Skip to contentSkip to conversation

soc: ti: omap-prm: use atomic iopoll instead of sleeping one

PATCH In review Sasha Levin 5 participants 61 messages Jul 2, 2020 – Jul 9, 2020

Subsystems (19)
lore
Share & tools
lore
author maintainer
#7 lore raw

PATCH 8/53 gpu: host1x: Clean up debugfs in error handling path

In reply to Sasha Levin

From: Christophe JAILLET <[email protected]>

[ Upstream commit 109be8b23fb2ec8e2d309325ee3b7a49eab63961 ]

host1x_debug_init() must be reverted in an error handling path.

This is already fixed in the remove function since commit 44156eee91ba ("gpu: host1x: Clean up debugfs on removal")

---
 drivers/gpu/host1x/dev.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
drivers/gpu/host1x/dev.c +3 -2
index d24344e919227..3c0f151847bae 100644
@@ -468,11 +468,12 @@ static int host1x_probe(struct platform_device *pdev)
 
 	err = host1x_register(host);
 	if (err < 0)
-		goto deinit_intr;
+		goto deinit_debugfs;
 
 	return 0;
 
-deinit_intr:
+deinit_debugfs:
+	host1x_debug_deinit(host);
 	host1x_intr_deinit(host);
 deinit_syncpt:
 	host1x_syncpt_deinit(host);
Signature · 1 line
2.25.1
author maintainer
#8 lore raw

PATCH 9/53 drm/tegra: hub: Do not enable orphaned window group

In reply to Sasha Levin

From: Nicolin Chen <[email protected]>

[ Upstream commit ef4e417eb3ec7fe657928f10ac1d2154d8a5fb38 ]

Though the unconditional enable/disable code is not a final solution, we don't want to run into a NULL pointer situation when window group doesn't link to its DC parent if the DC is disabled in Device Tree.

So this patch simply adds a check to make sure that window group has a valid parent before running into tegra_windowgroup_enable/disable.

---
 drivers/gpu/drm/tegra/hub.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)
drivers/gpu/drm/tegra/hub.c +6 -2
index 8183e617bf6b8..a2ef8f218d4ec 100644
@@ -149,7 +149,9 @@ int tegra_display_hub_prepare(struct tegra_display_hub *hub)
 	for (i = 0; i < hub->soc->num_wgrps; i++) {
 		struct tegra_windowgroup *wgrp = &hub->wgrps[i];
 
-		tegra_windowgroup_enable(wgrp);
+		/* Skip orphaned window group whose parent DC is disabled */
+		if (wgrp->parent)
+			tegra_windowgroup_enable(wgrp);
 	}
 
 	return 0;
@@ -166,7 +168,9 @@ void tegra_display_hub_cleanup(struct tegra_display_hub *hub)
 	for (i = 0; i < hub->soc->num_wgrps; i++) {
 		struct tegra_windowgroup *wgrp = &hub->wgrps[i];
 
-		tegra_windowgroup_disable(wgrp);
+		/* Skip orphaned window group whose parent DC is disabled */
+		if (wgrp->parent)
+			tegra_windowgroup_disable(wgrp);
 	}
 }
 
Signature · 1 line
2.25.1
author maintainer
#9 lore raw

PATCH 10/53 gpu: host1x: Detach driver on unregister

In reply to Sasha Levin

From: Thierry Reding <[email protected]>

[ Upstream commit d9a0a05bf8c76e6dc79230669a8b5d685b168c30 ]

Currently when a host1x device driver is unregistered, it is not detached from the host1x controller, which means that the device will stay around and when the driver is registered again, it may bind to the old, stale device rather than the new one that was created from scratch upon driver registration. This in turn can cause various weird crashes within the driver core because it is confronted with a device that was already deleted.

Fix this by detaching the driver from the host1x controller when it is unregistered. This ensures that the deleted device also is no longer present in the device list that drivers will bind to.

---
 drivers/gpu/host1x/bus.c | 9 +++++++++
 1 file changed, 9 insertions(+)
drivers/gpu/host1x/bus.c +9 -0
index 6a995db51d6d8..e201f62d62c0c 100644
@@ -686,8 +686,17 @@ EXPORT_SYMBOL(host1x_driver_register_full);
  */
 void host1x_driver_unregister(struct host1x_driver *driver)
 {
+	struct host1x *host1x;
+
 	driver_unregister(&driver->driver);
 
+	mutex_lock(&devices_lock);
+
+	list_for_each_entry(host1x, &devices, list)
+		host1x_detach_driver(host1x, driver);
+
+	mutex_unlock(&devices_lock);
+
 	mutex_lock(&drivers_lock);
 	list_del_init(&driver->list);
 	mutex_unlock(&drivers_lock);
Signature · 1 line
2.25.1
author maintainer
#30 lore raw

PATCH 25/53 ixgbe: protect ring accesses with READ- and WRITE_ONCE

In reply to Sasha Levin

From: Ciara Loftus <[email protected]>

[ Upstream commit f140ad9fe2ae16f385f8fe4dc9cf67bb4c51d794 ]

READ_ONCE should be used when reading rings prior to accessing the statistics pointer. Introduce this as well as the corresponding WRITE_ONCE usage when allocating and freeing the rings, to ensure protected access.

---
 drivers/net/ethernet/intel/ixgbe/ixgbe_lib.c  | 12 ++++++------
 drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 14 +++++++++++---
 2 files changed, 17 insertions(+), 9 deletions(-)
drivers/net/ethernet/intel/ixgbe/ixgbe_lib.c +6 -6
index fd9f5d41b5942..2e35c5706cf10 100644
@@ -921,7 +921,7 @@ static int ixgbe_alloc_q_vector(struct ixgbe_adapter *adapter,
 		ring->queue_index = txr_idx;
 
 		/* assign ring to adapter */
-		adapter->tx_ring[txr_idx] = ring;
+		WRITE_ONCE(adapter->tx_ring[txr_idx], ring);
 
 		/* update count and index */
 		txr_count--;
@@ -948,7 +948,7 @@ static int ixgbe_alloc_q_vector(struct ixgbe_adapter *adapter,
 		set_ring_xdp(ring);
 
 		/* assign ring to adapter */
-		adapter->xdp_ring[xdp_idx] = ring;
+		WRITE_ONCE(adapter->xdp_ring[xdp_idx], ring);
 
 		/* update count and index */
 		xdp_count--;
@@ -991,7 +991,7 @@ static int ixgbe_alloc_q_vector(struct ixgbe_adapter *adapter,
 		ring->queue_index = rxr_idx;
 
 		/* assign ring to adapter */
-		adapter->rx_ring[rxr_idx] = ring;
+		WRITE_ONCE(adapter->rx_ring[rxr_idx], ring);
 
 		/* update count and index */
 		rxr_count--;
@@ -1020,13 +1020,13 @@ static void ixgbe_free_q_vector(struct ixgbe_adapter *adapter, int v_idx)
 
 	ixgbe_for_each_ring(ring, q_vector->tx) {
 		if (ring_is_xdp(ring))
-			adapter->xdp_ring[ring->queue_index] = NULL;
+			WRITE_ONCE(adapter->xdp_ring[ring->queue_index], NULL);
 		else
-			adapter->tx_ring[ring->queue_index] = NULL;
+			WRITE_ONCE(adapter->tx_ring[ring->queue_index], NULL);
 	}
 
 	ixgbe_for_each_ring(ring, q_vector->rx)
-		adapter->rx_ring[ring->queue_index] = NULL;
+		WRITE_ONCE(adapter->rx_ring[ring->queue_index], NULL);
 
 	adapter->q_vector[v_idx] = NULL;
 	napi_hash_del(&q_vector->napi);
drivers/net/ethernet/intel/ixgbe/ixgbe_main.c +11 -3
index ea6834bae04c0..a32a072761aa2 100644
@@ -7065,7 +7065,10 @@ void ixgbe_update_stats(struct ixgbe_adapter *adapter)
 	}
 
 	for (i = 0; i < adapter->num_rx_queues; i++) {
-		struct ixgbe_ring *rx_ring = adapter->rx_ring[i];
+		struct ixgbe_ring *rx_ring = READ_ONCE(adapter->rx_ring[i]);
+
+		if (!rx_ring)
+			continue;
 		non_eop_descs += rx_ring->rx_stats.non_eop_descs;
 		alloc_rx_page += rx_ring->rx_stats.alloc_rx_page;
 		alloc_rx_page_failed += rx_ring->rx_stats.alloc_rx_page_failed;
@@ -7086,15 +7089,20 @@ void ixgbe_update_stats(struct ixgbe_adapter *adapter)
 	packets = 0;
 	/* gather some stats to the adapter struct that are per queue */
 	for (i = 0; i < adapter->num_tx_queues; i++) {
-		struct ixgbe_ring *tx_ring = adapter->tx_ring[i];
+		struct ixgbe_ring *tx_ring = READ_ONCE(adapter->tx_ring[i]);
+
+		if (!tx_ring)
+			continue;
 		restart_queue += tx_ring->tx_stats.restart_queue;
 		tx_busy += tx_ring->tx_stats.tx_busy;
 		bytes += tx_ring->stats.bytes;
 		packets += tx_ring->stats.packets;
 	}
 	for (i = 0; i < adapter->num_xdp_queues; i++) {
-		struct ixgbe_ring *xdp_ring = adapter->xdp_ring[i];
+		struct ixgbe_ring *xdp_ring = READ_ONCE(adapter->xdp_ring[i]);
 
+		if (!xdp_ring)
+			continue;
 		restart_queue += xdp_ring->tx_stats.restart_queue;
 		tx_busy += xdp_ring->tx_stats.tx_busy;
 		bytes += xdp_ring->stats.bytes;
Signature · 1 line
2.25.1
author maintainer
#31 lore raw

PATCH 26/53 i40e: protect ring accesses with READ- and WRITE_ONCE

In reply to Sasha Levin

From: Ciara Loftus <[email protected]>

[ Upstream commit d59e267912cd90b0adf33b4659050d831e746317 ]

READ_ONCE should be used when reading rings prior to accessing the statistics pointer. Introduce this as well as the corresponding WRITE_ONCE usage when allocating and freeing the rings, to ensure protected access.

---
 drivers/net/ethernet/intel/i40e/i40e_main.c | 29 ++++++++++++++-------
 1 file changed, 19 insertions(+), 10 deletions(-)
drivers/net/ethernet/intel/i40e/i40e_main.c +19 -10
index 2a037ec244b94..80dc5fcb82db7 100644
@@ -439,11 +439,15 @@ static void i40e_get_netdev_stats_struct(struct net_device *netdev,
 		i40e_get_netdev_stats_struct_tx(ring, stats);
 
 		if (i40e_enabled_xdp_vsi(vsi)) {
-			ring++;
+			ring = READ_ONCE(vsi->xdp_rings[i]);
+			if (!ring)
+				continue;
 			i40e_get_netdev_stats_struct_tx(ring, stats);
 		}
 
-		ring++;
+		ring = READ_ONCE(vsi->rx_rings[i]);
+		if (!ring)
+			continue;
 		do {
 			start   = u64_stats_fetch_begin_irq(&ring->syncp);
 			packets = ring->stats.packets;
@@ -787,6 +791,8 @@ static void i40e_update_vsi_stats(struct i40e_vsi *vsi)
 	for (q = 0; q < vsi->num_queue_pairs; q++) {
 		/* locate Tx ring */
 		p = READ_ONCE(vsi->tx_rings[q]);
+		if (!p)
+			continue;
 
 		do {
 			start = u64_stats_fetch_begin_irq(&p->syncp);
@@ -800,8 +806,11 @@ static void i40e_update_vsi_stats(struct i40e_vsi *vsi)
 		tx_linearize += p->tx_stats.tx_linearize;
 		tx_force_wb += p->tx_stats.tx_force_wb;
 
-		/* Rx queue is part of the same block as Tx queue */
-		p = &p[1];
+		/* locate Rx ring */
+		p = READ_ONCE(vsi->rx_rings[q]);
+		if (!p)
+			continue;
+
 		do {
 			start = u64_stats_fetch_begin_irq(&p->syncp);
 			packets = p->stats.packets;
@@ -10816,10 +10825,10 @@ static void i40e_vsi_clear_rings(struct i40e_vsi *vsi)
 	if (vsi->tx_rings && vsi->tx_rings[0]) {
 		for (i = 0; i < vsi->alloc_queue_pairs; i++) {
 			kfree_rcu(vsi->tx_rings[i], rcu);
-			vsi->tx_rings[i] = NULL;
-			vsi->rx_rings[i] = NULL;
+			WRITE_ONCE(vsi->tx_rings[i], NULL);
+			WRITE_ONCE(vsi->rx_rings[i], NULL);
 			if (vsi->xdp_rings)
-				vsi->xdp_rings[i] = NULL;
+				WRITE_ONCE(vsi->xdp_rings[i], NULL);
 		}
 	}
 }
@@ -10853,7 +10862,7 @@ static int i40e_alloc_rings(struct i40e_vsi *vsi)
 		if (vsi->back->hw_features & I40E_HW_WB_ON_ITR_CAPABLE)
 			ring->flags = I40E_TXR_FLAGS_WB_ON_ITR;
 		ring->itr_setting = pf->tx_itr_default;
-		vsi->tx_rings[i] = ring++;
+		WRITE_ONCE(vsi->tx_rings[i], ring++);
 
 		if (!i40e_enabled_xdp_vsi(vsi))
 			goto setup_rx;
@@ -10871,7 +10880,7 @@ static int i40e_alloc_rings(struct i40e_vsi *vsi)
 			ring->flags = I40E_TXR_FLAGS_WB_ON_ITR;
 		set_ring_xdp(ring);
 		ring->itr_setting = pf->tx_itr_default;
-		vsi->xdp_rings[i] = ring++;
+		WRITE_ONCE(vsi->xdp_rings[i], ring++);
 
 setup_rx:
 		ring->queue_index = i;
@@ -10884,7 +10893,7 @@ static int i40e_alloc_rings(struct i40e_vsi *vsi)
 		ring->size = 0;
 		ring->dcb_tc = 0;
 		ring->itr_setting = pf->rx_itr_default;
-		vsi->rx_rings[i] = ring;
+		WRITE_ONCE(vsi->rx_rings[i], ring);
 	}
 
 	return 0;
Signature · 1 line
2.25.1
author maintainer
#32 lore raw

PATCH 27/53 ice: protect ring accesses with WRITE_ONCE

In reply to Sasha Levin

From: Ciara Loftus <[email protected]>

[ Upstream commit b1d95cc2391ffac0c5b27256a4fb0d2cfb021a29 ]

The READ_ONCE macro is used when reading rings prior to accessing the statistics pointer. The corresponding WRITE_ONCE usage when allocating and freeing the rings to ensure protected access was not in place. Introduce this.

---
 drivers/net/ethernet/intel/ice/ice_lib.c  | 8 ++++----
 drivers/net/ethernet/intel/ice/ice_main.c | 2 +-
 2 files changed, 5 insertions(+), 5 deletions(-)
drivers/net/ethernet/intel/ice/ice_lib.c +4 -4
index 2f256bf45efcf..6dd839b325259 100644
@@ -1063,7 +1063,7 @@ static void ice_vsi_clear_rings(struct ice_vsi *vsi)
 		for (i = 0; i < vsi->alloc_txq; i++) {
 			if (vsi->tx_rings[i]) {
 				kfree_rcu(vsi->tx_rings[i], rcu);
-				vsi->tx_rings[i] = NULL;
+				WRITE_ONCE(vsi->tx_rings[i], NULL);
 			}
 		}
 	}
@@ -1071,7 +1071,7 @@ static void ice_vsi_clear_rings(struct ice_vsi *vsi)
 		for (i = 0; i < vsi->alloc_rxq; i++) {
 			if (vsi->rx_rings[i]) {
 				kfree_rcu(vsi->rx_rings[i], rcu);
-				vsi->rx_rings[i] = NULL;
+				WRITE_ONCE(vsi->rx_rings[i], NULL);
 			}
 		}
 	}
@@ -1104,7 +1104,7 @@ static int ice_vsi_alloc_rings(struct ice_vsi *vsi)
 		ring->vsi = vsi;
 		ring->dev = dev;
 		ring->count = vsi->num_tx_desc;
-		vsi->tx_rings[i] = ring;
+		WRITE_ONCE(vsi->tx_rings[i], ring);
 	}
 
 	/* Allocate Rx rings */
@@ -1123,7 +1123,7 @@ static int ice_vsi_alloc_rings(struct ice_vsi *vsi)
 		ring->netdev = vsi->netdev;
 		ring->dev = dev;
 		ring->count = vsi->num_rx_desc;
-		vsi->rx_rings[i] = ring;
+		WRITE_ONCE(vsi->rx_rings[i], ring);
 	}
 
 	return 0;
drivers/net/ethernet/intel/ice/ice_main.c +1 -1
index 69e50331e08e9..7fd2ec63f128e 100644
@@ -1701,7 +1701,7 @@ static int ice_xdp_alloc_setup_rings(struct ice_vsi *vsi)
 		xdp_ring->netdev = NULL;
 		xdp_ring->dev = dev;
 		xdp_ring->count = vsi->num_tx_desc;
-		vsi->xdp_rings[i] = xdp_ring;
+		WRITE_ONCE(vsi->xdp_rings[i], xdp_ring);
 		if (ice_setup_tx_ring(xdp_ring))
 			goto free_xdp_rings;
 		ice_set_ring_xdp(xdp_ring);
Signature · 1 line
2.25.1
author maintainer
#34 lore raw

PATCH 29/53 powerpc/kvm/book3s64: Fix kernel crash with nested kvm & DEBUG_VIRTUAL

In reply to Sasha Levin

From: "Aneesh Kumar K.V" <[email protected]>

[ Upstream commit c1ed1754f271f6b7acb1bfdc8cfb62220fbed423 ]

With CONFIG_DEBUG_VIRTUAL=y, __pa() checks for addr value and if it's less than PAGE_OFFSET it leads to a BUG().

  #define __pa(x)
  ({
  	VIRTUAL_BUG_ON((unsigned long)(x) < PAGE_OFFSET);
  	(unsigned long)(x) & 0x0fffffffffffffffUL;
  })

kernel BUG at arch/powerpc/kvm/book3s_64_mmu_radix.c:43! cpu 0x70: Vector: 700 (Program Check) at [c0000018a2187360]

      pc: c000000000161b30: __kvmhv_copy_tofrom_guest_radix+0x130/0x1f0
      lr: c000000000161d5c: kvmhv_copy_from_guest_radix+0x3c/0x80
  ...
  kvmhv_copy_from_guest_radix+0x3c/0x80
  kvmhv_load_from_eaddr+0x48/0xc0
  kvmppc_ld+0x98/0x1e0
  kvmppc_load_last_inst+0x50/0x90
  kvmppc_hv_emulate_mmio+0x288/0x2b0
  kvmppc_book3s_radix_page_fault+0xd8/0x2b0
  kvmppc_book3s_hv_page_fault+0x37c/0x1050
  kvmppc_vcpu_run_hv+0xbb8/0x1080
  kvmppc_vcpu_run+0x34/0x50
  kvm_arch_vcpu_ioctl_run+0x2fc/0x410
  kvm_vcpu_ioctl+0x2b4/0x8f0
  ksys_ioctl+0xf4/0x150
  sys_ioctl+0x28/0x80
  system_call_exception+0x104/0x1d0
  system_call_common+0xe8/0x214

kvmhv_copy_tofrom_guest_radix() uses a NULL value for to/from to indicate direction of copy.

Avoid calling __pa() if the value is NULL to avoid the BUG().

