eccec8a..a0eee69e52ea 100644
@@ -229,11 +229,6 @@ static int tcan4x5x_init(struct m_can_classdev *cdev)
 	if (ret)
 		return ret;
 
-	/* Zero out the MCAN buffers */
-	ret = m_can_init_ram(cdev);
-	if (ret)
-		return ret;
-
 	ret = regmap_update_bits(tcan4x5x->regmap, TCAN4X5X_CONFIG,
 				 TCAN4X5X_MODE_SEL_MASK, TCAN4X5X_MODE_NORMAL);
 	if (ret)
Signature · 1 line
2.35.1
author maintainer
#345 lore raw

PATCH 342/731 can: tcan4x5x: Fix use of register error status mask

In reply to Greg Kroah-Hartman

From: Markus Schneider-Pargmann <[email protected]>

[ Upstream commit 67727a17a6b375d68fe569b77e6516b034b834c0 ]

TCAN4X5X_ERROR_STATUS is not a status register that needs clearing during interrupt handling. Instead this is a masking register that masks error interrupts. Writing TCAN4X5X_CLEAR_ALL_INT to this register effectively masks everything.

Rename the register and mask all error interrupts only once by writing to the register in tcan4x5x_init.

Fixes 5443c226ba91 ("can: tcan4x5x: Add tcan4x5x driver to the kernel")
---
 drivers/net/can/m_can/tcan4x5x-core.c | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)
drivers/net/can/m_can/tcan4x5x-core.c +7 -7
index a0eee69e52ea..c83b347be1cf 100644
@@ -10,7 +10,7 @@
 #define TCAN4X5X_DEV_ID1 0x04
 #define TCAN4X5X_REV 0x08
 #define TCAN4X5X_STATUS 0x0C
-#define TCAN4X5X_ERROR_STATUS 0x10
+#define TCAN4X5X_ERROR_STATUS_MASK 0x10
 #define TCAN4X5X_CONTROL 0x14
 
 #define TCAN4X5X_CONFIG 0x800
@@ -204,12 +204,7 @@ static int tcan4x5x_clear_interrupts(struct m_can_classdev *cdev)
 	if (ret)
 		return ret;
 
-	ret = tcan4x5x_write_tcan_reg(cdev, TCAN4X5X_INT_FLAGS,
-				      TCAN4X5X_CLEAR_ALL_INT);
-	if (ret)
-		return ret;
-
-	return tcan4x5x_write_tcan_reg(cdev, TCAN4X5X_ERROR_STATUS,
+	return tcan4x5x_write_tcan_reg(cdev, TCAN4X5X_INT_FLAGS,
 				       TCAN4X5X_CLEAR_ALL_INT);
 }
 
@@ -229,6 +224,11 @@ static int tcan4x5x_init(struct m_can_classdev *cdev)
 	if (ret)
 		return ret;
 
+	ret = tcan4x5x_write_tcan_reg(cdev, TCAN4X5X_ERROR_STATUS_MASK,
+				      TCAN4X5X_CLEAR_ALL_INT);
+	if (ret)
+		return ret;
+
 	ret = regmap_update_bits(tcan4x5x->regmap, TCAN4X5X_CONFIG,
 				 TCAN4X5X_MODE_SEL_MASK, TCAN4X5X_MODE_NORMAL);
 	if (ret)
Signature · 1 line
2.35.1
author maintainer
#346 lore raw

PATCH 343/731 net: lan9303: Fix read error execution path

In reply to Greg Kroah-Hartman

From: Jerry Ray <[email protected]>

[ Upstream commit 8964916d206071b058c6351f88b1966bd58cbde0 ]

This patch fixes an issue where a read failure of a port statistic counter will return unknown results. While it is highly unlikely the read will ever fail, it is much cleaner to return a zero for the stat count.

Fixes a1292595e006 ("net: dsa: add new DSA switch driver for the SMSC-LAN9303")
---
 drivers/net/dsa/lan9303-core.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/net/dsa/lan9303-core.c +3 -1
index 59a803e3c8d0..22547b10dfe5 100644
@@ -1003,9 +1003,11 @@ static void lan9303_get_ethtool_stats(struct dsa_switch *ds, int port,
 		ret = lan9303_read_switch_port(
 			chip, port, lan9303_mib[u].offset, &reg);
 
-		if (ret)
+		if (ret) {
 			dev_warn(chip->dev, "Reading status port %d reg %u failed\n",
 				 port, lan9303_mib[u].offset);
+			reg = 0;
+		}
 		data[u] = reg;
 	}
 }
Signature · 1 line
2.35.1
author maintainer
#347 lore raw

PATCH 344/731 ntb_netdev: Use dev_kfree_skb_any() in interrupt context

In reply to Greg Kroah-Hartman

From: Eric Pilmore <[email protected]>

[ Upstream commit 5f7d78b2b12a9d561f48fa00bab29b40f4616dad ]

TX/RX callback handlers (ntb_netdev_tx_handler(), ntb_netdev_rx_handler()) can be called in interrupt context via the DMA framework when the respective DMA operations have completed. As such, any calls by these routines to free skb's, should use the interrupt context safe dev_kfree_skb_any() function.

Previously, these callback handlers would call the interrupt unsafe version of dev_kfree_skb(). This has not presented an issue on Intel IOAT DMA engines as that driver utilizes tasklets rather than a hard interrupt handler, like the AMD PTDMA DMA driver. On AMD systems, a kernel WARNING message is encountered, which is being issued from skb_release_head_state() due to in_hardirq() being true.

Besides the user visible WARNING from the kernel, the other symptom of this bug was that TCP/IP performance across the ntb_netdev interface was very poor, i.e. approximately an order of magnitude below what was expected. With the repair to use dev_kfree_skb_any(), kernel WARNINGs from skb_release_head_state() ceased and TCP/IP performance, as measured by iperf, was on par with expected results, approximately 20 Gb/s on AMD Milan based server. Note that this performance is comparable with Intel based servers.

Fixes 765ccc7bc3d91 ("ntb_netdev: correct skb leak")
Fixes 548c237c0a997 ("net: Add support for NTB virtual ethernet device")
---
 drivers/net/ntb_netdev.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/net/ntb_netdev.c +2 -2
index 1b7d588ff3c5..b701ee83e64a 100644
@@ -137,7 +137,7 @@ static void ntb_netdev_rx_handler(struct ntb_transport_qp *qp, void *qp_data,
 enqueue_again:
 	rc = ntb_transport_rx_enqueue(qp, skb, skb->data, ndev->mtu + ETH_HLEN);
 	if (rc) {
-		dev_kfree_skb(skb);
+		dev_kfree_skb_any(skb);
 		ndev->stats.rx_errors++;
 		ndev->stats.rx_fifo_errors++;
 	}
@@ -192,7 +192,7 @@ static void ntb_netdev_tx_handler(struct ntb_transport_qp *qp, void *qp_data,
 		ndev->stats.tx_aborted_errors++;
 	}
 
-	dev_kfree_skb(skb);
+	dev_kfree_skb_any(skb);
 
 	if (ntb_transport_tx_free_entry(dev->qp) >= tx_start) {
 		/* Make sure anybody stopping the queue after this sees the new
Signature · 1 line
2.35.1
author maintainer
#348 lore raw

PATCH 345/731 sctp: sysctl: make extra pointers netns aware

In reply to Greg Kroah-Hartman

From: Firo Yang <[email protected]>

[ Upstream commit da05cecc4939c0410d56c29e252998b192756318 ]

Recently, a customer reported that from their container whose net namespace is different to the host's init_net, they can't set the container's net.sctp.rto_max to any value smaller than init_net.sctp.rto_min.

For instance,
Host:
sudo sysctl net.sctp.rto_min
net.sctp.rto_min = 1000

Container:
echo 100 > /mnt/proc-net/sctp/rto_min
echo 400 > /mnt/proc-net/sctp/rto_max
echo: write error: Invalid argument

This is caused by the check made from this'commit 4f3fdf3bc59c ("sctp: add check rto_min and rto_max in sysctl")'
When validating the input value, it's always referring the boundary value set for the init_net namespace.

Having container's rto_max smaller than host's init_net.sctp.rto_min does make sense. Consider that the rto between two containers on the same host is very likely smaller than it for two hosts.

So to fix this problem, as suggested by Marcelo, this patch makes the extra pointers of rto_min, rto_max, pf_retrans, and ps_retrans point to the corresponding variables from the newly created net namespace while the new net namespace is being registered in sctp_sysctl_net_register.

---
 net/sctp/sysctl.c | 73 ++++++++++++++++++++++++++++-------------------
 1 file changed, 44 insertions(+), 29 deletions(-)
net/sctp/sysctl.c +44 -29
index b46a416787ec..43ebf090029d 100644
@@ -84,17 +84,18 @@ static struct ctl_table sctp_table[] = {
 	{ /* sentinel */ }
 };
 
+/* The following index defines are used in sctp_sysctl_net_register().
+ * If you add new items to the sctp_net_table, please ensure that
+ * the index values of these defines hold the same meaning indicated by
+ * their macro names when they appear in sctp_net_table.
+ */
+#define SCTP_RTO_MIN_IDX       0
+#define SCTP_RTO_MAX_IDX       1
+#define SCTP_PF_RETRANS_IDX    2
+#define SCTP_PS_RETRANS_IDX    3
+
 static struct ctl_table sctp_net_table[] = {
-	{
-		.procname	= "rto_initial",
-		.data		= &init_net.sctp.rto_initial,
-		.maxlen		= sizeof(unsigned int),
-		.mode		= 0644,
-		.proc_handler	= proc_dointvec_minmax,
-		.extra1         = SYSCTL_ONE,
-		.extra2         = &timer_max
-	},
-	{
+	[SCTP_RTO_MIN_IDX] = {
 		.procname	= "rto_min",
 		.data		= &init_net.sctp.rto_min,
 		.maxlen		= sizeof(unsigned int),
@@ -103,7 +104,7 @@ static struct ctl_table sctp_net_table[] = {
 		.extra1         = SYSCTL_ONE,
 		.extra2         = &init_net.sctp.rto_max
 	},
-	{
+	[SCTP_RTO_MAX_IDX] =  {
 		.procname	= "rto_max",
 		.data		= &init_net.sctp.rto_max,
 		.maxlen		= sizeof(unsigned int),
@@ -112,6 +113,33 @@ static struct ctl_table sctp_net_table[] = {
 		.extra1         = &init_net.sctp.rto_min,
 		.extra2         = &timer_max
 	},
+	[SCTP_PF_RETRANS_IDX] = {
+		.procname	= "pf_retrans",
+		.data		= &init_net.sctp.pf_retrans,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec_minmax,
+		.extra1		= SYSCTL_ZERO,
+		.extra2		= &init_net.sctp.ps_retrans,
+	},
+	[SCTP_PS_RETRANS_IDX] = {
+		.procname	= "ps_retrans",
+		.data		= &init_net.sctp.ps_retrans,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec_minmax,
+		.extra1		= &init_net.sctp.pf_retrans,
+		.extra2		= &ps_retrans_max,
+	},
+	{
+		.procname	= "rto_initial",
+		.data		= &init_net.sctp.rto_initial,
+		.maxlen		= sizeof(unsigned int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec_minmax,
+		.extra1         = SYSCTL_ONE,
+		.extra2         = &timer_max
+	},
 	{
 		.procname	= "rto_alpha_exp_divisor",
 		.data		= &init_net.sctp.rto_alpha,
@@ -207,24 +235,6 @@ static struct ctl_table sctp_net_table[] = {
 		.extra1		= SYSCTL_ONE,
 		.extra2		= SYSCTL_INT_MAX,
 	},
-	{
-		.procname	= "pf_retrans",
-		.data		= &init_net.sctp.pf_retrans,
-		.maxlen		= sizeof(int),
-		.mode		= 0644,
-		.proc_handler	= proc_dointvec_minmax,
-		.extra1		= SYSCTL_ZERO,
-		.extra2		= &init_net.sctp.ps_retrans,
-	},
-	{
-		.procname	= "ps_retrans",
-		.data		= &init_net.sctp.ps_retrans,
-		.maxlen		= sizeof(int),
-		.mode		= 0644,
-		.proc_handler	= proc_dointvec_minmax,
-		.extra1		= &init_net.sctp.pf_retrans,
-		.extra2		= &ps_retrans_max,
-	},
 	{
 		.procname	= "sndbuf_policy",
 		.data		= &init_net.sctp.sndbuf_policy,
@@ -586,6 +596,11 @@ int sctp_sysctl_net_register(struct net *net)
 	for (i = 0; table[i].data; i++)
 		table[i].data += (char *)(&net->sctp) - (char *)&init_net.sctp;
 
+	table[SCTP_RTO_MIN_IDX].extra2 = &net->sctp.rto_max;
+	table[SCTP_RTO_MAX_IDX].extra1 = &net->sctp.rto_min;
+	table[SCTP_PF_RETRANS_IDX].extra2 = &net->sctp.ps_retrans;
+	table[SCTP_PS_RETRANS_IDX].extra1 = &net->sctp.pf_retrans;
+
 	net->sctp.sysctl_header = register_net_sysctl(net, "net/sctp", table);
 	if (net->sctp.sysctl_header == NULL) {
 		kfree(table);
Signature · 1 line
2.35.1
author maintainer
#349 lore raw

PATCH 346/731 Bluetooth: MGMT: Fix error report for ADD_EXT_ADV_PARAMS

In reply to Greg Kroah-Hartman

From: Inga Stotland <[email protected]>

[ Upstream commit 3b1c7c00b8c22b3cb79532252c59eb0b287bb86d ]

When validating the parameter length for MGMT_OP_ADD_EXT_ADV_PARAMS command, use the correct op code in error status report:
was MGMT_OP_ADD_ADVERTISING, changed to MGMT_OP_ADD_EXT_ADV_PARAMS.

Fixes 12410572833a2 ("Bluetooth: Break add adv into two mgmt commands")
---
 net/bluetooth/mgmt.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
net/bluetooth/mgmt.c +1 -1
index f09f0a78eb7b..04000499f4a2 100644
@@ -7971,7 +7971,7 @@ static int add_ext_adv_params(struct sock *sk, struct hci_dev *hdev,
 	 * extra parameters we don't know about will be ignored in this request.
 	 */
 	if (data_len < MGMT_ADD_EXT_ADV_PARAMS_MIN_SIZE)
-		return mgmt_cmd_status(sk, hdev->id, MGMT_OP_ADD_ADVERTISING,
+		return mgmt_cmd_status(sk, hdev->id, MGMT_OP_ADD_EXT_ADV_PARAMS,
 				       MGMT_STATUS_INVALID_PARAMS);
 
 	flags = __le32_to_cpu(cp->flags);
Signature · 1 line
2.35.1
author maintainer
#350 lore raw

PATCH 347/731 Bluetooth: btintel: Fix missing free skb in btintel_setup_combined()

In reply to Greg Kroah-Hartman

From: Wang ShaoBo <[email protected]>

[ Upstream commit cee50ce899de415baf4da3ed38b7d4f13c3170d1 ]

skb allocated by __hci_cmd_sync would not be used whether in checking for supported iBT hardware variants or after, we should free it in all error branches, this patch makes the case read version failed or default error case free skb before return.

Fixes c86c7285bb08 ("Bluetooth: btintel: Fix the legacy bootloader returns tlv based version")
Fixes 019a1caa7fd2 ("Bluetooth: btintel: Refactoring setup routine for bootloader devices")
---
 drivers/bluetooth/btintel.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
drivers/bluetooth/btintel.c +3 -2
index de3d851d85e7..d707aa63e944 100644
@@ -2353,7 +2353,7 @@ static int btintel_setup_combined(struct hci_dev *hdev)
 		 */
 		err = btintel_read_version(hdev, &ver);
 		if (err)
-			return err;
+			break;
 
 		/* Apply the device specific HCI quirks
 		 *
@@ -2394,7 +2394,8 @@ static int btintel_setup_combined(struct hci_dev *hdev)
 	default:
 		bt_dev_err(hdev, "Unsupported Intel hw variant (%u)",
 			   INTEL_HW_VARIANT(ver_tlv.cnvi_bt));
-		return -EINVAL;
+		err = -EINVAL;
+		break;
 	}
 
 exit_error:
Signature · 1 line
2.35.1
author maintainer
#351 lore raw

PATCH 348/731 Bluetooth: btusb: dont call kfree_skb() under spin_lock_irqsave()

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit b15a6bd3c80c77faec8317319b97f976b1a08332 ]

It is not allowed to call kfree_skb() from hardware interrupt context or with interrupts being disabled. So replace kfree_skb() with dev_kfree_skb_irq() under spin_lock_irqsave().

Fixes 803b58367ffb ("Bluetooth: btusb: Implement driver internal packet reassembly")
---
 drivers/bluetooth/btusb.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
drivers/bluetooth/btusb.c +3 -3
index 69380cb03dd3..9c32263f872b 100644
@@ -735,13 +735,13 @@ static inline void btusb_free_frags(struct btusb_data *data)
 
 	spin_lock_irqsave(&data->rxlock, flags);
 
-	kfree_skb(data->evt_skb);
+	dev_kfree_skb_irq(data->evt_skb);
 	data->evt_skb = NULL;
 
-	kfree_skb(data->acl_skb);
+	dev_kfree_skb_irq(data->acl_skb);
 	data->acl_skb = NULL;
 
-	kfree_skb(data->sco_skb);
+	dev_kfree_skb_irq(data->sco_skb);
 	data->sco_skb = NULL;
 
 	spin_unlock_irqrestore(&data->rxlock, flags);
Signature · 1 line
2.35.1
author maintainer
#352 lore raw

PATCH 349/731 Bluetooth: hci_qca: dont call kfree_skb() under spin_lock_irqsave()

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit df4cfc91208e0a98f078223793f5871b1a82cc54 ]

It is not allowed to call kfree_skb() from hardware interrupt context or with interrupts being disabled. So replace kfree_skb() with dev_kfree_skb_irq() under spin_lock_irqsave().

Fixes 0ff252c1976d ("Bluetooth: hciuart: Add support QCA chipset for UART")
---
 drivers/bluetooth/hci_qca.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/bluetooth/hci_qca.c +1 -1
index ed91af4319b5..5d0428fc854f 100644
@@ -912,7 +912,7 @@ static int qca_enqueue(struct hci_uart *hu, struct sk_buff *skb)
 	default:
 		BT_ERR("Illegal tx state: %d (losing packet)",
 		       qca->tx_ibs_state);
-		kfree_skb(skb);
+		dev_kfree_skb_irq(skb);
 		break;
 	}
 
Signature · 1 line
2.35.1
author maintainer
#353 lore raw

PATCH 350/731 Bluetooth: hci_ll: dont call kfree_skb() under spin_lock_irqsave()

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit 8f458f783dfbb19c1f1cb58ed06eeb701f52091b ]

It is not allowed to call kfree_skb() from hardware interrupt context or with interrupts being disabled. So replace kfree_skb() with dev_kfree_skb_irq() under spin_lock_irqsave().

Fixes 166d2f6a4332 ("[Bluetooth] Add UART driver for Texas Instruments' BRF63xx chips")
---
 drivers/bluetooth/hci_ll.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/bluetooth/hci_ll.c +1 -1
index eb1e736efeeb..e4e5b26e2c33 100644
@@ -345,7 +345,7 @@ static int ll_enqueue(struct hci_uart *hu, struct sk_buff *skb)
 	default:
 		BT_ERR("illegal hcill state: %ld (losing packet)",
 		       ll->hcill_state);
-		kfree_skb(skb);
+		dev_kfree_skb_irq(skb);
 		break;
 	}
 
Signature · 1 line
2.35.1
author maintainer
#354 lore raw

PATCH 351/731 Bluetooth: hci_h5: dont call kfree_skb() under spin_lock_irqsave()

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit 383630cc6758d619874c2e8bb2f68a61f3f9ef6e ]

It is not allowed to call kfree_skb() from hardware interrupt context or with interrupts being disabled. So replace kfree_skb() with dev_kfree_skb_irq() under spin_lock_irqsave().

Fixes 43eb12d78960 ("Bluetooth: Fix/implement Three-wire reliable packet sending")
---
 drivers/bluetooth/hci_h5.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/bluetooth/hci_h5.c +1 -1
index e0ea9d25bb39..1363b21c81b7 100644
@@ -313,7 +313,7 @@ static void h5_pkt_cull(struct h5 *h5)
 			break;
 
 		__skb_unlink(skb, &h5->unack);
-		kfree_skb(skb);
+		dev_kfree_skb_irq(skb);
 	}
 
 	if (skb_queue_empty(&h5->unack))
Signature · 1 line
2.35.1
author maintainer
#355 lore raw

PATCH 352/731 Bluetooth: hci_bcsp: dont call kfree_skb() under spin_lock_irqsave()

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit 7b503e339c1a80bf0051ec2d19c3bc777014ac61 ]

It is not allowed to call kfree_skb() from hardware interrupt context or with interrupts being disabled. So replace kfree_skb() with dev_kfree_skb_irq() under spin_lock_irqsave().

Fixes 1da177e4c3f4 ("Linux-2.6.12-rc2")
---
 drivers/bluetooth/hci_bcsp.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/bluetooth/hci_bcsp.c +1 -1
index cf4a56095817..8055f63603f4 100644
@@ -378,7 +378,7 @@ static void bcsp_pkt_cull(struct bcsp_struct *bcsp)
 		i++;
 
 		__skb_unlink(skb, &bcsp->unack);
-		kfree_skb(skb);
+		dev_kfree_skb_irq(skb);
 	}
 
 	if (skb_queue_empty(&bcsp->unack))
Signature · 1 line
2.35.1
author maintainer
#356 lore raw

PATCH 353/731 Bluetooth: hci_core: dont call kfree_skb() under spin_lock_irqsave()

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit 39c1eb6fcbae8ce9bb71b2ac5cb609355a2b181b ]

It is not allowed to call kfree_skb() from hardware interrupt context or with interrupts being disabled. So replace kfree_skb() with dev_kfree_skb_irq() under spin_lock_irqsave().

Fixes 9238f36a5a50 ("Bluetooth: Add request cmd_complete and cmd_status functions")
---
 net/bluetooth/hci_core.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
net/bluetooth/hci_core.c +1 -1
index bb84ff5fb98a..a41b4dcf1a7a 100644
@@ -5101,7 +5101,7 @@ void hci_req_cmd_complete(struct hci_dev *hdev, u16 opcode, u8 status,
 			*req_complete_skb = bt_cb(skb)->hci.req_complete_skb;
 		else
 			*req_complete = bt_cb(skb)->hci.req_complete;
-		kfree_skb(skb);
+		dev_kfree_skb_irq(skb);
 	}
 	spin_unlock_irqrestore(&hdev->cmd_q.lock, flags);
 }
Signature · 1 line
2.35.1
author maintainer
#357 lore raw

PATCH 354/731 Bluetooth: RFCOMM: dont call kfree_skb() under spin_lock_irqsave()

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit 0ba18967d4544955b2eff2fbc4f2a8750c4df90a ]

It is not allowed to call kfree_skb() from hardware interrupt context or with interrupts being disabled. So replace kfree_skb() with dev_kfree_skb_irq() under spin_lock_irqsave().

Fixes 81be03e026dc ("Bluetooth: RFCOMM: Replace use of memcpy_from_msg with bt_skb_sendmmsg")
---
 net/bluetooth/rfcomm/core.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
net/bluetooth/rfcomm/core.c +1 -1
index 7324764384b6..8d6fce9005bd 100644
@@ -590,7 +590,7 @@ int rfcomm_dlc_send(struct rfcomm_dlc *d, struct sk_buff *skb)
 
 		ret = rfcomm_dlc_send_frag(d, frag);
 		if (ret < 0) {
-			kfree_skb(frag);
+			dev_kfree_skb_irq(frag);
 			goto unlock;
 		}
 
Signature · 1 line
2.35.1
author maintainer
#358 lore raw

PATCH 355/731 stmmac: fix potential division by 0

In reply to Greg Kroah-Hartman

From: Piergiorgio Beruto <[email protected]>

[ Upstream commit ede5a389852d3640a28e7187fb32b7f204380901 ]

When the MAC is connected to a 10 Mb/s PHY and the PTP clock is derived from the MAC reference clock (default), the clk_ptp_rate becomes too small and the calculated sub second increment becomes 0 when computed by the stmmac_config_sub_second_increment() function within stmmac_init_tstamp_counter().

Therefore, the subsequent div_u64 in stmmac_init_tstamp_counter() operation triggers a divide by 0 exception as shown below.

[   95.062067] socfpga-dwmac ff700000.ethernet eth0: Register MEM_TYPE_PAGE_POOL RxQ-0
[   95.076440] socfpga-dwmac ff700000.ethernet eth0: PHY [stmmac-0:08] driver [NCN26000] (irq=49)
[   95.095964] dwmac1000: Master AXI performs any burst length
[   95.101588] socfpga-dwmac ff700000.ethernet eth0: No Safety Features support found
[   95.109428] Division by zero in kernel.
[   95.113447] CPU: 0 PID: 239 Comm: ifconfig Not tainted 6.1.0-rc7-centurion3-1.0.3.0-01574-gb624218205b7-dirty #77
[   95.123686] Hardware name: Altera SOCFPGA
[   95.127695]  unwind_backtrace from show_stack+0x10/0x14
[   95.132938]  show_stack from dump_stack_lvl+0x40/0x4c
[   95.137992]  dump_stack_lvl from Ldiv0+0x8/0x10
[   95.142527]  Ldiv0 from __aeabi_uidivmod+0x8/0x18
[   95.147232]  __aeabi_uidivmod from div_u64_rem+0x1c/0x40
[   95.152552]  div_u64_rem from stmmac_init_tstamp_counter+0xd0/0x164
[   95.158826]  stmmac_init_tstamp_counter from stmmac_hw_setup+0x430/0xf00
[   95.165533]  stmmac_hw_setup from __stmmac_open+0x214/0x2d4
[   95.171117]  __stmmac_open from stmmac_open+0x30/0x44
[   95.176182]  stmmac_open from __dev_open+0x11c/0x134
[   95.181172]  __dev_open from __dev_change_flags+0x168/0x17c
[   95.186750]  __dev_change_flags from dev_change_flags+0x14/0x50
[   95.192662]  dev_change_flags from devinet_ioctl+0x2b4/0x604
[   95.198321]  devinet_ioctl from inet_ioctl+0x1ec/0x214
[   95.203462]  inet_ioctl from sock_ioctl+0x14c/0x3c4
[   95.208354]  sock_ioctl from vfs_ioctl+0x20/0x38
[   95.212984]  vfs_ioctl from sys_ioctl+0x250/0x844
[   95.217691]  sys_ioctl from ret_fast_syscall+0x0/0x4c
[   95.222743] Exception stack(0xd0ee1fa8 to 0xd0ee1ff0)
[   95.227790] 1fa0:                   00574c4f be9aeca4 00000003 00008914 be9aeca4 be9aec50
[   95.235945] 1fc0: 00574c4f be9aeca4 0059f078 00000036 be9aee8c be9aef7a 00000015 00000000
[   95.244096] 1fe0: 005a01f0 be9aec38 004d7484 b6e67d74
---
 drivers/net/ethernet/stmicro/stmmac/stmmac_hwtstamp.c | 3 ++-
 drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.h      | 2 +-
 2 files changed, 3 insertions(+), 2 deletions(-)
drivers/net/ethernet/stmicro/stmmac/stmmac_hwtstamp.c +2 -1
index d68ef72dcdde..4538e4fd8189 100644
@@ -47,7 +47,8 @@ static void config_sub_second_increment(void __iomem *ioaddr,
 	if (!(value & PTP_TCR_TSCTRLSSR))
 		data = (data * 1000) / 465;
 
-	data &= PTP_SSIR_SSINC_MASK;
+	if (data > PTP_SSIR_SSINC_MAX)
+		data = PTP_SSIR_SSINC_MAX;
 
 	reg_value = data;
 	if (gmac4)
drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.h +1 -1
index 53172a439810..bf619295d079 100644
@@ -64,7 +64,7 @@
 #define	PTP_TCR_TSENMACADDR	BIT(18)
 
 /* SSIR defines */
-#define	PTP_SSIR_SSINC_MASK		0xff
+#define	PTP_SSIR_SSINC_MAX		0xff
 #define	GMAC4_PTP_SSIR_SSINC_SHIFT	16
 
 /* Auxiliary Control defines */
Signature · 1 line
2.35.1
author maintainer
#359 lore raw

PATCH 356/731 i40e: Fix the inability to attach XDP program on downed interface

In reply to Greg Kroah-Hartman

From: Bartosz Staszewski <[email protected]>

[ Upstream commit 0c87b545a2ed5cd8a6318011f1c92b188c2d74bc ]

Whenever trying to load XDP prog on downed interface, function i40e_xdp was passing vsi->rx_buf_len field to i40e_xdp_setup() which was equal 0. i40e_open() calls i40e_vsi_configure_rx() which configures that field, but that only happens when interface is up. When it is down, i40e_open() is not being called, thus vsi->rx_buf_len is not set.

