4.19.235-rc1 review

This is the start of the stable review cycle for the 4.19.235 release. There are 30 patches in this series, all will be posted as a response to this one. If anyone has any issues with these being applied, please let me know.

Responses should be made by Wed, 16 Mar 2022 11:27:22 +0000. Anything received after that time might be too late.

The whole patch series can be found in one patch at:

	https://www.kernel.org/pub/linux/kernel/v4.x/stable-review/patch-4.19.235-rc1.gz

or in the git tree and branch at:
git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable-rc.git linux-4.19.y
and the diffstat can be found below.

thanks,

greg k-h

-------------
Pseudo-Shortlog of commits:

Greg Kroah-Hartman <[email protected]>

    Linux 4.19.235-rc1

Valentin Schneider <[email protected]>
ia64: ensure proper NUMA distance and possible map initialization

Dietmar Eggemann <[email protected]>
sched/topology: Fix sched_domain_topology_level alloc in sched_init_numa()

Valentin Schneider <[email protected]>
sched/topology: Make sched_init_numa() use a set for the deduplicating sort

Qu Wenruo <[email protected]>
btrfs: unlock newly allocated extent buffer after error

Josh Triplett <[email protected]>
ext4: add check to prevent attempting to resize an fs with sparse_super2

Russell King (Oracle) <[email protected]>
ARM: fix Thumb2 regression with Spectre BHB

Michael S. Tsirkin <[email protected]>
virtio: acknowledge all features before access

Michael S. Tsirkin <[email protected]>
virtio: unexport virtio_finalize_features

James Morse <[email protected]>
KVM: arm64: Reset PMC_EL0 to avoid a panic() on systems with no PMU

Emil Renner Berthing <[email protected]>
riscv: Fix auipc+jalr relocation range checks

Robert Hancock <[email protected]>
net: macb: Fix lost RX packet wakeup race in NAPI receive

Dan Carpenter <[email protected]>
staging: gdm724x: fix use after free in gdm_lte_rx()

Randy Dunlap <[email protected]>
ARM: Spectre-BHB: provide empty stub for non-config

Mike Kravetz <[email protected]>
selftests/memfd: clean up mapping in mfd_fail_write

Sven Schnelle <[email protected]>
tracing: Ensure trace buffer is at least 4096 bytes large

Marek Marczykowski-Górecki <[email protected]>
Revert "xen-netback: Check for hotplug-status existence before watching"

Marek Marczykowski-Górecki <[email protected]>
Revert "xen-netback: remove 'hotplug-status' once it has served its purpose"

suresh kumar <[email protected]>
net-sysfs: add check for netdevice being present to speed_show

Eric Dumazet <[email protected]>
sctp: fix kernel-infoleak for SCTP sockets

Clément Léger <[email protected]>
net: phy: DP83822: clear MISR2 register to disable interrupts

Miaoqian Lin <[email protected]>
gianfar: ethtool: Fix refcount leak in gfar_get_ts_info

Mark Featherston <[email protected]>
gpio: ts4900: Do not set DAT and OE together

Pavel Skripkin <[email protected]>
NFC: port100: fix use-after-free in port100_send_complete

Mohammad Kabat <[email protected]>
net/mlx5: Fix size field in bufferx_reg struct

Duoming Zhou <[email protected]>
ax25: Fix NULL pointer dereference in ax25_kill_by_device

Jiasheng Jiang <[email protected]>
net: ethernet: lpc_eth: Handle error for clk_enable

Jiasheng Jiang <[email protected]>
net: ethernet: ti: cpts: Handle error for clk_enable

Miaoqian Lin <[email protected]>
ethernet: Fix error handling in xemaclite_of_probe

Tom Rix <[email protected]>
qed: return status of qed_iov_get_link

Jia-Ju Bai <[email protected]>
net: qlogic: check the return value of dma_alloc_coherent() in qed_vf_hw_prepare()

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

Diffstat:

 Makefile                                         |  4 +-
 arch/arm/include/asm/spectre.h                   |  6 ++
 arch/arm/kernel/entry-armv.S                     |  4 +-
 arch/arm64/kvm/sys_regs.c                        |  4 +-
 arch/ia64/kernel/acpi.c                          |  7 +-
 arch/riscv/kernel/module.c                       | 21 +++--
 drivers/gpio/gpio-ts4900.c                       | 24 ++++--
 drivers/net/ethernet/cadence/macb_main.c         | 25 +++++-
 drivers/net/ethernet/freescale/gianfar_ethtool.c |  1 +
 drivers/net/ethernet/nxp/lpc_eth.c               |  5 +-
 drivers/net/ethernet/qlogic/qed/qed_sriov.c      | 18 +++--
 drivers/net/ethernet/qlogic/qed/qed_vf.c         |  7 ++
 drivers/net/ethernet/ti/cpts.c                   |  4 +-
 drivers/net/ethernet/xilinx/xilinx_emaclite.c    |  4 +-
 drivers/net/phy/dp83822.c                        |  2 +-
 drivers/net/xen-netback/xenbus.c                 | 13 ++--
 drivers/nfc/port100.c                            |  2 +
 drivers/staging/gdm724x/gdm_lte.c                |  5 +-
 drivers/virtio/virtio.c                          | 40 +++++-----
 fs/btrfs/extent-tree.c                           |  1 +
 fs/ext4/resize.c                                 |  5 ++
 include/linux/mlx5/mlx5_ifc.h                    |  4 +-
 include/linux/topology.h                         |  1 +
 include/linux/virtio.h                           |  1 -
 include/linux/virtio_config.h                    |  3 +-
 kernel/sched/topology.c                          | 99 ++++++++++++------------
 kernel/trace/trace.c                             | 10 ++-
 net/ax25/af_ax25.c                               |  7 ++
 net/core/net-sysfs.c                             |  2 +-
 net/sctp/diag.c                                  |  9 +--
 tools/testing/selftests/memfd/memfd_test.c       |  1 +
 31 files changed, 215 insertions(+), 124 deletions(-)

From: Jia-Ju Bai <[email protected]>

[ Upstream commit e0058f0fa80f6e09c4d363779c241c45a3c56b94 ]

The function dma_alloc_coherent() in qed_vf_hw_prepare() can fail, so its return value should be checked.