[mpe: Massage change log a bit to mention CONFIG_DEBUG_VIRTUAL]
---
 arch/powerpc/kvm/book3s_64_mmu_radix.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
arch/powerpc/kvm/book3s_64_mmu_radix.c +2 -1
index bc6c1aa3d0e92..ef40addd52c65 100644
@@ -40,7 +40,8 @@ unsigned long __kvmhv_copy_tofrom_guest_radix(int lpid, int pid,
 	/* Can't access quadrants 1 or 2 in non-HV mode, call the HV to do it */
 	if (kvmhv_on_pseries())
 		return plpar_hcall_norets(H_COPY_TOFROM_GUEST, lpid, pid, eaddr,
-					  __pa(to), __pa(from), n);
+					  (to != NULL) ? __pa(to): 0,
+					  (from != NULL) ? __pa(from): 0, n);
 
 	quadrant = 1;
 	if (!pid)
Signature · 1 line
2.25.1
author maintainer
#41 lore raw

PATCH 36/53 s390/debug: avoid kernel warning on too large number of pages

In reply to Sasha Levin

From: Christian Borntraeger <[email protected]>

[ Upstream commit 827c4913923e0b441ba07ba4cc41e01181102303 ]

When specifying insanely large debug buffers a kernel warning is printed. The debug code does handle the error gracefully, though. Instead of duplicating the check let us silence the warning to avoid crashes when panic_on_warn is used.

---
 arch/s390/kernel/debug.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
arch/s390/kernel/debug.c +2 -1
index 6d321f5f101d6..7184d55d87aae 100644
@@ -198,9 +198,10 @@ static debug_entry_t ***debug_areas_alloc(int pages_per_area, int nr_areas)
 	if (!areas)
 		goto fail_malloc_areas;
 	for (i = 0; i < nr_areas; i++) {
+		/* GFP_NOWARN to avoid user triggerable WARN, we handle fails */
 		areas[i] = kmalloc_array(pages_per_area,
 					 sizeof(debug_entry_t *),
-					 GFP_KERNEL);
+					 GFP_KERNEL | __GFP_NOWARN);
 		if (!areas[i])
 			goto fail_malloc_areas2;
 		for (j = 0; j < pages_per_area; j++) {
Signature · 1 line
2.25.1

Sasha Levin

author maintainer
#0 lore raw

PATCH 1/53 soc: ti: omap-prm: use atomic iopoll instead of sleeping one

From: Tero Kristo <[email protected]>

[ Upstream commit 98ece19f247159a51003796ede7112fef2df5d7f ]

The reset handling APIs for omap-prm can be invoked PM runtime which runs in atomic context. For this to work properly, switch to atomic iopoll version instead of the current which can sleep. Otherwise, this throws a "BUG: scheduling while atomic" warning. Issue is seen rather easily when CONFIG_PREEMPT is enabled.

---
 drivers/soc/ti/omap_prm.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)
drivers/soc/ti/omap_prm.c +4 -4
index 96c6f777519c0..c9b3f9ebf0bbf 100644
@@ -256,10 +256,10 @@ static int omap_reset_deassert(struct reset_controller_dev *rcdev,
 		goto exit;
 
 	/* wait for the status to be set */
-	ret = readl_relaxed_poll_timeout(reset->prm->base +
-					 reset->prm->data->rstst,
-					 v, v & BIT(st_bit), 1,
-					 OMAP_RESET_MAX_WAIT);
+	ret = readl_relaxed_poll_timeout_atomic(reset->prm->base +
+						 reset->prm->data->rstst,
+						 v, v & BIT(st_bit), 1,
+						 OMAP_RESET_MAX_WAIT);
 	if (ret)
 		pr_err("%s: timedout waiting for %s:%lu\n", __func__,
 		       reset->prm->data->name, id);
Signature · 1 line
2.25.1
author maintainer
#2 lore raw

PATCH 3/53 perf/x86/rapl: Move RAPL support to common x86 code

From: Stephane Eranian <[email protected]>

[ Upstream commit fd3ae1e1587d64ef8cc8e361903d33625458073e ]

To prepare for support of both Intel and AMD RAPL.

As per the AMD PPR, Fam17h support Package RAPL counters to monitor power usage. The RAPL counter operates as with Intel RAPL, and as such it is beneficial to share the code.

No change in functionality.

---
 arch/x86/events/Kconfig            | 6 +++---
 arch/x86/events/Makefile           | 1 +
 arch/x86/events/intel/Makefile     | 2 --
 arch/x86/events/{intel => }/rapl.c | 9 ++++++---
 4 files changed, 10 insertions(+), 8 deletions(-)
 rename arch/x86/events/{intel => }/rapl.c (98%)
arch/x86/events/Kconfig +3 -3
index 9a7a1446cb3a0..4a809c6cbd2f5 100644
@@ -10,11 +10,11 @@ config PERF_EVENTS_INTEL_UNCORE
 	available on NehalemEX and more modern processors.
 
 config PERF_EVENTS_INTEL_RAPL
-	tristate "Intel rapl performance events"
-	depends on PERF_EVENTS && CPU_SUP_INTEL && PCI
+	tristate "Intel/AMD rapl performance events"
+	depends on PERF_EVENTS && (CPU_SUP_INTEL || CPU_SUP_AMD) && PCI
 	default y
 	---help---
-	Include support for Intel rapl performance events for power
+	Include support for Intel and AMD rapl performance events for power
 	monitoring on modern processors.
 
 config PERF_EVENTS_INTEL_CSTATE
arch/x86/events/Makefile +1 -0
index 9e07f554333fb..b418ef6878796 100644
@@ -1,5 +1,6 @@
 # SPDX-License-Identifier: GPL-2.0-only
 obj-y					+= core.o probe.o
+obj-$(PERF_EVENTS_INTEL_RAPL)		+= rapl.o
 obj-y					+= amd/
 obj-$(CONFIG_X86_LOCAL_APIC)            += msr.o
 obj-$(CONFIG_CPU_SUP_INTEL)		+= intel/
arch/x86/events/intel/Makefile +0 -2
index 3468b0c1dc7c9..e67a5886336c1 100644
@@ -2,8 +2,6 @@
 obj-$(CONFIG_CPU_SUP_INTEL)		+= core.o bts.o
 obj-$(CONFIG_CPU_SUP_INTEL)		+= ds.o knc.o
 obj-$(CONFIG_CPU_SUP_INTEL)		+= lbr.o p4.o p6.o pt.o
-obj-$(CONFIG_PERF_EVENTS_INTEL_RAPL)	+= intel-rapl-perf.o
-intel-rapl-perf-objs			:= rapl.o
 obj-$(CONFIG_PERF_EVENTS_INTEL_UNCORE)	+= intel-uncore.o
 intel-uncore-objs			:= uncore.o uncore_nhmex.o uncore_snb.o uncore_snbep.o
 obj-$(CONFIG_PERF_EVENTS_INTEL_CSTATE)	+= intel-cstate.o
arch/x86/events/rapl.c renamed from arch/x86/events/intel/rapl.c +6 -3
similarity index 98%
rename from arch/x86/events/intel/rapl.c
rename to arch/x86/events/rapl.c
index a5dbd25852cb7..ece043fb7b494 100644
@@ -1,11 +1,14 @@
 // SPDX-License-Identifier: GPL-2.0-only
 /*
- * Support Intel RAPL energy consumption counters
+ * Support Intel/AMD RAPL energy consumption counters
  * Copyright (C) 2013 Google, Inc., Stephane Eranian
  *
  * Intel RAPL interface is specified in the IA-32 Manual Vol3b
  * section 14.7.1 (September 2013)
  *
+ * AMD RAPL interface for Fam17h is described in the public PPR:
+ * https://bugzilla.kernel.org/show_bug.cgi?id=206537
+ *
  * RAPL provides more controls than just reporting energy consumption
  * however here we only expose the 3 energy consumption free running
  * counters (pp0, pkg, dram).
@@ -58,8 +61,8 @@
 #include <linux/nospec.h>
 #include <asm/cpu_device_id.h>
 #include <asm/intel-family.h>
-#include "../perf_event.h"
-#include "../probe.h"
+#include "perf_event.h"
+#include "probe.h"
 
 MODULE_LICENSE("GPL");
 
Signature · 1 line
2.25.1
author maintainer
#1 lore raw

PATCH 2/53 regmap: fix alignment issue

In reply to Sasha Levin

From: Jens Thoms Toerring <[email protected]>

[ Upstream commit 53d860952c8215cf9ae1ea33409c8cb71ad6ad3d ]

The assembly and disassembly of data to be sent to or received from a device invoke functions regmap_format_XX() and regmap_parse_XX() that extract or insert data items from or into a buffer, using assignments. In some cases the functions are called with a buffer pointer with an odd address. On architectures with strict alignment requirements this can result in a kernel crash. The assignments have been replaced by functions that take alignment into account.

---
 drivers/base/regmap/regmap.c | 100 ++++++++++++++++-------------------
 1 file changed, 46 insertions(+), 54 deletions(-)
drivers/base/regmap/regmap.c +46 -54
index 59f911e577192..d3112665c7d08 100644
@@ -17,6 +17,7 @@
 #include <linux/delay.h>
 #include <linux/log2.h>
 #include <linux/hwspinlock.h>
+#include <asm/unaligned.h>
 
 #define CREATE_TRACE_POINTS
 #include "trace.h"
@@ -249,22 +250,20 @@ static void regmap_format_8(void *buf, unsigned int val, unsigned int shift)
 
 static void regmap_format_16_be(void *buf, unsigned int val, unsigned int shift)
 {
-	__be16 *b = buf;
-
-	b[0] = cpu_to_be16(val << shift);
+	put_unaligned_be16(val << shift, buf);
 }
 
 static void regmap_format_16_le(void *buf, unsigned int val, unsigned int shift)
 {
-	__le16 *b = buf;
-
-	b[0] = cpu_to_le16(val << shift);
+	put_unaligned_le16(val << shift, buf);
 }
 
 static void regmap_format_16_native(void *buf, unsigned int val,
 				    unsigned int shift)
 {
-	*(u16 *)buf = val << shift;
+	u16 v = val << shift;
+
+	memcpy(buf, &v, sizeof(v));
 }
 
 static void regmap_format_24(void *buf, unsigned int val, unsigned int shift)
@@ -280,43 +279,39 @@ static void regmap_format_24(void *buf, unsigned int val, unsigned int shift)
 
 static void regmap_format_32_be(void *buf, unsigned int val, unsigned int shift)
 {
-	__be32 *b = buf;
-
-	b[0] = cpu_to_be32(val << shift);
+	put_unaligned_be32(val << shift, buf);
 }
 
 static void regmap_format_32_le(void *buf, unsigned int val, unsigned int shift)
 {
-	__le32 *b = buf;
-
-	b[0] = cpu_to_le32(val << shift);
+	put_unaligned_le32(val << shift, buf);
 }
 
 static void regmap_format_32_native(void *buf, unsigned int val,
 				    unsigned int shift)
 {
-	*(u32 *)buf = val << shift;
+	u32 v = val << shift;
+
+	memcpy(buf, &v, sizeof(v));
 }
 
 #ifdef CONFIG_64BIT
 static void regmap_format_64_be(void *buf, unsigned int val, unsigned int shift)
 {
-	__be64 *b = buf;
-
-	b[0] = cpu_to_be64((u64)val << shift);
+	put_unaligned_be64((u64) val << shift, buf);
 }
 
 static void regmap_format_64_le(void *buf, unsigned int val, unsigned int shift)
 {
-	__le64 *b = buf;
-
-	b[0] = cpu_to_le64((u64)val << shift);
+	put_unaligned_le64((u64) val << shift, buf);
 }
 
 static void regmap_format_64_native(void *buf, unsigned int val,
 				    unsigned int shift)
 {
-	*(u64 *)buf = (u64)val << shift;
+	u64 v = (u64) val << shift;
+
+	memcpy(buf, &v, sizeof(v));
 }
 #endif
 
@@ -333,35 +328,34 @@ static unsigned int regmap_parse_8(const void *buf)
 
 static unsigned int regmap_parse_16_be(const void *buf)
 {
-	const __be16 *b = buf;
-
-	return be16_to_cpu(b[0]);
+	return get_unaligned_be16(buf);
 }
 
 static unsigned int regmap_parse_16_le(const void *buf)
 {
-	const __le16 *b = buf;
-
-	return le16_to_cpu(b[0]);
+	return get_unaligned_le16(buf);
 }
 
 static void regmap_parse_16_be_inplace(void *buf)
 {
-	__be16 *b = buf;
+	u16 v = get_unaligned_be16(buf);
 
-	b[0] = be16_to_cpu(b[0]);
+	memcpy(buf, &v, sizeof(v));
 }
 
 static void regmap_parse_16_le_inplace(void *buf)
 {
-	__le16 *b = buf;
+	u16 v = get_unaligned_le16(buf);
 
-	b[0] = le16_to_cpu(b[0]);
+	memcpy(buf, &v, sizeof(v));
 }
 
 static unsigned int regmap_parse_16_native(const void *buf)
 {
-	return *(u16 *)buf;
+	u16 v;
+
+	memcpy(&v, buf, sizeof(v));
+	return v;
 }
 
 static unsigned int regmap_parse_24(const void *buf)
@@ -376,69 +370,67 @@ static unsigned int regmap_parse_24(const void *buf)
 
 static unsigned int regmap_parse_32_be(const void *buf)
 {
-	const __be32 *b = buf;
-
-	return be32_to_cpu(b[0]);
+	return get_unaligned_be32(buf);
 }
 
 static unsigned int regmap_parse_32_le(const void *buf)
 {
-	const __le32 *b = buf;
-
-	return le32_to_cpu(b[0]);
+	return get_unaligned_le32(buf);
 }
 
 static void regmap_parse_32_be_inplace(void *buf)
 {
-	__be32 *b = buf;
+	u32 v = get_unaligned_be32(buf);
 
-	b[0] = be32_to_cpu(b[0]);
+	memcpy(buf, &v, sizeof(v));
 }
 
 static void regmap_parse_32_le_inplace(void *buf)
 {
-	__le32 *b = buf;
+	u32 v = get_unaligned_le32(buf);
 
-	b[0] = le32_to_cpu(b[0]);
+	memcpy(buf, &v, sizeof(v));
 }
 
 static unsigned int regmap_parse_32_native(const void *buf)
 {
-	return *(u32 *)buf;
+	u32 v;
+
+	memcpy(&v, buf, sizeof(v));
+	return v;
 }
 
 #ifdef CONFIG_64BIT
 static unsigned int regmap_parse_64_be(const void *buf)
 {
-	const __be64 *b = buf;
-
-	return be64_to_cpu(b[0]);
+	return get_unaligned_be64(buf);
 }
 
 static unsigned int regmap_parse_64_le(const void *buf)
 {
-	const __le64 *b = buf;
-
-	return le64_to_cpu(b[0]);
+	return get_unaligned_le64(buf);
 }
 
 static void regmap_parse_64_be_inplace(void *buf)
 {
-	__be64 *b = buf;
+	u64 v =  get_unaligned_be64(buf);
 
-	b[0] = be64_to_cpu(b[0]);
+	memcpy(buf, &v, sizeof(v));
 }
 
 static void regmap_parse_64_le_inplace(void *buf)
 {
-	__le64 *b = buf;
+	u64 v = get_unaligned_le64(buf);
 
-	b[0] = le64_to_cpu(b[0]);
+	memcpy(buf, &v, sizeof(v));
 }
 
 static unsigned int regmap_parse_64_native(const void *buf)
 {
-	return *(u64 *)buf;
+	u64 v;
+
+	memcpy(&v, buf, sizeof(v));
+	return v;
 }
 #endif
 
Signature · 1 line
2.25.1
author maintainer
#3 lore raw

PATCH 4/53 perf/x86/rapl: Fix RAPL config variable bug

In reply to Sasha Levin

From: Stephane Eranian <[email protected]>

[ Upstream commit 16accae3d97f97d7f61c4ee5d0002bccdef59088 ]

This patch fixes a bug introduced by:

fd3ae1e1587d6 ("perf/x86/rapl: Move RAPL support to common x86 code")

The Kconfig variable name was wrong. It was missing the CONFIG_ prefix.

---
 arch/x86/events/Makefile | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/x86/events/Makefile +1 -1
index b418ef6878796..726e83c0a31a1 100644
@@ -1,6 +1,6 @@
 # SPDX-License-Identifier: GPL-2.0-only
 obj-y					+= core.o probe.o
-obj-$(PERF_EVENTS_INTEL_RAPL)		+= rapl.o
+obj-$(CONFIG_PERF_EVENTS_INTEL_RAPL)	+= rapl.o
 obj-y					+= amd/
 obj-$(CONFIG_X86_LOCAL_APIC)            += msr.o
 obj-$(CONFIG_CPU_SUP_INTEL)		+= intel/
Signature · 1 line
2.25.1
author maintainer
#5 lore raw

PATCH 6/53 drm/ttm: Fix dma_fence refcnt leak in ttm_bo_vm_fault_reserved

In reply to Sasha Levin

From: Xiyu Yang <[email protected]>

[ Upstream commit 37cc4b95d13f311c04aa8e9daacca3905ad45ca7 ]

ttm_bo_vm_fault_reserved() invokes dma_fence_get(), which returns a reference of the specified dma_fence object to "moving" with increased refcnt.

When ttm_bo_vm_fault_reserved() returns, local variable "moving" becomes invalid, so the refcount should be decreased to keep refcount balanced.

The reference counting issue happens in several exception handling paths of ttm_bo_vm_fault_reserved(). When those error scenarios occur such as "err" equals to -EBUSY, the function forgets to decrease the refcnt increased by dma_fence_get(), causing a refcnt leak.

Fix this issue by calling dma_fence_put() when no_wait_gpu flag is equals to true.

---
 drivers/gpu/drm/ttm/ttm_bo_vm.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/gpu/drm/ttm/ttm_bo_vm.c +2 -0
index 0ad30b1129821..72100b84c7a90 100644
@@ -300,8 +300,10 @@ vm_fault_t ttm_bo_vm_fault_reserved(struct vm_fault *vmf,
 			break;
 		case -EBUSY:
 		case -ERESTARTSYS:
+			dma_fence_put(moving);
 			return VM_FAULT_NOPAGE;
 		default:
+			dma_fence_put(moving);
 			return VM_FAULT_SIGBUS;
 		}
 
Signature · 6 lines
2.25.1

_______________________________________________
dri-devel mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dri-devel
author maintainer
#4 lore raw

PATCH 5/53 ARM: dts: omap4-droid4: Fix spi configuration and increase rate

In reply to Sasha Levin

From: Tony Lindgren <[email protected]>

[ Upstream commit 0df12a01f4857495816b05f048c4c31439446e35 ]

We can currently sometimes get "RXS timed out" errors and "EOT timed out" errors with spi transfers.

These errors can be made easy to reproduce by reading the cpcap iio values in a loop while keeping the CPUs busy by also reading /dev/urandom.

The "RXS timed out" errors we can fix by adding spi-cpol and spi-cpha in addition to the spi-cs-high property we already have.

The "EOT timed out" errors we can fix by increasing the spi clock rate to 9.6 MHz. Looks similar MC13783 PMIC says it works at spi clock rates up to 20 MHz, so let's assume we can pick any rate up to 20 MHz also for cpcap.

---
 arch/arm/boot/dts/motorola-cpcap-mapphone.dtsi | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
arch/arm/boot/dts/motorola-cpcap-mapphone.dtsi +3 -1
index e39eee628afd6..08a7d3ce383f2 100644
@@ -13,8 +13,10 @@ cpcap: pmic@0 {
 		#interrupt-cells = <2>;
 		#address-cells = <1>;
 		#size-cells = <0>;
-		spi-max-frequency = <3000000>;
+		spi-max-frequency = <9600000>;
 		spi-cs-high;
+		spi-cpol;
+		spi-cpha;
 
 		cpcap_adc: adc {
 			compatible = "motorola,mapphone-cpcap-adc";
Signature · 1 line
2.25.1
author maintainer
#6 lore raw

PATCH 7/53 drm/ttm: Fix dma_fence refcnt leak when adding move fence

In reply to Sasha Levin

From: Xiyu Yang <[email protected]>

[ Upstream commit 11425c4519e2c974a100fc984867046d905b9380 ]

ttm_bo_add_move_fence() invokes dma_fence_get(), which returns a reference of the specified dma_fence object to "fence" with increased refcnt.