Solution for this is calculate buffer length in newly created function - i40e_calculate_vsi_rx_buf_len() that return actual buffer length. Buffer length is being calculated based on the same rules applied previously in i40e_vsi_configure_rx() function.

Fixes 613142b0bb88 ("i40e: Log error for oversized MTU on device")
Fixes 0c8493d90b6b ("i40e: add XDP support for pass and drop actions")
---
 drivers/net/ethernet/intel/i40e/i40e_main.c | 36 ++++++++++++++-------
 1 file changed, 24 insertions(+), 12 deletions(-)
drivers/net/ethernet/intel/i40e/i40e_main.c +24 -12
index ed2c961902b6..c013d86559af 100644
@@ -3548,6 +3548,24 @@ static int i40e_vsi_configure_tx(struct i40e_vsi *vsi)
 	return err;
 }
 
+/**
+ * i40e_calculate_vsi_rx_buf_len - Calculates buffer length
+ *
+ * @vsi: VSI to calculate rx_buf_len from
+ */
+static u16 i40e_calculate_vsi_rx_buf_len(struct i40e_vsi *vsi)
+{
+	if (!vsi->netdev || (vsi->back->flags & I40E_FLAG_LEGACY_RX))
+		return I40E_RXBUFFER_2048;
+
+#if (PAGE_SIZE < 8192)
+	if (!I40E_2K_TOO_SMALL_WITH_PADDING && vsi->netdev->mtu <= ETH_DATA_LEN)
+		return I40E_RXBUFFER_1536 - NET_IP_ALIGN;
+#endif
+
+	return PAGE_SIZE < 8192 ? I40E_RXBUFFER_3072 : I40E_RXBUFFER_2048;
+}
+
 /**
  * i40e_vsi_configure_rx - Configure the VSI for Rx
  * @vsi: the VSI being configured
@@ -3559,20 +3577,14 @@ static int i40e_vsi_configure_rx(struct i40e_vsi *vsi)
 	int err = 0;
 	u16 i;
 
-	if (!vsi->netdev || (vsi->back->flags & I40E_FLAG_LEGACY_RX)) {
-		vsi->max_frame = I40E_MAX_RXBUFFER;
-		vsi->rx_buf_len = I40E_RXBUFFER_2048;
+	vsi->max_frame = I40E_MAX_RXBUFFER;
+	vsi->rx_buf_len = i40e_calculate_vsi_rx_buf_len(vsi);
+
 #if (PAGE_SIZE < 8192)
-	} else if (!I40E_2K_TOO_SMALL_WITH_PADDING &&
-		   (vsi->netdev->mtu <= ETH_DATA_LEN)) {
+	if (vsi->netdev && !I40E_2K_TOO_SMALL_WITH_PADDING &&
+	    vsi->netdev->mtu <= ETH_DATA_LEN)
 		vsi->max_frame = I40E_RXBUFFER_1536 - NET_IP_ALIGN;
-		vsi->rx_buf_len = I40E_RXBUFFER_1536 - NET_IP_ALIGN;
 #endif
-	} else {
-		vsi->max_frame = I40E_MAX_RXBUFFER;
-		vsi->rx_buf_len = (PAGE_SIZE < 8192) ? I40E_RXBUFFER_3072 :
-						       I40E_RXBUFFER_2048;
-	}
 
 	/* set up individual rings */
 	for (i = 0; i < vsi->num_queue_pairs && !err; i++)
@@ -13147,7 +13159,7 @@ static int i40e_xdp_setup(struct i40e_vsi *vsi, struct bpf_prog *prog,
 	int i;
 
 	/* Don't allow frames that span over multiple buffers */
-	if (frame_size > vsi->rx_buf_len) {
+	if (frame_size > i40e_calculate_vsi_rx_buf_len(vsi)) {
 		NL_SET_ERR_MSG_MOD(extack, "MTU too large to enable XDP");
 		return -EINVAL;
 	}
Signature · 1 line
2.35.1
author maintainer
#360 lore raw

PATCH 357/731 net: dsa: tag_8021q: avoid leaking ctx on dsa_tag_8021q_register() error path

In reply to Greg Kroah-Hartman

From: Vladimir Oltean <[email protected]>

[ Upstream commit e095493091e850d5292ad01d8fbf5cde1d89ac53 ]

If dsa_tag_8021q_setup() fails, for example due to the inability of the device to install a VLAN, the tag_8021q context of the switch will leak. Make sure it is freed on the error path.

Fixes 328621f6131f ("net: dsa: tag_8021q: absorb dsa_8021q_setup into dsa_tag_8021q_{,un}register")
---
 net/dsa/tag_8021q.c | 11 ++++++++++-
 1 file changed, 10 insertions(+), 1 deletion(-)
net/dsa/tag_8021q.c +10 -1
index f8f7b7c34e7d..e443088ab0f6 100644
@@ -529,6 +529,7 @@ static void dsa_tag_8021q_teardown(struct dsa_switch *ds)
 int dsa_tag_8021q_register(struct dsa_switch *ds, __be16 proto)
 {
 	struct dsa_8021q_context *ctx;
+	int err;
 
 	ctx = kzalloc(sizeof(*ctx), GFP_KERNEL);
 	if (!ctx)
@@ -541,7 +542,15 @@ int dsa_tag_8021q_register(struct dsa_switch *ds, __be16 proto)
 
 	ds->tag_8021q_ctx = ctx;
 
-	return dsa_tag_8021q_setup(ds);
+	err = dsa_tag_8021q_setup(ds);
+	if (err)
+		goto err_free;
+
+	return 0;
+
+err_free:
+	kfree(ctx);
+	return err;
 }
 EXPORT_SYMBOL_GPL(dsa_tag_8021q_register);
 
Signature · 1 line
2.35.1
author maintainer
#361 lore raw

PATCH 358/731 apparmor: fix a memleak in multi_transaction_new()

In reply to Greg Kroah-Hartman

From: Gaosheng Cui <[email protected]>

[ Upstream commit c73275cf6834787ca090317f1d20dbfa3b7f05aa ]

In multi_transaction_new(), the variable t is not freed or passed out on the failure of copy_from_user(t->data, buf, size), which could lead to a memleak.

Fix this bug by adding a put_multi_transaction(t) in the error path.

Fixes 1dea3b41e84c5 ("apparmor: speed up transactional queries")
---
 security/apparmor/apparmorfs.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
security/apparmor/apparmorfs.c +3 -1
index a891705b1d57..8c7719108d7f 100644
@@ -867,8 +867,10 @@ static struct multi_transaction *multi_transaction_new(struct file *file,
 	if (!t)
 		return ERR_PTR(-ENOMEM);
 	kref_init(&t->count);
-	if (copy_from_user(t->data, buf, size))
+	if (copy_from_user(t->data, buf, size)) {
+		put_multi_transaction(t);
 		return ERR_PTR(-EFAULT);
+	}
 
 	return t;
 }
Signature · 1 line
2.35.1
author maintainer
#362 lore raw

PATCH 359/731 apparmor: fix lockdep warning when removing a namespace

In reply to Greg Kroah-Hartman

From: John Johansen <[email protected]>

[ Upstream commit 9c4557efc558a68e4cd973490fd936d6e3414db8 ]

Fix the following lockdep warning

[ 1119.158984] ============================================
[ 1119.158988] WARNING: possible recursive locking detected
[ 1119.158996] 6.0.0-rc1+ #257 Tainted: G            E    N
[ 1119.158999] --------------------------------------------
[ 1119.159001] bash/80100 is trying to acquire lock:
[ 1119.159007] ffff88803e79b4a0 (&ns->lock/1){+.+.}-{4:4}, at: destroy_ns.part.0+0x43/0x140
[ 1119.159028]
               but task is already holding lock:
[ 1119.159030] ffff8881009764a0 (&ns->lock/1){+.+.}-{4:4}, at: aa_remove_profiles+0x3f0/0x640
[ 1119.159040]
               other info that might help us debug this:
[ 1119.159042]  Possible unsafe locking scenario:
[ 1119.159043]        CPU0
[ 1119.159045]        ----
[ 1119.159047]   lock(&ns->lock/1);
[ 1119.159051]   lock(&ns->lock/1);
[ 1119.159055]
                *** DEADLOCK ***

Which is caused by an incorrect lockdep nesting notation

Fixes feb3c766a3ab ("apparmor: fix possible recursive lock warning in __aa_create_ns")
---
 security/apparmor/policy.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
security/apparmor/policy.c +1 -1
index 4c010c9a6af1..fcf22577f606 100644
@@ -1125,7 +1125,7 @@ ssize_t aa_remove_profiles(struct aa_ns *policy_ns, struct aa_label *subj,
 
 	if (!name) {
 		/* remove namespace - can only happen if fqname[0] == ':' */
-		mutex_lock_nested(&ns->parent->lock, ns->level);
+		mutex_lock_nested(&ns->parent->lock, ns->parent->level);
 		__aa_bump_ns_revision(ns);
 		__aa_remove_ns(ns);
 		mutex_unlock(&ns->parent->lock);
Signature · 1 line
2.35.1
author maintainer
#363 lore raw

PATCH 360/731 apparmor: Fix abi check to include v8 abi

In reply to Greg Kroah-Hartman

From: John Johansen <[email protected]>

[ Upstream commit 1b5a6198f5a9d0aa5497da0dc4bcd4fc166ee516 ]

The v8 abi is supported by the kernel but the userspace supported version check does not allow for it. This was missed when v8 was added due to a bug in the userspace compiler which was setting an older abi version for v8 encoding (which is forward compatible except on the network encoding). However it is possible to detect the network encoding by checking the policydb network support which the code does. The end result was that missing the abi flag worked until userspace was fixed and began correctly checking for the v8 abi version.

Fixes 56974a6fcfef ("apparmor: add base infastructure for socket mediation")
---
 security/apparmor/policy_unpack.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
security/apparmor/policy_unpack.c +1 -1
index 03c9609ca407..d5b3a062d1d1 100644
@@ -964,7 +964,7 @@ static int verify_header(struct aa_ext *e, int required, const char **ns)
 	 * if not specified use previous version
 	 * Mask off everything that is not kernel abi version
 	 */
-	if (VERSION_LT(e->version, v5) || VERSION_GT(e->version, v7)) {
+	if (VERSION_LT(e->version, v5) || VERSION_GT(e->version, v8)) {
 		audit_iface(NULL, NULL, NULL, "unsupported interface version",
 			    e, error);
 		return error;
Signature · 1 line
2.35.1
author maintainer
#364 lore raw

PATCH 361/731 crypto: hisilicon/qm - fix missing destroy qp_idr

In reply to Greg Kroah-Hartman

From: Weili Qian <[email protected]>

[ Upstream commit 116be08f6e4e385733d42360a33c3d883d2dd702 ]

In the function hisi_qm_memory_init(), if resource alloc fails after idr_init, the initialized qp_idr needs to be destroyed.

---
 drivers/crypto/hisilicon/qm.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)
drivers/crypto/hisilicon/qm.c +4 -3
index b8900a5dbf6e..fd89918abd19 100644
@@ -5727,8 +5727,8 @@ static int hisi_qm_memory_init(struct hisi_qm *qm)
 					 GFP_ATOMIC);
 	dev_dbg(dev, "allocate qm dma buf size=%zx)\n", qm->qdma.size);
 	if (!qm->qdma.va) {
-		ret =  -ENOMEM;
-		goto err_alloc_qdma;
+		ret = -ENOMEM;
+		goto err_destroy_idr;
 	}
 
 	QM_INIT_BUF(qm, eqe, QM_EQ_DEPTH);
@@ -5744,7 +5744,8 @@ static int hisi_qm_memory_init(struct hisi_qm *qm)
 
 err_alloc_qp_array:
 	dma_free_coherent(dev, qm->qdma.size, qm->qdma.va, qm->qdma.dma);
-err_alloc_qdma:
+err_destroy_idr:
+	idr_destroy(&qm->qp_idr);
 	kfree(qm->factor);
 
 	return ret;
Signature · 1 line
2.35.1
author maintainer
#365 lore raw

PATCH 362/731 crypto: sun8i-ss - use dma_addr instead u32

In reply to Greg Kroah-Hartman

From: Corentin Labbe <[email protected]>

[ Upstream commit 839b8ae2fc10f205317bcc32c9de18456756e1f5 ]

The DMA address need to be stored in a dma_addr_t

Fixes 359e893e8af4 ("crypto: sun8i-ss - rework handling of IV")
---
 drivers/crypto/allwinner/sun8i-ss/sun8i-ss-cipher.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/crypto/allwinner/sun8i-ss/sun8i-ss-cipher.c +1 -1
index 3c46ad8c3a1c..005eefecfdf5 100644
@@ -105,7 +105,7 @@ static int sun8i_ss_setup_ivs(struct skcipher_request *areq)
 	unsigned int ivsize = crypto_skcipher_ivsize(tfm);
 	struct sun8i_ss_flow *sf = &ss->flows[rctx->flow];
 	int i = 0;
-	u32 a;
+	dma_addr_t a;
 	int err;
 
 	rctx->ivlen = ivsize;
Signature · 1 line
2.35.1
author maintainer
#366 lore raw

PATCH 363/731 crypto: nitrox - avoid double free on error path in nitrox_sriov_init()

In reply to Greg Kroah-Hartman

From: Natalia Petrova <[email protected]>

[ Upstream commit 094528b6a5a755b1195a01e10b13597d67d1a0e6 ]

If alloc_workqueue() fails in nitrox_mbox_init() it deallocates ndev->iov.vfdev and returns error code, but then nitrox_sriov_init() calls nitrox_sriov_cleanup() where ndev->iov.vfdev is deallocated again.

Fix this by nulling ndev->iov.vfdev after the first deallocation.

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

Fixes 9e5de3e06e54 ("crypto: cavium/nitrox - Add mailbox...")
---
 drivers/crypto/cavium/nitrox/nitrox_mbx.c | 1 +
 1 file changed, 1 insertion(+)
drivers/crypto/cavium/nitrox/nitrox_mbx.c +1 -0
index 2e9c0d214363..199fcec9b8d0 100644
@@ -191,6 +191,7 @@ int nitrox_mbox_init(struct nitrox_device *ndev)
 	ndev->iov.pf2vf_wq = alloc_workqueue("nitrox_pf2vf", 0, 0);
 	if (!ndev->iov.pf2vf_wq) {
 		kfree(ndev->iov.vfdev);
+		ndev->iov.vfdev = NULL;
 		return -ENOMEM;
 	}
 	/* enable pf2vf mailbox interrupts */
Signature · 1 line
2.35.1
author maintainer
#367 lore raw

PATCH 364/731 scsi: core: Fix a race between scsi_done() and scsi_timeout()

In reply to Greg Kroah-Hartman

From: Bart Van Assche <[email protected]>

[ Upstream commit 978b7922d3dca672b41bb4b8ce6c06ab77112741 ]

If there is a race between scsi_done() and scsi_timeout() and if scsi_timeout() loses the race, scsi_timeout() should not reset the request timer. Hence change the return value for this case from BLK_EH_RESET_TIMER into BLK_EH_DONE.

Although the block layer holds a reference on a request (req->ref) while calling a timeout handler, restarting the timer (blk_add_timer()) while a request is being completed is racy.

Cc Keith Busch <[email protected]>
Cc Christoph Hellwig <[email protected]>
Cc Ming Lei <[email protected]>
Cc John Garry <[email protected]>
Cc Hannes Reinecke <[email protected]>
Fixes 15f73f5b3e59 ("blk-mq: move failure injection out of blk_mq_complete_request")
---
 drivers/scsi/scsi_error.c | 14 +++-----------
 1 file changed, 3 insertions(+), 11 deletions(-)
drivers/scsi/scsi_error.c +3 -11
index bb5a6e0fa49a..dd9f5778f687 100644
@@ -343,19 +343,11 @@ enum blk_eh_timer_return scsi_times_out(struct request *req)
 
 	if (rtn == BLK_EH_DONE) {
 		/*
-		 * Set the command to complete first in order to prevent a real
-		 * completion from releasing the command while error handling
-		 * is using it. If the command was already completed, then the
-		 * lower level driver beat the timeout handler, and it is safe
-		 * to return without escalating error recovery.
-		 *
-		 * If timeout handling lost the race to a real completion, the
-		 * block layer may ignore that due to a fake timeout injection,
-		 * so return RESET_TIMER to allow error handling another shot
-		 * at this command.
+		 * If scsi_done() has already set SCMD_STATE_COMPLETE, do not
+		 * modify *scmd.
 		 */
 		if (test_and_set_bit(SCMD_STATE_COMPLETE, &scmd->state))
-			return BLK_EH_RESET_TIMER;
+			return BLK_EH_DONE;
 		if (scsi_abort_command(scmd) != SUCCESS) {
 			set_host_byte(scmd, DID_TIME_OUT);
 			scsi_eh_scmd_add(scmd);
Signature · 1 line
2.35.1
author maintainer
#368 lore raw

PATCH 365/731 apparmor: Use pointer to struct aa_label for lbs_cred

In reply to Greg Kroah-Hartman

From: Xiu Jianfeng <[email protected]>

[ Upstream commit 37923d4321b1e38170086da2c117f78f2b0f49c6 ]

According to the implementations of cred_label() and set_cred_label(), we should use pointer to struct aa_label for lbs_cred instead of struct aa_task_ctx, this patch fixes it.

Fixes bbd3662a8348 ("Infrastructure management of the cred security blob")
---
 security/apparmor/lsm.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
security/apparmor/lsm.c +2 -2
index f72406fe1bf2..10274eb90fa3 100644
@@ -1170,10 +1170,10 @@ static int apparmor_inet_conn_request(const struct sock *sk, struct sk_buff *skb
 #endif
 
 /*
- * The cred blob is a pointer to, not an instance of, an aa_task_ctx.
+ * The cred blob is a pointer to, not an instance of, an aa_label.
  */
 struct lsm_blob_sizes apparmor_blob_sizes __lsm_ro_after_init = {
-	.lbs_cred = sizeof(struct aa_task_ctx *),
+	.lbs_cred = sizeof(struct aa_label *),
 	.lbs_file = sizeof(struct aa_file_ctx),
 	.lbs_task = sizeof(struct aa_task_ctx),
 };
Signature · 1 line
2.35.1
author maintainer
#369 lore raw

PATCH 366/731 PCI: dwc: Fix n_fts[] array overrun

In reply to Greg Kroah-Hartman

From: Vidya Sagar <[email protected]>

[ Upstream commit 66110361281b2f7da0c8bd51eaf1f152f4236035 ]

commit aeaa0bfe89654 ("PCI: dwc: Move N_FTS setup to common setup") incorrectly uses pci->link_gen in deriving the index to the n_fts[] array also introducing the issue of accessing beyond the boundaries of array for greater than Gen-2 speeds. This change fixes that issue.

---
 drivers/pci/controller/dwc/pcie-designware.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/pci/controller/dwc/pcie-designware.c +1 -1
index e408ebf5bd73..00972a7bc976 100644
@@ -730,7 +730,7 @@ void dw_pcie_setup(struct dw_pcie *pci)
 	if (pci->n_fts[1]) {
 		val = dw_pcie_readl_dbi(pci, PCIE_LINK_WIDTH_SPEED_CONTROL);
 		val &= ~PORT_LOGIC_N_FTS_MASK;
-		val |= pci->n_fts[pci->link_gen - 1];
+		val |= pci->n_fts[1];
 		dw_pcie_writel_dbi(pci, PCIE_LINK_WIDTH_SPEED_CONTROL, val);
 	}
 
Signature · 1 line
2.35.1
author maintainer
#370 lore raw

PATCH 367/731 RDMA/core: Fix order of nldev_exit call

In reply to Greg Kroah-Hartman

From: Leon Romanovsky <[email protected]>

[ Upstream commit 4508d32ccced24c972bc4592104513e1ff8439b5 ]

Create symmetrical exit flow by calling to nldev_exit() after call to rdma_nl_unregister(RDMA_NL_LS).

---
 drivers/infiniband/core/device.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/infiniband/core/device.c +1 -1
index 1519379b116e..ab2106a09f9c 100644
@@ -2848,8 +2848,8 @@ static int __init ib_core_init(void)
 static void __exit ib_core_cleanup(void)
 {
 	roce_gid_mgmt_cleanup();
-	nldev_exit();
 	rdma_nl_unregister(RDMA_NL_LS);
+	nldev_exit();
 	unregister_pernet_device(&rdma_dev_net_ops);
 	unregister_blocking_lsm_notifier(&ibdev_lsm_nb);
 	ib_sa_cleanup();
Signature · 1 line
2.35.1
author maintainer
#371 lore raw

PATCH 368/731 PCI: pci-epf-test: Register notifier if only core_init_notifier is enabled

In reply to Greg Kroah-Hartman

From: Kunihiko Hayashi <[email protected]>

[ Upstream commit 6acd25cc98ce0c9ee4fefdaf44fc8bca534b26e5 ]

The pci_epf_test_notifier function should be installed also if only core_init_notifier is enabled. Fix the current logic.

---
 drivers/pci/endpoint/functions/pci-epf-test.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/pci/endpoint/functions/pci-epf-test.c +1 -1
index a5ed779b0a51..45535d4ae644 100644
@@ -883,7 +883,7 @@ static int pci_epf_test_bind(struct pci_epf *epf)
 	if (ret)
 		epf_test->dma_supported = false;
 
-	if (linkup_notifier) {
+	if (linkup_notifier || core_init_notifier) {
 		epf->nb.notifier_call = pci_epf_test_notifier;
 		pci_epc_register_notifier(epc, &epf->nb);
 	} else {
Signature · 1 line
2.35.1
author maintainer
#372 lore raw

PATCH 369/731 f2fs: Fix the race condition of resize flag between resizefs

In reply to Greg Kroah-Hartman

From: Zhang Qilong <[email protected]>

[ Upstream commit 28fc4e9077ce59ab28c89c20dc6be5154473218f ]

Because the set/clear SBI_IS_RESIZEFS flag not between any locks, In the following case:
thread1 thread2
->ioctl(resizefs)

    ->set RESIZEFS flag		 ->ioctl(resizefs)
    ...                   	  ->set RESIZEFS flag
    ->clear RESIZEFS flag
    				  ->resizefs stream
				    # No RESIZEFS flag in the stream

Also before freeze_super, the resizefs not started, we should not set the SBI_IS_RESIZEFS flag.

So move the set/clear SBI_IS_RESIZEFS flag between the cp_mutex and gc_lock.

Fixes b4b10061ef98 ("f2fs: refactor resize_fs to avoid meta updates in progress")
---
 fs/f2fs/gc.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)
fs/f2fs/gc.c +2 -3
index e75a276f5b9c..4cbaa6ab083f 100644
@@ -2051,8 +2051,6 @@ int f2fs_resize_fs(struct f2fs_sb_info *sbi, __u64 block_count)
 	if (err)
 		return err;
 
-	set_sbi_flag(sbi, SBI_IS_RESIZEFS);
-
 	freeze_super(sbi->sb);
 	down_write(&sbi->gc_lock);
 	down_write(&sbi->cp_global_sem);
@@ -2068,6 +2066,7 @@ int f2fs_resize_fs(struct f2fs_sb_info *sbi, __u64 block_count)
 	if (err)
 		goto out_err;
 
+	set_sbi_flag(sbi, SBI_IS_RESIZEFS);
 	err = free_segment_range(sbi, secs, false);
 	if (err)
 		goto recover_out;
@@ -2091,6 +2090,7 @@ int f2fs_resize_fs(struct f2fs_sb_info *sbi, __u64 block_count)
 		f2fs_commit_super(sbi, false);
 	}
 recover_out:
+	clear_sbi_flag(sbi, SBI_IS_RESIZEFS);
 	if (err) {
 		set_sbi_flag(sbi, SBI_NEED_FSCK);
 		f2fs_err(sbi, "resize_fs failed, should run fsck to repair!");
@@ -2103,6 +2103,5 @@ int f2fs_resize_fs(struct f2fs_sb_info *sbi, __u64 block_count)
 	up_write(&sbi->cp_global_sem);
 	up_write(&sbi->gc_lock);
 	thaw_super(sbi->sb);
-	clear_sbi_flag(sbi, SBI_IS_RESIZEFS);
 	return err;
 }
Signature · 1 line
2.35.1
author maintainer
#373 lore raw

PATCH 370/731 crypto: rockchip - do not do custom power management

In reply to Greg Kroah-Hartman

From: Corentin Labbe <[email protected]>

[ Upstream commit c50ef1411c8cbad0c7db100c477126076b6e3348 ]

The clock enable/disable at tfm init/exit is fragile, if 2 tfm are init in the same time and one is removed just after, it will leave the hardware uncloked even if a user remains.

Instead simply enable clocks at probe time.
We will do PM later.

Fixes ce0183cb6464b ("crypto: rockchip - switch to skcipher API")
---
 drivers/crypto/rockchip/rk3288_crypto.c          | 4 ++--
 drivers/crypto/rockchip/rk3288_crypto.h          | 2 --
 drivers/crypto/rockchip/rk3288_crypto_ahash.c    | 3 +--
 drivers/crypto/rockchip/rk3288_crypto_skcipher.c | 5 +++--
 4 files changed, 6 insertions(+), 8 deletions(-)
drivers/crypto/rockchip/rk3288_crypto.c +2 -2
index 35d73061d156..5f8444b9633a 100644
@@ -395,8 +395,7 @@ static int rk_crypto_probe(struct platform_device *pdev)
 		     rk_crypto_done_task_cb, (unsigned long)crypto_info);
 	crypto_init_queue(&crypto_info->queue, 50);
 
-	crypto_info->enable_clk = rk_crypto_enable_clk;
-	crypto_info->disable_clk = rk_crypto_disable_clk;
+	rk_crypto_enable_clk(crypto_info);
 	crypto_info->load_data = rk_load_data;
 	crypto_info->unload_data = rk_unload_data;
 	crypto_info->enqueue = rk_crypto_enqueue;
@@ -423,6 +422,7 @@ static int rk_crypto_remove(struct platform_device *pdev)
 	struct rk_crypto_info *crypto_tmp = platform_get_drvdata(pdev);
 
 	rk_crypto_unregister();
+	rk_crypto_disable_clk(crypto_tmp);
 	tasklet_kill(&crypto_tmp->done_task);
 	tasklet_kill(&crypto_tmp->queue_task);
 	return 0;
drivers/crypto/rockchip/rk3288_crypto.h +0 -2
index 97278c2574ff..2fa7131e4060 100644
@@ -220,8 +220,6 @@ struct rk_crypto_info {
 	int (*start)(struct rk_crypto_info *dev);
 	int (*update)(struct rk_crypto_info *dev);
 	void (*complete)(struct crypto_async_request *base, int err);
-	int (*enable_clk)(struct rk_crypto_info *dev);
-	void (*disable_clk)(struct rk_crypto_info *dev);
 	int (*load_data)(struct rk_crypto_info *dev,
 			 struct scatterlist *sg_src,
 			 struct scatterlist *sg_dst);
drivers/crypto/rockchip/rk3288_crypto_ahash.c +1 -2
index ed03058497bc..49017d1fb510 100644
@@ -301,7 +301,7 @@ static int rk_cra_hash_init(struct crypto_tfm *tfm)
 				 sizeof(struct rk_ahash_rctx) +
 				 crypto_ahash_reqsize(tctx->fallback_tfm));
 
-	return tctx->dev->enable_clk(tctx->dev);
+	return 0;
 }
 
 static void rk_cra_hash_exit(struct crypto_tfm *tfm)
@@ -309,7 +309,6 @@ static void rk_cra_hash_exit(struct crypto_tfm *tfm)
 	struct rk_ahash_ctx *tctx = crypto_tfm_ctx(tfm);
 
 	free_page((unsigned long)tctx->dev->addr_vir);
-	return tctx->dev->disable_clk(tctx->dev);
 }
 
 struct rk_crypto_tmp rk_ahash_sha1 = {
drivers/crypto/rockchip/rk3288_crypto_skcipher.c +3 -2
index 5bbf0d2722e1..8c44a19eab75 100644
@@ -388,8 +388,10 @@ static int rk_ablk_init_tfm(struct crypto_skcipher *tfm)
 	ctx->dev->update = rk_ablk_rx;
 	ctx->dev->complete = rk_crypto_complete;
 	ctx->dev->addr_vir = (char *)__get_free_page(GFP_KERNEL);
+	if (!ctx->dev->addr_vir)
+		return -ENOMEM;
 
-	return ctx->dev->addr_vir ? ctx->dev->enable_clk(ctx->dev) : -ENOMEM;
+	return 0;
 }
 
 static void rk_ablk_exit_tfm(struct crypto_skcipher *tfm)
@@ -397,7 +399,6 @@ static void rk_ablk_exit_tfm(struct crypto_skcipher *tfm)
 	struct rk_cipher_ctx *ctx = crypto_skcipher_ctx(tfm);
 
 	free_page((unsigned long)ctx->dev->addr_vir);
-	ctx->dev->disable_clk(ctx->dev);
 }
 
 struct rk_crypto_tmp rk_ecb_aes_alg = {
Signature · 1 line
2.35.1
author maintainer
#374 lore raw

PATCH 371/731 crypto: rockchip - do not store mode globally

In reply to Greg Kroah-Hartman

From: Corentin Labbe <[email protected]>

[ Upstream commit 87e356c4966444866186f68f05832fdcc0f351a3 ]

Storing the mode globally does not work if 2 requests are handled in the same time.
We should store it in a request context.

Fixes ce0183cb6464b ("crypto: rockchip - switch to skcipher API")
---
 drivers/crypto/rockchip/rk3288_crypto.h       |  5 +-
 .../crypto/rockchip/rk3288_crypto_skcipher.c  | 58 ++++++++++++-------
 2 files changed, 41 insertions(+), 22 deletions(-)
drivers/crypto/rockchip/rk3288_crypto.h +4 -1
index 2fa7131e4060..3e60e3dca1b5 100644
@@ -245,10 +245,13 @@ struct rk_ahash_rctx {
 struct rk_cipher_ctx {
 	struct rk_crypto_info		*dev;
 	unsigned int			keylen;
-	u32				mode;
 	u8				iv[AES_BLOCK_SIZE];
 };
 
+struct rk_cipher_rctx {
+	u32				mode;
+};
+
 enum alg_type {
 	ALG_TYPE_HASH,
 	ALG_TYPE_CIPHER,
drivers/crypto/rockchip/rk3288_crypto_skcipher.c +37 -21

208 lines not on this page · raw message

Signature · 1 line
2.35.1
author maintainer
#375 lore raw

PATCH 372/731 crypto: rockchip - add fallback for cipher

In reply to Greg Kroah-Hartman

From: Corentin Labbe <[email protected]>

[ Upstream commit 68ef8af09a1a912a5ed2cfaa4cca7606f52cef90 ]

The hardware does not handle 0 size length request, let's add a fallback.
Furthermore fallback will be used for all unaligned case the hardware cannot handle.

Fixes ce0183cb6464b ("crypto: rockchip - switch to skcipher API")
---
 drivers/crypto/Kconfig                        |  4 +
 drivers/crypto/rockchip/rk3288_crypto.h       |  2 +
 .../crypto/rockchip/rk3288_crypto_skcipher.c  | 97 ++++++++++++++++---
 3 files changed, 90 insertions(+), 13 deletions(-)
drivers/crypto/Kconfig +4 -0
index 51690e73153a..9432375988fe 100644
@@ -772,6 +772,10 @@ config CRYPTO_DEV_IMGTEC_HASH
 config CRYPTO_DEV_ROCKCHIP
 	tristate "Rockchip's Cryptographic Engine driver"
 	depends on OF && ARCH_ROCKCHIP
+	depends on PM
+	select CRYPTO_ECB
+	select CRYPTO_CBC
+	select CRYPTO_DES
 	select CRYPTO_AES
 	select CRYPTO_LIB_DES
 	select CRYPTO_MD5
drivers/crypto/rockchip/rk3288_crypto.h +2 -0
index 3e60e3dca1b5..dfff0e2a83e4 100644
@@ -246,10 +246,12 @@ struct rk_cipher_ctx {
 	struct rk_crypto_info		*dev;
 	unsigned int			keylen;
 	u8				iv[AES_BLOCK_SIZE];
+	struct crypto_skcipher *fallback_tfm;
 };
 
 struct rk_cipher_rctx {
 	u32				mode;
+	struct skcipher_request fallback_req;   // keep at the end
 };
 
 enum alg_type {
drivers/crypto/rockchip/rk3288_crypto_skcipher.c +84 -13

181 lines not on this page · raw message

Signature · 1 line
2.35.1
author maintainer
#376 lore raw

PATCH 373/731 crypto: rockchip - add fallback for ahash

In reply to Greg Kroah-Hartman

From: Corentin Labbe <[email protected]>

[ Upstream commit 816600485cb597b3ff7d6806a95a78512839f775 ]

Adds a fallback for all case hardware cannot handle.

Fixes ce0183cb6464b ("crypto: rockchip - switch to skcipher API")
---
 drivers/crypto/rockchip/rk3288_crypto_ahash.c | 38 +++++++++++++++++++
 1 file changed, 38 insertions(+)
drivers/crypto/rockchip/rk3288_crypto_ahash.c +38 -0
index 49017d1fb510..16009bb0bf16 100644
@@ -16,6 +16,40 @@
  * so we put the fixed hash out when met zero message.
  */
 
+static bool rk_ahash_need_fallback(struct ahash_request *req)
+{
+	struct scatterlist *sg;
+
+	sg = req->src;
+	while (sg) {
+		if (!IS_ALIGNED(sg->offset, sizeof(u32))) {
+			return true;
+		}
+		if (sg->length % 4) {
+			return true;
+		}
+		sg = sg_next(sg);
+	}
+	return false;
+}
+
+static int rk_ahash_digest_fb(struct ahash_request *areq)
+{
+	struct rk_ahash_rctx *rctx = ahash_request_ctx(areq);
+	struct crypto_ahash *tfm = crypto_ahash_reqtfm(areq);
+	struct rk_ahash_ctx *tfmctx = crypto_ahash_ctx(tfm);
+
+	ahash_request_set_tfm(&rctx->fallback_req, tfmctx->fallback_tfm);
+	rctx->fallback_req.base.flags = areq->base.flags &
+					CRYPTO_TFM_REQ_MAY_SLEEP;
+
+	rctx->fallback_req.nbytes = areq->nbytes;
+	rctx->fallback_req.src = areq->src;
+	rctx->fallback_req.result = areq->result;
+
+	return crypto_ahash_digest(&rctx->fallback_req);
+}
+
 static int zero_message_process(struct ahash_request *req)
 {
 	struct crypto_ahash *tfm = crypto_ahash_reqtfm(req);
@@ -167,6 +201,9 @@ static int rk_ahash_digest(struct ahash_request *req)
 	struct rk_ahash_ctx *tctx = crypto_tfm_ctx(req->base.tfm);
 	struct rk_crypto_info *dev = tctx->dev;
 
+	if (rk_ahash_need_fallback(req))
+		return rk_ahash_digest_fb(req);
+
 	if (!req->nbytes)
 		return zero_message_process(req);
 	else
@@ -309,6 +346,7 @@ static void rk_cra_hash_exit(struct crypto_tfm *tfm)
 	struct rk_ahash_ctx *tctx = crypto_tfm_ctx(tfm);
 
 	free_page((unsigned long)tctx->dev->addr_vir);
+	crypto_free_ahash(tctx->fallback_tfm);
 }
 
 struct rk_crypto_tmp rk_ahash_sha1 = {
Signature · 1 line
2.35.1
author maintainer
#377 lore raw

PATCH 374/731 crypto: rockchip - better handle cipher key

In reply to Greg Kroah-Hartman

From: Corentin Labbe <[email protected]>

[ Upstream commit d6b23ccef82816050c2fd458c9dabfa0e0af09b9 ]

The key should not be set in hardware too much in advance, this will fail it 2 TFM with different keys generate alternative requests. The key should be stored and used just before doing cipher operations.

Fixes ce0183cb6464b ("crypto: rockchip - switch to skcipher API")
---
 drivers/crypto/rockchip/rk3288_crypto.h          |  1 +
 drivers/crypto/rockchip/rk3288_crypto_skcipher.c | 10 +++++++---
 2 files changed, 8 insertions(+), 3 deletions(-)
drivers/crypto/rockchip/rk3288_crypto.h +1 -0
index dfff0e2a83e4..665cc0bb2264 100644
@@ -245,6 +245,7 @@ struct rk_ahash_rctx {
 struct rk_cipher_ctx {
 	struct rk_crypto_info		*dev;
 	unsigned int			keylen;
+	u8				key[AES_MAX_KEY_SIZE];
 	u8				iv[AES_BLOCK_SIZE];
 	struct crypto_skcipher *fallback_tfm;
 };
drivers/crypto/rockchip/rk3288_crypto_skcipher.c +7 -3
index eac5bba66e25..1ef94f8db2c5 100644
@@ -95,7 +95,7 @@ static int rk_aes_setkey(struct crypto_skcipher *cipher,
 	    keylen != AES_KEYSIZE_256)
 		return -EINVAL;
 	ctx->keylen = keylen;
-	memcpy_toio(ctx->dev->reg + RK_CRYPTO_AES_KEY_0, key, keylen);
+	memcpy(ctx->key, key, keylen);
 
 	return crypto_skcipher_setkey(ctx->fallback_tfm, key, keylen);
 }
@@ -111,7 +111,7 @@ static int rk_des_setkey(struct crypto_skcipher *cipher,
 		return err;
 
 	ctx->keylen = keylen;
-	memcpy_toio(ctx->dev->reg + RK_CRYPTO_TDES_KEY1_0, key, keylen);
+	memcpy(ctx->key, key, keylen);
 
 	return crypto_skcipher_setkey(ctx->fallback_tfm, key, keylen);
 }
@@ -127,7 +127,8 @@ static int rk_tdes_setkey(struct crypto_skcipher *cipher,
 		return err;
 
 	ctx->keylen = keylen;
-	memcpy_toio(ctx->dev->reg + RK_CRYPTO_TDES_KEY1_0, key, keylen);
+	memcpy(ctx->key, key, keylen);
+
 	return crypto_skcipher_setkey(ctx->fallback_tfm, key, keylen);
 }
 
@@ -283,6 +284,7 @@ static void rk_ablk_hw_init(struct rk_crypto_info *dev)
 			     RK_CRYPTO_TDES_BYTESWAP_IV;
 		CRYPTO_WRITE(dev, RK_CRYPTO_TDES_CTRL, rctx->mode);
 		memcpy_toio(dev->reg + RK_CRYPTO_TDES_IV_0, req->iv, ivsize);
+		memcpy_toio(ctx->dev->reg + RK_CRYPTO_TDES_KEY1_0, ctx->key, ctx->keylen);
 		conf_reg = RK_CRYPTO_DESSEL;
 	} else {
 		rctx->mode |= RK_CRYPTO_AES_FIFO_MODE |
@@ -295,6 +297,7 @@ static void rk_ablk_hw_init(struct rk_crypto_info *dev)
 			rctx->mode |= RK_CRYPTO_AES_256BIT_key;
 		CRYPTO_WRITE(dev, RK_CRYPTO_AES_CTRL, rctx->mode);
 		memcpy_toio(dev->reg + RK_CRYPTO_AES_IV_0, req->iv, ivsize);
+		memcpy_toio(ctx->dev->reg + RK_CRYPTO_AES_KEY_0, ctx->key, ctx->keylen);
 	}
 	conf_reg |= RK_CRYPTO_BYTESWAP_BTFIFO |
 		    RK_CRYPTO_BYTESWAP_BRFIFO;
@@ -484,6 +487,7 @@ static void rk_ablk_exit_tfm(struct crypto_skcipher *tfm)
 {
 	struct rk_cipher_ctx *ctx = crypto_skcipher_ctx(tfm);
 
+	memzero_explicit(ctx->key, ctx->keylen);
 	free_page((unsigned long)ctx->dev->addr_vir);
 	crypto_free_skcipher(ctx->fallback_tfm);
 }
Signature · 1 line
2.35.1
author maintainer
#378 lore raw

PATCH 375/731 crypto: rockchip - remove non-aligned handling

In reply to Greg Kroah-Hartman

From: Corentin Labbe <[email protected]>

[ Upstream commit bb3c7b73363c9a149b12b74c44ae94b73a8fddf8 ]

Now driver have fallback for un-aligned cases, remove all code handling those cases.

Fixes ce0183cb6464b ("crypto: rockchip - switch to skcipher API")
---
 drivers/crypto/rockchip/rk3288_crypto.c       | 69 +++++--------------
 drivers/crypto/rockchip/rk3288_crypto.h       |  4 --
 drivers/crypto/rockchip/rk3288_crypto_ahash.c | 22 ++----
 .../crypto/rockchip/rk3288_crypto_skcipher.c  | 39 +++--------
 4 files changed, 31 insertions(+), 103 deletions(-)
drivers/crypto/rockchip/rk3288_crypto.c +16 -53
index 5f8444b9633a..31453257ab11 100644
@@ -88,63 +88,26 @@ static int rk_load_data(struct rk_crypto_info *dev,
 {
 	unsigned int count;
 
-	dev->aligned = dev->aligned ?
-		check_alignment(sg_src, sg_dst, dev->align_size) :
-		dev->aligned;
-	if (dev->aligned) {
-		count = min(dev->left_bytes, sg_src->length);
-		dev->left_bytes -= count;
-
-		if (!dma_map_sg(dev->dev, sg_src, 1, DMA_TO_DEVICE)) {
-			dev_err(dev->dev, "[%s:%d] dma_map_sg(src)  error\n",
+	count = min(dev->left_bytes, sg_src->length);
+	dev->left_bytes -= count;
+
+	if (!dma_map_sg(dev->dev, sg_src, 1, DMA_TO_DEVICE)) {
+		dev_err(dev->dev, "[%s:%d] dma_map_sg(src)  error\n",
 				__func__, __LINE__);
-			return -EINVAL;
-		}
-		dev->addr_in = sg_dma_address(sg_src);
+		return -EINVAL;
+	}
+	dev->addr_in = sg_dma_address(sg_src);
 
-		if (sg_dst) {
-			if (!dma_map_sg(dev->dev, sg_dst, 1, DMA_FROM_DEVICE)) {
-				dev_err(dev->dev,
+	if (sg_dst) {
+		if (!dma_map_sg(dev->dev, sg_dst, 1, DMA_FROM_DEVICE)) {
+			dev_err(dev->dev,
 					"[%s:%d] dma_map_sg(dst)  error\n",
 					__func__, __LINE__);
-				dma_unmap_sg(dev->dev, sg_src, 1,
-					     DMA_TO_DEVICE);
-				return -EINVAL;
-			}
-			dev->addr_out = sg_dma_address(sg_dst);
-		}
-	} else {
-		count = (dev->left_bytes > PAGE_SIZE) ?
-			PAGE_SIZE : dev->left_bytes;
-
-		if (!sg_pcopy_to_buffer(dev->first, dev->src_nents,
-					dev->addr_vir, count,
-					dev->total - dev->left_bytes)) {
-			dev_err(dev->dev, "[%s:%d] pcopy err\n",
-				__func__, __LINE__);
+			dma_unmap_sg(dev->dev, sg_src, 1,
+					DMA_TO_DEVICE);
 			return -EINVAL;
 		}
-		dev->left_bytes -= count;
-		sg_init_one(&dev->sg_tmp, dev->addr_vir, count);
-		if (!dma_map_sg(dev->dev, &dev->sg_tmp, 1, DMA_TO_DEVICE)) {
-			dev_err(dev->dev, "[%s:%d] dma_map_sg(sg_tmp)  error\n",
-				__func__, __LINE__);
-			return -ENOMEM;
-		}
-		dev->addr_in = sg_dma_address(&dev->sg_tmp);
-
-		if (sg_dst) {
-			if (!dma_map_sg(dev->dev, &dev->sg_tmp, 1,
-					DMA_FROM_DEVICE)) {
-				dev_err(dev->dev,
-					"[%s:%d] dma_map_sg(sg_tmp)  error\n",
-					__func__, __LINE__);
-				dma_unmap_sg(dev->dev, &dev->sg_tmp, 1,
-					     DMA_TO_DEVICE);
-				return -ENOMEM;
-			}
-			dev->addr_out = sg_dma_address(&dev->sg_tmp);
-		}
+		dev->addr_out = sg_dma_address(sg_dst);
 	}
 	dev->count = count;
 	return 0;
@@ -154,11 +117,11 @@ static void rk_unload_data(struct rk_crypto_info *dev)
 {
 	struct scatterlist *sg_in, *sg_out;
 
-	sg_in = dev->aligned ? dev->sg_src : &dev->sg_tmp;
+	sg_in = dev->sg_src;
 	dma_unmap_sg(dev->dev, sg_in, 1, DMA_TO_DEVICE);
 
 	if (dev->sg_dst) {
-		sg_out = dev->aligned ? dev->sg_dst : &dev->sg_tmp;
+		sg_out = dev->sg_dst;
 		dma_unmap_sg(dev->dev, sg_out, 1, DMA_FROM_DEVICE);
 	}
 }
drivers/crypto/rockchip/rk3288_crypto.h +0 -4
index 665cc0bb2264..df4db59fa13e 100644
@@ -204,12 +204,8 @@ struct rk_crypto_info {
 	/* the public variable */
 	struct scatterlist		*sg_src;
 	struct scatterlist		*sg_dst;
-	struct scatterlist		sg_tmp;
 	struct scatterlist		*first;
 	unsigned int			left_bytes;
-	void				*addr_vir;
-	int				aligned;
-	int				align_size;
 	size_t				src_nents;
 	size_t				dst_nents;
 	unsigned int			total;
drivers/crypto/rockchip/rk3288_crypto_ahash.c +6 -16
index 16009bb0bf16..c762e462eb57 100644
@@ -236,8 +236,6 @@ static int rk_ahash_start(struct rk_crypto_info *dev)
 
 	dev->total = req->nbytes;
 	dev->left_bytes = req->nbytes;
-	dev->aligned = 0;
-	dev->align_size = 4;
 	dev->sg_dst = NULL;
 	dev->sg_src = req->src;
 	dev->first = req->src;
@@ -272,15 +270,13 @@ static int rk_ahash_crypto_rx(struct rk_crypto_info *dev)
 
 	dev->unload_data(dev);
 	if (dev->left_bytes) {
-		if (dev->aligned) {
-			if (sg_is_last(dev->sg_src)) {
-				dev_warn(dev->dev, "[%s:%d], Lack of data\n",
-					 __func__, __LINE__);
-				err = -ENOMEM;
-				goto out_rx;
-			}
-			dev->sg_src = sg_next(dev->sg_src);
+		if (sg_is_last(dev->sg_src)) {
+			dev_warn(dev->dev, "[%s:%d], Lack of data\n",
+					__func__, __LINE__);
+			err = -ENOMEM;
+			goto out_rx;
 		}
+		dev->sg_src = sg_next(dev->sg_src);
 		err = rk_ahash_set_data_start(dev);
 	} else {
 		/*
@@ -318,11 +314,6 @@ static int rk_cra_hash_init(struct crypto_tfm *tfm)
 	algt = container_of(alg, struct rk_crypto_tmp, alg.hash);
 
 	tctx->dev = algt->dev;
-	tctx->dev->addr_vir = (void *)__get_free_page(GFP_KERNEL);
-	if (!tctx->dev->addr_vir) {
-		dev_err(tctx->dev->dev, "failed to kmalloc for addr_vir\n");
-		return -ENOMEM;
-	}
 	tctx->dev->start = rk_ahash_start;
 	tctx->dev->update = rk_ahash_crypto_rx;
 	tctx->dev->complete = rk_ahash_crypto_complete;
@@ -345,7 +336,6 @@ static void rk_cra_hash_exit(struct crypto_tfm *tfm)
 {
 	struct rk_ahash_ctx *tctx = crypto_tfm_ctx(tfm);
 
-	free_page((unsigned long)tctx->dev->addr_vir);
 	crypto_free_ahash(tctx->fallback_tfm);
 }
 
drivers/crypto/rockchip/rk3288_crypto_skcipher.c +9 -30
index 1ef94f8db2c5..d067b7f09165 100644
@@ -356,7 +356,6 @@ static int rk_ablk_start(struct rk_crypto_info *dev)
 	dev->src_nents = sg_nents(req->src);
 	dev->sg_dst = req->dst;
 	dev->dst_nents = sg_nents(req->dst);
-	dev->aligned = 1;
 
 	spin_lock_irqsave(&dev->lock, flags);
 	rk_ablk_hw_init(dev);
@@ -376,13 +375,9 @@ static void rk_iv_copyback(struct rk_crypto_info *dev)
 
 	/* Update the IV buffer to contain the next IV for encryption mode. */
 	if (!(rctx->mode & RK_CRYPTO_DEC)) {
-		if (dev->aligned) {
-			memcpy(req->iv, sg_virt(dev->sg_dst) +
-				dev->sg_dst->length - ivsize, ivsize);
-		} else {
-			memcpy(req->iv, dev->addr_vir +
-				dev->count - ivsize, ivsize);
-		}
+		memcpy(req->iv,
+		       sg_virt(dev->sg_dst) + dev->sg_dst->length - ivsize,
+		       ivsize);
 	}
 }
 
@@ -420,27 +415,16 @@ static int rk_ablk_rx(struct rk_crypto_info *dev)
 		skcipher_request_cast(dev->async_req);
 
 	dev->unload_data(dev);
-	if (!dev->aligned) {
-		if (!sg_pcopy_from_buffer(req->dst, dev->dst_nents,
-					  dev->addr_vir, dev->count,
-					  dev->total - dev->left_bytes -
-					  dev->count)) {
-			err = -EINVAL;
-			goto out_rx;
-		}
-	}
 	if (dev->left_bytes) {
 		rk_update_iv(dev);
-		if (dev->aligned) {
-			if (sg_is_last(dev->sg_src)) {
-				dev_err(dev->dev, "[%s:%d] Lack of data\n",
+		if (sg_is_last(dev->sg_src)) {
+			dev_err(dev->dev, "[%s:%d] Lack of data\n",
 					__func__, __LINE__);
-				err = -ENOMEM;
-				goto out_rx;
-			}
-			dev->sg_src = sg_next(dev->sg_src);
-			dev->sg_dst = sg_next(dev->sg_dst);
+			err = -ENOMEM;
+			goto out_rx;
 		}
+		dev->sg_src = sg_next(dev->sg_src);
+		dev->sg_dst = sg_next(dev->sg_dst);
 		err = rk_set_data_start(dev);
 	} else {
 		rk_iv_copyback(dev);
@@ -462,13 +446,9 @@ static int rk_ablk_init_tfm(struct crypto_skcipher *tfm)
 	algt = container_of(alg, struct rk_crypto_tmp, alg.skcipher);
 
 	ctx->dev = algt->dev;
-	ctx->dev->align_size = crypto_tfm_alg_alignmask(crypto_skcipher_tfm(tfm)) + 1;
 	ctx->dev->start = rk_ablk_start;
 	ctx->dev->update = rk_ablk_rx;
 	ctx->dev->complete = rk_crypto_complete;
-	ctx->dev->addr_vir = (char *)__get_free_page(GFP_KERNEL);
-	if (!ctx->dev->addr_vir)
-		return -ENOMEM;
 
 	ctx->fallback_tfm = crypto_alloc_skcipher(name, 0, CRYPTO_ALG_NEED_FALLBACK);
 	if (IS_ERR(ctx->fallback_tfm)) {
@@ -488,7 +468,6 @@ static void rk_ablk_exit_tfm(struct crypto_skcipher *tfm)
 	struct rk_cipher_ctx *ctx = crypto_skcipher_ctx(tfm);
 
 	memzero_explicit(ctx->key, ctx->keylen);
-	free_page((unsigned long)ctx->dev->addr_vir);
 	crypto_free_skcipher(ctx->fallback_tfm);
 }
 
Signature · 1 line
2.35.1
author maintainer
#379 lore raw

PATCH 376/731 crypto: rockchip - rework by using crypto_engine

In reply to Greg Kroah-Hartman

From: Corentin Labbe <[email protected]>

[ Upstream commit 57d67c6e8219b2a034c16d6149e30fb40fd39935 ]

Instead of doing manual queue management, let's use the crypto/engine for that.
In the same time, rework the requests handling to be easier to understand (and fix all bugs related to them).

Fixes ce0183cb6464b ("crypto: rockchip - switch to skcipher API")
---
 drivers/crypto/Kconfig                        |   1 +
 drivers/crypto/rockchip/rk3288_crypto.c       | 152 +----------
 drivers/crypto/rockchip/rk3288_crypto.h       |  39 +--
 drivers/crypto/rockchip/rk3288_crypto_ahash.c | 144 +++++-----
 .../crypto/rockchip/rk3288_crypto_skcipher.c  | 250 +++++++++---------
 5 files changed, 221 insertions(+), 365 deletions(-)
drivers/crypto/Kconfig +1 -0
index 9432375988fe..b46343b5c26b 100644
@@ -777,6 +777,7 @@ config CRYPTO_DEV_ROCKCHIP
 	select CRYPTO_CBC
 	select CRYPTO_DES
 	select CRYPTO_AES
+	select CRYPTO_ENGINE
 	select CRYPTO_LIB_DES
 	select CRYPTO_MD5
 	select CRYPTO_SHA1
drivers/crypto/rockchip/rk3288_crypto.c +9 -143

200 lines not on this page · raw message

drivers/crypto/rockchip/rk3288_crypto.h +10 -29
index df4db59fa13e..045e811b4af8 100644
@@ -5,9 +5,11 @@
 #include <crypto/aes.h>
 #include <crypto/internal/des.h>
 #include <crypto/algapi.h>
+#include <linux/dma-mapping.h>
 #include <linux/interrupt.h>
 #include <linux/delay.h>
 #include <linux/scatterlist.h>
+#include <crypto/engine.h>
 #include <crypto/internal/hash.h>
 #include <crypto/internal/skcipher.h>
 
@@ -193,39 +195,15 @@ struct rk_crypto_info {
 	struct reset_control		*rst;
 	void __iomem			*reg;
 	int				irq;
-	struct crypto_queue		queue;
-	struct tasklet_struct		queue_task;
-	struct tasklet_struct		done_task;
-	struct crypto_async_request	*async_req;
-	int 				err;
-	/* device lock */
-	spinlock_t			lock;
-
-	/* the public variable */
-	struct scatterlist		*sg_src;
-	struct scatterlist		*sg_dst;
-	struct scatterlist		*first;
-	unsigned int			left_bytes;
-	size_t				src_nents;
-	size_t				dst_nents;
-	unsigned int			total;
-	unsigned int			count;
-	dma_addr_t			addr_in;
-	dma_addr_t			addr_out;
-	bool				busy;
-	int (*start)(struct rk_crypto_info *dev);
-	int (*update)(struct rk_crypto_info *dev);
-	void (*complete)(struct crypto_async_request *base, int err);
-	int (*load_data)(struct rk_crypto_info *dev,
-			 struct scatterlist *sg_src,
-			 struct scatterlist *sg_dst);
-	void (*unload_data)(struct rk_crypto_info *dev);
-	int (*enqueue)(struct rk_crypto_info *dev,
-		       struct crypto_async_request *async_req);
+
+	struct crypto_engine *engine;
+	struct completion complete;
+	int status;
 };
 
 /* the private variable of hash */
 struct rk_ahash_ctx {
+	struct crypto_engine_ctx enginectx;
 	struct rk_crypto_info		*dev;
 	/* for fallback */
 	struct crypto_ahash		*fallback_tfm;
@@ -235,10 +213,12 @@ struct rk_ahash_ctx {
 struct rk_ahash_rctx {
 	struct ahash_request		fallback_req;
 	u32				mode;
+	int nrsg;
 };
 
 /* the private variable of cipher */
 struct rk_cipher_ctx {
+	struct crypto_engine_ctx enginectx;
 	struct rk_crypto_info		*dev;
 	unsigned int			keylen;
 	u8				key[AES_MAX_KEY_SIZE];
@@ -247,6 +227,7 @@ struct rk_cipher_ctx {
 };
 
 struct rk_cipher_rctx {
+	u8 backup_iv[AES_BLOCK_SIZE];
 	u32				mode;
 	struct skcipher_request fallback_req;   // keep at the end
 };
drivers/crypto/rockchip/rk3288_crypto_ahash.c +78 -66

218 lines not on this page · raw message

drivers/crypto/rockchip/rk3288_crypto_skcipher.c +123 -127

328 lines not on this page · raw message

Signature · 1 line
2.35.1
author maintainer
#380 lore raw

PATCH 377/731 apparmor: Fix memleak in alloc_ns()

In reply to Greg Kroah-Hartman

From: Xiu Jianfeng <[email protected]>

[ Upstream commit e9e6fa49dbab6d84c676666f3fe7d360497fd65b ]

After changes in commit a1bd627b46d1 ("apparmor: share profile name on replacement"), the hname member of struct aa_policy is not valid slab object, but a subset of that, it can not be freed by kfree_sensitive(), use aa_policy_destroy() to fix it.