When ttm_bo_add_move_fence() returns, local variable "fence" becomes invalid, so the refcount should be decreased to keep refcount balanced.

The reference counting issue happens in one exception handling path of ttm_bo_add_move_fence(). When no_wait_gpu flag is equals to true, the function forgets to decrease the refcnt increased by dma_fence_get(), causing a refcnt leak.

Fix this issue by calling dma_fence_put() when no_wait_gpu flag is equals to true.

---
 drivers/gpu/drm/ttm/ttm_bo.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/gpu/drm/ttm/ttm_bo.c +3 -1
index 9e07c3f75156b..ef5bc00c73e23 100644
@@ -881,8 +881,10 @@ static int ttm_bo_add_move_fence(struct ttm_buffer_object *bo,
 	if (!fence)
 		return 0;
 
-	if (no_wait_gpu)
+	if (no_wait_gpu) {
+		dma_fence_put(fence);
 		return -EBUSY;
+	}
 
 	dma_resv_add_shared_fence(bo->base.resv, fence);
 
Signature · 6 lines
2.25.1

_______________________________________________
dri-devel mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dri-devel
author maintainer
#14 lore raw

PATCH 13/53 drm: mcde: Fix display initialization problem

In reply to Sasha Levin

From: Linus Walleij <[email protected]>

[ Upstream commit b984b6d8b52372b98cce0a6ff6c2787f50665b87 ]

The following bug appeared in the MCDE driver/display initialization during the recent merge window.

First the place we call drm_fbdev_generic_setup() in the wrong place: this needs to be called AFTER calling drm_dev_register() else we get this splat:

 ------------[ cut here ]------------
WARNING: CPU: 0 PID: 1 at ../drivers/gpu/drm/drm_fb_helper.c:2198 drm_fbdev_generic_setup+0x164/0x1a8
mcde a0350000.mcde: Device has not been registered.
Modules linked in:
Hardware name: ST-Ericsson Ux5x0 platform (Device Tree Support)
[<c010e704>] (unwind_backtrace) from [<c010a86c>] (show_stack+0x10/0x14)
[<c010a86c>] (show_stack) from [<c0414f38>] (dump_stack+0x9c/0xb0)
[<c0414f38>] (dump_stack) from [<c0121c8c>] (__warn+0xb8/0xd0)
[<c0121c8c>] (__warn) from [<c0121d18>] (warn_slowpath_fmt+0x74/0xb8)
[<c0121d18>] (warn_slowpath_fmt) from [<c04b154c>] (drm_fbdev_generic_setup+0x164/0x1a8)
[<c04b154c>] (drm_fbdev_generic_setup) from [<c04ed278>] (mcde_drm_bind+0xc4/0x160)
[<c04ed278>] (mcde_drm_bind) from [<c04f06b8>] (try_to_bring_up_master+0x15c/0x1a4)

(...)

---
 drivers/gpu/drm/mcde/mcde_drv.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/gpu/drm/mcde/mcde_drv.c +2 -1
index f28cb7a576ba4..1e7c5aa4d5e62 100644
@@ -208,7 +208,6 @@ static int mcde_modeset_init(struct drm_device *drm)
 
 	drm_mode_config_reset(drm);
 	drm_kms_helper_poll_init(drm);
-	drm_fbdev_generic_setup(drm, 32);
 
 	return 0;
 
@@ -275,6 +274,8 @@ static int mcde_drm_bind(struct device *dev)
 	if (ret < 0)
 		goto unbind;
 
+	drm_fbdev_generic_setup(drm, 32);
+
 	return 0;
 
 unbind:
Signature · 6 lines
2.25.1

_______________________________________________
dri-devel mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dri-devel
author maintainer
#22 lore raw

PATCH 17/53 net: usb: ax88179_178a: fix packet alignment padding

In reply to Sasha Levin

From: Jeremy Kerr <[email protected]>

[ Upstream commit e869e7a17798d85829fa7d4f9bbe1eebd4b2d3f6 ]

Using a AX88179 device (0b95:1790), I see two bytes of appended data on every RX packet. For example, this 48-byte ping, using 0xff as a payload byte:

04:20:22.528472 IP 192.168.1.1 > 192.168.1.2: ICMP echo request, id 2447, seq 1, length 64

	0x0000:  000a cd35 ea50 000a cd35 ea4f 0800 4500
	0x0010:  0054 c116 4000 4001 f63e c0a8 0101 c0a8
	0x0020:  0102 0800 b633 098f 0001 87ea cd5e 0000
	0x0030:  0000 dcf2 0600 0000 0000 ffff ffff ffff
	0x0040:  ffff ffff ffff ffff ffff ffff ffff ffff
	0x0050:  ffff ffff ffff ffff ffff ffff ffff ffff
	0x0060:  ffff 961f

Those last two bytes - 96 1f - aren't part of the original packet.

In the ax88179 RX path, the usbnet rx_fixup function trims a 2-byte 'alignment pseudo header' from the start of the packet, and sets the length from a per-packet field populated by hardware. It looks like that length field *includes* the 2-byte header; the current driver assumes that it's excluded.

This change trims the 2-byte alignment header after we've set the packet length, so the resulting packet length is correct. While we're moving the comment around, this also fixes the spelling of 'pseudo'.

---
 drivers/net/usb/ax88179_178a.c | 11 ++++++-----
 1 file changed, 6 insertions(+), 5 deletions(-)
drivers/net/usb/ax88179_178a.c +6 -5
index 93044cf1417a5..1fe4cc28d154d 100644
@@ -1414,10 +1414,10 @@ static int ax88179_rx_fixup(struct usbnet *dev, struct sk_buff *skb)
 		}
 
 		if (pkt_cnt == 0) {
-			/* Skip IP alignment psudo header */
-			skb_pull(skb, 2);
 			skb->len = pkt_len;
-			skb_set_tail_pointer(skb, pkt_len);
+			/* Skip IP alignment pseudo header */
+			skb_pull(skb, 2);
+			skb_set_tail_pointer(skb, skb->len);
 			skb->truesize = pkt_len + sizeof(struct sk_buff);
 			ax88179_rx_checksum(skb, pkt_hdr);
 			return 1;
@@ -1426,8 +1426,9 @@ static int ax88179_rx_fixup(struct usbnet *dev, struct sk_buff *skb)
 		ax_skb = skb_clone(skb, GFP_ATOMIC);
 		if (ax_skb) {
 			ax_skb->len = pkt_len;
-			ax_skb->data = skb->data + 2;
-			skb_set_tail_pointer(ax_skb, pkt_len);
+			/* Skip IP alignment pseudo header */
+			skb_pull(ax_skb, 2);
+			skb_set_tail_pointer(ax_skb, ax_skb->len);
 			ax_skb->truesize = pkt_len + sizeof(struct sk_buff);
 			ax88179_rx_checksum(ax_skb, pkt_hdr);
 			usbnet_skb_return(dev, ax_skb);
Signature · 1 line
2.25.1
author maintainer
#27 lore raw

PATCH 22/53 net: ethernet: mvneta: Fix Serdes configuration for SoCs without comphy

In reply to Sasha Levin

From: Sascha Hauer <[email protected]>

[ Upstream commit b4748553f53f2971e07d2619f13d461daac0f3bb ]

The MVNETA_SERDES_CFG register is only available on older SoCs like the Armada XP. On newer SoCs like the Armada 38x the fields are moved to comphy. This patch moves the writes to this register next to the comphy initialization, so that depending on the SoC either comphy or MVNETA_SERDES_CFG is configured.
With this we no longer write to the MVNETA_SERDES_CFG on SoCs where it doesn't exist.

---
 drivers/net/ethernet/marvell/mvneta.c | 80 +++++++++++++++------------
 1 file changed, 44 insertions(+), 36 deletions(-)
drivers/net/ethernet/marvell/mvneta.c +44 -36
index 43b44a1e8f69e..401eeeca89660 100644
@@ -106,6 +106,7 @@
 #define      MVNETA_TX_IN_PRGRS                  BIT(1)
 #define      MVNETA_TX_FIFO_EMPTY                BIT(8)
 #define MVNETA_RX_MIN_FRAME_SIZE                 0x247c
+/* Only exists on Armada XP and Armada 370 */
 #define MVNETA_SERDES_CFG			 0x24A0
 #define      MVNETA_SGMII_SERDES_PROTO		 0x0cc7
 #define      MVNETA_QSGMII_SERDES_PROTO		 0x0667
@@ -3523,26 +3524,55 @@ static int mvneta_setup_txqs(struct mvneta_port *pp)
 	return 0;
 }
 
-static int mvneta_comphy_init(struct mvneta_port *pp)
+static int mvneta_comphy_init(struct mvneta_port *pp, phy_interface_t interface)
 {
 	int ret;
 
-	if (!pp->comphy)
-		return 0;
-
-	ret = phy_set_mode_ext(pp->comphy, PHY_MODE_ETHERNET,
-			       pp->phy_interface);
+	ret = phy_set_mode_ext(pp->comphy, PHY_MODE_ETHERNET, interface);
 	if (ret)
 		return ret;
 
 	return phy_power_on(pp->comphy);
 }
 
+static int mvneta_config_interface(struct mvneta_port *pp,
+				   phy_interface_t interface)
+{
+	int ret = 0;
+
+	if (pp->comphy) {
+		if (interface == PHY_INTERFACE_MODE_SGMII ||
+		    interface == PHY_INTERFACE_MODE_1000BASEX ||
+		    interface == PHY_INTERFACE_MODE_2500BASEX) {
+			ret = mvneta_comphy_init(pp, interface);
+		}
+	} else {
+		switch (interface) {
+		case PHY_INTERFACE_MODE_QSGMII:
+			mvreg_write(pp, MVNETA_SERDES_CFG,
+				    MVNETA_QSGMII_SERDES_PROTO);
+			break;
+
+		case PHY_INTERFACE_MODE_SGMII:
+		case PHY_INTERFACE_MODE_1000BASEX:
+			mvreg_write(pp, MVNETA_SERDES_CFG,
+				    MVNETA_SGMII_SERDES_PROTO);
+			break;
+		default:
+			return -EINVAL;
+		}
+	}
+
+	pp->phy_interface = interface;
+
+	return ret;
+}
+
 static void mvneta_start_dev(struct mvneta_port *pp)
 {
 	int cpu;
 
-	WARN_ON(mvneta_comphy_init(pp));
+	WARN_ON(mvneta_config_interface(pp, pp->phy_interface));
 
 	mvneta_max_rx_size_set(pp, pp->pkt_size);
 	mvneta_txq_max_tx_size_set(pp, pp->pkt_size);
@@ -3920,14 +3950,10 @@ static void mvneta_mac_config(struct phylink_config *config, unsigned int mode,
 	if (state->speed == SPEED_2500)
 		new_ctrl4 |= MVNETA_GMAC4_SHORT_PREAMBLE_ENABLE;
 
-	if (pp->comphy && pp->phy_interface != state->interface &&
-	    (state->interface == PHY_INTERFACE_MODE_SGMII ||
-	     state->interface == PHY_INTERFACE_MODE_1000BASEX ||
-	     state->interface == PHY_INTERFACE_MODE_2500BASEX)) {
-		pp->phy_interface = state->interface;
-
-		WARN_ON(phy_power_off(pp->comphy));
-		WARN_ON(mvneta_comphy_init(pp));
+	if (pp->phy_interface != state->interface) {
+		if (pp->comphy)
+			WARN_ON(phy_power_off(pp->comphy));
+		WARN_ON(mvneta_config_interface(pp, state->interface));
 	}
 
 	if (new_ctrl0 != gmac_ctrl0)
@@ -4971,20 +4997,10 @@ static void mvneta_conf_mbus_windows(struct mvneta_port *pp,
 }
 
 /* Power up the port */
-static int mvneta_port_power_up(struct mvneta_port *pp, int phy_mode)
+static void mvneta_port_power_up(struct mvneta_port *pp, int phy_mode)
 {
 	/* MAC Cause register should be cleared */
 	mvreg_write(pp, MVNETA_UNIT_INTR_CAUSE, 0);
-
-	if (phy_mode == PHY_INTERFACE_MODE_QSGMII)
-		mvreg_write(pp, MVNETA_SERDES_CFG, MVNETA_QSGMII_SERDES_PROTO);
-	else if (phy_mode == PHY_INTERFACE_MODE_SGMII ||
-		 phy_interface_mode_is_8023z(phy_mode))
-		mvreg_write(pp, MVNETA_SERDES_CFG, MVNETA_SGMII_SERDES_PROTO);
-	else if (!phy_interface_mode_is_rgmii(phy_mode))
-		return -EINVAL;
-
-	return 0;
 }
 
 /* Device initialization routine */
@@ -5170,11 +5186,7 @@ static int mvneta_probe(struct platform_device *pdev)
 	if (err < 0)
 		goto err_netdev;
 
-	err = mvneta_port_power_up(pp, phy_mode);
-	if (err < 0) {
-		dev_err(&pdev->dev, "can't power up port\n");
-		goto err_netdev;
-	}
+	mvneta_port_power_up(pp, phy_mode);
 
 	/* Armada3700 network controller does not support per-cpu
 	 * operation, so only single NAPI should be initialized.
@@ -5328,11 +5340,7 @@ static int mvneta_resume(struct device *device)
 		}
 	}
 	mvneta_defaults_set(pp);
-	err = mvneta_port_power_up(pp, pp->phy_interface);
-	if (err < 0) {
-		dev_err(device, "can't power up port\n");
-		return err;
-	}
+	mvneta_port_power_up(pp, pp->phy_interface);
 
 	netif_device_attach(dev);
 
Signature · 1 line
2.25.1
author maintainer
#29 lore raw

PATCH 24/53 tg3: driver sleeps indefinitely when EEH errors exceed eeh_max_freezes

In reply to Sasha Levin

From: David Christensen <[email protected]>

[ Upstream commit 3a2656a211caf35e56afc9425e6e518fa52f7fbc ]

The driver function tg3_io_error_detected() calls napi_disable twice, without an intervening napi_enable, when the number of EEH errors exceeds eeh_max_freezes, resulting in an indefinite sleep while holding rtnl_lock.

Add check for pcierr_recovery which skips code already executed for the "Frozen" state.

---
 drivers/net/ethernet/broadcom/tg3.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/net/ethernet/broadcom/tg3.c +2 -2
index ff98a82b7bc47..d71ce7634ac19 100644
@@ -18170,8 +18170,8 @@ static pci_ers_result_t tg3_io_error_detected(struct pci_dev *pdev,
 
 	rtnl_lock();
 
-	/* We probably don't have netdev yet */
-	if (!netdev || !netif_running(netdev))
+	/* Could be second call or maybe we don't have netdev yet */
+	if (!netdev || tp->pcierr_recovery || !netif_running(netdev))
 		goto done;
 
 	/* We needn't recover from permanent error */
Signature · 1 line
2.25.1
author maintainer
#33 lore raw

PATCH 28/53 ibmvnic: continue to init in CRQ reset returns H_CLOSED

In reply to Sasha Levin

From: Dany Madden <[email protected]>

[ Upstream commit 8b40eb73509f5704a0e8cd25de0163876299f1a7 ]

Continue the reset path when partner adapter is not ready or H_CLOSED is returned from reset crq. This patch allows the CRQ init to proceed to establish a valid CRQ for traffic to flow after reset.

---
 drivers/net/ethernet/ibm/ibmvnic.c | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)
drivers/net/ethernet/ibm/ibmvnic.c +7 -2
index 1b4d04e4474bb..2dbcbdbb0e4ad 100644
@@ -1958,13 +1958,18 @@ static int do_reset(struct ibmvnic_adapter *adapter,
 			release_sub_crqs(adapter, 1);
 		} else {
 			rc = ibmvnic_reset_crq(adapter);
-			if (!rc)
+			if (rc == H_CLOSED || rc == H_SUCCESS) {
 				rc = vio_enable_interrupts(adapter->vdev);
+				if (rc)
+					netdev_err(adapter->netdev,
+						   "Reset failed to enable interrupts. rc=%d\n",
+						   rc);
+			}
 		}
 
 		if (rc) {
 			netdev_err(adapter->netdev,
-				   "Couldn't initialize crq. rc=%d\n", rc);
+				   "Reset couldn't initialize crq. rc=%d\n", rc);
 			goto out;
 		}
 
Signature · 1 line
2.25.1
author maintainer
#37 lore raw

PATCH 32/53 iommu/vt-d: Don't apply gfx quirks to untrusted devices

In reply to Sasha Levin

From: Rajat Jain <[email protected]>

[ Upstream commit 67e8a5b18d41af9298db5c17193f671f235cce01 ]

Currently, an external malicious PCI device can masquerade the VID:PID of faulty gfx devices, and thus apply iommu quirks to effectively disable the IOMMU restrictions for itself.

Thus we need to ensure that the device we are applying quirks to, is indeed an internal trusted device.

---
 drivers/iommu/intel-iommu.c | 37 +++++++++++++++++++++++++++++++++++++
 1 file changed, 37 insertions(+)
drivers/iommu/intel-iommu.c +37 -0
index fde7aba49b746..a29ed14839c41 100644
@@ -6200,6 +6200,23 @@ intel_iommu_domain_set_attr(struct iommu_domain *domain,
 	return ret;
 }
 
+/*
+ * Check that the device does not live on an external facing PCI port that is
+ * marked as untrusted. Such devices should not be able to apply quirks and
+ * thus not be able to bypass the IOMMU restrictions.
+ */
+static bool risky_device(struct pci_dev *pdev)
+{
+	if (pdev->untrusted) {
+		pci_info(pdev,
+			 "Skipping IOMMU quirk for dev [%04X:%04X] on untrusted PCI link\n",
+			 pdev->vendor, pdev->device);
+		pci_info(pdev, "Please check with your BIOS/Platform vendor about this\n");
+		return true;
+	}
+	return false;
+}
+
 const struct iommu_ops intel_iommu_ops = {
 	.capable		= intel_iommu_capable,
 	.domain_alloc		= intel_iommu_domain_alloc,
@@ -6229,6 +6246,9 @@ const struct iommu_ops intel_iommu_ops = {
 
 static void quirk_iommu_igfx(struct pci_dev *dev)
 {
+	if (risky_device(dev))
+		return;
+
 	pci_info(dev, "Disabling IOMMU for graphics on this chipset\n");
 	dmar_map_gfx = 0;
 }
@@ -6270,6 +6290,9 @@ DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_INTEL, 0x163D, quirk_iommu_igfx);
 
 static void quirk_iommu_rwbf(struct pci_dev *dev)
 {
+	if (risky_device(dev))
+		return;
+
 	/*
 	 * Mobile 4 Series Chipset neglects to set RWBF capability,
 	 * but needs it. Same seems to hold for the desktop versions.
@@ -6300,6 +6323,9 @@ static void quirk_calpella_no_shadow_gtt(struct pci_dev *dev)
 {
 	unsigned short ggc;
 
+	if (risky_device(dev))
+		return;
+
 	if (pci_read_config_word(dev, GGC, &ggc))
 		return;
 
@@ -6333,6 +6359,12 @@ static void __init check_tylersburg_isoch(void)
 	pdev = pci_get_device(PCI_VENDOR_ID_INTEL, 0x3a3e, NULL);
 	if (!pdev)
 		return;
+
+	if (risky_device(pdev)) {
+		pci_dev_put(pdev);
+		return;
+	}
+
 	pci_dev_put(pdev);
 
 	/* System Management Registers. Might be hidden, in which case
@@ -6342,6 +6374,11 @@ static void __init check_tylersburg_isoch(void)
 	if (!pdev)
 		return;
 
+	if (risky_device(pdev)) {
+		pci_dev_put(pdev);
+		return;
+	}
+
 	if (pci_read_config_dword(pdev, 0x188, &vtisochctrl)) {
 		pci_dev_put(pdev);
 		return;
Signature · 6 lines
2.25.1

_______________________________________________
iommu mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/iommu
author maintainer
#38 lore raw

PATCH 33/53 drm: panel-orientation-quirks: Add quirk for Asus T101HA panel

In reply to Sasha Levin

From: Hans de Goede <[email protected]>

[ Upstream commit 6c22bc18a3b93a38018844636557ad02e588e055 ]

Like the Asus T100HA the Asus T101HA also uses a panel which has been mounted 90 degrees rotated, albeit in the opposite direction. Add a quirk for this.