Fixes a1bd627b46d1 ("apparmor: share profile name on replacement")
---
 security/apparmor/policy_ns.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
security/apparmor/policy_ns.c +1 -1
index 70921d95fb40..53d24cf63893 100644
@@ -121,7 +121,7 @@ static struct aa_ns *alloc_ns(const char *prefix, const char *name)
 	return ns;
 
 fail_unconfined:
-	kfree_sensitive(ns->base.hname);
+	aa_policy_destroy(&ns->base);
 fail_ns:
 	kfree_sensitive(ns);
 	return NULL;
Signature · 1 line
2.35.1
author maintainer
#381 lore raw

PATCH 378/731 f2fs: fix to invalidate dcc->f2fs_issue_discard in error path

In reply to Greg Kroah-Hartman

From: Chao Yu <[email protected]>

[ Upstream commit 91586ce0d39a05f88795aa8814fb99b1387236b3 ]

Syzbot reports a NULL pointer dereference issue as below:

 __refcount_add include/linux/refcount.h:193 [inline]
 __refcount_inc include/linux/refcount.h:250 [inline]
 refcount_inc include/linux/refcount.h:267 [inline]
 get_task_struct include/linux/sched/task.h:110 [inline]
 kthread_stop+0x34/0x1c0 kernel/kthread.c:703
 f2fs_stop_discard_thread+0x3c/0x5c fs/f2fs/segment.c:1638
 kill_f2fs_super+0x5c/0x194 fs/f2fs/super.c:4522
 deactivate_locked_super+0x70/0xe8 fs/super.c:332
 deactivate_super+0xd0/0xd4 fs/super.c:363
 cleanup_mnt+0x1f8/0x234 fs/namespace.c:1186
 __cleanup_mnt+0x20/0x30 fs/namespace.c:1193
 task_work_run+0xc4/0x14c kernel/task_work.c:177
 exit_task_work include/linux/task_work.h:38 [inline]
 do_exit+0x26c/0xbe0 kernel/exit.c:795
 do_group_exit+0x60/0xe8 kernel/exit.c:925
 __do_sys_exit_group kernel/exit.c:936 [inline]
 __se_sys_exit_group kernel/exit.c:934 [inline]
 __wake_up_parent+0x0/0x40 kernel/exit.c:934
 __invoke_syscall arch/arm64/kernel/syscall.c:38 [inline]
 invoke_syscall arch/arm64/kernel/syscall.c:52 [inline]
 el0_svc_common+0x138/0x220 arch/arm64/kernel/syscall.c:142
 do_el0_svc+0x48/0x164 arch/arm64/kernel/syscall.c:206
 el0_svc+0x58/0x150 arch/arm64/kernel/entry-common.c:636
 el0t_64_sync_handler+0x84/0xf0 arch/arm64/kernel/entry-common.c:654
 el0t_64_sync+0x18c/0x190 arch/arm64/kernel/entry.S:581

The root cause of this issue is in error path of f2fs_start_discard_thread(), it missed to invalidate dcc->f2fs_issue_discard, later kthread_stop() may access invalid pointer.

Fixes 4d67490498ac ("f2fs: Don't create discard thread when device doesn't support realtime discard")
Reported-by [email protected]
Reported-by [email protected]
---
 fs/f2fs/segment.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
fs/f2fs/segment.c +3 -1
index af810b2d5d90..28939e3573ea 100644
@@ -2129,8 +2129,10 @@ int f2fs_start_discard_thread(struct f2fs_sb_info *sbi)
 
 	dcc->f2fs_issue_discard = kthread_run(issue_discard_thread, sbi,
 				"f2fs_discard-%u:%u", MAJOR(dev), MINOR(dev));
-	if (IS_ERR(dcc->f2fs_issue_discard))
+	if (IS_ERR(dcc->f2fs_issue_discard)) {
 		err = PTR_ERR(dcc->f2fs_issue_discard);
+		dcc->f2fs_issue_discard = NULL;
+	}
 
 	return err;
 }
Signature · 1 line
2.35.1
author maintainer
#382 lore raw

PATCH 379/731 f2fs: fix normal discard process

In reply to Greg Kroah-Hartman

From: Dongdong Zhang <[email protected]>

[ Upstream commit b5f1a218ae5e4339130d6e733f0e63d623e09a2c ]

In the DPOLICY_BG mode, there is a conflict between the two conditions "i + 1 < dpolicy->granularity" and "i < DEFAULT_DISCARD_GRANULARITY". If i = 15, the first condition is false, it will enter the second condition and dispatch all small granularity discards in function
__issue_discard_cmd_orderly. The restrictive effect
of the first condition to small discards will be invalidated. These two conditions should align.

Fixes 20ee4382322c ("f2fs: issue small discard by LBA order")
---
 fs/f2fs/segment.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/f2fs/segment.c +1 -1
index 28939e3573ea..194c0811fbdf 100644
@@ -1551,7 +1551,7 @@ static int __issue_discard_cmd(struct f2fs_sb_info *sbi,
 		if (i + 1 < dpolicy->granularity)
 			break;
 
-		if (i < DEFAULT_DISCARD_GRANULARITY && dpolicy->ordered)
+		if (i + 1 < DEFAULT_DISCARD_GRANULARITY && dpolicy->ordered)
 			return __issue_discard_cmd_orderly(sbi, dpolicy);
 
 		pend_list = &dcc->pend_list[i];
Signature · 1 line
2.35.1
author maintainer
#383 lore raw

PATCH 380/731 f2fs: fix to destroy sbi->post_read_wq in error path of f2fs_fill_super()

In reply to Greg Kroah-Hartman

From: Chao Yu <[email protected]>

[ Upstream commit 7b02b2201893a71b881026cf574902019ab00db5 ]

In error path of f2fs_fill_super(), this patch fixes to call f2fs_destroy_post_read_wq() once if we fail in f2fs_start_ckpt_thread().

Fixes 261eeb9c1585 ("f2fs: introduce checkpoint_merge mount option")
---
 fs/f2fs/super.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/f2fs/super.c +1 -1
index a0d1ef73b83e..f4e8de1f4789 100644
@@ -4428,9 +4428,9 @@ static int f2fs_fill_super(struct super_block *sb, void *data, int silent)
 	f2fs_destroy_node_manager(sbi);
 free_sm:
 	f2fs_destroy_segment_manager(sbi);
-	f2fs_destroy_post_read_wq(sbi);
 stop_ckpt_thread:
 	f2fs_stop_ckpt_thread(sbi);
+	f2fs_destroy_post_read_wq(sbi);
 free_devices:
 	destroy_device_list(sbi);
 	kvfree(sbi->ckpt);
Signature · 1 line
2.35.1
author maintainer
#384 lore raw

PATCH 381/731 RDMA/irdma: Report the correct link speed

In reply to Greg Kroah-Hartman

From: Shiraz Saleem <[email protected]>

[ Upstream commit 4eace75e0853273755b878ffa9cce6de84df975a ]

The active link speed is currently hard-coded in irdma_query_port due to which the port rate in ibstatus does reflect the active link speed.

Call ib_get_eth_speed in irdma_query_port to get the active link speed.

Fixes b48c24c2d710 ("RDMA/irdma: Implement device supported verb APIs")
---
 drivers/infiniband/hw/irdma/verbs.c | 35 +++--------------------------
 1 file changed, 3 insertions(+), 32 deletions(-)
drivers/infiniband/hw/irdma/verbs.c +3 -32
index 911902d2b93e..c5971a840b87 100644
@@ -60,36 +60,6 @@ static int irdma_query_device(struct ib_device *ibdev,
 	return 0;
 }
 
-/**
- * irdma_get_eth_speed_and_width - Get IB port speed and width from netdev speed
- * @link_speed: netdev phy link speed
- * @active_speed: IB port speed
- * @active_width: IB port width
- */
-static void irdma_get_eth_speed_and_width(u32 link_speed, u16 *active_speed,
-					  u8 *active_width)
-{
-	if (link_speed <= SPEED_1000) {
-		*active_width = IB_WIDTH_1X;
-		*active_speed = IB_SPEED_SDR;
-	} else if (link_speed <= SPEED_10000) {
-		*active_width = IB_WIDTH_1X;
-		*active_speed = IB_SPEED_FDR10;
-	} else if (link_speed <= SPEED_20000) {
-		*active_width = IB_WIDTH_4X;
-		*active_speed = IB_SPEED_DDR;
-	} else if (link_speed <= SPEED_25000) {
-		*active_width = IB_WIDTH_1X;
-		*active_speed = IB_SPEED_EDR;
-	} else if (link_speed <= SPEED_40000) {
-		*active_width = IB_WIDTH_4X;
-		*active_speed = IB_SPEED_FDR10;
-	} else {
-		*active_width = IB_WIDTH_4X;
-		*active_speed = IB_SPEED_EDR;
-	}
-}
-
 /**
  * irdma_query_port - get port attributes
  * @ibdev: device pointer from stack
@@ -117,8 +87,9 @@ static int irdma_query_port(struct ib_device *ibdev, u32 port,
 		props->state = IB_PORT_DOWN;
 		props->phys_state = IB_PORT_PHYS_STATE_DISABLED;
 	}
-	irdma_get_eth_speed_and_width(SPEED_100000, &props->active_speed,
-				      &props->active_width);
+
+	ib_get_eth_speed(ibdev, port, &props->active_speed,
+			 &props->active_width);
 
 	if (rdma_protocol_roce(ibdev, 1)) {
 		props->gid_tbl_len = 32;
Signature · 1 line
2.35.1
author maintainer
#385 lore raw

PATCH 382/731 scsi: qla2xxx: Fix set-but-not-used variable warnings

In reply to Greg Kroah-Hartman

From: Bart Van Assche <[email protected]>

[ Upstream commit 4fb2169d66b837a2986f569f5d5b81f79e6e4a4c ]

Fix the following two compiler warnings:

drivers/scsi/qla2xxx/qla_init.c: In function ‘qla24xx_async_abort_cmd’:
drivers/scsi/qla2xxx/qla_init.c:171:17: warning: variable ‘bail’ set but not used [-Wunused-but-set-variable]
171 | uint8_t bail;
| ^~~~
drivers/scsi/qla2xxx/qla_init.c: In function ‘qla2x00_async_tm_cmd’:
drivers/scsi/qla2xxx/qla_init.c:2023:17: warning: variable ‘bail’ set but not used [-Wunused-but-set-variable]
2023 | uint8_t bail;
| ^~~~

Cc Arun Easi <[email protected]>
Cc Giridhar Malavali <[email protected]>
Fixes feafb7b1714c ("[SCSI] qla2xxx: Fix vport delete issues")
---
 drivers/scsi/qla2xxx/qla_def.h    | 22 +++++++++++-----------
 drivers/scsi/qla2xxx/qla_init.c   |  6 ++----
 drivers/scsi/qla2xxx/qla_inline.h |  4 +---
 drivers/scsi/qla2xxx/qla_os.c     |  4 +---
 4 files changed, 15 insertions(+), 21 deletions(-)
drivers/scsi/qla2xxx/qla_def.h +11 -11
index 51c7ce5f9792..307ffdfe048b 100644
@@ -5117,17 +5117,17 @@ struct secure_flash_update_block_pk {
 		(test_bit(ISP_ABORT_NEEDED, &ha->dpc_flags) || \
 			 test_bit(LOOP_RESYNC_NEEDED, &ha->dpc_flags))
 
-#define QLA_VHA_MARK_BUSY(__vha, __bail) do {		\
-	atomic_inc(&__vha->vref_count);			\
-	mb();						\
-	if (__vha->flags.delete_progress) {		\
-		atomic_dec(&__vha->vref_count);		\
-		wake_up(&__vha->vref_waitq);		\
-		__bail = 1;				\
-	} else {					\
-		__bail = 0;				\
-	}						\
-} while (0)
+static inline bool qla_vha_mark_busy(scsi_qla_host_t *vha)
+{
+	atomic_inc(&vha->vref_count);
+	mb();
+	if (vha->flags.delete_progress) {
+		atomic_dec(&vha->vref_count);
+		wake_up(&vha->vref_waitq);
+		return true;
+	}
+	return false;
+}
 
 #define QLA_VHA_MARK_NOT_BUSY(__vha) do {		\
 	atomic_dec(&__vha->vref_count);			\
drivers/scsi/qla2xxx/qla_init.c +2 -4
index b81797a3ab61..46b3b31a41bd 100644
@@ -168,7 +168,6 @@ int qla24xx_async_abort_cmd(srb_t *cmd_sp, bool wait)
 	struct srb_iocb *abt_iocb;
 	srb_t *sp;
 	int rval = QLA_FUNCTION_FAILED;
-	uint8_t bail;
 
 	/* ref: INIT for ABTS command */
 	sp = qla2xxx_get_qpair_sp(cmd_sp->vha, cmd_sp->qpair, cmd_sp->fcport,
@@ -176,7 +175,7 @@ int qla24xx_async_abort_cmd(srb_t *cmd_sp, bool wait)
 	if (!sp)
 		return QLA_MEMORY_ALLOC_FAILED;
 
-	QLA_VHA_MARK_BUSY(vha, bail);
+	qla_vha_mark_busy(vha);
 	abt_iocb = &sp->u.iocb_cmd;
 	sp->type = SRB_ABT_CMD;
 	sp->name = "abort";
@@ -2022,14 +2021,13 @@ qla2x00_async_tm_cmd(fc_port_t *fcport, uint32_t flags, uint32_t lun,
 	struct srb_iocb *tm_iocb;
 	srb_t *sp;
 	int rval = QLA_FUNCTION_FAILED;
-	uint8_t bail;
 
 	/* ref: INIT */
 	sp = qla2x00_get_sp(vha, fcport, GFP_KERNEL);
 	if (!sp)
 		goto done;
 
-	QLA_VHA_MARK_BUSY(vha, bail);
+	qla_vha_mark_busy(vha);
 	sp->type = SRB_TM_CMD;
 	sp->name = "tmf";
 	qla2x00_init_async_sp(sp, qla2x00_get_async_timeout(vha),
drivers/scsi/qla2xxx/qla_inline.h +1 -3
index db17f7f410cd..5185dc5daf80 100644
@@ -225,11 +225,9 @@ static inline srb_t *
 qla2x00_get_sp(scsi_qla_host_t *vha, fc_port_t *fcport, gfp_t flag)
 {
 	srb_t *sp = NULL;
-	uint8_t bail;
 	struct qla_qpair *qpair;
 
-	QLA_VHA_MARK_BUSY(vha, bail);
-	if (unlikely(bail))
+	if (unlikely(qla_vha_mark_busy(vha)))
 		return NULL;
 
 	qpair = vha->hw->base_qpair;
drivers/scsi/qla2xxx/qla_os.c +1 -3
index 00e97f0a07eb..05d827227d0b 100644
@@ -5043,13 +5043,11 @@ struct qla_work_evt *
 qla2x00_alloc_work(struct scsi_qla_host *vha, enum qla_work_type type)
 {
 	struct qla_work_evt *e;
-	uint8_t bail;
 
 	if (test_bit(UNLOADING, &vha->dpc_flags))
 		return NULL;
 
-	QLA_VHA_MARK_BUSY(vha, bail);
-	if (bail)
+	if (qla_vha_mark_busy(vha))
 		return NULL;
 
 	e = kzalloc(sizeof(struct qla_work_evt), GFP_ATOMIC);
Signature · 1 line
2.35.1
author maintainer
#386 lore raw

PATCH 383/731 RDMA/siw: Fix immediate work request flush to completion queue

In reply to Greg Kroah-Hartman

From: Bernard Metzler <[email protected]>

[ Upstream commit bdf1da5df9da680589a7f74448dd0a94dd3e1446 ]

Correctly set send queue element opcode during immediate work request flushing in post sendqueue operation, if the QP is in ERROR state. An undefined ocode value results in out-of-bounds access to an array for mapping the opcode between siw internal and RDMA core representation in work completion generation. It resulted in a KASAN BUG report of type 'global-out-of-bounds' during NFSoRDMA testing.

This patch further fixes a potential case of a malicious user which may write undefined values for completion queue elements status or opcode, if the CQ is memory mapped to user land. It avoids the same out-of-bounds access to arrays for status and opcode mapping as described above.

Fixes 303ae1cdfdf7 ("rdma/siw: application interface")
Fixes b0fff7317bb4 ("rdma/siw: completion queue methods")
---
 drivers/infiniband/sw/siw/siw_cq.c    | 24 ++++++++++++++--
 drivers/infiniband/sw/siw/siw_verbs.c | 40 ++++++++++++++++++++++++---
 2 files changed, 58 insertions(+), 6 deletions(-)
drivers/infiniband/sw/siw/siw_cq.c +22 -2
index d68e37859e73..acc7bcd538b5 100644
@@ -56,8 +56,6 @@ int siw_reap_cqe(struct siw_cq *cq, struct ib_wc *wc)
 	if (READ_ONCE(cqe->flags) & SIW_WQE_VALID) {
 		memset(wc, 0, sizeof(*wc));
 		wc->wr_id = cqe->id;
-		wc->status = map_cqe_status[cqe->status].ib;
-		wc->opcode = map_wc_opcode[cqe->opcode];
 		wc->byte_len = cqe->bytes;
 
 		/*
@@ -71,10 +69,32 @@ int siw_reap_cqe(struct siw_cq *cq, struct ib_wc *wc)
 				wc->wc_flags = IB_WC_WITH_INVALIDATE;
 			}
 			wc->qp = cqe->base_qp;
+			wc->opcode = map_wc_opcode[cqe->opcode];
+			wc->status = map_cqe_status[cqe->status].ib;
 			siw_dbg_cq(cq,
 				   "idx %u, type %d, flags %2x, id 0x%pK\n",
 				   cq->cq_get % cq->num_cqe, cqe->opcode,
 				   cqe->flags, (void *)(uintptr_t)cqe->id);
+		} else {
+			/*
+			 * A malicious user may set invalid opcode or
+			 * status in the user mmapped CQE array.
+			 * Sanity check and correct values in that case
+			 * to avoid out-of-bounds access to global arrays
+			 * for opcode and status mapping.
+			 */
+			u8 opcode = cqe->opcode;
+			u16 status = cqe->status;
+
+			if (opcode >= SIW_NUM_OPCODES) {
+				opcode = 0;
+				status = IB_WC_GENERAL_ERR;
+			} else if (status >= SIW_NUM_WC_STATUS) {
+				status = IB_WC_GENERAL_ERR;
+			}
+			wc->opcode = map_wc_opcode[opcode];
+			wc->status = map_cqe_status[status].ib;
+
 		}
 		WRITE_ONCE(cqe->flags, 0);
 		cq->cq_get++;
drivers/infiniband/sw/siw/siw_verbs.c +36 -4
index ff33659acffa..9c7fbda9e068 100644
@@ -674,13 +674,45 @@ static int siw_copy_inline_sgl(const struct ib_send_wr *core_wr,
 static int siw_sq_flush_wr(struct siw_qp *qp, const struct ib_send_wr *wr,
 			   const struct ib_send_wr **bad_wr)
 {
-	struct siw_sqe sqe = {};
 	int rv = 0;
 
 	while (wr) {
-		sqe.id = wr->wr_id;
-		sqe.opcode = wr->opcode;
-		rv = siw_sqe_complete(qp, &sqe, 0, SIW_WC_WR_FLUSH_ERR);
+		struct siw_sqe sqe = {};
+
+		switch (wr->opcode) {
+		case IB_WR_RDMA_WRITE:
+			sqe.opcode = SIW_OP_WRITE;
+			break;
+		case IB_WR_RDMA_READ:
+			sqe.opcode = SIW_OP_READ;
+			break;
+		case IB_WR_RDMA_READ_WITH_INV:
+			sqe.opcode = SIW_OP_READ_LOCAL_INV;
+			break;
+		case IB_WR_SEND:
+			sqe.opcode = SIW_OP_SEND;
+			break;
+		case IB_WR_SEND_WITH_IMM:
+			sqe.opcode = SIW_OP_SEND_WITH_IMM;
+			break;
+		case IB_WR_SEND_WITH_INV:
+			sqe.opcode = SIW_OP_SEND_REMOTE_INV;
+			break;
+		case IB_WR_LOCAL_INV:
+			sqe.opcode = SIW_OP_INVAL_STAG;
+			break;
+		case IB_WR_REG_MR:
+			sqe.opcode = SIW_OP_REG_MR;
+			break;
+		default:
+			rv = -EINVAL;
+			break;
+		}
+		if (!rv) {
+			sqe.id = wr->wr_id;
+			rv = siw_sqe_complete(qp, &sqe, 0,
+					      SIW_WC_WR_FLUSH_ERR);
+		}
 		if (rv) {
 			if (bad_wr)
 				*bad_wr = wr;
Signature · 1 line
2.35.1
author maintainer
#387 lore raw

PATCH 384/731 IB/mad: Dont call to function that might sleep while in atomic context

In reply to Greg Kroah-Hartman

From: Leonid Ravich <[email protected]>

[ Upstream commit 5c20311d76cbaeb7ed2ecf9c8b8322f8fc4a7ae3 ]

Tracepoints are not allowed to sleep, as such the following splat is generated due to call to ib_query_pkey() in atomic context.

WARNING: CPU: 0 PID: 1888000 at kernel/trace/ring_buffer.c:2492 rb_commit+0xc1/0x220
CPU: 0 PID: 1888000 Comm: kworker/u9:0 Kdump: loaded Tainted: G           OE    --------- -  - 4.18.0-305.3.1.el8.x86_64 #1
 Hardware name: Red Hat KVM, BIOS 1.13.0-2.module_el8.3.0+555+a55c8938 04/01/2014
 Workqueue: ib-comp-unb-wq ib_cq_poll_work [ib_core]
 RIP: 0010:rb_commit+0xc1/0x220
 RSP: 0000:ffffa8ac80f9bca0 EFLAGS: 00010202
 RAX: ffff8951c7c01300 RBX: ffff8951c7c14a00 RCX: 0000000000000246
 RDX: ffff8951c707c000 RSI: ffff8951c707c57c RDI: ffff8951c7c14a00
 RBP: 0000000000000000 R08: 0000000000000000 R09: 0000000000000000
 R10: ffff8951c7c01300 R11: 0000000000000001 R12: 0000000000000246
 R13: 0000000000000000 R14: ffffffff964c70c0 R15: 0000000000000000
 FS:  0000000000000000(0000) GS:ffff8951fbc00000(0000) knlGS:0000000000000000
 CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
 CR2: 00007f20e8f39010 CR3: 000000002ca10005 CR4: 0000000000170ef0
 Call Trace:
  ring_buffer_unlock_commit+0x1d/0xa0
  trace_buffer_unlock_commit_regs+0x3b/0x1b0
  trace_event_buffer_commit+0x67/0x1d0
  trace_event_raw_event_ib_mad_recv_done_handler+0x11c/0x160 [ib_core]
  ib_mad_recv_done+0x48b/0xc10 [ib_core]
  ? trace_event_raw_event_cq_poll+0x6f/0xb0 [ib_core]
  __ib_process_cq+0x91/0x1c0 [ib_core]
  ib_cq_poll_work+0x26/0x80 [ib_core]
  process_one_work+0x1a7/0x360
  ? create_worker+0x1a0/0x1a0
  worker_thread+0x30/0x390
  ? create_worker+0x1a0/0x1a0
  kthread+0x116/0x130
  ? kthread_flush_work_fn+0x10/0x10
  ret_from_fork+0x35/0x40
 ---[ end trace 78ba8509d3830a16 ]---
---
 drivers/infiniband/core/mad.c |  5 -----
 include/trace/events/ib_mad.h | 13 ++++---------
 2 files changed, 4 insertions(+), 14 deletions(-)
drivers/infiniband/core/mad.c +0 -5
index 1893aa613ad7..674344eb8e2f 100644
@@ -59,9 +59,6 @@ static void create_mad_addr_info(struct ib_mad_send_wr_private *mad_send_wr,
 			  struct ib_mad_qp_info *qp_info,
 			  struct trace_event_raw_ib_mad_send_template *entry)
 {
-	u16 pkey;
-	struct ib_device *dev = qp_info->port_priv->device;
-	u32 pnum = qp_info->port_priv->port_num;
 	struct ib_ud_wr *wr = &mad_send_wr->send_wr;
 	struct rdma_ah_attr attr = {};
 
@@ -69,8 +66,6 @@ static void create_mad_addr_info(struct ib_mad_send_wr_private *mad_send_wr,
 
 	/* These are common */
 	entry->sl = attr.sl;
-	ib_query_pkey(dev, pnum, wr->pkey_index, &pkey);
-	entry->pkey = pkey;
 	entry->rqpn = wr->remote_qpn;
 	entry->rqkey = wr->remote_qkey;
 	entry->dlid = rdma_ah_get_dlid(&attr);
include/trace/events/ib_mad.h +4 -9
index 59363a083ecb..d92691c78cff 100644
@@ -49,7 +49,6 @@ DECLARE_EVENT_CLASS(ib_mad_send_template,
 		__field(int,            retries_left)
 		__field(int,            max_retries)
 		__field(int,            retry)
-		__field(u16,            pkey)
 	),
 
 	TP_fast_assign(
@@ -89,7 +88,7 @@ DECLARE_EVENT_CLASS(ib_mad_send_template,
 		  "hdr : base_ver 0x%x class 0x%x class_ver 0x%x " \
 		  "method 0x%x status 0x%x class_specific 0x%x tid 0x%llx " \
 		  "attr_id 0x%x attr_mod 0x%x  => dlid 0x%08x sl %d "\
-		  "pkey 0x%x rpqn 0x%x rqpkey 0x%x",
+		  "rpqn 0x%x rqpkey 0x%x",
 		__entry->dev_index, __entry->port_num, __entry->qp_num,
 		__entry->agent_priv, be64_to_cpu(__entry->wrtid),
 		__entry->retries_left, __entry->max_retries,
@@ -100,7 +99,7 @@ DECLARE_EVENT_CLASS(ib_mad_send_template,
 		be16_to_cpu(__entry->class_specific),
 		be64_to_cpu(__entry->tid), be16_to_cpu(__entry->attr_id),
 		be32_to_cpu(__entry->attr_mod),
-		be32_to_cpu(__entry->dlid), __entry->sl, __entry->pkey,
+		be32_to_cpu(__entry->dlid), __entry->sl,
 		__entry->rqpn, __entry->rqkey
 	)
 );
@@ -204,7 +203,6 @@ TRACE_EVENT(ib_mad_recv_done_handler,
 		__field(u16,            wc_status)
 		__field(u32,            slid)
 		__field(u32,            dev_index)
-		__field(u16,            pkey)
 	),
 
 	TP_fast_assign(
@@ -224,9 +222,6 @@ TRACE_EVENT(ib_mad_recv_done_handler,
 		__entry->slid = wc->slid;
 		__entry->src_qp = wc->src_qp;
 		__entry->sl = wc->sl;
-		ib_query_pkey(qp_info->port_priv->device,
-			      qp_info->port_priv->port_num,
-			      wc->pkey_index, &__entry->pkey);
 		__entry->wc_status = wc->status;
 	),
 
@@ -234,7 +229,7 @@ TRACE_EVENT(ib_mad_recv_done_handler,
 		  "base_ver 0x%02x class 0x%02x class_ver 0x%02x " \
 		  "method 0x%02x status 0x%04x class_specific 0x%04x " \
 		  "tid 0x%016llx attr_id 0x%04x attr_mod 0x%08x " \
-		  "slid 0x%08x src QP%d, sl %d pkey 0x%04x",
+		  "slid 0x%08x src QP%d, sl %d",
 		__entry->dev_index, __entry->port_num, __entry->qp_num,
 		__entry->wc_status,
 		__entry->length,
@@ -244,7 +239,7 @@ TRACE_EVENT(ib_mad_recv_done_handler,
 		be16_to_cpu(__entry->class_specific),
 		be64_to_cpu(__entry->tid), be16_to_cpu(__entry->attr_id),
 		be32_to_cpu(__entry->attr_mod),
-		__entry->slid, __entry->src_qp, __entry->sl, __entry->pkey
+		__entry->slid, __entry->src_qp, __entry->sl
 	)
 );
 
Signature · 1 line
2.35.1
author maintainer
#388 lore raw

PATCH 385/731 PCI: vmd: Disable MSI remapping after suspend

In reply to Greg Kroah-Hartman

From: Nirmal Patel <[email protected]>

[ Upstream commit d899aa668498c07ff217b666ae9712990306e682 ]

MSI remapping is disabled by VMD driver for Intel's Icelake and newer systems in order to improve performance by setting VMCONFIG_MSI_REMAP. By design VMCONFIG_MSI_REMAP register is cleared by firmware during boot. The same register gets cleared when system is put in S3 power state. VMD driver needs to set this register again in order to avoid interrupt issues with devices behind VMD if MSI remapping was disabled before.