---
 drivers/gpu/drm/drm_panel_orientation_quirks.c | 6 ++++++
 1 file changed, 6 insertions(+)
drivers/gpu/drm/drm_panel_orientation_quirks.c +6 -0
index ffd95bfeaa94c..d11d83703931e 100644
@@ -121,6 +121,12 @@ static const struct dmi_system_id orientation_data[] = {
 		  DMI_EXACT_MATCH(DMI_PRODUCT_NAME, "T100HAN"),
 		},
 		.driver_data = (void *)&asus_t100ha,
+	}, {	/* Asus T101HA */
+		.matches = {
+		  DMI_EXACT_MATCH(DMI_SYS_VENDOR, "ASUSTeK COMPUTER INC."),
+		  DMI_EXACT_MATCH(DMI_PRODUCT_NAME, "T101HA"),
+		},
+		.driver_data = (void *)&lcd800x1280_rightside_up,
 	}, {	/* GPD MicroPC (generic strings, also match on bios date) */
 		.matches = {
 		  DMI_EXACT_MATCH(DMI_SYS_VENDOR, "Default string"),
Signature · 6 lines
2.25.1

_______________________________________________
dri-devel mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dri-devel
author maintainer
#39 lore raw

PATCH 34/53 drm: panel-orientation-quirks: Use generic orientation-data for Acer S1003

In reply to Sasha Levin

From: Hans de Goede <[email protected]>

[ Upstream commit a05caf9e62a85d12da27e814ac13195f4683f21c ]

The Acer S1003 has proper DMI strings for sys-vendor and product-name, so we do not need to match by BIOS-date.

This means that the Acer S1003 can use the generic lcd800x1280_rightside_up drm_dmi_panel_orientation_data struct which is also used by other quirks.

---
 drivers/gpu/drm/drm_panel_orientation_quirks.c | 8 +-------
 1 file changed, 1 insertion(+), 7 deletions(-)
drivers/gpu/drm/drm_panel_orientation_quirks.c +1 -7
index d11d83703931e..d00ea384dcbfe 100644
@@ -30,12 +30,6 @@ struct drm_dmi_panel_orientation_data {
 	int orientation;
 };
 
-static const struct drm_dmi_panel_orientation_data acer_s1003 = {
-	.width = 800,
-	.height = 1280,
-	.orientation = DRM_MODE_PANEL_ORIENTATION_RIGHT_UP,
-};
-
 static const struct drm_dmi_panel_orientation_data asus_t100ha = {
 	.width = 800,
 	.height = 1280,
@@ -114,7 +108,7 @@ static const struct dmi_system_id orientation_data[] = {
 		  DMI_EXACT_MATCH(DMI_SYS_VENDOR, "Acer"),
 		  DMI_EXACT_MATCH(DMI_PRODUCT_NAME, "One S1003"),
 		},
-		.driver_data = (void *)&acer_s1003,
+		.driver_data = (void *)&lcd800x1280_rightside_up,
 	}, {	/* Asus T100HA */
 		.matches = {
 		  DMI_EXACT_MATCH(DMI_SYS_VENDOR, "ASUSTeK COMPUTER INC."),
Signature · 6 lines
2.25.1

_______________________________________________
dri-devel mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dri-devel
author maintainer
#40 lore raw

PATCH 35/53 s390/kasan: fix early pgm check handler execution

In reply to Sasha Levin

From: Vasily Gorbik <[email protected]>

[ Upstream commit 998f5bbe3dbdab81c1cfb1aef7c3892f5d24f6c7 ]

Currently if early_pgm_check_handler is called it ends up in pgm check loop. The problem is that early_pgm_check_handler is instrumented by KASAN but executed without DAT flag enabled which leads to addressing exception when KASAN checks try to access shadow memory.

Fix that by executing early handlers with DAT flag on under KASAN as expected.

---
 arch/s390/kernel/early.c | 2 ++
 1 file changed, 2 insertions(+)
arch/s390/kernel/early.c +2 -0
index cd241ee66eff4..0782772318580 100644
@@ -170,6 +170,8 @@ static noinline __init void setup_lowcore_early(void)
 	psw_t psw;
 
 	psw.mask = PSW_MASK_BASE | PSW_DEFAULT_KEY | PSW_MASK_EA | PSW_MASK_BA;
+	if (IS_ENABLED(CONFIG_KASAN))
+		psw.mask |= PSW_MASK_DAT;
 	psw.addr = (unsigned long) s390_base_ext_handler;
 	S390_lowcore.external_new_psw = psw;
 	psw.addr = (unsigned long) s390_base_pgm_handler;
Signature · 1 line
2.25.1
author maintainer
#45 lore raw

PATCH 38/53 drm/sun4i: mixer: Call of_dma_configure if there's an IOMMU

In reply to Sasha Levin

From: Maxime Ripard <[email protected]>

[ Upstream commit 842ec61f4006a6477a9deaedd69131e9f46e4cb5 ]

The main DRM device is actually a virtual device so it doesn't have the iommus property, which is instead on the DMA masters, in this case the mixers.

Add a call to of_dma_configure with the mixers DT node but on the DRM virtual device to configure it in the same way than the mixers.

[Maxime: Applied to -fixes since it missed the merge window and display is
broken without it]

---
 drivers/gpu/drm/sun4i/sun8i_mixer.c | 13 +++++++++++++
 1 file changed, 13 insertions(+)
drivers/gpu/drm/sun4i/sun8i_mixer.c +13 -0
index 4a64f7ae437a8..19b3b41847046 100644
@@ -452,6 +452,19 @@ static int sun8i_mixer_bind(struct device *dev, struct device *master,
 	mixer->engine.ops = &sun8i_engine_ops;
 	mixer->engine.node = dev->of_node;
 
+	if (of_find_property(dev->of_node, "iommus", NULL)) {
+		/*
+		 * This assume we have the same DMA constraints for
+		 * all our the mixers in our pipeline. This sounds
+		 * bad, but it has always been the case for us, and
+		 * DRM doesn't do per-device allocation either, so we
+		 * would need to fix DRM first...
+		 */
+		ret = of_dma_configure(drm->dev, dev->of_node, true);
+		if (ret)
+			return ret;
+	}
+
 	/*
 	 * While this function can fail, we shouldn't do anything
 	 * if this happens. Some early DE2 DT entries don't provide
Signature · 6 lines
2.25.1

_______________________________________________
dri-devel mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dri-devel
author maintainer
#47 lore raw

PATCH 40/53 cifs: update ctime and mtime during truncate

In reply to Sasha Levin

From: Zhang Xiaoxu <[email protected]>

[ Upstream commit 5618303d8516f8ac5ecfe53ee8e8bc9a40eaf066 ]

As the man description of the truncate, if the size changed, then the st_ctime and st_mtime fields should be updated. But in cifs, we doesn't do it.

It lead the xfstests generic/313 failed.

So, add the ATTR_MTIME|ATTR_CTIME flags on attrs when change the file size

---
 fs/cifs/inode.c | 9 +++++++++
 1 file changed, 9 insertions(+)
fs/cifs/inode.c +9 -0
index 5d2965a237305..15f2cdc71ac98 100644
@@ -2344,6 +2344,15 @@ cifs_set_file_size(struct inode *inode, struct iattr *attrs,
 	if (rc == 0) {
 		cifsInode->server_eof = attrs->ia_size;
 		cifs_setsize(inode, attrs->ia_size);
+
+		/*
+		 * The man page of truncate says if the size changed,
+		 * then the st_ctime and st_mtime fields for the file
+		 * are updated.
+		 */
+		attrs->ia_ctime = attrs->ia_mtime = current_time(inode);
+		attrs->ia_valid |= ATTR_CTIME | ATTR_MTIME;
+
 		cifs_truncate_page(inode->i_mapping, inode->i_size);
 	}
 
Signature · 1 line
2.25.1
author maintainer
#48 lore raw

PATCH 41/53 ARM: imx6: add missing put_device() call in imx6q_suspend_init()

In reply to Sasha Levin

From: yu kuai <[email protected]>

[ Upstream commit 4845446036fc9c13f43b54a65c9b757c14f5141b ]

if of_find_device_by_node() succeed, imx6q_suspend_init() doesn't have a corresponding put_device(). Thus add a jump target to fix the exception handling for this function implementation.

---
 arch/arm/mach-imx/pm-imx6.c | 10 ++++++----
 1 file changed, 6 insertions(+), 4 deletions(-)
arch/arm/mach-imx/pm-imx6.c +6 -4
index dd34dff137626..40c74b4c4d730 100644
@@ -493,14 +493,14 @@ static int __init imx6q_suspend_init(const struct imx6_pm_socdata *socdata)
 	if (!ocram_pool) {
 		pr_warn("%s: ocram pool unavailable!\n", __func__);
 		ret = -ENODEV;
-		goto put_node;
+		goto put_device;
 	}
 
 	ocram_base = gen_pool_alloc(ocram_pool, MX6Q_SUSPEND_OCRAM_SIZE);
 	if (!ocram_base) {
 		pr_warn("%s: unable to alloc ocram!\n", __func__);
 		ret = -ENOMEM;
-		goto put_node;
+		goto put_device;
 	}
 
 	ocram_pbase = gen_pool_virt_to_phys(ocram_pool, ocram_base);
@@ -523,7 +523,7 @@ static int __init imx6q_suspend_init(const struct imx6_pm_socdata *socdata)
 	ret = imx6_pm_get_base(&pm_info->mmdc_base, socdata->mmdc_compat);
 	if (ret) {
 		pr_warn("%s: failed to get mmdc base %d!\n", __func__, ret);
-		goto put_node;
+		goto put_device;
 	}
 
 	ret = imx6_pm_get_base(&pm_info->src_base, socdata->src_compat);
@@ -570,7 +570,7 @@ static int __init imx6q_suspend_init(const struct imx6_pm_socdata *socdata)
 		&imx6_suspend,
 		MX6Q_SUSPEND_OCRAM_SIZE - sizeof(*pm_info));
 
-	goto put_node;
+	goto put_device;
 
 pl310_cache_map_failed:
 	iounmap(pm_info->gpc_base.vbase);
@@ -580,6 +580,8 @@ static int __init imx6q_suspend_init(const struct imx6_pm_socdata *socdata)
 	iounmap(pm_info->src_base.vbase);
 src_map_failed:
 	iounmap(pm_info->mmdc_base.vbase);
+put_device:
+	put_device(&pdev->dev);
 put_node:
 	of_node_put(node);
 
Signature · 1 line
2.25.1
author maintainer
#50 lore raw

PATCH 43/53 scsi: mptscsih: Fix read sense data size

In reply to Sasha Levin

From: Tomas Henzl <[email protected]>

[ Upstream commit afe89f115e84edbc76d316759e206580a06c6973 ]

The sense data buffer in sense_buf_pool is allocated with size of MPT_SENSE_BUFFER_ALLOC(64) (multiplied by req_depth) while SNS_LEN(sc)(96) is used when reading the data. That may lead to a read from unallocated area, sometimes from another (unallocated) page. To fix this, limit the read size to MPT_SENSE_BUFFER_ALLOC.

---
 drivers/message/fusion/mptscsih.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)
drivers/message/fusion/mptscsih.c +1 -3
index f0737c57ed5fc..1491561d2e5c9 100644
@@ -118,8 +118,6 @@ int 		mptscsih_suspend(struct pci_dev *pdev, pm_message_t state);
 int 		mptscsih_resume(struct pci_dev *pdev);
 #endif
 
-#define SNS_LEN(scp)	SCSI_SENSE_BUFFERSIZE
-
 
 /*=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=*/
 /*
@@ -2422,7 +2420,7 @@ mptscsih_copy_sense_data(struct scsi_cmnd *sc, MPT_SCSI_HOST *hd, MPT_FRAME_HDR
 		/* Copy the sense received into the scsi command block. */
 		req_index = le16_to_cpu(mf->u.frame.hwhdr.msgctxu.fld.req_idx);
 		sense_data = ((u8 *)ioc->sense_buf_pool + (req_index * MPT_SENSE_BUFFER_ALLOC));
-		memcpy(sc->sense_buffer, sense_data, SNS_LEN(sc));
+		memcpy(sc->sense_buffer, sense_data, MPT_SENSE_BUFFER_ALLOC);
 
 		/* Log SMART data (asc = 0x5D, non-IM case only) if required.
 		 */
Signature · 1 line
2.25.1
author maintainer
#56 lore raw

PATCH 49/53 io_uring: fix current->mm NULL dereference on exit

In reply to Sasha Levin

From: Pavel Begunkov <[email protected]>

[ Upstream commit d60b5fbc1ce8210759b568da49d149b868e7c6d3 ]

Don't reissue requests from io_iopoll_reap_events(), the task may not have mm, which ends up with NULL. It's better to kill everything off on exit anyway.

[  677.734670] RIP: 0010:io_iopoll_complete+0x27e/0x630
...
[  677.734679] Call Trace:
[  677.734695]  ? __send_signal+0x1f2/0x420
[  677.734698]  ? _raw_spin_unlock_irqrestore+0x24/0x40
[  677.734699]  ? send_signal+0xf5/0x140
[  677.734700]  io_iopoll_getevents+0x12f/0x1a0
[  677.734702]  io_iopoll_reap_events.part.0+0x5e/0xa0
[  677.734703]  io_ring_ctx_wait_and_kill+0x132/0x1c0
[  677.734704]  io_uring_release+0x20/0x30
[  677.734706]  __fput+0xcd/0x230
[  677.734707]  ____fput+0xe/0x10
[  677.734709]  task_work_run+0x67/0xa0
[  677.734710]  do_exit+0x35d/0xb70
[  677.734712]  do_group_exit+0x43/0xa0
[  677.734713]  get_signal+0x140/0x900
[  677.734715]  do_signal+0x37/0x780
[  677.734717]  ? enqueue_hrtimer+0x41/0xb0
[  677.734718]  ? recalibrate_cpu_khz+0x10/0x10
[  677.734720]  ? ktime_get+0x3e/0xa0
[  677.734721]  ? lapic_next_deadline+0x26/0x30
[  677.734723]  ? tick_program_event+0x4d/0x90
[  677.734724]  ? __hrtimer_get_next_event+0x4d/0x80
[  677.734726]  __prepare_exit_to_usermode+0x126/0x1c0
[  677.734741]  prepare_exit_to_usermode+0x9/0x40
[  677.734742]  idtentry_exit_cond_rcu+0x4c/0x60
[  677.734743]  sysvec_reschedule_ipi+0x92/0x160
[  677.734744]  ? asm_sysvec_reschedule_ipi+0xa/0x20
[  677.734745]  asm_sysvec_reschedule_ipi+0x12/0x20
---
 fs/io_uring.c | 9 +++++++++
 1 file changed, 9 insertions(+)
fs/io_uring.c +9 -0
index 6cf9d509371e2..43dc745727408 100644
@@ -858,6 +858,7 @@ static int __io_sqe_files_update(struct io_ring_ctx *ctx,
 				 struct io_uring_files_update *ip,
 				 unsigned nr_args);
 static int io_grab_files(struct io_kiocb *req);
+static void io_complete_rw_common(struct kiocb *kiocb, long res);
 static void io_cleanup_req(struct io_kiocb *req);
 static int io_file_get(struct io_submit_state *state, struct io_kiocb *req,
 		       int fd, struct file **out_file, bool fixed);
@@ -1697,6 +1698,14 @@ static void io_iopoll_queue(struct list_head *again)
 	do {
 		req = list_first_entry(again, struct io_kiocb, list);
 		list_del(&req->list);
+
+		/* shouldn't happen unless io_uring is dying, cancel reqs */
+		if (unlikely(!current->mm)) {
+			io_complete_rw_common(&req->rw.kiocb, -EAGAIN);
+			io_put_req(req);
+			continue;
+		}
+
 		refcount_inc(&req->refs);
 		io_queue_async_work(req);
 	} while (!list_empty(again));
Signature · 1 line
2.25.1
author maintainer
#59 lore raw

PATCH 52/53 kgdb: Avoid suspicious RCU usage warning

In reply to Sasha Levin

From: Douglas Anderson <[email protected]>

[ Upstream commit 440ab9e10e2e6e5fd677473ee6f9e3af0f6904d6 ]

At times when I'm using kgdb I see a splat on my console about suspicious RCU usage. I managed to come up with a case that could reproduce this that looked like this:

WARNING: suspicious RCU usage
5.7.0-rc4+ #609 Not tainted
-----------------------------
kernel/pid.c:395 find_task_by_pid_ns() needs rcu_read_lock() protection!

other info that might help us debug this:

rcu_scheduler_active = 2, debug_locks = 1
3 locks held by swapper/0/1:
#0: ffffff81b6b8e988 (&dev->mutex){....}-{3:3}, at: __device_attach+0x40/0x13c #1: ffffffd01109e9e8 (dbg_master_lock){....}-{2:2}, at: kgdb_cpu_enter+0x20c/0x7ac #2: ffffffd01109ea90 (dbg_slave_lock){....}-{2:2}, at: kgdb_cpu_enter+0x3ec/0x7ac

  stack backtrace:
  CPU: 7 PID: 1 Comm: swapper/0 Not tainted 5.7.0-rc4+ #609
  Hardware name: Google Cheza (rev3+) (DT)
  Call trace:
   dump_backtrace+0x0/0x1b8
   show_stack+0x1c/0x24
   dump_stack+0xd4/0x134
   lockdep_rcu_suspicious+0xf0/0x100
   find_task_by_pid_ns+0x5c/0x80
   getthread+0x8c/0xb0
   gdb_serial_stub+0x9d4/0xd04
   kgdb_cpu_enter+0x284/0x7ac
   kgdb_handle_exception+0x174/0x20c
   kgdb_brk_fn+0x24/0x30
   call_break_hook+0x6c/0x7c
   brk_handler+0x20/0x5c
   do_debug_exception+0x1c8/0x22c
   el1_sync_handler+0x3c/0xe4
   el1_sync+0x7c/0x100
   rpmh_rsc_probe+0x38/0x420
   platform_drv_probe+0x94/0xb4
   really_probe+0x134/0x300
   driver_probe_device+0x68/0x100
   __device_attach_driver+0x90/0xa8
   bus_for_each_drv+0x84/0xcc
   __device_attach+0xb4/0x13c
   device_initial_probe+0x18/0x20
   bus_probe_device+0x38/0x98
   device_add+0x38c/0x420

If I understand properly we should just be able to blanket kgdb under one big RCU read lock and the problem should go away. We'll add it to the beast-of-a-function known as kgdb_cpu_enter().

With this I no longer get any splats and things seem to work fine.