Fixes ee81ee84f873 ("PCI: vmd: Disable MSI-X remapping when possible")
---
 drivers/pci/controller/vmd.c | 5 +++++
 1 file changed, 5 insertions(+)
drivers/pci/controller/vmd.c +5 -0
index a5987e52700e..8dce71142e10 100644
@@ -900,6 +900,11 @@ static int vmd_resume(struct device *dev)
 	struct vmd_dev *vmd = pci_get_drvdata(pdev);
 	int err, i;
 
+       if (vmd->irq_domain)
+               vmd_set_msi_remapping(vmd, true);
+       else
+               vmd_set_msi_remapping(vmd, false);
+
 	for (i = 0; i < vmd->msix_count; i++) {
 		err = devm_request_irq(dev, pci_irq_vector(pdev, i),
 				       vmd_irq, IRQF_NO_THREAD,
Signature · 1 line
2.35.1
author maintainer
#389 lore raw

PATCH 386/731 RDMA/restrack: Release MR restrack when delete

In reply to Greg Kroah-Hartman

From: Mark Zhang <[email protected]>

[ Upstream commit dac153f2802db1ad46207283cb9b2aae3d707a45 ]

The MR restrack also needs to be released when delete it, otherwise it cause memory leak as the task struct won't be released.

---
 drivers/infiniband/core/restrack.c | 2 --
 1 file changed, 2 deletions(-)
drivers/infiniband/core/restrack.c +0 -2
index 1f935d9f6178..01a499a8b88d 100644
@@ -343,8 +343,6 @@ void rdma_restrack_del(struct rdma_restrack_entry *res)
 	rt = &dev->res[res->type];
 
 	old = xa_erase(&rt->xa, res->id);
-	if (res->type == RDMA_RESTRACK_MR)
-		return;
 	WARN_ON(old != res);
 
 out:
Signature · 1 line
2.35.1
author maintainer
#390 lore raw

PATCH 387/731 RDMA/core: Make sure "ib_port" is valid when access sysfs node

In reply to Greg Kroah-Hartman

From: Mark Zhang <[email protected]>

[ Upstream commit 5e15ff29b156bbbdeadae230c8ecd5ecd8ca2477 ]

The "ib_port" structure must be set before adding the sysfs kobject, and reset after removing it, otherwise it may crash when accessing the sysfs node:

  Unable to handle kernel NULL pointer dereference at virtual address 0000000000000050
  Mem abort info:
    ESR = 0x96000006
    Exception class = DABT (current EL), IL = 32 bits
    SET = 0, FnV = 0
    EA = 0, S1PTW = 0
  Data abort info:
    ISV = 0, ISS = 0x00000006
    CM = 0, WnR = 0
  user pgtable: 4k pages, 48-bit VAs, pgdp = 00000000e85f5ba5
  [0000000000000050] pgd=0000000848fd9003, pud=000000085b387003, pmd=0000000000000000
  Internal error: Oops: 96000006 [#2] PREEMPT SMP
  Modules linked in: ib_umad(O) mlx5_ib(O) nfnetlink_cttimeout(E) nfnetlink(E) act_gact(E) cls_flower(E) sch_ingress(E) openvswitch(E) nsh(E) nf_nat_ipv6(E) nf_nat_ipv4(E) nf_conncount(E) nf_nat(E) nf_conntrack(E) nf_defrag_ipv6(E) nf_defrag_ipv4(E) mst_pciconf(O) ipmi_devintf(E) ipmi_msghandler(E) ipmb_dev_int(OE) mlx5_core(O) mlxfw(O) mlxdevm(O) auxiliary(O) ib_uverbs(O) ib_core(O) mlx_compat(O) psample(E) sbsa_gwdt(E) uio_pdrv_genirq(E) uio(E) mlxbf_pmc(OE) mlxbf_gige(OE) mlxbf_tmfifo(OE) gpio_mlxbf2(OE) pwr_mlxbf(OE) mlx_trio(OE) i2c_mlxbf(OE) mlx_bootctl(OE) bluefield_edac(OE) knem(O) ip_tables(E) ipv6(E) crc_ccitt(E) [last unloaded: mst_pci]
  Process grep (pid: 3372, stack limit = 0x0000000022055c92)
  CPU: 5 PID: 3372 Comm: grep Tainted: G      D    OE     4.19.161-mlnx.47.gadcd9e3 #1
  Hardware name: https://www.mellanox.com BlueField SoC/BlueField SoC, BIOS BlueField:3.9.2-15-ga2403ab Sep  8 2022
  pstate: 40000005 (nZcv daif -PAN -UAO)
  pc : hw_stat_port_show+0x4c/0x80 [ib_core]
  lr : port_attr_show+0x40/0x58 [ib_core]
  sp : ffff000029f43b50
  x29: ffff000029f43b50 x28: 0000000019375000
  x27: ffff8007b821a540 x26: ffff000029f43e30
  x25: 0000000000008000 x24: ffff000000eaa958
  x23: 0000000000001000 x22: ffff8007a4ce3000
  x21: ffff8007baff8000 x20: ffff8007b9066ac0
  x19: ffff8007bae97578 x18: 0000000000000000
  x17: 0000000000000000 x16: 0000000000000000
  x15: 0000000000000000 x14: 0000000000000000
  x13: 0000000000000000 x12: 0000000000000000
  x11: 0000000000000000 x10: 0000000000000000
  x9 : 0000000000000000 x8 : ffff8007a4ce4000
  x7 : 0000000000000000 x6 : 000000000000003f
  x5 : ffff000000e6a280 x4 : ffff8007a4ce3000
  x3 : 0000000000000000 x2 : aaaaaaaaaaaaaaab
  x1 : ffff8007b9066a10 x0 : ffff8007baff8000
  Call trace:
   hw_stat_port_show+0x4c/0x80 [ib_core]
   port_attr_show+0x40/0x58 [ib_core]
   sysfs_kf_seq_show+0x8c/0x150
   kernfs_seq_show+0x44/0x50
   seq_read+0x1b4/0x45c
   kernfs_fop_read+0x148/0x1d8
   __vfs_read+0x58/0x180
   vfs_read+0x94/0x154
   ksys_read+0x68/0xd8
   __arm64_sys_read+0x28/0x34
   el0_svc_common+0x88/0x18c
   el0_svc_handler+0x78/0x94
   el0_svc+0x8/0xe8
  Code: f2955562 aa1603e4 aa1503e0 f9405683 (f9402861)
---
 drivers/infiniband/core/sysfs.c | 17 +++++++++++------
 1 file changed, 11 insertions(+), 6 deletions(-)
drivers/infiniband/core/sysfs.c +11 -6
index 8d709986b88c..253ccaf343f6 100644
@@ -1198,6 +1198,9 @@ static struct ib_port *setup_port(struct ib_core_device *coredev, int port_num,
 	p->port_num = port_num;
 	kobject_init(&p->kobj, &port_type);
 
+	if (device->port_data && is_full_dev)
+		device->port_data[port_num].sysfs = p;
+
 	cur_group = p->groups_list;
 	ret = alloc_port_table_group("gids", &p->groups[0], p->attrs_list,
 				     attr->gid_tbl_len, show_port_gid);
@@ -1243,9 +1246,6 @@ static struct ib_port *setup_port(struct ib_core_device *coredev, int port_num,
 	}
 
 	list_add_tail(&p->kobj.entry, &coredev->port_list);
-	if (device->port_data && is_full_dev)
-		device->port_data[port_num].sysfs = p;
-
 	return p;
 
 err_groups:
@@ -1253,6 +1253,8 @@ static struct ib_port *setup_port(struct ib_core_device *coredev, int port_num,
 err_del:
 	kobject_del(&p->kobj);
 err_put:
+	if (device->port_data && is_full_dev)
+		device->port_data[port_num].sysfs = NULL;
 	kobject_put(&p->kobj);
 	return ERR_PTR(ret);
 }
@@ -1261,14 +1263,17 @@ static void destroy_port(struct ib_core_device *coredev, struct ib_port *port)
 {
 	bool is_full_dev = &port->ibdev->coredev == coredev;
 
-	if (port->ibdev->port_data &&
-	    port->ibdev->port_data[port->port_num].sysfs == port)
-		port->ibdev->port_data[port->port_num].sysfs = NULL;
 	list_del(&port->kobj.entry);
 	if (is_full_dev)
 		sysfs_remove_groups(&port->kobj, port->ibdev->ops.port_groups);
+
 	sysfs_remove_groups(&port->kobj, port->groups_list);
 	kobject_del(&port->kobj);
+
+	if (port->ibdev->port_data &&
+	    port->ibdev->port_data[port->port_num].sysfs == port)
+		port->ibdev->port_data[port->port_num].sysfs = NULL;
+
 	kobject_put(&port->kobj);
 }
 
Signature · 1 line
2.35.1
author maintainer
#391 lore raw

PATCH 388/731 RDMA/nldev: Return "-EAGAIN" if the cm_id isnt from expected port

In reply to Greg Kroah-Hartman

From: Mark Zhang <[email protected]>

[ Upstream commit ecacb3751f254572af0009b9501e2cdc83a30b6a ]

When filling a cm_id entry, return "-EAGAIN" instead of 0 if the cm_id doesn'the have the same port as requested, otherwise an incomplete entry may be returned, which causes "rdam res show cm_id" to return an error.

For example on a machine with two rdma devices with "rping -C 1 -v -s" running background, the "rdma" command fails:

  $ rdma -V
  rdma utility, iproute2-5.19.0
  $ rdma res show cm_id
  link mlx5_0/- cm-idn 0 state LISTEN ps TCP pid 28056 comm rping src-addr 0.0.0.0:7174
  error: Protocol not available

While with this fix it succeeds:

  $ rdma res show cm_id
  link mlx5_0/- cm-idn 0 state LISTEN ps TCP pid 26395 comm rping src-addr 0.0.0.0:7174
  link mlx5_1/- cm-idn 0 state LISTEN ps TCP pid 26395 comm rping src-addr 0.0.0.0:7174
---
 drivers/infiniband/core/nldev.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/infiniband/core/nldev.c +1 -1
index cd89e59cbe33..6f3799698634 100644
@@ -550,7 +550,7 @@ static int fill_res_cm_id_entry(struct sk_buff *msg, bool has_cap_net_admin,
 	struct rdma_cm_id *cm_id = &id_priv->id;
 
 	if (port && port != cm_id->port_num)
-		return 0;
+		return -EAGAIN;
 
 	if (cm_id->port_num &&
 	    nla_put_u32(msg, RDMA_NLDEV_ATTR_PORT_INDEX, cm_id->port_num))
Signature · 1 line
2.35.1
author maintainer
#392 lore raw

PATCH 389/731 RDMA/siw: Set defined status for work completion with undefined status

In reply to Greg Kroah-Hartman

From: Bernard Metzler <[email protected]>

[ Upstream commit 60da2d11fcbc043304910e4d2ca82f9bab953e63 ]

A malicious user may write undefined values into memory mapped completion queue elements status or opcode. Undefined status or opcode values will result in out-of-bounds access to an array mapping siw internal representation of opcode and status to RDMA core representation when reaping CQ elements. While siw detects those undefined values, it did not correctly set completion status to a defined value, thus defeating the whole purpose of the check.

This bug leads to the following Smatch static checker warning:

drivers/infiniband/sw/siw/siw_cq.c:96 siw_reap_cqe()
error: buffer overflow 'map_cqe_status' 10 <= 21

Fixes bdf1da5df9da ("RDMA/siw: Fix immediate work request flush to completion queue")
---
 drivers/infiniband/sw/siw/siw_cq.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/infiniband/sw/siw/siw_cq.c +2 -2
index acc7bcd538b5..403029de6b92 100644
@@ -88,9 +88,9 @@ int siw_reap_cqe(struct siw_cq *cq, struct ib_wc *wc)
 
 			if (opcode >= SIW_NUM_OPCODES) {
 				opcode = 0;
-				status = IB_WC_GENERAL_ERR;
+				status = SIW_WC_GENERAL_ERR;
 			} else if (status >= SIW_NUM_WC_STATUS) {
-				status = IB_WC_GENERAL_ERR;
+				status = SIW_WC_GENERAL_ERR;
 			}
 			wc->opcode = map_wc_opcode[opcode];
 			wc->status = map_cqe_status[status].ib;
Signature · 1 line
2.35.1
author maintainer
#393 lore raw

PATCH 390/731 scsi: scsi_debug: Fix a warning in resp_write_scat()

In reply to Greg Kroah-Hartman

From: Harshit Mogalapalli <[email protected]>

[ Upstream commit 216e179724c1d9f57a8ababf8bd7aaabef67f01b ]

As 'lbdof_blen' is coming from user, if the size in kzalloc() is >= MAX_ORDER then we hit a warning.

Call trace:

sg_ioctl
sg_ioctl_common
scsi_ioctl

    sg_scsi_ioctl
     blk_execute_rq
      blk_mq_sched_insert_request
       blk_mq_run_hw_queue
        __blk_mq_delay_run_hw_queue
         __blk_mq_run_hw_queue
          blk_mq_sched_dispatch_requests
           __blk_mq_sched_dispatch_requests
            blk_mq_dispatch_rq_list
             scsi_queue_rq
              scsi_dispatch_cmd
               scsi_debug_queuecommand
                schedule_resp
                 resp_write_scat

If you try to allocate a memory larger than(>=) MAX_ORDER, then kmalloc() will definitely fail. It creates a stack trace and messes up dmesg. The user controls the size here so if they specify a too large size it will fail.

Add __GFP_NOWARN in order to avoid too large allocation warning. This is detected by static analysis using smatch.

---
 drivers/scsi/scsi_debug.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/scsi/scsi_debug.c +1 -1
index 0b16061d8da8..b44fd9b80934 100644
@@ -3640,7 +3640,7 @@ static int resp_write_scat(struct scsi_cmnd *scp,
 		mk_sense_buffer(scp, ILLEGAL_REQUEST, INVALID_FIELD_IN_CDB, 0);
 		return illegal_condition_result;
 	}
-	lrdp = kzalloc(lbdof_blen, GFP_ATOMIC);
+	lrdp = kzalloc(lbdof_blen, GFP_ATOMIC | __GFP_NOWARN);
 	if (lrdp == NULL)
 		return SCSI_MLQUEUE_HOST_BUSY;
 	if (sdebug_verbose)
Signature · 1 line
2.35.1
author maintainer
#394 lore raw

PATCH 391/731 crypto: ccree - Remove debugfs when platform_driver_register failed

In reply to Greg Kroah-Hartman

From: Gaosheng Cui <[email protected]>

[ Upstream commit 4f1c596df706c9aca662b6c214fad84047ae2a97 ]

When platform_driver_register failed, we need to remove debugfs, which will caused a resource leak, fix it.

Failed logs as follows:

[   32.606488] debugfs: Directory 'ccree' with parent '/' already present!
Fixes 4c3f97276e15 ("crypto: ccree - introduce CryptoCell driver")
---
 drivers/crypto/ccree/cc_driver.c | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)
drivers/crypto/ccree/cc_driver.c +9 -1
index 790fa9058a36..41f0a404bdf9 100644
@@ -656,9 +656,17 @@ static struct platform_driver ccree_driver = {
 
 static int __init ccree_init(void)
 {
+	int rc;
+
 	cc_debugfs_global_init();
 
-	return platform_driver_register(&ccree_driver);
+	rc = platform_driver_register(&ccree_driver);
+	if (rc) {
+		cc_debugfs_global_fini();
+		return rc;
+	}
+
+	return 0;
 }
 module_init(ccree_init);
 
Signature · 1 line
2.35.1
author maintainer
#395 lore raw

PATCH 392/731 crypto: cryptd - Use request context instead of stack for sub-request

In reply to Greg Kroah-Hartman

From: Herbert Xu <[email protected]>

[ Upstream commit 3a58c231172537f7b0e19d93ed33decd04f80eab ]

cryptd is buggy as it tries to use sync_skcipher without going through the proper sync_skcipher interface. In fact it doesn't even need sync_skcipher since it's already a proper skcipher and can easily access the request context instead of using something off the stack.

Fixes 36b3875a97b8 ("crypto: cryptd - Remove VLA usage of skcipher")
---
 crypto/cryptd.c | 36 +++++++++++++++++++-----------------
 1 file changed, 19 insertions(+), 17 deletions(-)
crypto/cryptd.c +19 -17
index 668095eca0fa..ca3a40fc7da9 100644
@@ -68,11 +68,12 @@ struct aead_instance_ctx {
 
 struct cryptd_skcipher_ctx {
 	refcount_t refcnt;
-	struct crypto_sync_skcipher *child;
+	struct crypto_skcipher *child;
 };
 
 struct cryptd_skcipher_request_ctx {
 	crypto_completion_t complete;
+	struct skcipher_request req;
 };
 
 struct cryptd_hash_ctx {
@@ -227,13 +228,13 @@ static int cryptd_skcipher_setkey(struct crypto_skcipher *parent,
 				  const u8 *key, unsigned int keylen)
 {
 	struct cryptd_skcipher_ctx *ctx = crypto_skcipher_ctx(parent);
-	struct crypto_sync_skcipher *child = ctx->child;
+	struct crypto_skcipher *child = ctx->child;
 
-	crypto_sync_skcipher_clear_flags(child, CRYPTO_TFM_REQ_MASK);
-	crypto_sync_skcipher_set_flags(child,
-				       crypto_skcipher_get_flags(parent) &
-					 CRYPTO_TFM_REQ_MASK);
-	return crypto_sync_skcipher_setkey(child, key, keylen);
+	crypto_skcipher_clear_flags(child, CRYPTO_TFM_REQ_MASK);
+	crypto_skcipher_set_flags(child,
+				  crypto_skcipher_get_flags(parent) &
+				  CRYPTO_TFM_REQ_MASK);
+	return crypto_skcipher_setkey(child, key, keylen);
 }
 
 static void cryptd_skcipher_complete(struct skcipher_request *req, int err)
@@ -258,13 +259,13 @@ static void cryptd_skcipher_encrypt(struct crypto_async_request *base,
 	struct cryptd_skcipher_request_ctx *rctx = skcipher_request_ctx(req);
 	struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(req);
 	struct cryptd_skcipher_ctx *ctx = crypto_skcipher_ctx(tfm);
-	struct crypto_sync_skcipher *child = ctx->child;
-	SYNC_SKCIPHER_REQUEST_ON_STACK(subreq, child);
+	struct skcipher_request *subreq = &rctx->req;
+	struct crypto_skcipher *child = ctx->child;
 
 	if (unlikely(err == -EINPROGRESS))
 		goto out;
 
-	skcipher_request_set_sync_tfm(subreq, child);
+	skcipher_request_set_tfm(subreq, child);
 	skcipher_request_set_callback(subreq, CRYPTO_TFM_REQ_MAY_SLEEP,
 				      NULL, NULL);
 	skcipher_request_set_crypt(subreq, req->src, req->dst, req->cryptlen,
@@ -286,13 +287,13 @@ static void cryptd_skcipher_decrypt(struct crypto_async_request *base,
 	struct cryptd_skcipher_request_ctx *rctx = skcipher_request_ctx(req);
 	struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(req);
 	struct cryptd_skcipher_ctx *ctx = crypto_skcipher_ctx(tfm);
-	struct crypto_sync_skcipher *child = ctx->child;
-	SYNC_SKCIPHER_REQUEST_ON_STACK(subreq, child);
+	struct skcipher_request *subreq = &rctx->req;
+	struct crypto_skcipher *child = ctx->child;
 
 	if (unlikely(err == -EINPROGRESS))
 		goto out;
 
-	skcipher_request_set_sync_tfm(subreq, child);
+	skcipher_request_set_tfm(subreq, child);
 	skcipher_request_set_callback(subreq, CRYPTO_TFM_REQ_MAY_SLEEP,
 				      NULL, NULL);
 	skcipher_request_set_crypt(subreq, req->src, req->dst, req->cryptlen,
@@ -343,9 +344,10 @@ static int cryptd_skcipher_init_tfm(struct crypto_skcipher *tfm)
 	if (IS_ERR(cipher))
 		return PTR_ERR(cipher);
 
-	ctx->child = (struct crypto_sync_skcipher *)cipher;
+	ctx->child = cipher;
 	crypto_skcipher_set_reqsize(
-		tfm, sizeof(struct cryptd_skcipher_request_ctx));
+		tfm, sizeof(struct cryptd_skcipher_request_ctx) +
+		     crypto_skcipher_reqsize(cipher));
 	return 0;
 }
 
@@ -353,7 +355,7 @@ static void cryptd_skcipher_exit_tfm(struct crypto_skcipher *tfm)
 {
 	struct cryptd_skcipher_ctx *ctx = crypto_skcipher_ctx(tfm);
 
-	crypto_free_sync_skcipher(ctx->child);
+	crypto_free_skcipher(ctx->child);
 }
 
 static void cryptd_skcipher_free(struct skcipher_instance *inst)
@@ -931,7 +933,7 @@ struct crypto_skcipher *cryptd_skcipher_child(struct cryptd_skcipher *tfm)
 {
 	struct cryptd_skcipher_ctx *ctx = crypto_skcipher_ctx(&tfm->base);
 
-	return &ctx->child->base;
+	return ctx->child;
 }
 EXPORT_SYMBOL_GPL(cryptd_skcipher_child);
 
Signature · 1 line
2.35.1
author maintainer
#396 lore raw

PATCH 393/731 crypto: hisilicon/qm - add missing pci_dev_put() in q_num_set()

In reply to Greg Kroah-Hartman

From: Xiongfeng Wang <[email protected]>

[ Upstream commit cc7710d0d4ebc6998f04035cde4f32c5ddbe9d7f ]

pci_get_device() will increase the reference count for the returned pci_dev. We need to use pci_dev_put() to decrease the reference count before q_num_set() returns.

Fixes c8b4b477079d ("crypto: hisilicon - add HiSilicon HPRE accelerator")
---
 drivers/crypto/hisilicon/qm.h | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/crypto/hisilicon/qm.h +4 -2
index 3068093229a5..bbb35de994eb 100644
@@ -318,14 +318,14 @@ struct hisi_qp {
 static inline int q_num_set(const char *val, const struct kernel_param *kp,
 			    unsigned int device)
 {
-	struct pci_dev *pdev = pci_get_device(PCI_VENDOR_ID_HUAWEI,
-					      device, NULL);
+	struct pci_dev *pdev;
 	u32 n, q_num;
 	int ret;
 
 	if (!val)
 		return -EINVAL;
 
+	pdev = pci_get_device(PCI_VENDOR_ID_HUAWEI, device, NULL);
 	if (!pdev) {
 		q_num = min_t(u32, QM_QNUM_V1, QM_QNUM_V2);
 		pr_info("No device found currently, suppose queue number is %u\n",
@@ -335,6 +335,8 @@ static inline int q_num_set(const char *val, const struct kernel_param *kp,
 			q_num = QM_QNUM_V1;
 		else
 			q_num = QM_QNUM_V2;
+
+		pci_dev_put(pdev);
 	}
 
 	ret = kstrtou32(val, 10, &n);
Signature · 1 line
2.35.1
author maintainer
#397 lore raw

PATCH 394/731 RDMA/hns: Repacing dseg_len by macros in fill_ext_sge_inl_data()

In reply to Greg Kroah-Hartman

From: Luoyouming <[email protected]>

[ Upstream commit 3b1f864c904915b3baebffb31ea05ee704b0df3c ]

The sge size is known to be constant, so it's unnecessary to use sizeof to calculate.

Stable-dep-of: 8eaa6f7d569b ("RDMA/hns: Fix ext_sge num error when post send")

---
 drivers/infiniband/hw/hns/hns_roce_hw_v2.c | 7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)
drivers/infiniband/hw/hns/hns_roce_hw_v2.c +3 -4
index 1421896abaf0..5875ccf86f66 100644
@@ -156,8 +156,7 @@ static int fill_ext_sge_inl_data(struct hns_roce_qp *qp,
 				 unsigned int *sge_idx, u32 msg_len)
 {
 	struct ib_device *ibdev = &(to_hr_dev(qp->ibqp.device))->ib_dev;
-	unsigned int dseg_len = sizeof(struct hns_roce_v2_wqe_data_seg);
-	unsigned int ext_sge_sz = qp->sq.max_gs * dseg_len;
+	unsigned int ext_sge_sz = qp->sq.max_gs * HNS_ROCE_SGE_SIZE;
 	unsigned int left_len_in_pg;
 	unsigned int idx = *sge_idx;
 	unsigned int i = 0;
@@ -185,7 +184,7 @@ static int fill_ext_sge_inl_data(struct hns_roce_qp *qp,
 		if (len <= left_len_in_pg) {
 			memcpy(dseg, addr, len);
 
-			idx += len / dseg_len;
+			idx += len / HNS_ROCE_SGE_SIZE;
 
 			i++;
 			if (i >= wr->num_sge)
@@ -200,7 +199,7 @@ static int fill_ext_sge_inl_data(struct hns_roce_qp *qp,
 
 			len -= left_len_in_pg;
 			addr += left_len_in_pg;
-			idx += left_len_in_pg / dseg_len;
+			idx += left_len_in_pg / HNS_ROCE_SGE_SIZE;
 			dseg = hns_roce_get_extend_sge(qp,
 						idx & (qp->sge.sge_cnt - 1));
 			left_len_in_pg = 1 << HNS_HW_PAGE_SHIFT;
Signature · 1 line
2.35.1
author maintainer
#398 lore raw

PATCH 395/731 RDMA/hns: Fix ext_sge num error when post send

In reply to Greg Kroah-Hartman

From: Luoyouming <[email protected]>

[ Upstream commit 8eaa6f7d569b4a22bfc1b0a3fdfeeb401feb65a4 ]

In the HNS ROCE driver, The sge is divided into standard sge and extended sge. There are 2 standard sge in RC/XRC, and the UD standard sge is 0. In the scenario of RC SQ inline, if the data does not exceed 32bytes, the standard sge will be used. If it exceeds, only the extended sge will be used to fill the data.

Currently, when filling the extended sge, max_gs is directly used as the number of the extended sge, which did not subtract the number of standard sge. There is a logical error. The new algorithm subtracts the number of standard sge from max_gs to get the actual number of extended sge.

Fixes 30b707886aeb ("RDMA/hns: Support inline data in extented sge space for RC")
---
 drivers/infiniband/hw/hns/hns_roce_hw_v2.c | 13 +++++++++++--
 1 file changed, 11 insertions(+), 2 deletions(-)
drivers/infiniband/hw/hns/hns_roce_hw_v2.c +11 -2
index 5875ccf86f66..94f3a0a87dfd 100644
@@ -151,20 +151,29 @@ static void set_atomic_seg(const struct ib_send_wr *wr,
 	hr_reg_write(rc_sq_wqe, RC_SEND_WQE_SGE_NUM, valid_num_sge);
 }
 
+static unsigned int get_std_sge_num(struct hns_roce_qp *qp)
+{
+	if (qp->ibqp.qp_type == IB_QPT_GSI || qp->ibqp.qp_type == IB_QPT_UD)
+		return 0;
+
+	return HNS_ROCE_SGE_IN_WQE;
+}
+
 static int fill_ext_sge_inl_data(struct hns_roce_qp *qp,
 				 const struct ib_send_wr *wr,
 				 unsigned int *sge_idx, u32 msg_len)
 {
 	struct ib_device *ibdev = &(to_hr_dev(qp->ibqp.device))->ib_dev;
-	unsigned int ext_sge_sz = qp->sq.max_gs * HNS_ROCE_SGE_SIZE;
 	unsigned int left_len_in_pg;
 	unsigned int idx = *sge_idx;
+	unsigned int std_sge_num;
 	unsigned int i = 0;
 	unsigned int len;
 	void *addr;
 	void *dseg;
 
-	if (msg_len > ext_sge_sz) {
+	std_sge_num = get_std_sge_num(qp);
+	if (msg_len > (qp->sq.max_gs - std_sge_num) * HNS_ROCE_SGE_SIZE) {
 		ibdev_err(ibdev,
 			  "no enough extended sge space for inline data.\n");
 		return -EINVAL;
Signature · 1 line
2.35.1
author maintainer
#399 lore raw

PATCH 396/731 PCI: Check for alloc failure in pci_request_irq()

In reply to Greg Kroah-Hartman

From: Zeng Heng <[email protected]>

[ Upstream commit 2d9cd957d40c3ac491b358e7cff0515bb07a3a9c ]

When kvasprintf() fails to allocate memory, it returns a NULL pointer. Return error from pci_request_irq() so we don't dereference it.

[bhelgaas: commit log]
Fixes 704e8953d3e9 ("PCI/irq: Add pci_request_irq() and pci_free_irq() helpers")
---
 drivers/pci/irq.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/pci/irq.c +2 -0
index 12ecd0aaa28d..0050e8f6814e 100644
@@ -44,6 +44,8 @@ int pci_request_irq(struct pci_dev *dev, unsigned int nr, irq_handler_t handler,
 	va_start(ap, fmt);
 	devname = kvasprintf(GFP_KERNEL, fmt, ap);
 	va_end(ap);
+	if (!devname)
+		return -ENOMEM;
 
 	ret = request_threaded_irq(pci_irq_vector(dev, nr), handler, thread_fn,
 				   irqflags, devname, dev_id);
Signature · 1 line
2.35.1
author maintainer
#400 lore raw

PATCH 397/731 RDMA/hfi: Decrease PCI device reference count in error path

In reply to Greg Kroah-Hartman

From: Xiongfeng Wang <[email protected]>

[ Upstream commit 9b51d072da1d27e1193e84708201c48e385ad912 ]

pci_get_device() will increase the reference count for the returned pci_dev, and also decrease the reference count for the input parameter *from* if it is not NULL.

If we break out the loop in node_affinity_init() with 'dev' not NULL, we need to call pci_dev_put() to decrease the reference count. Add missing pci_dev_put() in error path.

Fixes c513de490f80 ("IB/hfi1: Invalid NUMA node information can cause a divide by zero")
---
 drivers/infiniband/hw/hfi1/affinity.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/infiniband/hw/hfi1/affinity.c +2 -0
index 98c813ba4304..4c403d9e90cb 100644
@@ -178,6 +178,8 @@ int node_affinity_init(void)
 	for (node = 0; node < node_affinity.num_possible_nodes; node++)
 		hfi1_per_node_cntr[node] = 1;
 
+	pci_dev_put(dev);
+
 	return 0;
 }
 
Signature · 1 line
2.35.1
author maintainer
#401 lore raw

PATCH 398/731 crypto: ccree - Make cc_debugfs_global_fini() available for module init function

In reply to Greg Kroah-Hartman

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

[ Upstream commit 8e96729fc26c8967db45a3fb7a60387619f77a22 ]

ccree_init() calls cc_debugfs_global_fini(), the former is an init function and the latter an exit function though.

A modular build emits:

	WARNING: modpost: drivers/crypto/ccree/ccree.o: section mismatch in reference: init_module (section: .init.text) -> cc_debugfs_global_fini (section: .exit.text)

(with CONFIG_DEBUG_SECTION_MISMATCH=y).

Fixes 4f1c596df706 ("crypto: ccree - Remove debugfs when platform_driver_register failed")
---
 drivers/crypto/ccree/cc_debugfs.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/crypto/ccree/cc_debugfs.c +1 -1
index 7083767602fc..8f008f024f8f 100644
@@ -55,7 +55,7 @@ void __init cc_debugfs_global_init(void)
 	cc_debugfs_dir = debugfs_create_dir("ccree", NULL);
 }
 
-void __exit cc_debugfs_global_fini(void)
+void cc_debugfs_global_fini(void)
 {
 	debugfs_remove(cc_debugfs_dir);
 }
Signature · 1 line
2.35.1
author maintainer
#402 lore raw

PATCH 399/731 RDMA/hns: fix memory leak in hns_roce_alloc_mr()

In reply to Greg Kroah-Hartman

From: Zhengchao Shao <[email protected]>

[ Upstream commit a115aa00b18f7b8982b8f458149632caf64a862a ]

When hns_roce_mr_enable() failed in hns_roce_alloc_mr(), mr_key is not released. Compiled test only.

Fixes 9b2cf76c9f05 ("RDMA/hns: Optimize PBL buffer allocation process")
---
 drivers/infiniband/hw/hns/hns_roce_mr.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/infiniband/hw/hns/hns_roce_mr.c +2 -2
index 20360df25771..a593c142cd6b 100644
@@ -415,10 +415,10 @@ struct ib_mr *hns_roce_alloc_mr(struct ib_pd *pd, enum ib_mr_type mr_type,
 
 	return &mr->ibmr;
 
-err_key:
-	free_mr_key(hr_dev, mr);
 err_pbl:
 	free_mr_pbl(hr_dev, mr);
+err_key:
+	free_mr_key(hr_dev, mr);
 err_free:
 	kfree(mr);
 	return ERR_PTR(ret);
Signature · 1 line
2.35.1
author maintainer
#403 lore raw

PATCH 400/731 RDMA/rxe: Fix NULL-ptr-deref in rxe_qp_do_cleanup() when socket create failed

In reply to Greg Kroah-Hartman

From: Zhang Xiaoxu <[email protected]>

[ Upstream commit f67376d801499f4fa0838c18c1efcad8840e550d ]

There is a null-ptr-deref when mount.cifs over rdma:

  BUG: KASAN: null-ptr-deref in rxe_qp_do_cleanup+0x2f3/0x360 [rdma_rxe]
  Read of size 8 at addr 0000000000000018 by task mount.cifs/3046

  CPU: 2 PID: 3046 Comm: mount.cifs Not tainted 6.1.0-rc5+ #62
  Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.14.0-1.fc3
  Call Trace:
   <TASK>
   dump_stack_lvl+0x34/0x44
   kasan_report+0xad/0x130
   rxe_qp_do_cleanup+0x2f3/0x360 [rdma_rxe]
   execute_in_process_context+0x25/0x90
   __rxe_cleanup+0x101/0x1d0 [rdma_rxe]
   rxe_create_qp+0x16a/0x180 [rdma_rxe]
   create_qp.part.0+0x27d/0x340
   ib_create_qp_kernel+0x73/0x160
   rdma_create_qp+0x100/0x230
   _smbd_get_connection+0x752/0x20f0
   smbd_get_connection+0x21/0x40
   cifs_get_tcp_session+0x8ef/0xda0
   mount_get_conns+0x60/0x750
   cifs_mount+0x103/0xd00
   cifs_smb3_do_mount+0x1dd/0xcb0
   smb3_get_tree+0x1d5/0x300
   vfs_get_tree+0x41/0xf0
   path_mount+0x9b3/0xdd0
   __x64_sys_mount+0x190/0x1d0
   do_syscall_64+0x35/0x80
   entry_SYSCALL_64_after_hwframe+0x46/0xb0

The root cause of the issue is the socket create failed in rxe_qp_init_req().

So move the reset rxe_qp_do_cleanup() after the NULL ptr check.

---
 drivers/infiniband/sw/rxe/rxe_qp.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
drivers/infiniband/sw/rxe/rxe_qp.c +3 -3
index 57ebf4871608..d7a968356a9b 100644
@@ -830,12 +830,12 @@ static void rxe_qp_do_cleanup(struct work_struct *work)
 		qp->resp.mr = NULL;
 	}
 
-	if (qp_type(qp) == IB_QPT_RC)
-		sk_dst_reset(qp->sk->sk);
-
 	free_rd_atomic_resources(qp);
 
 	if (qp->sk) {
+		if (qp_type(qp) == IB_QPT_RC)
+			sk_dst_reset(qp->sk->sk);
+
 		kernel_sock_shutdown(qp->sk, SHUT_RDWR);
 		sock_release(qp->sk);
 	}
Signature · 1 line
2.35.1
author maintainer
#404 lore raw

PATCH 401/731 dt-bindings: imx6q-pcie: Fix clock names for imx6sx and imx8mq

In reply to Greg Kroah-Hartman

From: Serge Semin <[email protected]>

[ Upstream commit b8a83e600bdde93e7da41ea3204b2b3832a3c99b ]

Originally as it was defined the legacy bindings the pcie_inbound_axi and pcie_aux clock names were supposed to be used in the fsl,imx6sx-pcie and fsl,imx8mq-pcie devices respectively. But the bindings conversion has been incorrectly so now the fourth clock name is defined as "pcie_inbound_axi for imx6sx-pcie, pcie_aux for imx8mq-pcie", which is completely wrong. Let's fix that by conditionally apply the clock-names constraints based on the compatible string content.

Fixes 751ca492f131 ("dt-bindings: PCI: imx6: convert the imx pcie controller to dtschema")
---
 .../bindings/pci/fsl,imx6q-pcie.yaml          | 46 +++++++++++++++++--
 1 file changed, 42 insertions(+), 4 deletions(-)
Documentation/devicetree/bindings/pci/fsl,imx6q-pcie.yaml +42 -4
index acea1cd444fd..9b0548264a39 100644
@@ -14,9 +14,6 @@ description: |+
   This PCIe host controller is based on the Synopsys DesignWare PCIe IP
   and thus inherits all the common properties defined in snps,dw-pcie.yaml.
 
-allOf:
-  - $ref: /schemas/pci/snps,dw-pcie.yaml#
-
 properties:
   compatible:
     enum:
@@ -59,7 +56,7 @@ properties:
       - const: pcie
       - const: pcie_bus
       - const: pcie_phy
-      - const: pcie_inbound_axi for imx6sx-pcie, pcie_aux for imx8mq-pcie
+      - enum: [ pcie_inbound_axi, pcie_aux ]
 
   num-lanes:
     const: 1
@@ -166,6 +163,47 @@ required:
   - clocks
   - clock-names
 
+allOf:
+  - $ref: /schemas/pci/snps,dw-pcie.yaml#
+  - if:
+      properties:
+        compatible:
+          contains:
+            const: fsl,imx6sx-pcie
+    then:
+      properties:
+        clock-names:
+          items:
+            - {}
+            - {}
+            - {}
+            - const: pcie_inbound_axi
+  - if:
+      properties:
+        compatible:
+          contains:
+            const: fsl,imx8mq-pcie
+    then:
+      properties:
+        clock-names:
+          items:
+            - {}
+            - {}
+            - {}
+            - const: pcie_aux
+  - if:
+      properties:
+        compatible:
+          not:
+            contains:
+              enum:
+                - fsl,imx6sx-pcie
+                - fsl,imx8mq-pcie
+    then:
+      properties:
+        clock-names:
+          maxItems: 3
+
 unevaluatedProperties: false
 
 examples:
Signature · 1 line
2.35.1
author maintainer
#405 lore raw

PATCH 402/731 dt-bindings: visconti-pcie: Fix interrupts array max constraints

In reply to Greg Kroah-Hartman

From: Serge Semin <[email protected]>

[ Upstream commit 4cf4b9b70ab2785461190c08a3542d2d74c28b46 ]

In accordance with the way the device DT-node is actually defined in arch/arm64/boot/dts/toshiba/tmpv7708.dtsi and the way the device is probed by the DW PCIe driver there are two IRQs it actually has. It's MSI IRQ the DT-bindings lack. Let's extend the interrupts property constraints then and fix the schema example so one would be acceptable by the actual device DT-bindings.

Fixes 17c1b16340f0 ("dt-bindings: pci: Add DT binding for Toshiba Visconti PCIe controller")
---
 .../devicetree/bindings/pci/toshiba,visconti-pcie.yaml     | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)
Documentation/devicetree/bindings/pci/toshiba,visconti-pcie.yaml +4 -3
index 30b6396d83c8..aea0e2bcdd77 100644
@@ -36,7 +36,7 @@ properties:
       - const: mpu
 
   interrupts:
-    maxItems: 1
+    maxItems: 2
 
   clocks:
     items:
@@ -94,8 +94,9 @@ examples:
             #interrupt-cells = <1>;
             ranges = <0x81000000 0 0x40000000 0 0x40000000 0 0x00010000>,
                      <0x82000000 0 0x50000000 0 0x50000000 0 0x20000000>;
-            interrupts = <GIC_SPI 215 IRQ_TYPE_LEVEL_HIGH>;
-            interrupt-names = "intr";
+            interrupts = <GIC_SPI 211 IRQ_TYPE_LEVEL_HIGH>,
+                         <GIC_SPI 215 IRQ_TYPE_LEVEL_HIGH>;
+            interrupt-names = "msi", "intr";
             interrupt-map-mask = <0 0 0 7>;
             interrupt-map =
                 <0 0 0 1 &gic GIC_SPI 215 IRQ_TYPE_LEVEL_HIGH
Signature · 1 line
2.35.1
author maintainer
#406 lore raw

PATCH 403/731 scsi: hpsa: Fix possible memory leak in hpsa_init_one()

In reply to Greg Kroah-Hartman

From: Yuan Can <[email protected]>

[ Upstream commit 9c9ff300e0de07475796495d86f449340d454a0c ]

The hpda_alloc_ctlr_info() allocates h and its field reply_map. However, in hpsa_init_one(), if alloc_percpu() failed, the hpsa_init_one() jumps to clean1 directly, which frees h and leaks the h->reply_map.

Fix by calling hpda_free_ctlr_info() to release h->replay_map and h instead free h directly.

Fixes 8b834bff1b73 ("scsi: hpsa: fix selection of reply queue")
---
 drivers/scsi/hpsa.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/scsi/hpsa.c +1 -1
index 3faa87fa296a..74d563950c64 100644
@@ -8927,7 +8927,7 @@ static int hpsa_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
 		destroy_workqueue(h->monitor_ctlr_wq);
 		h->monitor_ctlr_wq = NULL;
 	}
-	kfree(h);
+	hpda_free_ctlr_info(h);
 	return rc;
 }
 
Signature · 1 line
2.35.1
author maintainer
#407 lore raw

PATCH 404/731 crypto: tcrypt - Fix multibuffer skcipher speed test mem leak

In reply to Greg Kroah-Hartman

From: Zhang Yiqun <[email protected]>

[ Upstream commit 1aa33fc8d4032227253ceb736f47c52b859d9683 ]

In the past, the data for mb-skcipher test has been allocated twice, that means the first allcated memory area is without free, which may cause a potential memory leakage. So this patch is to remove one allocation to fix this error.

Fixes e161c5930c15 ("crypto: tcrypt - add multibuf skcipher...")
---
 crypto/tcrypt.c | 9 ---------
 1 file changed, 9 deletions(-)
crypto/tcrypt.c +0 -9
index 3362897bf61b..4ada7e749390 100644
@@ -1295,15 +1295,6 @@ static void test_mb_skcipher_speed(const char *algo, int enc, int secs,
 			goto out_free_tfm;
 		}
 
-
-	for (i = 0; i < num_mb; ++i)
-		if (testmgr_alloc_buf(data[i].xbuf)) {
-			while (i--)
-				testmgr_free_buf(data[i].xbuf);
-			goto out_free_tfm;
-		}
-
-
 	for (i = 0; i < num_mb; ++i) {
 		data[i].req = skcipher_request_alloc(tfm, GFP_KERNEL);
 		if (!data[i].req) {
Signature · 1 line
2.35.1
author maintainer
#408 lore raw

PATCH 405/731 padata: Always leave BHs disabled when running ->parallel()

In reply to Greg Kroah-Hartman

From: Daniel Jordan <[email protected]>

[ Upstream commit 34c3a47d20ae55b3600fed733bf96eafe9c500d5 ]

A deadlock can happen when an overloaded system runs ->parallel() in the context of the current task:

    padata_do_parallel
      ->parallel()
        pcrypt_aead_enc/dec
          padata_do_serial
            spin_lock(&reorder->lock) // BHs still enabled
              <interrupt>
                ...
                  __do_softirq
                    ...
                      padata_do_serial
                        spin_lock(&reorder->lock)

It's a bug for BHs to be on in _do_serial as Steffen points out, so ensure they're off in the "current task" case like they are in padata_parallel_worker to avoid this situation.

Reported-by [email protected]
Fixes 4611ce224688 ("padata: allocate work structures for parallel jobs from a pool")
---
 kernel/padata.c | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)
kernel/padata.c +5 -3
index 18d3a5c699d8..9395b77fabb1 100644
@@ -207,14 +207,16 @@ int padata_do_parallel(struct padata_shell *ps,
 	pw = padata_work_alloc();
 	spin_unlock(&padata_works_lock);
 
+	if (!pw) {
+		/* Maximum works limit exceeded, run in the current task. */
+		padata->parallel(padata);
+	}
+
 	rcu_read_unlock_bh();
 
 	if (pw) {
 		padata_work_init(pw, padata_parallel_worker, padata, 0);
 		queue_work(pinst->parallel_wq, &pw->pw_work);
-	} else {
-		/* Maximum works limit exceeded, run in the current task. */
-		padata->parallel(padata);
 	}
 
 	return 0;
Signature · 1 line
2.35.1
author maintainer
#409 lore raw

PATCH 406/731 padata: Fix list iterator in padata_do_serial()

In reply to Greg Kroah-Hartman

From: Daniel Jordan <[email protected]>

[ Upstream commit 57ddfecc72a6c9941d159543e1c0c0a74fe9afdd ]

list_for_each_entry_reverse() assumes that the iterated list is nonempty and that every list_head is embedded in the same type, but its use in padata_do_serial() breaks both rules.

This doesn't cause any issues now because padata_priv and padata_list happen to have their list fields at the same offset, but we really shouldn't be relying on that.

Fixes bfde23ce200e ("padata: unbind parallel jobs from specific CPUs")
---
 kernel/padata.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)
kernel/padata.c +5 -2
index 9395b77fabb1..c17f772cc315 100644
@@ -390,13 +390,16 @@ void padata_do_serial(struct padata_priv *padata)
 	int hashed_cpu = padata_cpu_hash(pd, padata->seq_nr);
 	struct padata_list *reorder = per_cpu_ptr(pd->reorder_list, hashed_cpu);
 	struct padata_priv *cur;
+	struct list_head *pos;
 
 	spin_lock(&reorder->lock);
 	/* Sort in ascending order of sequence number. */
-	list_for_each_entry_reverse(cur, &reorder->list, list)
+	list_for_each_prev(pos, &reorder->list) {
+		cur = list_entry(pos, struct padata_priv, list);
 		if (cur->seq_nr < padata->seq_nr)
 			break;
-	list_add(&padata->list, &cur->list);
+	}
+	list_add(&padata->list, pos);
 	spin_unlock(&reorder->lock);
 
 	/*
Signature · 1 line
2.35.1
author maintainer
#410 lore raw

PATCH 407/731 scsi: mpt3sas: Fix possible resource leaks in mpt3sas_transport_port_add()

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit 78316e9dfc24906dd474630928ed1d3c562b568e ]

In mpt3sas_transport_port_add(), if sas_rphy_add() returns error, sas_rphy_free() needs be called to free the resource allocated in sas_end_device_alloc(). Otherwise a kernel crash will happen:

Unable to handle kernel NULL pointer dereference at virtual address 0000000000000108
CPU: 45 PID: 37020 Comm: bash Kdump: loaded Tainted: G        W          6.1.0-rc1+ #189
pstate: 60000005 (nZCv daif -PAN -UAO -TCO -DIT -SSBS BTYPE=--)
pc : device_del+0x54/0x3d0
lr : device_del+0x37c/0x3d0
Call trace:
 device_del+0x54/0x3d0
 attribute_container_class_device_del+0x28/0x38
 transport_remove_classdev+0x6c/0x80
 attribute_container_device_trigger+0x108/0x110
 transport_remove_device+0x28/0x38
 sas_rphy_remove+0x50/0x78 [scsi_transport_sas]
 sas_port_delete+0x30/0x148 [scsi_transport_sas]
 do_sas_phy_delete+0x78/0x80 [scsi_transport_sas]
 device_for_each_child+0x68/0xb0
 sas_remove_children+0x30/0x50 [scsi_transport_sas]
 sas_rphy_remove+0x38/0x78 [scsi_transport_sas]
 sas_port_delete+0x30/0x148 [scsi_transport_sas]
 do_sas_phy_delete+0x78/0x80 [scsi_transport_sas]
 device_for_each_child+0x68/0xb0
 sas_remove_children+0x30/0x50 [scsi_transport_sas]
 sas_remove_host+0x20/0x38 [scsi_transport_sas]
 scsih_remove+0xd8/0x420 [mpt3sas]

Because transport_add_device() is not called when sas_rphy_add() fails, the device is not added. When sas_rphy_remove() is subsequently called to remove the device in the remove() path, a NULL pointer dereference happens.

Fixes f92363d12359 ("[SCSI] mpt3sas: add new driver supporting 12GB SAS")
---
 drivers/scsi/mpt3sas/mpt3sas_transport.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/scsi/mpt3sas/mpt3sas_transport.c +2 -0
index 0681daee6c14..e5ecd6ada6cd 100644
@@ -829,6 +829,8 @@ mpt3sas_transport_port_add(struct MPT3SAS_ADAPTER *ioc, u16 handle,
 	if ((sas_rphy_add(rphy))) {
 		ioc_err(ioc, "failure at %s:%d/%s()!\n",
 			__FILE__, __LINE__, __func__);
+		sas_rphy_free(rphy);
+		rphy = NULL;
 	}
 
 	if (mpt3sas_port->remote_identify.device_type == SAS_END_DEVICE) {
Signature · 1 line
2.35.1
author maintainer
#411 lore raw

PATCH 408/731 scsi: hpsa: Fix error handling in hpsa_add_sas_host()

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit 4ef174a3ad9b5d73c1b6573e244ebba2b0d86eac ]

hpsa_sas_port_add_phy() does:
...
sas_phy_add() -> may return error here
sas_port_add_phy()
...

Whereas hpsa_free_sas_phy() does:
...
sas_port_delete_phy()
sas_phy_delete()
...

If hpsa_sas_port_add_phy() returns an error, hpsa_free_sas_phy() can not be called to free the memory because the port and the phy have not been added yet.

Replace hpsa_free_sas_phy() with sas_phy_free() and kfree() to avoid kernel crash in this case.

---
 drivers/scsi/hpsa.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/scsi/hpsa.c +2 -1
index 74d563950c64..a4a00944b290 100644
@@ -9788,7 +9788,8 @@ static int hpsa_add_sas_host(struct ctlr_info *h)
 	return 0;
 
 free_sas_phy:
-	hpsa_free_sas_phy(hpsa_sas_phy);
+	sas_phy_free(hpsa_sas_phy->phy);
+	kfree(hpsa_sas_phy);
 free_sas_port:
 	hpsa_free_sas_port(hpsa_sas_port);
 free_sas_node:
Signature · 1 line
2.35.1
author maintainer
#412 lore raw

PATCH 409/731 scsi: hpsa: Fix possible memory leak in hpsa_add_sas_device()

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit fda34a5d304d0b98cc967e8763b52221b66dc202 ]

If hpsa_sas_port_add_rphy() returns an error, the 'rphy' allocated in sas_end_device_alloc() needs to be freed. Address this by calling sas_rphy_free() in the error path.

---
 drivers/scsi/hpsa.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/scsi/hpsa.c +3 -1
index a4a00944b290..cf7988de7b90 100644
@@ -9825,10 +9825,12 @@ static int hpsa_add_sas_device(struct hpsa_sas_node *hpsa_sas_node,
 
 	rc = hpsa_sas_port_add_rphy(hpsa_sas_port, rphy);
 	if (rc)
-		goto free_sas_port;
+		goto free_sas_rphy;
 
 	return 0;
 
+free_sas_rphy:
+	sas_rphy_free(rphy);
 free_sas_port:
 	hpsa_free_sas_port(hpsa_sas_port);
 	device->sas_port = NULL;
Signature · 1 line
2.35.1
author maintainer
#413 lore raw

PATCH 410/731 scsi: efct: Fix possible memleak in efct_device_init()

In reply to Greg Kroah-Hartman

From: Chen Zhongjin <[email protected]>

[ Upstream commit bb0cd225dd37df1f4a22e36dad59ff33178ecdfc ]

In efct_device_init(), when efct_scsi_reg_fc_transport() fails, efct_scsi_tgt_driver_exit() is not called to release memory for efct_scsi_tgt_driver_init() and causes memleak:

unreferenced object 0xffff8881020ce000 (size 2048):
comm "modprobe", pid 465, jiffies 4294928222 (age 55.872s) backtrace:

    [<0000000021a1ef1b>] kmalloc_trace+0x27/0x110
    [<000000004c3ed51c>] target_register_template+0x4fd/0x7b0 [target_core_mod]
    [<00000000f3393296>] efct_scsi_tgt_driver_init+0x18/0x50 [efct]
    [<00000000115de533>] 0xffffffffc0d90011
    [<00000000d608f646>] do_one_initcall+0xd0/0x4e0
    [<0000000067828cf1>] do_init_module+0x1cc/0x6a0
    ...
Fixes 4df84e846624 ("scsi: elx: efct: Driver initialization routines")
---
 drivers/scsi/elx/efct/efct_driver.c | 1 +
 1 file changed, 1 insertion(+)
drivers/scsi/elx/efct/efct_driver.c +1 -0
index eab68fd9337a..37e1ab96ee5b 100644
@@ -42,6 +42,7 @@ efct_device_init(void)
 
 	rc = efct_scsi_reg_fc_transport();
 	if (rc) {
+		efct_scsi_tgt_driver_exit();
 		pr_err("failed to register to FC host\n");
 		return rc;
 	}
Signature · 1 line
2.35.1
author maintainer
#414 lore raw

PATCH 411/731 scsi: scsi_debug: Fix a warning in resp_verify()

In reply to Greg Kroah-Hartman

From: Harshit Mogalapalli <[email protected]>

[ Upstream commit ed0f17b748b20271cb568c7ca0b23b120316a47d ]

As 'vnum' is controlled by user, so if user tries to allocate memory larger than(>=) MAX_ORDER, then kcalloc() will fail, it creates a stack trace and messes up dmesg with a warning.

Add __GFP_NOWARN in order to avoid too large allocation warning. This is detected by static analysis using smatch.

Fixes c3e2fe9222d4 ("scsi: scsi_debug: Implement VERIFY(10), add VERIFY(16)")
---
 drivers/scsi/scsi_debug.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/scsi/scsi_debug.c +1 -1
index b44fd9b80934..6ca3205a0bf5 100644
@@ -4296,7 +4296,7 @@ static int resp_verify(struct scsi_cmnd *scp, struct sdebug_dev_info *devip)
 	if (ret)
 		return ret;
 
-	arr = kcalloc(lb_size, vnum, GFP_ATOMIC);
+	arr = kcalloc(lb_size, vnum, GFP_ATOMIC | __GFP_NOWARN);
 	if (!arr) {
 		mk_sense_buffer(scp, ILLEGAL_REQUEST, INSUFF_RES_ASC,
 				INSUFF_RES_ASCQ);
Signature · 1 line
2.35.1
author maintainer
#415 lore raw

PATCH 412/731 scsi: scsi_debug: Fix a warning in resp_report_zones()

In reply to Greg Kroah-Hartman

From: Harshit Mogalapalli <[email protected]>

[ Upstream commit 07f2ca139d9a7a1ba71c4c03997c8de161db2346 ]

As 'alloc_len' is user controlled data, if user tries to allocate memory larger than(>=) MAX_ORDER, then kcalloc() will fail, it creates a stack trace and messes up dmesg with a warning.

Add __GFP_NOWARN in order to avoid too large allocation warning. This is detected by static analysis using smatch.

Fixes 7db0e0c8190a ("scsi: scsi_debug: Fix buffer size of REPORT ZONES command")
---
 drivers/scsi/scsi_debug.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/scsi/scsi_debug.c +1 -1
index 6ca3205a0bf5..5624bb6a64a3 100644
@@ -4367,7 +4367,7 @@ static int resp_report_zones(struct scsi_cmnd *scp,
 	rep_max_zones = min((alloc_len - 64) >> ilog2(RZONES_DESC_HD),
 			    max_zones);
 
-	arr = kzalloc(alloc_len, GFP_ATOMIC);
+	arr = kzalloc(alloc_len, GFP_ATOMIC | __GFP_NOWARN);
 	if (!arr) {
 		mk_sense_buffer(scp, ILLEGAL_REQUEST, INSUFF_RES_ASC,
 				INSUFF_RES_ASCQ);
Signature · 1 line
2.35.1
author maintainer
#416 lore raw

PATCH 413/731 scsi: fcoe: Fix possible name leak when device_register() fails

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit 47b6a122c7b69a876c7ee2fc064a26b09627de9d ]

If device_register() returns an error, the name allocated by dev_set_name() needs to be freed. As the comment of device_register() says, one should use put_device() to give up the reference in the error path. Fix this by calling put_device(), then the name can be freed in kobject_cleanup().

The 'fcf' is freed in fcoe_fcf_device_release(), so the kfree() in the error path can be removed.

The 'ctlr' is freed in fcoe_ctlr_device_release(), so don't use the error label, just return NULL after calling put_device().