---
 kernel/debug/debug_core.c | 4 ++++
 1 file changed, 4 insertions(+)
kernel/debug/debug_core.c +4 -0
index d47c7d6656cd3..9be6accf8fe3d 100644
@@ -577,6 +577,7 @@ static int kgdb_cpu_enter(struct kgdb_state *ks, struct pt_regs *regs,
 		arch_kgdb_ops.disable_hw_break(regs);
 
 acquirelock:
+	rcu_read_lock();
 	/*
 	 * Interrupts will be restored by the 'trap return' code, except when
 	 * single stepping.
@@ -636,6 +637,7 @@ static int kgdb_cpu_enter(struct kgdb_state *ks, struct pt_regs *regs,
 			atomic_dec(&slaves_in_kgdb);
 			dbg_touch_watchdogs();
 			local_irq_restore(flags);
+			rcu_read_unlock();
 			return 0;
 		}
 		cpu_relax();
@@ -654,6 +656,7 @@ static int kgdb_cpu_enter(struct kgdb_state *ks, struct pt_regs *regs,
 		raw_spin_unlock(&dbg_master_lock);
 		dbg_touch_watchdogs();
 		local_irq_restore(flags);
+		rcu_read_unlock();
 
 		goto acquirelock;
 	}
@@ -777,6 +780,7 @@ static int kgdb_cpu_enter(struct kgdb_state *ks, struct pt_regs *regs,
 	raw_spin_unlock(&dbg_master_lock);
 	dbg_touch_watchdogs();
 	local_irq_restore(flags);
+	rcu_read_unlock();
 
 	return kgdb_info[cpu].ret_state;
 }
Signature · 1 line
2.25.1
author maintainer
#60 lore raw

PATCH 53/53 sched/core: Check cpus_mask, not cpus_ptr in __set_cpus_allowed_ptr(), to fix mask corruption

In reply to Sasha Levin

From: Scott Wood <[email protected]>

[ Upstream commit fd844ba9ae59b51e34e77105d79f8eca780b3bd6 ]

This function is concerned with the long-term CPU mask, not the transitory mask the task might have while migrate disabled. Before this patch, if a task was migrate-disabled at the time __set_cpus_allowed_ptr() was called, and the new mask happened to be equal to the CPU that the task was running on, then the mask update would be lost.

---
 kernel/sched/core.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
kernel/sched/core.c +1 -1
index 5eccfb816d23b..5f8130c27c7ab 100644
@@ -1637,7 +1637,7 @@ static int __set_cpus_allowed_ptr(struct task_struct *p,
 		goto out;
 	}
 
-	if (cpumask_equal(p->cpus_ptr, new_mask))
+	if (cpumask_equal(&p->cpus_mask, new_mask))
 		goto out;
 
 	/*
Signature · 1 line
2.25.1
author maintainer
#54 lore raw

PATCH 47/53 nvme-rdma: assign completion vector correctly

In reply to Sasha Levin

From: Max Gurtovoy <[email protected]>

[ Upstream commit 032a9966a22a3596addf81dacf0c1736dfedc32a ]

The completion vector index that is given during CQ creation can't exceed the number of support vectors by the underlying RDMA device. This violation currently can accure, for example, in case one will try to connect with N regular read/write queues and M poll queues and the sum of N + M > num_supported_vectors. This will lead to failure in establish a connection to remote target. Instead, in that case, share a completion vector between queues.

---
 drivers/nvme/host/rdma.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/nvme/host/rdma.c +1 -1
index cac8a930396a0..1f9a45145d0d3 100644
@@ -443,7 +443,7 @@ static int nvme_rdma_create_queue_ib(struct nvme_rdma_queue *queue)
 	 * Spread I/O queues completion vectors according their queue index.
 	 * Admin queues can always go on completion vector 0.
 	 */
-	comp_vector = idx == 0 ? idx : idx - 1;
+	comp_vector = (idx == 0 ? idx : idx - 1) % ibdev->num_comp_vectors;
 
 	/* Polling queues need direct cq polling context */
 	if (nvme_rdma_poll_queue(queue))
Signature · 7 lines
2.25.1


_______________________________________________
Linux-nvme mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-nvme
author maintainer
#55 lore raw

PATCH 48/53 x86/entry: Increase entry_stack size to a full page

In reply to Sasha Levin

From: Peter Zijlstra <[email protected]>

[ Upstream commit c7aadc09321d8f9a1d3bd1e6d8a47222ecddf6c5 ]

Marco crashed in bad_iret with a Clang11/KCSAN build due to overflowing the stack. Now that we run C code on it, expand it to a full page.

---
 arch/x86/include/asm/processor.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/x86/include/asm/processor.h +1 -1
index 3bcf27caf6c9f..d64b72346dd27 100644
@@ -369,7 +369,7 @@ struct x86_hw_tss {
 #define IO_BITMAP_OFFSET_INVALID	(__KERNEL_TSS_LIMIT + 1)
 
 struct entry_stack {
-	unsigned long		words[64];
+	char	stack[PAGE_SIZE];
 };
 
 struct entry_stack_page {
Signature · 1 line
2.25.1
author maintainer
#21 lore raw

PATCH 16/53 ASoC: hdac_hda: fix memleak with regmap not freed on remove

In reply to Sasha Levin

From: Pierre-Louis Bossart <[email protected]>

[ Upstream commit a94eaccefea1186947c5c5451fcae2245dd7e714 ]

kmemleak throws error reports on module load/unload tests, add snd_hdac_regmap_exit() in .remove().

While we are at it, also fix the error handling flow in .probe() to use snd_hdac_regmap_exit() if needed.

---
 sound/soc/codecs/hdac_hda.c | 16 ++++++++++------
 1 file changed, 10 insertions(+), 6 deletions(-)
sound/soc/codecs/hdac_hda.c +10 -6
index de003acb1951f..473efe9ef998a 100644
@@ -441,13 +441,13 @@ static int hdac_hda_codec_probe(struct snd_soc_component *component)
 	ret = snd_hda_codec_set_name(hcodec, hcodec->preset->name);
 	if (ret < 0) {
 		dev_err(&hdev->dev, "name failed %s\n", hcodec->preset->name);
-		goto error;
+		goto error_pm;
 	}
 
 	ret = snd_hdac_regmap_init(&hcodec->core);
 	if (ret < 0) {
 		dev_err(&hdev->dev, "regmap init failed\n");
-		goto error;
+		goto error_pm;
 	}
 
 	patch = (hda_codec_patch_t)hcodec->preset->driver_data;
@@ -455,7 +455,7 @@ static int hdac_hda_codec_probe(struct snd_soc_component *component)
 		ret = patch(hcodec);
 		if (ret < 0) {
 			dev_err(&hdev->dev, "patch failed %d\n", ret);
-			goto error;
+			goto error_regmap;
 		}
 	} else {
 		dev_dbg(&hdev->dev, "no patch file found\n");
@@ -467,7 +467,7 @@ static int hdac_hda_codec_probe(struct snd_soc_component *component)
 	ret = snd_hda_codec_parse_pcms(hcodec);
 	if (ret < 0) {
 		dev_err(&hdev->dev, "unable to map pcms to dai %d\n", ret);
-		goto error;
+		goto error_regmap;
 	}
 
 	/* HDMI controls need to be created in machine drivers */
@@ -476,7 +476,7 @@ static int hdac_hda_codec_probe(struct snd_soc_component *component)
 		if (ret < 0) {
 			dev_err(&hdev->dev, "unable to create controls %d\n",
 				ret);
-			goto error;
+			goto error_regmap;
 		}
 	}
 
@@ -496,7 +496,9 @@ static int hdac_hda_codec_probe(struct snd_soc_component *component)
 
 	return 0;
 
-error:
+error_regmap:
+	snd_hdac_regmap_exit(hdev);
+error_pm:
 	pm_runtime_put(&hdev->dev);
 error_no_pm:
 	snd_hdac_ext_bus_link_put(hdev->bus, hlink);
@@ -518,6 +520,8 @@ static void hdac_hda_codec_remove(struct snd_soc_component *component)
 
 	pm_runtime_disable(&hdev->dev);
 	snd_hdac_ext_bus_link_put(hdev->bus, hlink);
+
+	snd_hdac_regmap_exit(hdev);
 }
 
 static const struct snd_soc_dapm_route hdac_hda_dapm_routes[] = {
Signature · 1 line
2.25.1
author maintainer
#58 lore raw

PATCH 51/53 nfs: Fix memory leak of export_path

In reply to Sasha Levin

From: Tom Rix <[email protected]>

[ Upstream commit 4659ed7cc8514369043053463514408ca16ad6f3 ]

The try_location function is called within a loop by nfs_follow_referral. try_location calls nfs4_pathname_string to created the export_path. nfs4_pathname_string allocates the memory. export_path is stored in the nfs_fs_context/fs_context structure similarly as hostname and source. But whereas the ctx hostname and source are freed before assignment, export_path is not. So if there are multiple loops, the new export_path will overwrite the old without the old being freed.

So call kfree for export_path.

---
 fs/nfs/nfs4namespace.c | 1 +
 1 file changed, 1 insertion(+)
fs/nfs/nfs4namespace.c +1 -0
index a3ab6e219061b..873342308dc0d 100644
@@ -308,6 +308,7 @@ static int try_location(struct fs_context *fc,
 	if (IS_ERR(export_path))
 		return PTR_ERR(export_path);
 
+	kfree(ctx->nfs_server.export_path);
 	ctx->nfs_server.export_path = export_path;
 
 	source = kmalloc(len + 1 + ctx->nfs_server.export_path_len + 1,
Signature · 1 line
2.25.1
author maintainer
#57 lore raw

PATCH 50/53 arm64: Add KRYO{3,4}XX silver CPU cores to SSB safelist

In reply to Sasha Levin

From: Sai Prakash Ranjan <[email protected]>

[ Upstream commit 108447fd0d1a34b0929cd26dc637c917a734ebab ]

QCOM KRYO{3,4}XX silver/LITTLE CPU cores are based on Cortex-A55 and are SSB safe, hence add them to SSB safelist -> arm64_ssb_cpus[].

---
 arch/arm64/kernel/cpu_errata.c | 2 ++
 1 file changed, 2 insertions(+)
arch/arm64/kernel/cpu_errata.c +2 -0
index df56d2295d165..0f37045fafab3 100644
@@ -460,6 +460,8 @@ static const struct midr_range arm64_ssb_cpus[] = {
 	MIDR_ALL_VERSIONS(MIDR_CORTEX_A53),
 	MIDR_ALL_VERSIONS(MIDR_CORTEX_A55),
 	MIDR_ALL_VERSIONS(MIDR_BRAHMA_B53),
+	MIDR_ALL_VERSIONS(MIDR_QCOM_KRYO_3XX_SILVER),
+	MIDR_ALL_VERSIONS(MIDR_QCOM_KRYO_4XX_SILVER),
 	{},
 };
 
Signature · 1 line
2.25.1
author maintainer
#53 lore raw

PATCH 46/53 block: release bip in a right way in error path

In reply to Sasha Levin

From: Chengguang Xu <[email protected]>

[ Upstream commit 0b8eb629a700c0ef15a437758db8255f8444e76c ]

Release bip using kfree() in error path when that was allocated by kmalloc().

---
 block/bio-integrity.c | 23 ++++++++++++++---------
 1 file changed, 14 insertions(+), 9 deletions(-)
block/bio-integrity.c +14 -9
index bf62c25cde8f4..1d173feb38831 100644
@@ -24,6 +24,18 @@ void blk_flush_integrity(void)
 	flush_workqueue(kintegrityd_wq);
 }
 
+void __bio_integrity_free(struct bio_set *bs, struct bio_integrity_payload *bip)
+{
+	if (bs && mempool_initialized(&bs->bio_integrity_pool)) {
+		if (bip->bip_vec)
+			bvec_free(&bs->bvec_integrity_pool, bip->bip_vec,
+				  bip->bip_slab);
+		mempool_free(bip, &bs->bio_integrity_pool);
+	} else {
+		kfree(bip);
+	}
+}
+
 /**
  * bio_integrity_alloc - Allocate integrity payload and attach it to bio
  * @bio:	bio to attach integrity metadata to
@@ -75,7 +87,7 @@ struct bio_integrity_payload *bio_integrity_alloc(struct bio *bio,
 
 	return bip;
 err:
-	mempool_free(bip, &bs->bio_integrity_pool);
+	__bio_integrity_free(bs, bip);
 	return ERR_PTR(-ENOMEM);
 }
 EXPORT_SYMBOL(bio_integrity_alloc);
@@ -96,14 +108,7 @@ void bio_integrity_free(struct bio *bio)
 		kfree(page_address(bip->bip_vec->bv_page) +
 		      bip->bip_vec->bv_offset);
 
-	if (bs && mempool_initialized(&bs->bio_integrity_pool)) {
-		bvec_free(&bs->bvec_integrity_pool, bip->bip_vec, bip->bip_slab);
-
-		mempool_free(bip, &bs->bio_integrity_pool);
-	} else {
-		kfree(bip);
-	}
-
+	__bio_integrity_free(bs, bip);
 	bio->bi_integrity = NULL;
 	bio->bi_opf &= ~REQ_INTEGRITY;
 }
Signature · 1 line
2.25.1
author maintainer
#52 lore raw

PATCH 45/53 arm64: kpti: Add KRYO{3, 4}XX silver CPU cores to kpti safelist

In reply to Sasha Levin

From: Sai Prakash Ranjan <[email protected]>

[ Upstream commit f4617be35b4b547e82d30993f56d631dfc2d5f88 ]

QCOM KRYO{3,4}XX silver/LITTLE CPU cores are based on Cortex-A55 and are meltdown safe, hence add them to kpti_safe_list[].

---
 arch/arm64/kernel/cpufeature.c | 2 ++
 1 file changed, 2 insertions(+)
arch/arm64/kernel/cpufeature.c +2 -0
index 9fac745aa7bb2..b0fb1d5bf2235 100644
@@ -1059,6 +1059,8 @@ static bool unmap_kernel_at_el0(const struct arm64_cpu_capabilities *entry,
 		MIDR_ALL_VERSIONS(MIDR_CORTEX_A73),
 		MIDR_ALL_VERSIONS(MIDR_HISI_TSV110),
 		MIDR_ALL_VERSIONS(MIDR_NVIDIA_CARMEL),
+		MIDR_ALL_VERSIONS(MIDR_QCOM_KRYO_3XX_SILVER),
+		MIDR_ALL_VERSIONS(MIDR_QCOM_KRYO_4XX_SILVER),
 		{ /* sentinel */ }
 	};
 	char const *str = "kpti command line option";
Signature · 1 line
2.25.1
author maintainer
#15 lore raw

PATCH 14/53 ASoC: SOF: Intel: add PCI ID for CometLake-S

In reply to Sasha Levin

From: Pierre-Louis Bossart <[email protected]>

[ Upstream commit 258fb4f4c34a0db9d3834aba6784d7b322176bb9 ]

Mirror ID added for legacy HDaudio

---
 sound/soc/sof/sof-pci-dev.c | 2 ++
 1 file changed, 2 insertions(+)
sound/soc/sof/sof-pci-dev.c +2 -0
index cec631a1389b5..7b1846aeadd59 100644
@@ -427,6 +427,8 @@ static const struct pci_device_id sof_pci_ids[] = {
 #if IS_ENABLED(CONFIG_SND_SOC_SOF_COMETLAKE_H)
 	{ PCI_DEVICE(0x8086, 0x06c8),
 		.driver_data = (unsigned long)&cml_desc},
+	{ PCI_DEVICE(0x8086, 0xa3f0), /* CML-S */
+		.driver_data = (unsigned long)&cml_desc},
 #endif
 #if IS_ENABLED(CONFIG_SND_SOC_SOF_TIGERLAKE)
 	{ PCI_DEVICE(0x8086, 0xa0c8),
Signature · 1 line
2.25.1
author maintainer
#51 lore raw

PATCH 44/53 usb: dwc3: pci: Fix reference count leak in dwc3_pci_resume_work

In reply to Sasha Levin

From: Aditya Pakki <[email protected]>

[ Upstream commit 2655971ad4b34e97dd921df16bb0b08db9449df7 ]

dwc3_pci_resume_work() calls pm_runtime_get_sync() that increments the reference counter. In case of failure, decrement the reference before returning.

---
 drivers/usb/dwc3/dwc3-pci.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/usb/dwc3/dwc3-pci.c +3 -1
index b67372737dc9b..96c05b121fac8 100644
@@ -206,8 +206,10 @@ static void dwc3_pci_resume_work(struct work_struct *work)
 	int ret;
 
 	ret = pm_runtime_get_sync(&dwc3->dev);
-	if (ret)
+	if (ret) {
+		pm_runtime_put_sync_autosuspend(&dwc3->dev);
 		return;
+	}
 
 	pm_runtime_mark_last_busy(&dwc3->dev);
 	pm_runtime_put_sync_autosuspend(&dwc3->dev);
Signature · 1 line
2.25.1
author maintainer
#49 lore raw

PATCH 42/53 scsi: qla2xxx: Fix MPI failure AEN (8200) handling

In reply to Sasha Levin

From: Arun Easi <[email protected]>

[ Upstream commit cbb01c2f2f630f1497f703c51ff21538ae2d86b8 ]

Today, upon an MPI failure AEN, on top of collecting an MPI dump, a regular firmware dump is also taken and then chip reset. This is disruptive to IOs and not required. Make the firmware dump collection, followed by chip reset, optional (not done by default).

Firmware dump buffer and MPI dump buffer are independent of each other with this change and each can have dump that was taken at two different times for two different issues. The MPI dump is saved in a separate buffer and is retrieved differently from firmware dump.

To collect full dump on MPI failure AEN, a module parameter is introduced:
ql2xfulldump_on_mpifail (default: 0)

---
 drivers/scsi/qla2xxx/qla_attr.c |  30 +++++++-
 drivers/scsi/qla2xxx/qla_def.h  |  13 +++-
 drivers/scsi/qla2xxx/qla_gbl.h  |   3 +
 drivers/scsi/qla2xxx/qla_init.c |   2 +
 drivers/scsi/qla2xxx/qla_isr.c  |  54 +++++++++-----
 drivers/scsi/qla2xxx/qla_os.c   |   6 ++
 drivers/scsi/qla2xxx/qla_tmpl.c | 121 ++++++++++++++++++++++++++------
 7 files changed, 186 insertions(+), 43 deletions(-)
drivers/scsi/qla2xxx/qla_attr.c +27 -3
index 2c9e5ac24692d..c4917f441b10a 100644
@@ -26,7 +26,8 @@ qla2x00_sysfs_read_fw_dump(struct file *filp, struct kobject *kobj,
 	struct qla_hw_data *ha = vha->hw;
 	int rval = 0;
 
-	if (!(ha->fw_dump_reading || ha->mctp_dump_reading))
+	if (!(ha->fw_dump_reading || ha->mctp_dump_reading ||
+	      ha->mpi_fw_dump_reading))
 		return 0;
 
 	mutex_lock(&ha->optrom_mutex);
@@ -42,6 +43,10 @@ qla2x00_sysfs_read_fw_dump(struct file *filp, struct kobject *kobj,
 	} else if (ha->mctp_dumped && ha->mctp_dump_reading) {
 		rval = memory_read_from_buffer(buf, count, &off, ha->mctp_dump,
 		    MCTP_DUMP_SIZE);
+	} else if (ha->mpi_fw_dumped && ha->mpi_fw_dump_reading) {
+		rval = memory_read_from_buffer(buf, count, &off,
+					       ha->mpi_fw_dump,
+					       ha->mpi_fw_dump_len);
 	} else if (ha->fw_dump_reading) {
 		rval = memory_read_from_buffer(buf, count, &off, ha->fw_dump,
 					ha->fw_dump_len);
@@ -103,7 +108,6 @@ qla2x00_sysfs_write_fw_dump(struct file *filp, struct kobject *kobj,
 			qla82xx_set_reset_owner(vha);
 			qla8044_idc_unlock(ha);
 		} else {
-			ha->fw_dump_mpi = 1;
 			qla2x00_system_error(vha);
 		}
 		break;
@@ -137,6 +141,22 @@ qla2x00_sysfs_write_fw_dump(struct file *filp, struct kobject *kobj,
 			    vha->host_no);
 		}
 		break;
+	case 8:
+		if (!ha->mpi_fw_dump_reading)
+			break;
+		ql_log(ql_log_info, vha, 0x70e7,
+		       "MPI firmware dump cleared on (%ld).\n", vha->host_no);
+		ha->mpi_fw_dump_reading = 0;
+		ha->mpi_fw_dumped = 0;
+		break;
+	case 9:
+		if (ha->mpi_fw_dumped && !ha->mpi_fw_dump_reading) {
+			ha->mpi_fw_dump_reading = 1;
+			ql_log(ql_log_info, vha, 0x70e8,
+			       "Raw MPI firmware dump ready for read on (%ld).\n",
+			       vha->host_no);
+		}
+		break;
 	}
 	return count;
 }
@@ -706,7 +726,8 @@ qla2x00_sysfs_write_reset(struct file *filp, struct kobject *kobj,
 		scsi_unblock_requests(vha->host);
 		break;
 	case 0x2025d:
-		if (!IS_QLA81XX(ha) && !IS_QLA83XX(ha))
+		if (!IS_QLA81XX(ha) && !IS_QLA83XX(ha) &&
+		    !IS_QLA27XX(ha) && !IS_QLA28XX(ha))
 			return -EPERM;
 
 		ql_log(ql_log_info, vha, 0x706f,
@@ -724,6 +745,8 @@ qla2x00_sysfs_write_reset(struct file *filp, struct kobject *kobj,
 			qla83xx_idc_audit(vha, IDC_AUDIT_TIMESTAMP);
 			qla83xx_idc_unlock(vha, 0);
 			break;
+		} else if (IS_QLA27XX(ha) || IS_QLA28XX(ha)) {
+			qla27xx_reset_mpi(vha);
 		} else {
 			/* Make sure FC side is not in reset */
 			WARN_ON_ONCE(qla2x00_wait_for_hba_online(vha) !=
@@ -737,6 +760,7 @@ qla2x00_sysfs_write_reset(struct file *filp, struct kobject *kobj,
 			scsi_unblock_requests(vha->host);
 			break;
 		}
+		break;
 	case 0x2025e:
 		if (!IS_P3P_TYPE(ha) || vha != base_vha) {
 			ql_log(ql_log_info, vha, 0x7071,
drivers/scsi/qla2xxx/qla_def.h +12 -1
index 47c7a56438b54..daa9e936887bb 100644
@@ -3223,6 +3223,7 @@ struct isp_operations {
 		uint32_t);
 
 	void (*fw_dump) (struct scsi_qla_host *, int);
+	void (*mpi_fw_dump)(struct scsi_qla_host *, int);
 
 	int (*beacon_on) (struct scsi_qla_host *);
 	int (*beacon_off) (struct scsi_qla_host *);
@@ -3748,6 +3749,11 @@ struct qlt_hw_data {
 
 #define LEAK_EXCHG_THRESH_HOLD_PERCENT 75	/* 75 percent */
 
+struct qla_hw_data_stat {
+	u32 num_fw_dump;
+	u32 num_mpi_reset;
+};
+
 /*
  * Qlogic host adapter specific data structure.
 */
@@ -4230,7 +4236,6 @@ struct qla_hw_data {
 	uint32_t	fw_dump_len;
 	u32		fw_dump_alloc_len;
 	bool		fw_dumped;
-	bool		fw_dump_mpi;
 	unsigned long	fw_dump_cap_flags;
 #define RISC_PAUSE_CMPL		0
 #define DMA_SHUTDOWN_CMPL	1
@@ -4241,6 +4246,10 @@ struct qla_hw_data {
 #define ISP_MBX_RDY		6
 #define ISP_SOFT_RESET_CMPL	7
 	int		fw_dump_reading;
+	void		*mpi_fw_dump;
+	u32		mpi_fw_dump_len;
+	int		mpi_fw_dump_reading:1;
+	int		mpi_fw_dumped:1;
 	int		prev_minidump_failed;
 	dma_addr_t	eft_dma;
 	void		*eft;
@@ -4454,6 +4463,8 @@ struct qla_hw_data {
 	uint16_t last_zio_threshold;
 
 #define DEFAULT_ZIO_THRESHOLD 5
+
+	struct qla_hw_data_stat stat;
 };
 
 struct active_regions {
drivers/scsi/qla2xxx/qla_gbl.h +3 -0
index 1b93f5b4d77d9..b20c5fa122fb2 100644
@@ -173,6 +173,7 @@ extern int ql2xenablemsix;
 extern int qla2xuseresexchforels;
 extern int ql2xexlogins;
 extern int ql2xdifbundlinginternalbuffers;
+extern int ql2xfulldump_on_mpifail;
 
 extern int qla2x00_loop_reset(scsi_qla_host_t *);
 extern void qla2x00_abort_all_cmds(scsi_qla_host_t *, int);
@@ -645,6 +646,7 @@ extern void qla82xx_fw_dump(scsi_qla_host_t *, int);
 extern void qla8044_fw_dump(scsi_qla_host_t *, int);
 
 extern void qla27xx_fwdump(scsi_qla_host_t *, int);
+extern void qla27xx_mpi_fwdump(scsi_qla_host_t *, int);
 extern ulong qla27xx_fwdt_calculate_dump_size(struct scsi_qla_host *, void *);
 extern int qla27xx_fwdt_template_valid(void *);
 extern ulong qla27xx_fwdt_template_size(void *);
@@ -933,5 +935,6 @@ extern void qla24xx_process_purex_list(struct purex_list *);
 
 /* nvme.c */
 void qla_nvme_unregister_remote_port(struct fc_port *fcport);
+void qla27xx_reset_mpi(scsi_qla_host_t *vha);
 void qla_handle_els_plogi_done(scsi_qla_host_t *vha, struct event_arg *ea);
 #endif /* _QLA_GBL_H */
drivers/scsi/qla2xxx/qla_init.c +2 -0
index caa6b840e4594..496ead29b51e4 100644
@@ -3339,6 +3339,8 @@ qla2x00_alloc_fw_dump(scsi_qla_host_t *vha)
 				    dump_size / 1024);
 
 				if (IS_QLA27XX(ha) || IS_QLA28XX(ha)) {
+					ha->mpi_fw_dump = (char *)fw_dump +
+						ha->fwdt[1].dump_size;
 					mutex_unlock(&ha->optrom_mutex);
 					return;
 				}
drivers/scsi/qla2xxx/qla_isr.c +37 -17

72 lines not on this page · raw message

drivers/scsi/qla2xxx/qla_os.c +6 -0
index 9179bb4caed84..1120d133204c2 100644
@@ -35,6 +35,11 @@ static int apidev_major;
  */
 struct kmem_cache *srb_cachep;
 
+int ql2xfulldump_on_mpifail;
+module_param(ql2xfulldump_on_mpifail, int, S_IRUGO | S_IWUSR);
+MODULE_PARM_DESC(ql2xfulldump_on_mpifail,
+		 "Set this to take full dump on MPI hang.");
+
 /*
  * CT6 CTX allocation cache
  */
@@ -2518,6 +2523,7 @@ static struct isp_operations qla27xx_isp_ops = {
 	.read_nvram		= NULL,
 	.write_nvram		= NULL,
 	.fw_dump		= qla27xx_fwdump,
+	.mpi_fw_dump		= qla27xx_mpi_fwdump,
 	.beacon_on		= qla24xx_beacon_on,
 	.beacon_off		= qla24xx_beacon_off,
 	.beacon_blink		= qla83xx_beacon_blink,
drivers/scsi/qla2xxx/qla_tmpl.c +99 -22

148 lines not on this page · raw message

Signature · 1 line
2.25.1
author maintainer
#46 lore raw

PATCH 39/53 io_uring: fix io_sq_thread no schedule when busy

In reply to Sasha Levin

From: Xuan Zhuo <[email protected]>

[ Upstream commit b772f07add1c0b22e02c0f1e96f647560679d3a9 ]

When the user consumes and generates sqe at a fast rate, io_sqring_entries can always get sqe, and ret will not be equal to -EBUSY, so that io_sq_thread will never call cond_resched or schedule, and then we will get the following system error prompt:

rcu: INFO: rcu_sched self-detected stall on CPU
or
watchdog: BUG: soft lockup-CPU#23 stuck for 112s! [io_uring-sq:1863]

This patch checks whether need to call cond_resched() by checking the need_resched() function every cycle.

---
 fs/io_uring.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
fs/io_uring.c +2 -2
index 1829be7f63a35..6cf9d509371e2 100644
@@ -6071,7 +6071,7 @@ static int io_sq_thread(void *data)
 		 * If submit got -EBUSY, flag us as needing the application
 		 * to enter the kernel to reap and flush events.
 		 */
-		if (!to_submit || ret == -EBUSY) {
+		if (!to_submit || ret == -EBUSY || need_resched()) {
 			/*
 			 * Drop cur_mm before scheduling, we can't hold it for
 			 * long periods (or over schedule()). Do this before
@@ -6087,7 +6087,7 @@ static int io_sq_thread(void *data)
 			 * more IO, we should wait for the application to
 			 * reap events and wake us up.
 			 */
-			if (!list_empty(&ctx->poll_list) ||
+			if (!list_empty(&ctx->poll_list) || need_resched() ||
 			    (!time_after(jiffies, timeout) && ret != -EBUSY &&
 			    !percpu_ref_is_dying(&ctx->refs))) {
 				if (current->task_works)
Signature · 1 line
2.25.1
author maintainer
#16 lore raw

PATCH 15/53 ASoC: SOF: Intel: add PCI IDs for ICL-H and TGL-H

In reply to Sasha Levin

From: Pierre-Louis Bossart <[email protected]>

[ Upstream commit c8d2e2bfaeffa0f914330e8b4e45b986c8d30b58 ]

Usually the DSP is not traditionally enabled on H skews but this might be used moving forward.

---
 sound/soc/sof/sof-pci-dev.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)
sound/soc/sof/sof-pci-dev.c +8 -2
index 7b1846aeadd59..3a71f813fb563 100644
@@ -411,8 +411,11 @@ static const struct pci_device_id sof_pci_ids[] = {
 		.driver_data = (unsigned long)&cfl_desc},
 #endif
 #if IS_ENABLED(CONFIG_SND_SOC_SOF_ICELAKE)
-	{ PCI_DEVICE(0x8086, 0x34C8),
+	{ PCI_DEVICE(0x8086, 0x34C8), /* ICL-LP */
 		.driver_data = (unsigned long)&icl_desc},
+	{ PCI_DEVICE(0x8086, 0x3dc8), /* ICL-H */
+		.driver_data = (unsigned long)&icl_desc},
+
 #endif
 #if IS_ENABLED(CONFIG_SND_SOC_SOF_JASPERLAKE)
 	{ PCI_DEVICE(0x8086, 0x38c8),
@@ -431,8 +434,11 @@ static const struct pci_device_id sof_pci_ids[] = {
 		.driver_data = (unsigned long)&cml_desc},
 #endif
 #if IS_ENABLED(CONFIG_SND_SOC_SOF_TIGERLAKE)
-	{ PCI_DEVICE(0x8086, 0xa0c8),
+	{ PCI_DEVICE(0x8086, 0xa0c8), /* TGL-LP */
 		.driver_data = (unsigned long)&tgl_desc},
+	{ PCI_DEVICE(0x8086, 0x43c8), /* TGL-H */
+		.driver_data = (unsigned long)&tgl_desc},
+
 #endif
 #if IS_ENABLED(CONFIG_SND_SOC_SOF_ELKHARTLAKE)
 	{ PCI_DEVICE(0x8086, 0x4b55),
Signature · 1 line
2.25.1
author maintainer
#23 lore raw

PATCH 18/53 ALSA: hda: Intel: add missing PCI IDs for ICL-H, TGL-H and EKL

In reply to Sasha Levin

From: Pierre-Louis Bossart <[email protected]>

[ Upstream commit d50313a5a0d803bcf55121a2b82086633060d05e ]

Mirror PCI ids used for SOF.

---
 sound/pci/hda/hda_intel.c | 8 ++++++++
 1 file changed, 8 insertions(+)
sound/pci/hda/hda_intel.c +8 -0
index 41a03c61a74b3..11ec5c56c80e9 100644
@@ -2470,6 +2470,9 @@ static const struct pci_device_id azx_ids[] = {
 	/* Icelake */
 	{ PCI_DEVICE(0x8086, 0x34c8),
 	  .driver_data = AZX_DRIVER_SKL | AZX_DCAPS_INTEL_SKYLAKE},
+	/* Icelake-H */
+	{ PCI_DEVICE(0x8086, 0x3dc8),
+	  .driver_data = AZX_DRIVER_SKL | AZX_DCAPS_INTEL_SKYLAKE},
 	/* Jasperlake */
 	{ PCI_DEVICE(0x8086, 0x38c8),
 	  .driver_data = AZX_DRIVER_SKL | AZX_DCAPS_INTEL_SKYLAKE},
@@ -2478,9 +2481,14 @@ static const struct pci_device_id azx_ids[] = {
 	/* Tigerlake */
 	{ PCI_DEVICE(0x8086, 0xa0c8),
 	  .driver_data = AZX_DRIVER_SKL | AZX_DCAPS_INTEL_SKYLAKE},
+	/* Tigerlake-H */
+	{ PCI_DEVICE(0x8086, 0x43c8),
+	  .driver_data = AZX_DRIVER_SKL | AZX_DCAPS_INTEL_SKYLAKE},
 	/* Elkhart Lake */
 	{ PCI_DEVICE(0x8086, 0x4b55),
 	  .driver_data = AZX_DRIVER_SKL | AZX_DCAPS_INTEL_SKYLAKE},
+	{ PCI_DEVICE(0x8086, 0x4b58),
+	  .driver_data = AZX_DRIVER_SKL | AZX_DCAPS_INTEL_SKYLAKE},
 	/* Broxton-P(Apollolake) */
 	{ PCI_DEVICE(0x8086, 0x5a98),
 	  .driver_data = AZX_DRIVER_SKL | AZX_DCAPS_INTEL_BROXTON },
Signature · 1 line
2.25.1
author maintainer
#42 lore raw

PATCH 37/53 cifs: Fix double add page to memcg when cifs_readpages

In reply to Sasha Levin

From: Zhang Xiaoxu <[email protected]>

[ Upstream commit 95a3d8f3af9b0d63b43f221b630beaab9739d13a ]

When xfstests generic/451, there is an BUG at mm/memcontrol.c:
page:ffffea000560f2c0 refcount:2 mapcount:0 mapping:000000008544e0ea

       index:0xf
  mapping->aops:cifs_addr_ops dentry name:"tst-aio-dio-cycle-write.451"
  flags: 0x2fffff80000001(locked)
  raw: 002fffff80000001 ffffc90002023c50 ffffea0005280088 ffff88815cda0210
  raw: 000000000000000f 0000000000000000 00000002ffffffff ffff88817287d000
  page dumped because: VM_BUG_ON_PAGE(page->mem_cgroup)
  page->mem_cgroup:ffff88817287d000
  ------------[ cut here ]------------
  kernel BUG at mm/memcontrol.c:2659!
  invalid opcode: 0000 [#1] SMP
  CPU: 2 PID: 2038 Comm: xfs_io Not tainted 5.8.0-rc1 #44
  Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS ?-20190727_
    073836-buildvm-ppc64le-16.ppc.4
  RIP: 0010:commit_charge+0x35/0x50
  Code: 0d 48 83 05 54 b2 02 05 01 48 89 77 38 c3 48 c7
        c6 78 4a ea ba 48 83 05 38 b2 02 05 01 e8 63 0d9
  RSP: 0018:ffffc90002023a50 EFLAGS: 00010202
  RAX: 0000000000000000 RBX: ffff88817287d000 RCX: 0000000000000000
  RDX: 0000000000000000 RSI: ffff88817ac97ea0 RDI: ffff88817ac97ea0
  RBP: ffffea000560f2c0 R08: 0000000000000203 R09: 0000000000000005
  R10: 0000000000000030 R11: ffffc900020237a8 R12: 0000000000000000
  R13: 0000000000000001 R14: 0000000000000001 R15: ffff88815a1272c0
  FS:  00007f5071ab0800(0000) GS:ffff88817ac80000(0000) knlGS:0000000000000000
  CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
  CR2: 000055efcd5ca000 CR3: 000000015d312000 CR4: 00000000000006e0
  DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
  DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
  Call Trace:
   mem_cgroup_charge+0x166/0x4f0
   __add_to_page_cache_locked+0x4a9/0x710
   add_to_page_cache_locked+0x15/0x20
   cifs_readpages+0x217/0x1270
   read_pages+0x29a/0x670
   page_cache_readahead_unbounded+0x24f/0x390
   __do_page_cache_readahead+0x3f/0x60
   ondemand_readahead+0x1f1/0x470
   page_cache_async_readahead+0x14c/0x170
   generic_file_buffered_read+0x5df/0x1100
   generic_file_read_iter+0x10c/0x1d0
   cifs_strict_readv+0x139/0x170
   new_sync_read+0x164/0x250
   __vfs_read+0x39/0x60
   vfs_read+0xb5/0x1e0
   ksys_pread64+0x85/0xf0
   __x64_sys_pread64+0x22/0x30
   do_syscall_64+0x69/0x150
   entry_SYSCALL_64_after_hwframe+0x44/0xa9
  RIP: 0033:0x7f5071fcb1af
  Code: Bad RIP value.
  RSP: 002b:00007ffde2cdb8e0 EFLAGS: 00000293 ORIG_RAX: 0000000000000011
  RAX: ffffffffffffffda RBX: 00007ffde2cdb990 RCX: 00007f5071fcb1af
  RDX: 0000000000001000 RSI: 000055efcd5ca000 RDI: 0000000000000003
  RBP: 0000000000000003 R08: 0000000000000000 R09: 0000000000000000
  R10: 0000000000001000 R11: 0000000000000293 R12: 0000000000000001
  R13: 000000000009f000 R14: 0000000000000000 R15: 0000000000001000
  Modules linked in:
  ---[ end trace 725fa14a3e1af65c ]---

Since commit 3fea5a499d57 ("mm: memcontrol: convert page cache to a new mem_cgroup_charge() API") not cancel the page charge, the pages maybe double add to pagecache:
thread1 | thread2
cifs_readpages
readpages_get_pages
add_to_page_cache_locked(head,index=n)=0

                              | readpages_get_pages
                              | add_to_page_cache_locked(head,index=n+1)=0

add_to_page_cache_locked(head, index=n+1)=-EEXIST then, will next loop with list head page's index=n+1 and the page->mapping not NULL
readpages_get_pages add_to_page_cache_locked(head, index=n+1)
commit_charge
VM_BUG_ON_PAGE

So, we should not do the next loop when any page add to page cache failed.