---
 drivers/scsi/fcoe/fcoe_sysfs.c | 19 ++++++++++---------
 1 file changed, 10 insertions(+), 9 deletions(-)
drivers/scsi/fcoe/fcoe_sysfs.c +10 -9
index af658aa38fed..6260aa5ea6af 100644
@@ -830,14 +830,15 @@ struct fcoe_ctlr_device *fcoe_ctlr_device_add(struct device *parent,
 
 	dev_set_name(&ctlr->dev, "ctlr_%d", ctlr->id);
 	error = device_register(&ctlr->dev);
-	if (error)
-		goto out_del_q2;
+	if (error) {
+		destroy_workqueue(ctlr->devloss_work_q);
+		destroy_workqueue(ctlr->work_q);
+		put_device(&ctlr->dev);
+		return NULL;
+	}
 
 	return ctlr;
 
-out_del_q2:
-	destroy_workqueue(ctlr->devloss_work_q);
-	ctlr->devloss_work_q = NULL;
 out_del_q:
 	destroy_workqueue(ctlr->work_q);
 	ctlr->work_q = NULL;
@@ -1036,16 +1037,16 @@ struct fcoe_fcf_device *fcoe_fcf_device_add(struct fcoe_ctlr_device *ctlr,
 	fcf->selected = new_fcf->selected;
 
 	error = device_register(&fcf->dev);
-	if (error)
-		goto out_del;
+	if (error) {
+		put_device(&fcf->dev);
+		goto out;
+	}
 
 	fcf->state = FCOE_FCF_STATE_CONNECTED;
 	list_add_tail(&fcf->peers, &ctlr->fcfs);
 
 	return fcf;
 
-out_del:
-	kfree(fcf);
 out:
 	return NULL;
 }
Signature · 1 line
2.35.1
author maintainer
#417 lore raw

PATCH 414/731 scsi: scsi_debug: Fix possible name leak in sdebug_add_host_helper()

In reply to Greg Kroah-Hartman

From: Yang Yingliang <[email protected]>

[ Upstream commit e6d773f93a49e0eda88a903a2a6542ca83380eb1 ]

Afer commit 1fa5ae857bb1 ("driver core: get rid of struct device's bus_id string array"), the name of device is allocated dynamically, it needs be freed when device_register() returns error.

As comment of device_register() says, one should use put_device() to give up the reference in the error path. Fix this by calling put_device(), then the name can be freed in kobject_cleanup(), and sdbg_host is freed in sdebug_release_adapter().

When the device release is not set, it means the device is not initialized. We can not call put_device() in this case. Use kfree() to free memory.

Fixes 1fa5ae857bb1 ("driver core: get rid of struct device's bus_id string array")
---
 drivers/scsi/scsi_debug.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
drivers/scsi/scsi_debug.c +4 -1
index 5624bb6a64a3..591df0a91057 100644
@@ -7156,7 +7156,10 @@ static int sdebug_add_host_helper(int per_host_idx)
 		kfree(sdbg_devinfo->zstate);
 		kfree(sdbg_devinfo);
 	}
-	kfree(sdbg_host);
+	if (sdbg_host->dev.release)
+		put_device(&sdbg_host->dev);
+	else
+		kfree(sdbg_host);
 	pr_warn("%s: failed, errno=%d\n", __func__, -error);
 	return error;
 }
Signature · 1 line
2.35.1
author maintainer
#418 lore raw

PATCH 415/731 scsi: ipr: Fix WARNING in ipr_init()

In reply to Greg Kroah-Hartman

From: Shang XiaoJing <[email protected]>

[ Upstream commit e6f108bffc3708ddcff72324f7d40dfcd0204894 ]

ipr_init() will not call unregister_reboot_notifier() when pci_register_driver() fails, which causes a WARNING. Call unregister_reboot_notifier() when pci_register_driver() fails.

notifier callback ipr_halt [ipr] already registered

WARNING: CPU: 3 PID: 299 at kernel/notifier.c:29
notifier_chain_register+0x16d/0x230
Modules linked in: ipr(+) xhci_pci_renesas xhci_hcd ehci_hcd usbcore
led_class gpu_sched drm_buddy video wmi drm_ttm_helper ttm
drm_display_helper drm_kms_helper drm drm_panel_orientation_quirks
agpgart cfbft
CPU: 3 PID: 299 Comm: modprobe Tainted: G        W
6.1.0-rc1-00190-g39508d23b672-dirty #332
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS
rel-1.15.0-0-g2dd4b9b3f840-prebuilt.qemu.org 04/01/2014
RIP: 0010:notifier_chain_register+0x16d/0x230
Call Trace:
 <TASK>
 __blocking_notifier_chain_register+0x73/0xb0
 ipr_init+0x30/0x1000 [ipr]
 do_one_initcall+0xdb/0x480
 do_init_module+0x1cf/0x680
 load_module+0x6a50/0x70a0
 __do_sys_finit_module+0x12f/0x1c0
 do_syscall_64+0x3f/0x90
 entry_SYSCALL_64_after_hwframe+0x63/0xcd
Fixes f72919ec2bbb ("[SCSI] ipr: implement shutdown changes and remove obsolete write cache parameter")
---
 drivers/scsi/ipr.c | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)
drivers/scsi/ipr.c +9 -1
index 56b8a2d6ffe4..04fb7fc01226 100644
@@ -10869,11 +10869,19 @@ static struct notifier_block ipr_notifier = {
  **/
 static int __init ipr_init(void)
 {
+	int rc;
+
 	ipr_info("IBM Power RAID SCSI Device Driver version: %s %s\n",
 		 IPR_DRIVER_VERSION, IPR_DRIVER_DATE);
 
 	register_reboot_notifier(&ipr_notifier);
-	return pci_register_driver(&ipr_driver);
+	rc = pci_register_driver(&ipr_driver);
+	if (rc) {
+		unregister_reboot_notifier(&ipr_notifier);
+		return rc;
+	}
+
+	return 0;
 }
 
 /**
Signature · 1 line
2.35.1
author maintainer
#419 lore raw

PATCH 416/731 scsi: fcoe: Fix transport not deattached when fcoe_if_init() fails

In reply to Greg Kroah-Hartman

From: Chen Zhongjin <[email protected]>

[ Upstream commit 4155658cee394b22b24c6d64e49247bf26d95b92 ]

fcoe_init() calls fcoe_transport_attach(&fcoe_sw_transport), but when fcoe_if_init() fails, &fcoe_sw_transport is not detached and leaves freed &fcoe_sw_transport on fcoe_transports list. This causes panic when reinserting module.

 BUG: unable to handle page fault for address: fffffbfff82e2213
 RIP: 0010:fcoe_transport_attach+0xe1/0x230 [libfcoe]
 Call Trace:
  <TASK>
  do_one_initcall+0xd0/0x4e0
  load_module+0x5eee/0x7210
  ...
Fixes 78a582463c1e ("[SCSI] fcoe: convert fcoe.ko to become an fcoe transport provider driver")
---
 drivers/scsi/fcoe/fcoe.c | 1 +
 1 file changed, 1 insertion(+)
drivers/scsi/fcoe/fcoe.c +1 -0
index 5ae6c207d3ac..76dbdae0e987 100644
@@ -2501,6 +2501,7 @@ static int __init fcoe_init(void)
 
 out_free:
 	mutex_unlock(&fcoe_config_mutex);
+	fcoe_transport_detach(&fcoe_sw_transport);
 out_destroy:
 	destroy_workqueue(fcoe_wq);
 	return rc;
Signature · 1 line
2.35.1
author maintainer
#420 lore raw

PATCH 417/731 scsi: snic: Fix possible UAF in snic_tgt_create()

In reply to Greg Kroah-Hartman

From: Gaosheng Cui <[email protected]>

[ Upstream commit e118df492320176af94deec000ae034cc92be754 ]

Smatch reports a warning as follows:

drivers/scsi/snic/snic_disc.c:307 snic_tgt_create() warn:
'&tgt->list' not removed from list

If device_add() fails in snic_tgt_create(), tgt will be freed, but tgt->list will not be removed from snic->disc.tgt_list, then list traversal may cause UAF.

Remove from snic->disc.tgt_list before free().

---
 drivers/scsi/snic/snic_disc.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/scsi/snic/snic_disc.c +3 -0
index e9ccfb97773f..7cf871323b2c 100644
@@ -318,6 +318,9 @@ snic_tgt_create(struct snic *snic, struct snic_tgt_id *tgtid)
 			      ret);
 
 		put_device(&snic->shost->shost_gendev);
+		spin_lock_irqsave(snic->shost->host_lock, flags);
+		list_del(&tgt->list);
+		spin_unlock_irqrestore(snic->shost->host_lock, flags);
 		kfree(tgt);
 		tgt = NULL;
 
Signature · 1 line
2.35.1
author maintainer
#421 lore raw

PATCH 418/731 RDMA/nldev: Add checks for nla_nest_start() in fill_stat_counter_qps()

In reply to Greg Kroah-Hartman

From: Yuan Can <[email protected]>

[ Upstream commit ea5ef136e215fdef35f14010bc51fcd6686e6922 ]

As the nla_nest_start() may fail with NULL returned, the return value needs to be checked.

Fixes c4ffee7c9bdb ("RDMA/netlink: Implement counter dumpit calback")
---
 drivers/infiniband/core/nldev.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/infiniband/core/nldev.c +2 -0
index 6f3799698634..cefe12964f3e 100644
@@ -892,6 +892,8 @@ static int fill_stat_counter_qps(struct sk_buff *msg,
 	int ret = 0;
 
 	table_attr = nla_nest_start(msg, RDMA_NLDEV_ATTR_RES_QP);
+	if (!table_attr)
+		return -EMSGSIZE;
 
 	rt = &counter->device->res[RDMA_RESTRACK_QP];
 	xa_lock(&rt->xa);
Signature · 1 line
2.35.1
author maintainer
#422 lore raw

PATCH 419/731 f2fs: avoid victim selection from previous victim section

In reply to Greg Kroah-Hartman

From: Yonggil Song <[email protected]>

[ Upstream commit e219aecfd4b766c4e878a3769057e9809f7fcadc ]

When f2fs chooses GC victim in large section & LFS mode, next_victim_seg[gc_type] is referenced first. After segment is freed, next_victim_seg[gc_type] has the next segment number.
However, next_victim_seg[gc_type] still has the last segment number even after the last segment of section is freed. In this case, when f2fs chooses a victim for the next GC round, the last segment of previous victim section is chosen as a victim.

Initialize next_victim_seg[gc_type] to NULL_SEGNO for the last segment in large section.

Fixes e3080b0120a1 ("f2fs: support subsectional garbage collection")
---
 fs/f2fs/gc.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
fs/f2fs/gc.c +3 -2
index 4cbaa6ab083f..7863f8fd3b95 100644
@@ -1673,8 +1673,9 @@ static int do_garbage_collect(struct f2fs_sb_info *sbi,
 				get_valid_blocks(sbi, segno, false) == 0)
 			seg_freed++;
 
-		if (__is_large_section(sbi) && segno + 1 < end_segno)
-			sbi->next_victim_seg[gc_type] = segno + 1;
+		if (__is_large_section(sbi))
+			sbi->next_victim_seg[gc_type] =
+				(segno + 1 < end_segno) ? segno + 1 : NULL_SEGNO;
 skip:
 		f2fs_put_page(sum_page, 0);
 	}
Signature · 1 line
2.35.1
author maintainer
#423 lore raw

PATCH 420/731 RDMA/nldev: Fix failure to send large messages

In reply to Greg Kroah-Hartman

From: Mark Zhang <[email protected]>

[ Upstream commit fc8f93ad3e5485d45c992233c96acd902992dfc4 ]

Return "-EMSGSIZE" instead of "-EINVAL" when filling a QP entry, so that new SKBs will be allocated if there's not enough room in current SKB.

---
 drivers/infiniband/core/nldev.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/infiniband/core/nldev.c +1 -1
index cefe12964f3e..7ad3ba7d5a0a 100644
@@ -511,7 +511,7 @@ static int fill_res_qp_entry(struct sk_buff *msg, bool has_cap_net_admin,
 
 	/* In create_qp() port is not set yet */
 	if (qp->port && nla_put_u32(msg, RDMA_NLDEV_ATTR_PORT_INDEX, qp->port))
-		return -EINVAL;
+		return -EMSGSIZE;
 
 	ret = nla_put_u32(msg, RDMA_NLDEV_ATTR_RES_LQPN, qp->qp_num);
 	if (ret)
Signature · 1 line
2.35.1
author maintainer
#424 lore raw

PATCH 421/731 crypto: amlogic - Remove kcalloc without check

In reply to Greg Kroah-Hartman

From: Christophe JAILLET <[email protected]>

[ Upstream commit 3d780c8a9850ad60dee47a8d971ba7888f3d1bd3 ]

There is no real point in allocating dedicated memory for the irqs array. MAXFLOW is only 2, so it is easier to allocated the needed space directly within the 'meson_dev' structure.

This saves some memory allocation and avoids an indirection when using the irqs array.

Fixes 48fe583fe541 ("crypto: amlogic - Add crypto accelerator...")
---
 drivers/crypto/amlogic/amlogic-gxl-core.c | 1 -
 drivers/crypto/amlogic/amlogic-gxl.h      | 2 +-
 2 files changed, 1 insertion(+), 2 deletions(-)
drivers/crypto/amlogic/amlogic-gxl-core.c +0 -1
index 6e7ae896717c..937187027ad5 100644
@@ -237,7 +237,6 @@ static int meson_crypto_probe(struct platform_device *pdev)
 		return err;
 	}
 
-	mc->irqs = devm_kcalloc(mc->dev, MAXFLOW, sizeof(int), GFP_KERNEL);
 	for (i = 0; i < MAXFLOW; i++) {
 		mc->irqs[i] = platform_get_irq(pdev, i);
 		if (mc->irqs[i] < 0)
drivers/crypto/amlogic/amlogic-gxl.h +1 -1
index dc0f142324a3..8c0746a1d6d4 100644
@@ -95,7 +95,7 @@ struct meson_dev {
 	struct device *dev;
 	struct meson_flow *chanlist;
 	atomic_t flow;
-	int *irqs;
+	int irqs[MAXFLOW];
 #ifdef CONFIG_CRYPTO_DEV_AMLOGIC_GXL_DEBUG
 	struct dentry *dbgfs_dir;
 #endif
Signature · 1 line
2.35.1
author maintainer
#425 lore raw

PATCH 422/731 crypto: omap-sham - Use pm_runtime_resume_and_get() in omap_sham_probe()

In reply to Greg Kroah-Hartman

From: Shang XiaoJing <[email protected]>

[ Upstream commit 7bcceb4c9896b1b672b636ae70fe75110d6bf1ad ]

omap_sham_probe() calls pm_runtime_get_sync() and calls pm_runtime_put_sync() latter to put usage_counter. However, pm_runtime_get_sync() will increment usage_counter even it failed. Fix it by replacing it with pm_runtime_resume_and_get() to keep usage counter balanced.

Fixes b359f034c8bf ("crypto: omap-sham - Convert to use pm_runtime API")
---
 drivers/crypto/omap-sham.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/crypto/omap-sham.c +1 -1
index f6bf53c00b61..4ec6949a7ca9 100644
@@ -2114,7 +2114,7 @@ static int omap_sham_probe(struct platform_device *pdev)
 
 	pm_runtime_enable(dev);
 
-	err = pm_runtime_get_sync(dev);
+	err = pm_runtime_resume_and_get(dev);
 	if (err < 0) {
 		dev_err(dev, "failed to get sync: %d\n", err);
 		goto err_pm;
Signature · 1 line
2.35.1
author maintainer
#426 lore raw

PATCH 423/731 riscv/mm: add arch hook arch_clear_hugepage_flags

In reply to Greg Kroah-Hartman

From: Tong Tiangen <[email protected]>

[ Upstream commit d8bf77a1dc3079692f54be3087a5fd16d90027b0 ]

With the PG_arch_1 we keep track if the page's data cache is clean, architecture rely on this property to treat new pages as dirty with respect to the data cache and perform the flushing before mapping the pages into userspace.

This patch adds a new architecture hook, arch_clear_hugepage_flags,so that architectures which rely on the page flags being in a particular state for fresh allocations can adjust the flags accordingly when a page is freed into the pool.

Fixes 9e953cda5cdf ("riscv: Introduce huge page support for 32/64bit kernel")
---
 arch/riscv/include/asm/hugetlb.h | 6 ++++++
 1 file changed, 6 insertions(+)
arch/riscv/include/asm/hugetlb.h +6 -0
index a5c2ca1d1cd8..ec19d6afc896 100644
@@ -5,4 +5,10 @@
 #include <asm-generic/hugetlb.h>
 #include <asm/page.h>
 
+static inline void arch_clear_hugepage_flags(struct page *page)
+{
+	clear_bit(PG_dcache_clean, &page->flags);
+}
+#define arch_clear_hugepage_flags arch_clear_hugepage_flags
+
 #endif /* _ASM_RISCV_HUGETLB_H */
Signature · 1 line
2.35.1
author maintainer
#427 lore raw

PATCH 424/731 RDMA/hfi1: Fix error return code in parse_platform_config()

In reply to Greg Kroah-Hartman

From: Wang Yufen <[email protected]>

[ Upstream commit 725349f8ba1e78a146c6ff8f3ee5e2712e517106 ]

In the previous iteration of the while loop, the "ret" may have been assigned a value of 0, so the error return code -EINVAL may have been incorrectly set to 0. To fix set valid return code before calling to goto.

Fixes 97167e813415 ("staging/rdma/hfi1: Tune for unknown channel if configuration file is absent")
---
 drivers/infiniband/hw/hfi1/firmware.c | 6 ++++++
 1 file changed, 6 insertions(+)
drivers/infiniband/hw/hfi1/firmware.c +6 -0
index 31e63e245ea9..ddf3217893f8 100644
@@ -1744,6 +1744,7 @@ int parse_platform_config(struct hfi1_devdata *dd)
 
 	if (!dd->platform_config.data) {
 		dd_dev_err(dd, "%s: Missing config file\n", __func__);
+		ret = -EINVAL;
 		goto bail;
 	}
 	ptr = (u32 *)dd->platform_config.data;
@@ -1752,6 +1753,7 @@ int parse_platform_config(struct hfi1_devdata *dd)
 	ptr++;
 	if (magic_num != PLATFORM_CONFIG_MAGIC_NUM) {
 		dd_dev_err(dd, "%s: Bad config file\n", __func__);
+		ret = -EINVAL;
 		goto bail;
 	}
 
@@ -1775,6 +1777,7 @@ int parse_platform_config(struct hfi1_devdata *dd)
 	if (file_length > dd->platform_config.size) {
 		dd_dev_info(dd, "%s:File claims to be larger than read size\n",
 			    __func__);
+		ret = -EINVAL;
 		goto bail;
 	} else if (file_length < dd->platform_config.size) {
 		dd_dev_info(dd,
@@ -1795,6 +1798,7 @@ int parse_platform_config(struct hfi1_devdata *dd)
 			dd_dev_err(dd, "%s: Failed validation at offset %ld\n",
 				   __func__, (ptr - (u32 *)
 					      dd->platform_config.data));
+			ret = -EINVAL;
 			goto bail;
 		}
 
@@ -1838,6 +1842,7 @@ int parse_platform_config(struct hfi1_devdata *dd)
 					   __func__, table_type,
 					   (ptr - (u32 *)
 					    dd->platform_config.data));
+				ret = -EINVAL;
 				goto bail; /* We don't trust this file now */
 			}
 			pcfgcache->config_tables[table_type].table = ptr;
@@ -1857,6 +1862,7 @@ int parse_platform_config(struct hfi1_devdata *dd)
 					   __func__, table_type,
 					   (ptr -
 					    (u32 *)dd->platform_config.data));
+				ret = -EINVAL;
 				goto bail; /* We don't trust this file now */
 			}
 			pcfgcache->config_tables[table_type].table_metadata =
Signature · 1 line
2.35.1
author maintainer
#428 lore raw

PATCH 425/731 RDMA/srp: Fix error return code in srp_parse_options()

In reply to Greg Kroah-Hartman

From: Wang Yufen <[email protected]>

[ Upstream commit ed461b30b22c8fa85c25189c14cb89f29595cd14 ]

In the previous iteration of the while loop, the "ret" may have been assigned a value of 0, so the error return code -EINVAL may have been incorrectly set to 0. To fix set valid return code before calling to goto. Also investigate each case separately as Andy suggessted.

Fixes e711f968c49c ("IB/srp: replace custom implementation of hex2bin()")
Fixes 2a174df0c602 ("IB/srp: Use kstrtoull() instead of simple_strtoull()")
Fixes 19f313438c77 ("IB/srp: Add RDMA/CM support")
---
 drivers/infiniband/ulp/srp/ib_srp.c | 96 ++++++++++++++++++++++++-----
 1 file changed, 82 insertions(+), 14 deletions(-)
drivers/infiniband/ulp/srp/ib_srp.c +82 -14
index 2f4991cea98c..a6117a7d0ab1 100644
@@ -3397,7 +3397,8 @@ static int srp_parse_options(struct net *net, const char *buf,
 			break;
 
 		case SRP_OPT_PKEY:
-			if (match_hex(args, &token)) {
+			ret = match_hex(args, &token);
+			if (ret) {
 				pr_warn("bad P_Key parameter '%s'\n", p);
 				goto out;
 			}
@@ -3457,7 +3458,8 @@ static int srp_parse_options(struct net *net, const char *buf,
 			break;
 
 		case SRP_OPT_MAX_SECT:
-			if (match_int(args, &token)) {
+			ret = match_int(args, &token);
+			if (ret) {
 				pr_warn("bad max sect parameter '%s'\n", p);
 				goto out;
 			}
@@ -3465,8 +3467,15 @@ static int srp_parse_options(struct net *net, const char *buf,
 			break;
 
 		case SRP_OPT_QUEUE_SIZE:
-			if (match_int(args, &token) || token < 1) {
+			ret = match_int(args, &token);
+			if (ret) {
+				pr_warn("match_int() failed for queue_size parameter '%s', Error %d\n",
+					p, ret);
+				goto out;
+			}
+			if (token < 1) {
 				pr_warn("bad queue_size parameter '%s'\n", p);
+				ret = -EINVAL;
 				goto out;
 			}
 			target->scsi_host->can_queue = token;
@@ -3477,25 +3486,40 @@ static int srp_parse_options(struct net *net, const char *buf,
 			break;
 
 		case SRP_OPT_MAX_CMD_PER_LUN:
-			if (match_int(args, &token) || token < 1) {
+			ret = match_int(args, &token);
+			if (ret) {
+				pr_warn("match_int() failed for max cmd_per_lun parameter '%s', Error %d\n",
+					p, ret);
+				goto out;
+			}
+			if (token < 1) {
 				pr_warn("bad max cmd_per_lun parameter '%s'\n",
 					p);
+				ret = -EINVAL;
 				goto out;
 			}
 			target->scsi_host->cmd_per_lun = token;
 			break;
 
 		case SRP_OPT_TARGET_CAN_QUEUE:
-			if (match_int(args, &token) || token < 1) {
+			ret = match_int(args, &token);
+			if (ret) {
+				pr_warn("match_int() failed for max target_can_queue parameter '%s', Error %d\n",
+					p, ret);
+				goto out;
+			}
+			if (token < 1) {
 				pr_warn("bad max target_can_queue parameter '%s'\n",
 					p);
+				ret = -EINVAL;
 				goto out;
 			}
 			target->target_can_queue = token;
 			break;
 
 		case SRP_OPT_IO_CLASS:
-			if (match_hex(args, &token)) {
+			ret = match_hex(args, &token);
+			if (ret) {
 				pr_warn("bad IO class parameter '%s'\n", p);
 				goto out;
 			}
@@ -3504,6 +3528,7 @@ static int srp_parse_options(struct net *net, const char *buf,
 				pr_warn("unknown IO class parameter value %x specified (use %x or %x).\n",
 					token, SRP_REV10_IB_IO_CLASS,
 					SRP_REV16A_IB_IO_CLASS);
+				ret = -EINVAL;
 				goto out;
 			}
 			target->io_class = token;
@@ -3526,16 +3551,24 @@ static int srp_parse_options(struct net *net, const char *buf,
 			break;
 
 		case SRP_OPT_CMD_SG_ENTRIES:
-			if (match_int(args, &token) || token < 1 || token > 255) {
+			ret = match_int(args, &token);
+			if (ret) {
+				pr_warn("match_int() failed for max cmd_sg_entries parameter '%s', Error %d\n",
+					p, ret);
+				goto out;
+			}
+			if (token < 1 || token > 255) {
 				pr_warn("bad max cmd_sg_entries parameter '%s'\n",
 					p);
+				ret = -EINVAL;
 				goto out;
 			}
 			target->cmd_sg_cnt = token;
 			break;
 
 		case SRP_OPT_ALLOW_EXT_SG:
-			if (match_int(args, &token)) {
+			ret = match_int(args, &token);
+			if (ret) {
 				pr_warn("bad allow_ext_sg parameter '%s'\n", p);
 				goto out;
 			}
@@ -3543,43 +3576,77 @@ static int srp_parse_options(struct net *net, const char *buf,
 			break;
 
 		case SRP_OPT_SG_TABLESIZE:
-			if (match_int(args, &token) || token < 1 ||
-					token > SG_MAX_SEGMENTS) {
+			ret = match_int(args, &token);
+			if (ret) {
+				pr_warn("match_int() failed for max sg_tablesize parameter '%s', Error %d\n",
+					p, ret);
+				goto out;
+			}
+			if (token < 1 || token > SG_MAX_SEGMENTS) {
 				pr_warn("bad max sg_tablesize parameter '%s'\n",
 					p);
+				ret = -EINVAL;
 				goto out;
 			}
 			target->sg_tablesize = token;
 			break;
 
 		case SRP_OPT_COMP_VECTOR:
-			if (match_int(args, &token) || token < 0) {
+			ret = match_int(args, &token);
+			if (ret) {
+				pr_warn("match_int() failed for comp_vector parameter '%s', Error %d\n",
+					p, ret);
+				goto out;
+			}
+			if (token < 0) {
 				pr_warn("bad comp_vector parameter '%s'\n", p);
+				ret = -EINVAL;
 				goto out;
 			}
 			target->comp_vector = token;
 			break;
 
 		case SRP_OPT_TL_RETRY_COUNT:
-			if (match_int(args, &token) || token < 2 || token > 7) {
+			ret = match_int(args, &token);
+			if (ret) {
+				pr_warn("match_int() failed for tl_retry_count parameter '%s', Error %d\n",
+					p, ret);
+				goto out;
+			}
+			if (token < 2 || token > 7) {
 				pr_warn("bad tl_retry_count parameter '%s' (must be a number between 2 and 7)\n",
 					p);
+				ret = -EINVAL;
 				goto out;
 			}
 			target->tl_retry_count = token;
 			break;
 
 		case SRP_OPT_MAX_IT_IU_SIZE:
-			if (match_int(args, &token) || token < 0) {
+			ret = match_int(args, &token);
+			if (ret) {
+				pr_warn("match_int() failed for max it_iu_size parameter '%s', Error %d\n",
+					p, ret);
+				goto out;
+			}
+			if (token < 0) {
 				pr_warn("bad maximum initiator to target IU size '%s'\n", p);
+				ret = -EINVAL;
 				goto out;
 			}
 			target->max_it_iu_size = token;
 			break;
 
 		case SRP_OPT_CH_COUNT:
-			if (match_int(args, &token) || token < 1) {
+			ret = match_int(args, &token);
+			if (ret) {
+				pr_warn("match_int() failed for channel count parameter '%s', Error %d\n",
+					p, ret);
+				goto out;
+			}
+			if (token < 1) {
 				pr_warn("bad channel count %s\n", p);
+				ret = -EINVAL;
 				goto out;
 			}
 			target->ch_count = token;
@@ -3588,6 +3655,7 @@ static int srp_parse_options(struct net *net, const char *buf,
 		default:
 			pr_warn("unknown parameter or missing value '%s' in target creation request\n",
 				p);
+			ret = -EINVAL;
 			goto out;
 		}
 	}
Signature · 1 line
2.35.1
author maintainer
#429 lore raw

PATCH 426/731 PCI: mt7621: Rename mt7621_pci_ to mt7621_pcie_

In reply to Greg Kroah-Hartman

From: Bjorn Helgaas <[email protected]>

[ Upstream commit 4793895f597d42eb54a0f54711b61263b6a8dd03 ]

Rename mt7621_pci_* structs and functions to mt7621_pcie_* for consistency with the rest of the file.