---
 fs/cifs/file.c | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)
fs/cifs/file.c +7 -4
index 75ddce8ef456d..bc958fdb0143f 100644
@@ -4332,7 +4332,8 @@ readpages_get_pages(struct address_space *mapping, struct list_head *page_list,
 			break;
 
 		__SetPageLocked(page);
-		if (add_to_page_cache_locked(page, mapping, page->index, gfp)) {
+		rc = add_to_page_cache_locked(page, mapping, page->index, gfp);
+		if (rc) {
 			__ClearPageLocked(page);
 			break;
 		}
@@ -4348,6 +4349,7 @@ static int cifs_readpages(struct file *file, struct address_space *mapping,
 	struct list_head *page_list, unsigned num_pages)
 {
 	int rc;
+	int err = 0;
 	struct list_head tmplist;
 	struct cifsFileInfo *open_file = file->private_data;
 	struct cifs_sb_info *cifs_sb = CIFS_FILE_SB(file);
@@ -4392,7 +4394,7 @@ static int cifs_readpages(struct file *file, struct address_space *mapping,
 	 * the order of declining indexes. When we put the pages in
 	 * the rdata->pages, then we want them in increasing order.
 	 */
-	while (!list_empty(page_list)) {
+	while (!list_empty(page_list) && !err) {
 		unsigned int i, nr_pages, bytes, rsize;
 		loff_t offset;
 		struct page *page, *tpage;
@@ -4425,9 +4427,10 @@ static int cifs_readpages(struct file *file, struct address_space *mapping,
 			return 0;
 		}
 
-		rc = readpages_get_pages(mapping, page_list, rsize, &tmplist,
+		nr_pages = 0;
+		err = readpages_get_pages(mapping, page_list, rsize, &tmplist,
 					 &nr_pages, &offset, &bytes);
-		if (rc) {
+		if (!nr_pages) {
 			add_credits_and_wake_if(server, credits, 0);
 			break;
 		}
Signature · 1 line
2.25.1
author maintainer
#36 lore raw

PATCH 31/53 usbnet: smsc95xx: Fix use-after-free after removal

In reply to Sasha Levin

From: Tuomas Tynkkynen <[email protected]>

[ Upstream commit b835a71ef64a61383c414d6bf2896d2c0161deca ]

Syzbot reports an use-after-free in workqueue context:

BUG: KASAN: use-after-free in mutex_unlock+0x19/0x40 kernel/locking/mutex.c:737
 mutex_unlock+0x19/0x40 kernel/locking/mutex.c:737
 __smsc95xx_mdio_read drivers/net/usb/smsc95xx.c:217 [inline]
 smsc95xx_mdio_read+0x583/0x870 drivers/net/usb/smsc95xx.c:278
 check_carrier+0xd1/0x2e0 drivers/net/usb/smsc95xx.c:644
 process_one_work+0x777/0xf90 kernel/workqueue.c:2274
 worker_thread+0xa8f/0x1430 kernel/workqueue.c:2420
 kthread+0x2df/0x300 kernel/kthread.c:255

It looks like that smsc95xx_unbind() is freeing the structures that are still in use by the concurrently running workqueue callback. Thus switch to using cancel_delayed_work_sync() to ensure the work callback really is no longer active.

---
 drivers/net/usb/smsc95xx.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/usb/smsc95xx.c +1 -1
index 355be77f42418..3cf4dc3433f91 100644
@@ -1324,7 +1324,7 @@ static void smsc95xx_unbind(struct usbnet *dev, struct usb_interface *intf)
 	struct smsc95xx_priv *pdata = (struct smsc95xx_priv *)(dev->data[0]);
 
 	if (pdata) {
-		cancel_delayed_work(&pdata->carrier_check);
+		cancel_delayed_work_sync(&pdata->carrier_check);
 		netif_dbg(dev, ifdown, dev->net, "free pdata\n");
 		kfree(pdata);
 		pdata = NULL;
Signature · 1 line
2.25.1
author maintainer
#35 lore raw

PATCH 30/53 xprtrdma: Prevent dereferencing r_xprt->rx_ep after it is freed

In reply to Sasha Levin

From: Chuck Lever <[email protected]>

[ Upstream commit 2acc5cae292355f5f18ad377a2a966e7f03c8fec ]

r_xprt->rx_ep is known to be good while the transport's send lock is held. Otherwise additional references on rx_ep must be held when it is used outside of that lock's critical sections.

For now, bump the rx_ep reference count once whenever there is at least one outstanding Receive WR. This avoids the memory bandwidth overhead of taking and releasing the reference count for every ib_post_recv() and Receive completion.

---
 net/sunrpc/xprtrdma/verbs.c | 33 +++++++++++++++++++++++----------
 1 file changed, 23 insertions(+), 10 deletions(-)
net/sunrpc/xprtrdma/verbs.c +23 -10
index 05c4d3a9cda27..db0259c6467ef 100644
@@ -84,7 +84,8 @@ static void rpcrdma_rep_destroy(struct rpcrdma_rep *rep);
 static void rpcrdma_reps_unmap(struct rpcrdma_xprt *r_xprt);
 static void rpcrdma_mrs_create(struct rpcrdma_xprt *r_xprt);
 static void rpcrdma_mrs_destroy(struct rpcrdma_xprt *r_xprt);
-static int rpcrdma_ep_destroy(struct rpcrdma_ep *ep);
+static void rpcrdma_ep_get(struct rpcrdma_ep *ep);
+static int rpcrdma_ep_put(struct rpcrdma_ep *ep);
 static struct rpcrdma_regbuf *
 rpcrdma_regbuf_alloc(size_t size, enum dma_data_direction direction,
 		     gfp_t flags);
@@ -97,7 +98,8 @@ static void rpcrdma_regbuf_free(struct rpcrdma_regbuf *rb);
  */
 static void rpcrdma_xprt_drain(struct rpcrdma_xprt *r_xprt)
 {
-	struct rdma_cm_id *id = r_xprt->rx_ep->re_id;
+	struct rpcrdma_ep *ep = r_xprt->rx_ep;
+	struct rdma_cm_id *id = ep->re_id;
 
 	/* Flush Receives, then wait for deferred Reply work
 	 * to complete.
@@ -108,6 +110,8 @@ static void rpcrdma_xprt_drain(struct rpcrdma_xprt *r_xprt)
 	 * local invalidations.
 	 */
 	ib_drain_sq(id->qp);
+
+	rpcrdma_ep_put(ep);
 }
 
 /**
@@ -267,7 +271,7 @@ rpcrdma_cm_event_handler(struct rdma_cm_id *id, struct rdma_cm_event *event)
 		xprt_force_disconnect(xprt);
 		goto disconnected;
 	case RDMA_CM_EVENT_ESTABLISHED:
-		kref_get(&ep->re_kref);
+		rpcrdma_ep_get(ep);
 		ep->re_connect_status = 1;
 		rpcrdma_update_cm_private(ep, &event->param.conn);
 		trace_xprtrdma_inline_thresh(ep);
@@ -290,7 +294,7 @@ rpcrdma_cm_event_handler(struct rdma_cm_id *id, struct rdma_cm_event *event)
 		ep->re_connect_status = -ECONNABORTED;
 disconnected:
 		xprt_force_disconnect(xprt);
-		return rpcrdma_ep_destroy(ep);
+		return rpcrdma_ep_put(ep);
 	default:
 		break;
 	}
@@ -346,7 +350,7 @@ static struct rdma_cm_id *rpcrdma_create_id(struct rpcrdma_xprt *r_xprt,
 	return ERR_PTR(rc);
 }
 
-static void rpcrdma_ep_put(struct kref *kref)
+static void rpcrdma_ep_destroy(struct kref *kref)
 {
 	struct rpcrdma_ep *ep = container_of(kref, struct rpcrdma_ep, re_kref);
 
@@ -370,13 +374,18 @@ static void rpcrdma_ep_put(struct kref *kref)
 	module_put(THIS_MODULE);
 }
 
+static noinline void rpcrdma_ep_get(struct rpcrdma_ep *ep)
+{
+	kref_get(&ep->re_kref);
+}
+
 /* Returns:
  *     %0 if @ep still has a positive kref count, or
  *     %1 if @ep was destroyed successfully.
  */
-static int rpcrdma_ep_destroy(struct rpcrdma_ep *ep)
+static noinline int rpcrdma_ep_put(struct rpcrdma_ep *ep)
 {
-	return kref_put(&ep->re_kref, rpcrdma_ep_put);
+	return kref_put(&ep->re_kref, rpcrdma_ep_destroy);
 }
 
 static int rpcrdma_ep_create(struct rpcrdma_xprt *r_xprt)
@@ -493,7 +502,7 @@ static int rpcrdma_ep_create(struct rpcrdma_xprt *r_xprt)
 	return 0;
 
 out_destroy:
-	rpcrdma_ep_destroy(ep);
+	rpcrdma_ep_put(ep);
 	rdma_destroy_id(id);
 out_free:
 	kfree(ep);
@@ -522,8 +531,12 @@ int rpcrdma_xprt_connect(struct rpcrdma_xprt *r_xprt)
 
 	ep->re_connect_status = 0;
 	xprt_clear_connected(xprt);
-
 	rpcrdma_reset_cwnd(r_xprt);
+
+	/* Bump the ep's reference count while there are
+	 * outstanding Receives.
+	 */
+	rpcrdma_ep_get(ep);
 	rpcrdma_post_recvs(r_xprt, true);
 
 	rc = rpcrdma_sendctxs_create(r_xprt);
@@ -588,7 +601,7 @@ void rpcrdma_xprt_disconnect(struct rpcrdma_xprt *r_xprt)
 	rpcrdma_mrs_destroy(r_xprt);
 	rpcrdma_sendctxs_destroy(r_xprt);
 
-	if (rpcrdma_ep_destroy(ep))
+	if (rpcrdma_ep_put(ep))
 		rdma_destroy_id(id);
 
 	r_xprt->rx_ep = NULL;
Signature · 1 line
2.25.1
author maintainer
#28 lore raw

PATCH 23/53 net: ethernet: mvneta: Add 2500BaseX support for SoCs without comphy

In reply to Sasha Levin

From: Sascha Hauer <[email protected]>

[ Upstream commit 1a642ca7f38992b086101fe204a1ae3c90ed8016 ]

The older SoCs like Armada XP support a 2500BaseX mode in the datasheets referred to as DR-SGMII (Double rated SGMII) or HS-SGMII (High Speed SGMII). This is an upclocked 1000BaseX mode, thus PHY_INTERFACE_MODE_2500BASEX is the appropriate mode define for it. adding support for it merely means writing the correct magic value into the MVNETA_SERDES_CFG register.

---
 drivers/net/ethernet/marvell/mvneta.c | 6 ++++++
 1 file changed, 6 insertions(+)
drivers/net/ethernet/marvell/mvneta.c +6 -0
index 401eeeca89660..af578a5813bd2 100644
@@ -110,6 +110,7 @@
 #define MVNETA_SERDES_CFG			 0x24A0
 #define      MVNETA_SGMII_SERDES_PROTO		 0x0cc7
 #define      MVNETA_QSGMII_SERDES_PROTO		 0x0667
+#define      MVNETA_HSGMII_SERDES_PROTO		 0x1107
 #define MVNETA_TYPE_PRIO                         0x24bc
 #define      MVNETA_FORCE_UNI                    BIT(21)
 #define MVNETA_TXQ_CMD_1                         0x24e4
@@ -3558,6 +3559,11 @@ static int mvneta_config_interface(struct mvneta_port *pp,
 			mvreg_write(pp, MVNETA_SERDES_CFG,
 				    MVNETA_SGMII_SERDES_PROTO);
 			break;
+
+		case PHY_INTERFACE_MODE_2500BASEX:
+			mvreg_write(pp, MVNETA_SERDES_CFG,
+				    MVNETA_HSGMII_SERDES_PROTO);
+			break;
 		default:
 			return -EINVAL;
 		}
Signature · 1 line
2.25.1
author maintainer
#26 lore raw

PATCH 21/53 spi: spidev: fix a potential use-after-free in spidev_release()

In reply to Sasha Levin

From: Zhenzhong Duan <[email protected]>

[ Upstream commit 06096cc6c5a84ced929634b0d79376b94c65a4bd ]

If an spi device is unbounded from the driver before the release process, there will be an NULL pointer reference when it's referenced in spi_slave_abort().

Fix it by checking it's already freed before reference.

---
 drivers/spi/spidev.c | 20 ++++++++++----------
 1 file changed, 10 insertions(+), 10 deletions(-)
drivers/spi/spidev.c +10 -10
index 82e6481fdf950..012a89123067c 100644
@@ -608,15 +608,20 @@ static int spidev_open(struct inode *inode, struct file *filp)
 static int spidev_release(struct inode *inode, struct file *filp)
 {
 	struct spidev_data	*spidev;
+	int			dofree;
 
 	mutex_lock(&device_list_lock);
 	spidev = filp->private_data;
 	filp->private_data = NULL;
 
+	spin_lock_irq(&spidev->spi_lock);
+	/* ... after we unbound from the underlying device? */
+	dofree = (spidev->spi == NULL);
+	spin_unlock_irq(&spidev->spi_lock);
+
 	/* last close? */
 	spidev->users--;
 	if (!spidev->users) {
-		int		dofree;
 
 		kfree(spidev->tx_buffer);
 		spidev->tx_buffer = NULL;
@@ -624,19 +629,14 @@ static int spidev_release(struct inode *inode, struct file *filp)
 		kfree(spidev->rx_buffer);
 		spidev->rx_buffer = NULL;
 
-		spin_lock_irq(&spidev->spi_lock);
-		if (spidev->spi)
-			spidev->speed_hz = spidev->spi->max_speed_hz;
-
-		/* ... after we unbound from the underlying device? */
-		dofree = (spidev->spi == NULL);
-		spin_unlock_irq(&spidev->spi_lock);
-
 		if (dofree)
 			kfree(spidev);
+		else
+			spidev->speed_hz = spidev->spi->max_speed_hz;
 	}
 #ifdef CONFIG_SPI_SLAVE
-	spi_slave_abort(spidev->spi);
+	if (!dofree)
+		spi_slave_abort(spidev->spi);
 #endif
 	mutex_unlock(&device_list_lock);
 
Signature · 1 line
2.25.1
author maintainer
#13 lore raw

PATCH 12/53 staging: wfx: fix coherency of hif_scan() prototype

In reply to Sasha Levin

From: Jérôme Pouiller <[email protected]>

[ Upstream commit 29de523a6270a308d12d21f4fecf52dac491e226 ]

The function hif_scan() return the timeout for the completion of the scan request. It is the only function from hif_tx.c that return another thing than just an error code. This behavior is not coherent with the rest of file. Worse, if value returned is positive, the caller can't make say if it is a timeout or the value returned by the hardware.

Uniformize API with other HIF functions, only return the error code and pass timeout with parameters.

---
 drivers/staging/wfx/hif_tx.c | 6 ++++--
 drivers/staging/wfx/hif_tx.h | 2 +-
 drivers/staging/wfx/scan.c   | 6 +++---
 3 files changed, 8 insertions(+), 6 deletions(-)
drivers/staging/wfx/hif_tx.c +4 -2
index 20b3045d76674..15ff60a584668 100644
@@ -222,7 +222,7 @@ int hif_write_mib(struct wfx_dev *wdev, int vif_id, u16 mib_id, void *val,
 }
 
 int hif_scan(struct wfx_vif *wvif, struct cfg80211_scan_request *req,
-	     int chan_start_idx, int chan_num)
+	     int chan_start_idx, int chan_num, int *timeout)
 {
 	int ret, i;
 	struct hif_msg *hif;
@@ -269,11 +269,13 @@ int hif_scan(struct wfx_vif *wvif, struct cfg80211_scan_request *req,
 	tmo_chan_fg = 512 * USEC_PER_TU + body->probe_delay;
 	tmo_chan_fg *= body->num_of_probe_requests;
 	tmo = chan_num * max(tmo_chan_bg, tmo_chan_fg) + 512 * USEC_PER_TU;
+	if (timeout)
+		*timeout = usecs_to_jiffies(tmo);
 
 	wfx_fill_header(hif, wvif->id, HIF_REQ_ID_START_SCAN, buf_len);
 	ret = wfx_cmd_send(wvif->wdev, hif, NULL, 0, false);
 	kfree(hif);
-	return ret ? ret : usecs_to_jiffies(tmo);
+	return ret;
 }
 
 int hif_stop_scan(struct wfx_vif *wvif)
drivers/staging/wfx/hif_tx.h +1 -1
index f8520a14c14cd..7a21338470eeb 100644
@@ -43,7 +43,7 @@ int hif_read_mib(struct wfx_dev *wdev, int vif_id, u16 mib_id,
 int hif_write_mib(struct wfx_dev *wdev, int vif_id, u16 mib_id,
 		  void *buf, size_t buf_size);
 int hif_scan(struct wfx_vif *wvif, struct cfg80211_scan_request *req80211,
-	     int chan_start, int chan_num);
+	     int chan_start, int chan_num, int *timeout);
 int hif_stop_scan(struct wfx_vif *wvif);
 int hif_join(struct wfx_vif *wvif, const struct ieee80211_bss_conf *conf,
 	     struct ieee80211_channel *channel, const u8 *ssid, int ssidlen);
drivers/staging/wfx/scan.c +3 -3
index 9aa14331affd6..d47b8a3ba403c 100644
@@ -56,10 +56,10 @@ static int send_scan_req(struct wfx_vif *wvif,
 	wfx_tx_lock_flush(wvif->wdev);
 	wvif->scan_abort = false;
 	reinit_completion(&wvif->scan_complete);
-	timeout = hif_scan(wvif, req, start_idx, i - start_idx);
-	if (timeout < 0) {
+	ret = hif_scan(wvif, req, start_idx, i - start_idx, &timeout);
+	if (ret) {
 		wfx_tx_unlock(wvif->wdev);
-		return timeout;
+		return -EIO;
 	}
 	ret = wait_for_completion_timeout(&wvif->scan_complete, timeout);
 	if (req->channels[start_idx]->max_power != wvif->vif->bss_conf.txpower)
Signature · 1 line
2.25.1
author maintainer
#25 lore raw

PATCH 20/53 spi: spidev: fix a race between spidev_release and spidev_remove

In reply to Sasha Levin

From: Zhenzhong Duan <[email protected]>

[ Upstream commit abd42781c3d2155868821f1b947ae45bbc33330d ]

Imagine below scene, spidev is referenced after it's freed.

spidev_release() spidev_remove()
...

                                spin_lock_irq(&spidev->spi_lock);
                                    spidev->spi = NULL;
                                spin_unlock_irq(&spidev->spi_lock);
mutex_lock(&device_list_lock);
dofree = (spidev->spi == NULL);
if (dofree)
    kfree(spidev);
mutex_unlock(&device_list_lock);
                                mutex_lock(&device_list_lock);
                                list_del(&spidev->device_entry);
                                device_destroy(spidev_class, spidev->devt);
                                clear_bit(MINOR(spidev->devt), minors);
                                if (spidev->users == 0)
                                    kfree(spidev);
                                mutex_unlock(&device_list_lock);

Fix it by resetting spidev->spi in device_list_lock's protection.

---
 drivers/spi/spidev.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/spi/spidev.c +2 -2
index 80dd1025b9530..82e6481fdf950 100644
@@ -786,13 +786,13 @@ static int spidev_remove(struct spi_device *spi)
 {
 	struct spidev_data	*spidev = spi_get_drvdata(spi);
 
+	/* prevent new opens */
+	mutex_lock(&device_list_lock);
 	/* make sure ops on existing fds can abort cleanly */
 	spin_lock_irq(&spidev->spi_lock);
 	spidev->spi = NULL;
 	spin_unlock_irq(&spidev->spi_lock);
 
-	/* prevent new opens */
-	mutex_lock(&device_list_lock);
 	list_del(&spidev->device_entry);
 	device_destroy(spidev_class, spidev->devt);
 	clear_bit(MINOR(spidev->devt), minors);
Signature · 1 line
2.25.1
author maintainer
#24 lore raw

PATCH 19/53 usb: usbtest: fix missing kfree(dev->buf) in usbtest_disconnect

In reply to Sasha Levin

From: Zqiang <[email protected]>

[ Upstream commit 28ebeb8db77035e058a510ce9bd17c2b9a009dba ]

BUG: memory leak
unreferenced object 0xffff888055046e00 (size 256):
  comm "kworker/2:9", pid 2570, jiffies 4294942129 (age 1095.500s)
  hex dump (first 32 bytes):
    00 70 04 55 80 88 ff ff 18 bb 5a 81 ff ff ff ff  .p.U......Z.....
    f5 96 78 81 ff ff ff ff 37 de 8e 81 ff ff ff ff  ..x.....7.......
  backtrace:
    [<00000000d121dccf>] kmemleak_alloc_recursive
include/linux/kmemleak.h:43 [inline]
    [<00000000d121dccf>] slab_post_alloc_hook mm/slab.h:586 [inline]
    [<00000000d121dccf>] slab_alloc_node mm/slub.c:2786 [inline]
    [<00000000d121dccf>] slab_alloc mm/slub.c:2794 [inline]
    [<00000000d121dccf>] kmem_cache_alloc_trace+0x15e/0x2d0 mm/slub.c:2811
    [<000000005c3c3381>] kmalloc include/linux/slab.h:555 [inline]
    [<000000005c3c3381>] usbtest_probe+0x286/0x19d0
drivers/usb/misc/usbtest.c:2790
    [<000000001cec6910>] usb_probe_interface+0x2bd/0x870
drivers/usb/core/driver.c:361
    [<000000007806c118>] really_probe+0x48d/0x8f0 drivers/base/dd.c:551
    [<00000000a3308c3e>] driver_probe_device+0xfc/0x2a0 drivers/base/dd.c:724
    [<000000003ef66004>] __device_attach_driver+0x1b6/0x240
drivers/base/dd.c:831
    [<00000000eee53e97>] bus_for_each_drv+0x14e/0x1e0 drivers/base/bus.c:431
    [<00000000bb0648d0>] __device_attach+0x1f9/0x350 drivers/base/dd.c:897
    [<00000000838b324a>] device_initial_probe+0x1a/0x20 drivers/base/dd.c:944
    [<0000000030d501c1>] bus_probe_device+0x1e1/0x280 drivers/base/bus.c:491
    [<000000005bd7adef>] device_add+0x131d/0x1c40 drivers/base/core.c:2504
    [<00000000a0937814>] usb_set_configuration+0xe84/0x1ab0
drivers/usb/core/message.c:2030
    [<00000000e3934741>] generic_probe+0x6a/0xe0 drivers/usb/core/generic.c:210
    [<0000000098ade0f1>] usb_probe_device+0x90/0xd0
drivers/usb/core/driver.c:266
    [<000000007806c118>] really_probe+0x48d/0x8f0 drivers/base/dd.c:551
    [<00000000a3308c3e>] driver_probe_device+0xfc/0x2a0 drivers/base/dd.c:724
---
 drivers/usb/misc/usbtest.c | 1 +
 1 file changed, 1 insertion(+)
drivers/usb/misc/usbtest.c +1 -0
index 98ada1a3425c6..bae88893ee8e3 100644
@@ -2873,6 +2873,7 @@ static void usbtest_disconnect(struct usb_interface *intf)
 
 	usb_set_intfdata(intf, NULL);
 	dev_dbg(&intf->dev, "disconnect\n");
+	kfree(dev->buf);
 	kfree(dev);
 }
 
Signature · 1 line
2.25.1
author maintainer
#10 lore raw

PATCH 11/53 btrfs: use kfree() in btrfs_ioctl_get_subvol_info()

In reply to Sasha Levin

From: Waiman Long <[email protected]>

[ Upstream commit b091f7fede97cc64f7aaad3eeb37965aebee3082 ]

In btrfs_ioctl_get_subvol_info(), there is a classic case where kzalloc() was incorrectly paired with kzfree(). According to David Sterba, there isn't any sensitive information in the subvol_info that needs to be cleared before freeing. So kzfree() isn't really needed, use kfree() instead.

---
 fs/btrfs/ioctl.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/btrfs/ioctl.c +1 -1
index 40b729dce91cd..5070bd2436b76 100644
@@ -2691,7 +2691,7 @@ static int btrfs_ioctl_get_subvol_info(struct file *file, void __user *argp)
 	btrfs_put_root(root);
 out_free:
 	btrfs_free_path(path);
-	kzfree(subvol_info);
+	kfree(subvol_info);
 	return ret;
 }
 
Signature · 1 line
2.25.1
#11 lore raw

Re: [PATCH AUTOSEL 5.7 11/53] btrfs: use kfree() in btrfs_ioctl_get_subvol_info()

Quoted 9 lines · Sasha Levin

From: Waiman Long <[email protected]>

[ Upstream commit b091f7fede97cc64f7aaad3eeb37965aebee3082 ]

In btrfs_ioctl_get_subvol_info(), there is a classic case where kzalloc() was incorrectly paired with kzfree(). According to David Sterba, there isn't any sensitive information in the subvol_info that needs to be cleared before freeing. So kzfree() isn't really needed, use kfree() instead.

I don't think this patch is necessary for any stable tree, it's meant only to ease merging a tree-wide patchset to rename kzfree. In btrfs code there was no point using it so it's plain kfree.

maintainer
#17 lore raw

Re: [PATCH AUTOSEL 5.7 15/53] ASoC: SOF: Intel: add PCI IDs for ICL-H and TGL-H

In reply to Sasha Levin
Quoted 6 lines · Sasha Levin

From: Pierre-Louis Bossart <[email protected]>

[ Upstream commit c8d2e2bfaeffa0f914330e8b4e45b986c8d30b58 ]

Usually the DSP is not traditionally enabled on H skews but this might be used moving forward.

"This might be used moving forward"?

#18 lore raw

Re: [PATCH AUTOSEL 5.7 15/53] ASoC: SOF: Intel: add PCI IDs for ICL-H and TGL-H

Quoted 9 lines · Mark Brown
Quoted 6 lines · Sasha Levin

From: Pierre-Louis Bossart <[email protected]>

[ Upstream commit c8d2e2bfaeffa0f914330e8b4e45b986c8d30b58 ]

Usually the DSP is not traditionally enabled on H skews but this might be used moving forward.

"This might be used moving forward"?

There are two conditions for the SOF driver to be used in a distro:
a) the DSP needs to be enabled (as reported by the pci class info)
b) sound/hda/intel-dsp-config.c needs to contain a quirk to select SOF over the legacy HDaudio, such as presence of DMIC/SoundWire or a known vendor DMI.

Traditionally for desktops neither a) nor b) are true, but this is changing: we will start adding quirks for specific product lines as requested by OEMs.

Does this answer to your question?

maintainer
#19 lore raw

Re: [PATCH AUTOSEL 5.7 15/53] ASoC: SOF: Intel: add PCI IDs for ICL-H and TGL-H

Quoted 17 lines · Pierre-Louis Bossart
Quoted 9 lines · Mark Brown
Quoted 6 lines · Sasha Levin

From: Pierre-Louis Bossart <[email protected]>

[ Upstream commit c8d2e2bfaeffa0f914330e8b4e45b986c8d30b58 ]

Usually the DSP is not traditionally enabled on H skews but this might be used moving forward.

"This might be used moving forward"?

There are two conditions for the SOF driver to be used in a distro:
a) the DSP needs to be enabled (as reported by the pci class info)
b) sound/hda/intel-dsp-config.c needs to contain a quirk to select SOF over the legacy HDaudio, such as presence of DMIC/SoundWire or a known vendor DMI.

Quoted 4 lines

Traditionally for desktops neither a) nor b) are true, but this is changing:
we will start adding quirks for specific product lines as requested by OEMs.

Does this answer to your question?

The question was more of a why is this being backported one - without those extra quirks this does nothing, and frankly with them it seems like a *major* change someone might see in stable if they update their kernel and it suddenly switches to an entirely different DSP software stack.

#43 lore raw

Re: [PATCH AUTOSEL 5.7 37/53] cifs: Fix double add page to memcg when cifs_readpages

In reply to Sasha Levin

ср, 1 июл. 2020 г. в 18:35, Sasha Levin <[email protected]>:

Quoted 82 lines

From: Zhang Xiaoxu <[email protected]>

[ Upstream commit 95a3d8f3af9b0d63b43f221b630beaab9739d13a ]

When xfstests generic/451, there is an BUG at mm/memcontrol.c:
page:ffffea000560f2c0 refcount:2 mapcount:0 mapping:000000008544e0ea

       index:0xf
  mapping->aops:cifs_addr_ops dentry name:"tst-aio-dio-cycle-write.451"
  flags: 0x2fffff80000001(locked)
  raw: 002fffff80000001 ffffc90002023c50 ffffea0005280088 ffff88815cda0210
  raw: 000000000000000f 0000000000000000 00000002ffffffff ffff88817287d000
  page dumped because: VM_BUG_ON_PAGE(page->mem_cgroup)
  page->mem_cgroup:ffff88817287d000
  ------------[ cut here ]------------
  kernel BUG at mm/memcontrol.c:2659!
  invalid opcode: 0000 [#1] SMP
  CPU: 2 PID: 2038 Comm: xfs_io Not tainted 5.8.0-rc1 #44
  Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS ?-20190727_
    073836-buildvm-ppc64le-16.ppc.4
  RIP: 0010:commit_charge+0x35/0x50
  Code: 0d 48 83 05 54 b2 02 05 01 48 89 77 38 c3 48 c7
        c6 78 4a ea ba 48 83 05 38 b2 02 05 01 e8 63 0d9
  RSP: 0018:ffffc90002023a50 EFLAGS: 00010202
  RAX: 0000000000000000 RBX: ffff88817287d000 RCX: 0000000000000000
  RDX: 0000000000000000 RSI: ffff88817ac97ea0 RDI: ffff88817ac97ea0
  RBP: ffffea000560f2c0 R08: 0000000000000203 R09: 0000000000000005
  R10: 0000000000000030 R11: ffffc900020237a8 R12: 0000000000000000
  R13: 0000000000000001 R14: 0000000000000001 R15: ffff88815a1272c0
  FS:  00007f5071ab0800(0000) GS:ffff88817ac80000(0000) knlGS:0000000000000000
  CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
  CR2: 000055efcd5ca000 CR3: 000000015d312000 CR4: 00000000000006e0
  DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
  DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
  Call Trace:
   mem_cgroup_charge+0x166/0x4f0
   __add_to_page_cache_locked+0x4a9/0x710
   add_to_page_cache_locked+0x15/0x20
   cifs_readpages+0x217/0x1270
   read_pages+0x29a/0x670
   page_cache_readahead_unbounded+0x24f/0x390
   __do_page_cache_readahead+0x3f/0x60
   ondemand_readahead+0x1f1/0x470
   page_cache_async_readahead+0x14c/0x170
   generic_file_buffered_read+0x5df/0x1100
   generic_file_read_iter+0x10c/0x1d0
   cifs_strict_readv+0x139/0x170
   new_sync_read+0x164/0x250
   __vfs_read+0x39/0x60
   vfs_read+0xb5/0x1e0
   ksys_pread64+0x85/0xf0
   __x64_sys_pread64+0x22/0x30
   do_syscall_64+0x69/0x150
   entry_SYSCALL_64_after_hwframe+0x44/0xa9
  RIP: 0033:0x7f5071fcb1af
  Code: Bad RIP value.
  RSP: 002b:00007ffde2cdb8e0 EFLAGS: 00000293 ORIG_RAX: 0000000000000011
  RAX: ffffffffffffffda RBX: 00007ffde2cdb990 RCX: 00007f5071fcb1af
  RDX: 0000000000001000 RSI: 000055efcd5ca000 RDI: 0000000000000003
  RBP: 0000000000000003 R08: 0000000000000000 R09: 0000000000000000
  R10: 0000000000001000 R11: 0000000000000293 R12: 0000000000000001
  R13: 000000000009f000 R14: 0000000000000000 R15: 0000000000001000
  Modules linked in:
  ---[ end trace 725fa14a3e1af65c ]---

Since commit 3fea5a499d57 ("mm: memcontrol: convert page cache to a new mem_cgroup_charge() API") not cancel the page charge, the pages maybe double add to pagecache:
thread1 | thread2
cifs_readpages
readpages_get_pages
add_to_page_cache_locked(head,index=n)=0

                              | readpages_get_pages
                              | add_to_page_cache_locked(head,index=n+1)=0

add_to_page_cache_locked(head, index=n+1)=-EEXIST then, will next loop with list head page's index=n+1 and the page->mapping not NULL
readpages_get_pages add_to_page_cache_locked(head, index=n+1)
commit_charge
VM_BUG_ON_PAGE

Quoted 8 lines

So, we should not do the next loop when any page add to page cache failed.

Hi Sasha,

The patch description mentions the commit 3fea5a499d57 that changed the behavior and was introduced in v5.8-rc1. I noticed that you are targeting this patch for 4.9, 4.14, 4.19, 5.4 and 5.7 stable branches. Are you going to backport the commit 3fea5a499d57 as well?

--
Best regards,
Pavel Shilovsky

author maintainer
#12 lore raw

Re: [PATCH AUTOSEL 5.7 11/53] btrfs: use kfree() in btrfs_ioctl_get_subvol_info()

In reply to David Sterba
Quoted 11 lines · David Sterba
Quoted 9 lines · Sasha Levin

From: Waiman Long <[email protected]>

[ Upstream commit b091f7fede97cc64f7aaad3eeb37965aebee3082 ]

In btrfs_ioctl_get_subvol_info(), there is a classic case where kzalloc() was incorrectly paired with kzfree(). According to David Sterba, there isn't any sensitive information in the subvol_info that needs to be cleared before freeing. So kzfree() isn't really needed, use kfree() instead.

Quoted 3 lines

I don't think this patch is necessary for any stable tree, it's meant only to ease merging a tree-wide patchset to rename kzfree. In btrfs code there was no point using it so it's plain kfree.

I've dropped it, thanks!

Signature · 2 lines
Thanks,
Sasha
author maintainer
#20 lore raw

Re: [PATCH AUTOSEL 5.7 15/53] ASoC: SOF: Intel: add PCI IDs for ICL-H and TGL-H

In reply to Mark Brown
Quoted 21 lines · Mark Brown
Quoted 19 lines · Pierre-Louis Bossart
Quoted 9 lines · Mark Brown
Quoted 6 lines · Sasha Levin

From: Pierre-Louis Bossart <[email protected]>

[ Upstream commit c8d2e2bfaeffa0f914330e8b4e45b986c8d30b58 ]

Usually the DSP is not traditionally enabled on H skews but this might be used moving forward.

"This might be used moving forward"?

There are two conditions for the SOF driver to be used in a distro:
a) the DSP needs to be enabled (as reported by the pci class info)
b) sound/hda/intel-dsp-config.c needs to contain a quirk to select SOF over the legacy HDaudio, such as presence of DMIC/SoundWire or a known vendor DMI.

Traditionally for desktops neither a) nor b) are true, but this is changing:
we will start adding quirks for specific product lines as requested by OEMs.

Quoted 7 lines
Quoted 1 line

Does this answer to your question?

The question was more of a why is this being backported one - without those extra quirks this does nothing, and frankly with them it seems like a *major* change someone might see in stable if they update their kernel and it suddenly switches to an entirely different DSP software stack.

Hey Mark,

It got picked up because the stable rules explicitly call out PCI IDs as something we take into the stable trees, and so the AUTOSEL brains picked up on this...

I'll drop this patch as it doesn't do much.

Signature · 2 lines
Thanks,
Sasha
author maintainer
#44 lore raw

Re: [PATCH AUTOSEL 5.7 37/53] cifs: Fix double add page to memcg when cifs_readpages

In reply to Pavel Shilovsky
Quoted 92 lines · Pavel Shilovsky

ср, 1 июл. 2020 г. в 18:35, Sasha Levin <[email protected]>:

Quoted 82 lines

From: Zhang Xiaoxu <[email protected]>

[ Upstream commit 95a3d8f3af9b0d63b43f221b630beaab9739d13a ]

When xfstests generic/451, there is an BUG at mm/memcontrol.c:
page:ffffea000560f2c0 refcount:2 mapcount:0 mapping:000000008544e0ea

       index:0xf
  mapping->aops:cifs_addr_ops dentry name:"tst-aio-dio-cycle-write.451"
  flags: 0x2fffff80000001(locked)
  raw: 002fffff80000001 ffffc90002023c50 ffffea0005280088 ffff88815cda0210
  raw: 000000000000000f 0000000000000000 00000002ffffffff ffff88817287d000
  page dumped because: VM_BUG_ON_PAGE(page->mem_cgroup)
  page->mem_cgroup:ffff88817287d000
  ------------[ cut here ]------------
  kernel BUG at mm/memcontrol.c:2659!
  invalid opcode: 0000 [#1] SMP
  CPU: 2 PID: 2038 Comm: xfs_io Not tainted 5.8.0-rc1 #44
  Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS ?-20190727_
    073836-buildvm-ppc64le-16.ppc.4
  RIP: 0010:commit_charge+0x35/0x50
  Code: 0d 48 83 05 54 b2 02 05 01 48 89 77 38 c3 48 c7
        c6 78 4a ea ba 48 83 05 38 b2 02 05 01 e8 63 0d9
  RSP: 0018:ffffc90002023a50 EFLAGS: 00010202
  RAX: 0000000000000000 RBX: ffff88817287d000 RCX: 0000000000000000
  RDX: 0000000000000000 RSI: ffff88817ac97ea0 RDI: ffff88817ac97ea0
  RBP: ffffea000560f2c0 R08: 0000000000000203 R09: 0000000000000005
  R10: 0000000000000030 R11: ffffc900020237a8 R12: 0000000000000000
  R13: 0000000000000001 R14: 0000000000000001 R15: ffff88815a1272c0
  FS:  00007f5071ab0800(0000) GS:ffff88817ac80000(0000) knlGS:0000000000000000
  CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
  CR2: 000055efcd5ca000 CR3: 000000015d312000 CR4: 00000000000006e0
  DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
  DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
  Call Trace:
   mem_cgroup_charge+0x166/0x4f0
   __add_to_page_cache_locked+0x4a9/0x710
   add_to_page_cache_locked+0x15/0x20
   cifs_readpages+0x217/0x1270
   read_pages+0x29a/0x670
   page_cache_readahead_unbounded+0x24f/0x390
   __do_page_cache_readahead+0x3f/0x60
   ondemand_readahead+0x1f1/0x470
   page_cache_async_readahead+0x14c/0x170
   generic_file_buffered_read+0x5df/0x1100
   generic_file_read_iter+0x10c/0x1d0
   cifs_strict_readv+0x139/0x170
   new_sync_read+0x164/0x250
   __vfs_read+0x39/0x60
   vfs_read+0xb5/0x1e0
   ksys_pread64+0x85/0xf0
   __x64_sys_pread64+0x22/0x30
   do_syscall_64+0x69/0x150
   entry_SYSCALL_64_after_hwframe+0x44/0xa9
  RIP: 0033:0x7f5071fcb1af
  Code: Bad RIP value.
  RSP: 002b:00007ffde2cdb8e0 EFLAGS: 00000293 ORIG_RAX: 0000000000000011
  RAX: ffffffffffffffda RBX: 00007ffde2cdb990 RCX: 00007f5071fcb1af
  RDX: 0000000000001000 RSI: 000055efcd5ca000 RDI: 0000000000000003
  RBP: 0000000000000003 R08: 0000000000000000 R09: 0000000000000000
  R10: 0000000000001000 R11: 0000000000000293 R12: 0000000000000001
  R13: 000000000009f000 R14: 0000000000000000 R15: 0000000000001000
  Modules linked in:
  ---[ end trace 725fa14a3e1af65c ]---

Since commit 3fea5a499d57 ("mm: memcontrol: convert page cache to a new mem_cgroup_charge() API") not cancel the page charge, the pages maybe double add to pagecache:
thread1 | thread2
cifs_readpages
readpages_get_pages
add_to_page_cache_locked(head,index=n)=0

                              | readpages_get_pages
                              | add_to_page_cache_locked(head,index=n+1)=0

add_to_page_cache_locked(head, index=n+1)=-EEXIST then, will next loop with list head page's index=n+1 and the page->mapping not NULL
readpages_get_pages add_to_page_cache_locked(head, index=n+1)
commit_charge
VM_BUG_ON_PAGE

Quoted 8 lines

So, we should not do the next loop when any page add to page cache failed.

Quoted 6 lines

Hi Sasha,

The patch description mentions the commit 3fea5a499d57 that changed the behavior and was introduced in v5.8-rc1. I noticed that you are targeting this patch for 4.9, 4.14, 4.19, 5.4 and 5.7 stable branches. Are you going to backport the commit 3fea5a499d57 as well?

Nope, I'll drop this one. Thanks!

Signature · 2 lines
Thanks,
Sasha