Stable-dep-of: 19098934f910 ("PCI: mt7621: Add sentinel to quirks table")

---
 drivers/staging/mt7621-pci/pci-mt7621.c | 36 ++++++++++++-------------
 1 file changed, 18 insertions(+), 18 deletions(-)
drivers/staging/mt7621-pci/pci-mt7621.c +18 -18
index 6acfc94a16e7..c4f57bb63482 100644
@@ -93,8 +93,8 @@ struct mt7621_pcie_port {
  * reset lines are inverted.
  */
 struct mt7621_pcie {
-	void __iomem *base;
 	struct device *dev;
+	void __iomem *base;
 	struct list_head ports;
 	bool resets_inverted;
 };
@@ -129,7 +129,7 @@ static inline void pcie_port_write(struct mt7621_pcie_port *port,
 	writel_relaxed(val, port->base + reg);
 }
 
-static inline u32 mt7621_pci_get_cfgaddr(unsigned int bus, unsigned int slot,
+static inline u32 mt7621_pcie_get_cfgaddr(unsigned int bus, unsigned int slot,
 					 unsigned int func, unsigned int where)
 {
 	return (((where & 0xF00) >> 8) << 24) | (bus << 16) | (slot << 11) |
@@ -140,7 +140,7 @@ static void __iomem *mt7621_pcie_map_bus(struct pci_bus *bus,
 					 unsigned int devfn, int where)
 {
 	struct mt7621_pcie *pcie = bus->sysdata;
-	u32 address = mt7621_pci_get_cfgaddr(bus->number, PCI_SLOT(devfn),
+	u32 address = mt7621_pcie_get_cfgaddr(bus->number, PCI_SLOT(devfn),
 					     PCI_FUNC(devfn), where);
 
 	writel_relaxed(address, pcie->base + RALINK_PCI_CONFIG_ADDR);
@@ -148,7 +148,7 @@ static void __iomem *mt7621_pcie_map_bus(struct pci_bus *bus,
 	return pcie->base + RALINK_PCI_CONFIG_DATA + (where & 3);
 }
 
-struct pci_ops mt7621_pci_ops = {
+struct pci_ops mt7621_pcie_ops = {
 	.map_bus	= mt7621_pcie_map_bus,
 	.read		= pci_generic_config_read,
 	.write		= pci_generic_config_write,
@@ -156,7 +156,7 @@ struct pci_ops mt7621_pci_ops = {
 
 static u32 read_config(struct mt7621_pcie *pcie, unsigned int dev, u32 reg)
 {
-	u32 address = mt7621_pci_get_cfgaddr(0, dev, 0, reg);
+	u32 address = mt7621_pcie_get_cfgaddr(0, dev, 0, reg);
 
 	pcie_write(pcie, address, RALINK_PCI_CONFIG_ADDR);
 	return pcie_read(pcie, RALINK_PCI_CONFIG_DATA);
@@ -165,7 +165,7 @@ static u32 read_config(struct mt7621_pcie *pcie, unsigned int dev, u32 reg)
 static void write_config(struct mt7621_pcie *pcie, unsigned int dev,
 			 u32 reg, u32 val)
 {
-	u32 address = mt7621_pci_get_cfgaddr(0, dev, 0, reg);
+	u32 address = mt7621_pcie_get_cfgaddr(0, dev, 0, reg);
 
 	pcie_write(pcie, address, RALINK_PCI_CONFIG_ADDR);
 	pcie_write(pcie, val, RALINK_PCI_CONFIG_DATA);
@@ -505,16 +505,16 @@ static int mt7621_pcie_register_host(struct pci_host_bridge *host)
 {
 	struct mt7621_pcie *pcie = pci_host_bridge_priv(host);
 
-	host->ops = &mt7621_pci_ops;
+	host->ops = &mt7621_pcie_ops;
 	host->sysdata = pcie;
 	return pci_host_probe(host);
 }
 
-static const struct soc_device_attribute mt7621_pci_quirks_match[] = {
+static const struct soc_device_attribute mt7621_pcie_quirks_match[] = {
 	{ .soc_id = "mt7621", .revision = "E2" }
 };
 
-static int mt7621_pci_probe(struct platform_device *pdev)
+static int mt7621_pcie_probe(struct platform_device *pdev)
 {
 	struct device *dev = &pdev->dev;
 	const struct soc_device_attribute *attr;
@@ -535,7 +535,7 @@ static int mt7621_pci_probe(struct platform_device *pdev)
 	platform_set_drvdata(pdev, pcie);
 	INIT_LIST_HEAD(&pcie->ports);
 
-	attr = soc_device_match(mt7621_pci_quirks_match);
+	attr = soc_device_match(mt7621_pcie_quirks_match);
 	if (attr)
 		pcie->resets_inverted = true;
 
@@ -572,7 +572,7 @@ static int mt7621_pci_probe(struct platform_device *pdev)
 	return err;
 }
 
-static int mt7621_pci_remove(struct platform_device *pdev)
+static int mt7621_pcie_remove(struct platform_device *pdev)
 {
 	struct mt7621_pcie *pcie = platform_get_drvdata(pdev);
 	struct mt7621_pcie_port *port;
@@ -583,18 +583,18 @@ static int mt7621_pci_remove(struct platform_device *pdev)
 	return 0;
 }
 
-static const struct of_device_id mt7621_pci_ids[] = {
+static const struct of_device_id mt7621_pcie_ids[] = {
 	{ .compatible = "mediatek,mt7621-pci" },
 	{},
 };
-MODULE_DEVICE_TABLE(of, mt7621_pci_ids);
+MODULE_DEVICE_TABLE(of, mt7621_pcie_ids);
 
-static struct platform_driver mt7621_pci_driver = {
-	.probe = mt7621_pci_probe,
-	.remove = mt7621_pci_remove,
+static struct platform_driver mt7621_pcie_driver = {
+	.probe = mt7621_pcie_probe,
+	.remove = mt7621_pcie_remove,
 	.driver = {
 		.name = "mt7621-pci",
-		.of_match_table = of_match_ptr(mt7621_pci_ids),
+		.of_match_table = of_match_ptr(mt7621_pcie_ids),
 	},
 };
-builtin_platform_driver(mt7621_pci_driver);
+builtin_platform_driver(mt7621_pcie_driver);
Signature · 1 line
2.35.1
author maintainer
#430 lore raw

PATCH 427/731 PCI: mt7621: Add sentinel to quirks table

In reply to Greg Kroah-Hartman

From: John Thomson <[email protected]>

[ Upstream commit 19098934f910b4d47cb30251dd39ffa57bef9523 ]

Current driver is missing a sentinel in the struct soc_device_attribute array, which causes an oops when assessed by the soc_device_match(mt7621_pcie_quirks_match) call.

This was only exposed once the CONFIG_SOC_MT7621 mt7621 soc_dev_attr was fixed to register the SOC as a device, in:

commit 7c18b64bba3b ("mips: ralink: mt7621: do not use kzalloc too early")

Fix it by adding the required sentinel.

---
 drivers/staging/mt7621-pci/pci-mt7621.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/staging/mt7621-pci/pci-mt7621.c +2 -1
index c4f57bb63482..b520d1e0edd1 100644
@@ -511,7 +511,8 @@ static int mt7621_pcie_register_host(struct pci_host_bridge *host)
 }
 
 static const struct soc_device_attribute mt7621_pcie_quirks_match[] = {
-	{ .soc_id = "mt7621", .revision = "E2" }
+	{ .soc_id = "mt7621", .revision = "E2" },
+	{ /* sentinel */ }
 };
 
 static int mt7621_pcie_probe(struct platform_device *pdev)
Signature · 1 line
2.35.1
author maintainer
#431 lore raw

PATCH 428/731 orangefs: Fix sysfs not cleanup when dev init failed

In reply to Greg Kroah-Hartman

From: Zhang Xiaoxu <[email protected]>

[ Upstream commit ea60a4ad0cf88b411cde6888b8c890935686ecd7 ]

When the dev init failed, should cleanup the sysfs, otherwise, the module will never be loaded since can not create duplicate sysfs directory:

sysfs: cannot create duplicate filename '/fs/orangefs'

  CPU: 1 PID: 6549 Comm: insmod Tainted: G        W          6.0.0+ #44
  Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.14.0-1.fc33 04/01/2014
  Call Trace:
   <TASK>
   dump_stack_lvl+0x34/0x44
   sysfs_warn_dup.cold+0x17/0x24
   sysfs_create_dir_ns+0x16d/0x180
   kobject_add_internal+0x156/0x3a0
   kobject_init_and_add+0xcf/0x120
   orangefs_sysfs_init+0x7e/0x3a0 [orangefs]
   orangefs_init+0xfe/0x1000 [orangefs]
   do_one_initcall+0x87/0x2a0
   do_init_module+0xdf/0x320
   load_module+0x2f98/0x3330
   __do_sys_finit_module+0x113/0x1b0
   do_syscall_64+0x35/0x80
   entry_SYSCALL_64_after_hwframe+0x46/0xb0

kobject_add_internal failed for orangefs with -EEXIST, don't try to register things with the same name in the same directory.

Fixes 2f83ace37181 ("orangefs: put register_chrdev immediately before register_filesystem")
---
 fs/orangefs/orangefs-mod.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)
fs/orangefs/orangefs-mod.c +4 -4
index cd7297815f91..5ab741c60b7e 100644
@@ -141,7 +141,7 @@ static int __init orangefs_init(void)
 		gossip_err("%s: could not initialize device subsystem %d!\n",
 			   __func__,
 			   ret);
-		goto cleanup_device;
+		goto cleanup_sysfs;
 	}
 
 	ret = register_filesystem(&orangefs_fs_type);
@@ -152,11 +152,11 @@ static int __init orangefs_init(void)
 		goto out;
 	}
 
-	orangefs_sysfs_exit();
-
-cleanup_device:
 	orangefs_dev_cleanup();
 
+cleanup_sysfs:
+	orangefs_sysfs_exit();
+
 sysfs_init_failed:
 	orangefs_debugfs_cleanup();
 
Signature · 1 line
2.35.1
author maintainer
#432 lore raw

PATCH 429/731 RDMA/hns: Fix AH attr queried by query_qp

In reply to Greg Kroah-Hartman

From: Chengchang Tang <[email protected]>

[ Upstream commit bc34c04f7b97c3794dec5a6d6d27ffd5f0e4f5c8 ]

The queried AH attr is invalid. This patch fix it.

This problem is found by rdma-core test test_mr_rereg_pd

ERROR: test_mr_rereg_pd (tests.test_mr.MRTest)
Test that cover rereg MR's PD with this flow:
----------------------------------------------------------------------
Traceback (most recent call last):
File "./tests/test_mr.py", line 157, in test_mr_rereg_pd
self.restate_qps()
File "./tests/test_mr.py", line 113, in restate_qps
self.server.qp.to_rts(self.server_qp_attr)
File "qp.pyx", line 1137, in pyverbs.qp.QP.to_rts
File "qp.pyx", line 1123, in pyverbs.qp.QP.to_rtr
pyverbs.pyverbs_error.PyverbsRDMAError: Failed to modify QP state to RTR.
Errno: 22, Invalid argument

Fixes 926a01dc000d ("RDMA/hns: Add QP operations support for hip08 SoC")
---
 drivers/infiniband/hw/hns/hns_roce_hw_v2.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/infiniband/hw/hns/hns_roce_hw_v2.c +2 -0
index 94f3a0a87dfd..155ae202b1ec 100644
@@ -5129,6 +5129,8 @@ static int hns_roce_v2_query_qp(struct ib_qp *ibqp, struct ib_qp_attr *qp_attr,
 
 		rdma_ah_set_sl(&qp_attr->ah_attr,
 			       hr_reg_read(&context, QPC_SL));
+		rdma_ah_set_port_num(&qp_attr->ah_attr, hr_qp->port + 1);
+		rdma_ah_set_ah_flags(&qp_attr->ah_attr, IB_AH_GRH);
 		grh->flow_label = hr_reg_read(&context, QPC_FL);
 		grh->sgid_index = hr_reg_read(&context, QPC_GMV_IDX);
 		grh->hop_limit = hr_reg_read(&context, QPC_HOPLIMIT);
Signature · 1 line
2.35.1
author maintainer
#433 lore raw

PATCH 430/731 RDMA/hns: Fix PBL page MTR find

In reply to Greg Kroah-Hartman

From: Chengchang Tang <[email protected]>

[ Upstream commit 9fb39ef2ff3e18f1740625ba04093dfbef086d2b ]

Now, The address of the first two pages in the MR will be searched, which use to speed up the lookup of the pbl table for hardware. An exception will occur when there is only one page in this MR. This patch fix the number of page to search.

Fixes 9b2cf76c9f05 ("RDMA/hns: Optimize PBL buffer allocation process")
---
 drivers/infiniband/hw/hns/hns_roce_hw_v2.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/infiniband/hw/hns/hns_roce_hw_v2.c +2 -1
index 155ae202b1ec..b9557be812b7 100644
@@ -3024,7 +3024,8 @@ static int set_mtpt_pbl(struct hns_roce_dev *hr_dev,
 	int i, count;
 
 	count = hns_roce_mtr_find(hr_dev, &mr->pbl_mtr, 0, pages,
-				  ARRAY_SIZE(pages), &pbl_ba);
+				  min_t(int, ARRAY_SIZE(pages), mr->npages),
+				  &pbl_ba);
 	if (count < 1) {
 		ibdev_err(ibdev, "failed to find PBL mtr, count = %d.\n",
 			  count);
Signature · 1 line
2.35.1
author maintainer
#434 lore raw

PATCH 431/731 RDMA/hns: Fix page size cap from firmware

In reply to Greg Kroah-Hartman

From: Chengchang Tang <[email protected]>

[ Upstream commit 99dc5a0712883d5d13b620d25b3759d429577bc8 ]

Add verification to make sure the roce page size cap is supported by the system page size.

Fixes ba6bb7e97421 ("RDMA/hns: Add interfaces to get pf capabilities from firmware")
---
 drivers/infiniband/hw/hns/hns_roce_hw_v2.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/infiniband/hw/hns/hns_roce_hw_v2.c +3 -0
index b9557be812b7..16f39321b319 100644
@@ -2371,6 +2371,9 @@ static int hns_roce_query_pf_caps(struct hns_roce_dev *hr_dev)
 					  V2_QUERY_PF_CAPS_D_RQWQE_HOP_NUM_M,
 					  V2_QUERY_PF_CAPS_D_RQWQE_HOP_NUM_S);
 
+	if (!(caps->page_size_cap & PAGE_SIZE))
+		caps->page_size_cap = HNS_ROCE_V2_PAGE_SIZE_SUPPORTED;
+
 	return 0;
 }
 
Signature · 1 line
2.35.1
author maintainer
#435 lore raw

PATCH 432/731 RDMA/hns: Fix error code of CMD

In reply to Greg Kroah-Hartman

From: Chengchang Tang <[email protected]>

[ Upstream commit 667d6164b84884c64de3fc18670cd5a98b0b10cf ]

The error code is fixed to EIO when CMD fails to excute. This patch converts the error status reported by firmware to linux errno.

Fixes a04ff739f2a9 ("RDMA/hns: Add command queue support for hip08 RoCE driver")
---
 drivers/infiniband/hw/hns/hns_roce_hw_v2.c | 26 +++++++++++++++++++++-
 drivers/infiniband/hw/hns/hns_roce_hw_v2.h |  5 +++++
 2 files changed, 30 insertions(+), 1 deletion(-)
drivers/infiniband/hw/hns/hns_roce_hw_v2.c +25 -1
index 16f39321b319..79d92b789984 100644
@@ -1257,6 +1257,30 @@ static void update_cmdq_status(struct hns_roce_dev *hr_dev)
 		hr_dev->cmd.state = HNS_ROCE_CMDQ_STATE_FATAL_ERR;
 }
 
+static int hns_roce_cmd_err_convert_errno(u16 desc_ret)
+{
+	struct hns_roce_cmd_errcode errcode_table[] = {
+		{CMD_EXEC_SUCCESS, 0},
+		{CMD_NO_AUTH, -EPERM},
+		{CMD_NOT_EXIST, -EOPNOTSUPP},
+		{CMD_CRQ_FULL, -EXFULL},
+		{CMD_NEXT_ERR, -ENOSR},
+		{CMD_NOT_EXEC, -ENOTBLK},
+		{CMD_PARA_ERR, -EINVAL},
+		{CMD_RESULT_ERR, -ERANGE},
+		{CMD_TIMEOUT, -ETIME},
+		{CMD_HILINK_ERR, -ENOLINK},
+		{CMD_INFO_ILLEGAL, -ENXIO},
+		{CMD_INVALID, -EBADR},
+	};
+	u16 i;
+
+	for (i = 0; i < ARRAY_SIZE(errcode_table); i++)
+		if (desc_ret == errcode_table[i].return_status)
+			return errcode_table[i].errno;
+	return -EIO;
+}
+
 static int __hns_roce_cmq_send(struct hns_roce_dev *hr_dev,
 			       struct hns_roce_cmq_desc *desc, int num)
 {
@@ -1301,7 +1325,7 @@ static int __hns_roce_cmq_send(struct hns_roce_dev *hr_dev,
 			dev_err_ratelimited(hr_dev->dev,
 					    "Cmdq IO error, opcode = %x, return = %x\n",
 					    desc->opcode, desc_ret);
-			ret = -EIO;
+			ret = hns_roce_cmd_err_convert_errno(desc_ret);
 		}
 	} else {
 		/* FW/HW reset or incorrect number of desc */
drivers/infiniband/hw/hns/hns_roce_hw_v2.h +5 -0
index 2f4a0019a716..67f5b6fcfa1b 100644
@@ -276,6 +276,11 @@ enum hns_roce_cmd_return_status {
 	CMD_OTHER_ERR = 0xff
 };
 
+struct hns_roce_cmd_errcode {
+	enum hns_roce_cmd_return_status return_status;
+	int errno;
+};
+
 enum hns_roce_sgid_type {
 	GID_TYPE_FLAG_ROCE_V1 = 0,
 	GID_TYPE_FLAG_ROCE_V2_IPV4,
Signature · 1 line
2.35.1
author maintainer
#436 lore raw

PATCH 433/731 crypto: img-hash - Fix variable dereferenced before check hdev->req

In reply to Greg Kroah-Hartman

From: Gaosheng Cui <[email protected]>

[ Upstream commit 04ba54e5af8f8f0137b08cb51a0b3a2e1ea46c94 ]

Smatch report warning as follows:

drivers/crypto/img-hash.c:366 img_hash_dma_task() warn: variable dereferenced before check 'hdev->req'

Variable dereferenced should be done after check 'hdev->req', fix it.

Fixes d358f1abbf71 ("crypto: img-hash - Add Imagination Technologies hw hash accelerator")
Fixes 10badea259fa ("crypto: img-hash - Fix null pointer exception")
---
 drivers/crypto/img-hash.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)
drivers/crypto/img-hash.c +6 -2
index aa4c7b2af3e2..34b41cbcfa8d 100644
@@ -358,12 +358,16 @@ static int img_hash_dma_init(struct img_hash_dev *hdev)
 static void img_hash_dma_task(unsigned long d)
 {
 	struct img_hash_dev *hdev = (struct img_hash_dev *)d;
-	struct img_hash_request_ctx *ctx = ahash_request_ctx(hdev->req);
+	struct img_hash_request_ctx *ctx;
 	u8 *addr;
 	size_t nbytes, bleft, wsend, len, tbc;
 	struct scatterlist tsg;
 
-	if (!hdev->req || !ctx->sg)
+	if (!hdev->req)
+		return;
+
+	ctx = ahash_request_ctx(hdev->req);
+	if (!ctx->sg)
 		return;
 
 	addr = sg_virt(ctx->sg);
Signature · 1 line
2.35.1
author maintainer
#437 lore raw

PATCH 434/731 hwrng: amd - Fix PCI device refcount leak

In reply to Greg Kroah-Hartman

From: Xiongfeng Wang <[email protected]>

[ Upstream commit ecadb5b0111ea19fc7c240bb25d424a94471eb7d ]

for_each_pci_dev() is implemented by pci_get_device(). The comment of pci_get_device() says that it will increase the reference count for the returned pci_dev and also decrease the reference count for the input pci_dev @from if it is not NULL.

If we break for_each_pci_dev() loop with pdev not NULL, we need to call pci_dev_put() to decrease the reference count. Add the missing pci_dev_put() for the normal and error path.

Fixes 96d63c0297cc ("[PATCH] Add AMD HW RNG driver")
---
 drivers/char/hw_random/amd-rng.c | 18 +++++++++++++-----
 1 file changed, 13 insertions(+), 5 deletions(-)
drivers/char/hw_random/amd-rng.c +13 -5
index c22d4184bb61..0555e3838bce 100644
@@ -143,15 +143,19 @@ static int __init amd_rng_mod_init(void)
 found:
 	err = pci_read_config_dword(pdev, 0x58, &pmbase);
 	if (err)
-		return err;
+		goto put_dev;
 
 	pmbase &= 0x0000FF00;
-	if (pmbase == 0)
-		return -EIO;
+	if (pmbase == 0) {
+		err = -EIO;
+		goto put_dev;
+	}
 
 	priv = kzalloc(sizeof(*priv), GFP_KERNEL);
-	if (!priv)
-		return -ENOMEM;
+	if (!priv) {
+		err = -ENOMEM;
+		goto put_dev;
+	}
 
 	if (!request_region(pmbase + PMBASE_OFFSET, PMBASE_SIZE, DRV_NAME)) {
 		dev_err(&pdev->dev, DRV_NAME " region 0x%x already in use!\n",
@@ -185,6 +189,8 @@ static int __init amd_rng_mod_init(void)
 	release_region(pmbase + PMBASE_OFFSET, PMBASE_SIZE);
 out:
 	kfree(priv);
+put_dev:
+	pci_dev_put(pdev);
 	return err;
 }
 
@@ -200,6 +206,8 @@ static void __exit amd_rng_mod_exit(void)
 
 	release_region(priv->pmbase + PMBASE_OFFSET, PMBASE_SIZE);
 
+	pci_dev_put(priv->pcidev);
+
 	kfree(priv);
 }
 
Signature · 1 line
2.35.1
author maintainer
#438 lore raw

PATCH 435/731 hwrng: geode - Fix PCI device refcount leak

In reply to Greg Kroah-Hartman

From: Xiongfeng Wang <[email protected]>

[ Upstream commit 9f6ec8dc574efb7f4f3d7ee9cd59ae307e78f445 ]

for_each_pci_dev() is implemented by pci_get_device(). The comment of pci_get_device() says that it will increase the reference count for the returned pci_dev and also decrease the reference count for the input pci_dev @from if it is not NULL.

If we break for_each_pci_dev() loop with pdev not NULL, we need to call pci_dev_put() to decrease the reference count. We add a new struct 'amd_geode_priv' to record pointer of the pci_dev and membase, and then add missing pci_dev_put() for the normal and error path.

Fixes ef5d862734b8 ("[PATCH] Add Geode HW RNG driver")
---
 drivers/char/hw_random/geode-rng.c | 36 +++++++++++++++++++++++-------
 1 file changed, 28 insertions(+), 8 deletions(-)
drivers/char/hw_random/geode-rng.c +28 -8
index 138ce434f86b..12fbe8091831 100644
@@ -51,6 +51,10 @@ static const struct pci_device_id pci_tbl[] = {
 };
 MODULE_DEVICE_TABLE(pci, pci_tbl);
 
+struct amd_geode_priv {
+	struct pci_dev *pcidev;
+	void __iomem *membase;
+};
 
 static int geode_rng_data_read(struct hwrng *rng, u32 *data)
 {
@@ -90,6 +94,7 @@ static int __init geode_rng_init(void)
 	const struct pci_device_id *ent;
 	void __iomem *mem;
 	unsigned long rng_base;
+	struct amd_geode_priv *priv;
 
 	for_each_pci_dev(pdev) {
 		ent = pci_match_id(pci_tbl, pdev);
@@ -97,17 +102,26 @@ static int __init geode_rng_init(void)
 			goto found;
 	}
 	/* Device not found. */
-	goto out;
+	return err;
 
 found:
+	priv = kzalloc(sizeof(*priv), GFP_KERNEL);
+	if (!priv) {
+		err = -ENOMEM;
+		goto put_dev;
+	}
+
 	rng_base = pci_resource_start(pdev, 0);
 	if (rng_base == 0)
-		goto out;
+		goto free_priv;
 	err = -ENOMEM;
 	mem = ioremap(rng_base, 0x58);
 	if (!mem)
-		goto out;
-	geode_rng.priv = (unsigned long)mem;
+		goto free_priv;
+
+	geode_rng.priv = (unsigned long)priv;
+	priv->membase = mem;
+	priv->pcidev = pdev;
 
 	pr_info("AMD Geode RNG detected\n");
 	err = hwrng_register(&geode_rng);
@@ -116,20 +130,26 @@ static int __init geode_rng_init(void)
 		       err);
 		goto err_unmap;
 	}
-out:
 	return err;
 
 err_unmap:
 	iounmap(mem);
-	goto out;
+free_priv:
+	kfree(priv);
+put_dev:
+	pci_dev_put(pdev);
+	return err;
 }
 
 static void __exit geode_rng_exit(void)
 {
-	void __iomem *mem = (void __iomem *)geode_rng.priv;
+	struct amd_geode_priv *priv;
 
+	priv = (struct amd_geode_priv *)geode_rng.priv;
 	hwrng_unregister(&geode_rng);
-	iounmap(mem);
+	iounmap(priv->membase);
+	pci_dev_put(priv->pcidev);
+	kfree(priv);
 }
 
 module_init(geode_rng_init);
Signature · 1 line
2.35.1
author maintainer
#439 lore raw

PATCH 436/731 IB/IPoIB: Fix queue count inconsistency for PKEY child interfaces

In reply to Greg Kroah-Hartman

From: Dragos Tatulea <[email protected]>

[ Upstream commit dbc94a0fb81771a38733c0e8f2ea8c4fa6934dc1 ]

There are 2 ways to create IPoIB PKEY child interfaces:
1) Writing a PKEY to /sys/class/net/<ib parent interface>/create_child.
2) Using netlink with iproute.

While with sysfs the child interface has the same number of tx and rx queues as the parent, with netlink there will always be 1 tx and 1 rx queue for the child interface. That's because the get_num_tx/rx_queues() netlink ops are missing and the default value of 1 is taken for the number of queues (in rtnl_create_link()).

This change adds the get_num_tx/rx_queues() ops which allows for interfaces with multiple queues to be created over netlink. This constant only represents the max number of tx and rx queues on that net device.

---
 drivers/infiniband/ulp/ipoib/ipoib_netlink.c | 7 +++++++
 1 file changed, 7 insertions(+)
drivers/infiniband/ulp/ipoib/ipoib_netlink.c +7 -0
index 5b05cf3837da..28e9b70844e4 100644
@@ -42,6 +42,11 @@ static const struct nla_policy ipoib_policy[IFLA_IPOIB_MAX + 1] = {
 	[IFLA_IPOIB_UMCAST]	= { .type = NLA_U16 },
 };
 
+static unsigned int ipoib_get_max_num_queues(void)
+{
+	return min_t(unsigned int, num_possible_cpus(), 128);
+}
+
 static int ipoib_fill_info(struct sk_buff *skb, const struct net_device *dev)
 {
 	struct ipoib_dev_priv *priv = ipoib_priv(dev);
@@ -173,6 +178,8 @@ static struct rtnl_link_ops ipoib_link_ops __read_mostly = {
 	.changelink	= ipoib_changelink,
 	.get_size	= ipoib_get_size,
 	.fill_info	= ipoib_fill_info,
+	.get_num_rx_queues = ipoib_get_max_num_queues,
+	.get_num_tx_queues = ipoib_get_max_num_queues,
 };
 
 struct rtnl_link_ops *ipoib_get_link_ops(void)
Signature · 1 line
2.35.1
author maintainer
#440 lore raw

PATCH 437/731 RISC-V: Align the shadow stack

In reply to Greg Kroah-Hartman

From: Palmer Dabbelt <[email protected]>

[ Upstream commit b003b3b77d65133a0011ae3b7b255347438c12f6 ]

The standard RISC-V ABIs all require 16-byte stack alignment. We're only calling that one function on the shadow stack so I doubt it'd result in a real issue, but might as well keep this lined up.

Fixes 31da94c25aea ("riscv: add VMAP_STACK overflow detection")
---
 arch/riscv/kernel/traps.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/riscv/kernel/traps.c +1 -1
index 2f4cd85fb651..4102c97309cc 100644
@@ -211,7 +211,7 @@ static DEFINE_PER_CPU(unsigned long [OVERFLOW_STACK_SIZE/sizeof(long)],
  * shadow stack, handled_ kernel_ stack_ overflow(in kernel/entry.S) is used
  * to get per-cpu overflow stack(get_overflow_stack).
  */
-long shadow_stack[SHADOW_OVERFLOW_STACK_SIZE/sizeof(long)];
+long shadow_stack[SHADOW_OVERFLOW_STACK_SIZE/sizeof(long)] __aligned(16);
 asmlinkage unsigned long get_overflow_stack(void)
 {
 	return (unsigned long)this_cpu_ptr(overflow_stack) +
Signature · 1 line
2.35.1