---
 drivers/net/ethernet/qlogic/qed/qed_vf.c | 7 +++++++
 1 file changed, 7 insertions(+)

index 93a0fbf6a132..e12338abaf0a 100644
@@ -539,6 +539,9 @@ int qed_vf_hw_prepare(struct qed_hwfn *p_hwfn)
 						    p_iov->bulletin.size,
 						    &p_iov->bulletin.phys,
 						    GFP_KERNEL);
+	if (!p_iov->bulletin.p_virt)
+		goto free_pf2vf_reply;
+
 	DP_VERBOSE(p_hwfn, QED_MSG_IOV,
 		   "VF's bulletin Board [%p virt 0x%llx phys 0x%08x bytes]\n",
 		   p_iov->bulletin.p_virt,

@@ -578,6 +581,10 @@ int qed_vf_hw_prepare(struct qed_hwfn *p_hwfn)
 
 	return rc;
 
+free_pf2vf_reply:
+	dma_free_coherent(&p_hwfn->cdev->pdev->dev,
+			  sizeof(union pfvf_tlvs),
+			  p_iov->pf2vf_reply, p_iov->pf2vf_reply_phys);
 free_vf2pf_request:
 	dma_free_coherent(&p_hwfn->cdev->pdev->dev,
 			  sizeof(union vfpf_tlvs),

2.34.1

From: Tom Rix <[email protected]>

[ Upstream commit d9dc0c84ad2d4cc911ba252c973d1bf18d5eb9cf ]

Clang static analysis reports this issue
qed_sriov.c:4727:19: warning: Assigned value is
garbage or undefined

  ivi->max_tx_rate = tx_rate ? tx_rate : link.speed;
                   ^ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

link is only sometimes set by the call to qed_iov_get_link() qed_iov_get_link fails without setting link or returning status. So change the decl to return status.

---
 drivers/net/ethernet/qlogic/qed/qed_sriov.c | 18 +++++++++++-------
 1 file changed, 11 insertions(+), 7 deletions(-)

index 886c7aae662f..285231bc318b 100644
@@ -3800,11 +3800,11 @@ bool qed_iov_mark_vf_flr(struct qed_hwfn *p_hwfn, u32 *p_disabled_vfs)
 	return found;
 }
 
-static void qed_iov_get_link(struct qed_hwfn *p_hwfn,
-			     u16 vfid,
-			     struct qed_mcp_link_params *p_params,
-			     struct qed_mcp_link_state *p_link,
-			     struct qed_mcp_link_capabilities *p_caps)
+static int qed_iov_get_link(struct qed_hwfn *p_hwfn,
+			    u16 vfid,
+			    struct qed_mcp_link_params *p_params,
+			    struct qed_mcp_link_state *p_link,
+			    struct qed_mcp_link_capabilities *p_caps)
 {
 	struct qed_vf_info *p_vf = qed_iov_get_vf_info(p_hwfn,
 						       vfid,

@@ -3812,7 +3812,7 @@ static void qed_iov_get_link(struct qed_hwfn *p_hwfn,
 	struct qed_bulletin_content *p_bulletin;
 
 	if (!p_vf)
-		return;
+		return -EINVAL;
 
 	p_bulletin = p_vf->bulletin.p_virt;
 

@@ -3822,6 +3822,7 @@ static void qed_iov_get_link(struct qed_hwfn *p_hwfn,
 		__qed_vf_get_link_state(p_hwfn, p_link, p_bulletin);
 	if (p_caps)
 		__qed_vf_get_link_caps(p_hwfn, p_caps, p_bulletin);
+	return 0;
 }
 
 static int

@@ -4676,6 +4677,7 @@ static int qed_get_vf_config(struct qed_dev *cdev,
 	struct qed_public_vf_info *vf_info;
 	struct qed_mcp_link_state link;
 	u32 tx_rate;
+	int ret;
 
 	/* Sanitize request */
 	if (IS_VF(cdev))

@@ -4689,7 +4691,9 @@ static int qed_get_vf_config(struct qed_dev *cdev,
 
 	vf_info = qed_iov_get_public_vf_info(hwfn, vf_id, true);
 
-	qed_iov_get_link(hwfn, vf_id, NULL, &link, NULL);
+	ret = qed_iov_get_link(hwfn, vf_id, NULL, &link, NULL);
+	if (ret)
+		return ret;
 
 	/* Fill information about VF */
 	ivi->vf = vf_id;

2.34.1

From: Miaoqian Lin <[email protected]>

[ Upstream commit b19ab4b38b06aae12442b2de95ccf58b5dc53584 ]

This node pointer is returned by of_parse_phandle() with refcount incremented in this function. Calling of_node_put() to avoid the refcount leak. As the remove function do.

---
 drivers/net/ethernet/xilinx/xilinx_emaclite.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

index edb2215f9993..23a4f9061072 100644
@@ -1173,7 +1173,7 @@ static int xemaclite_of_probe(struct platform_device *ofdev)
 	if (rc) {
 		dev_err(dev,
 			"Cannot register network device, aborting\n");
-		goto error;
+		goto put_node;
 	}
 
 	dev_info(dev,

@@ -1181,6 +1181,8 @@ static int xemaclite_of_probe(struct platform_device *ofdev)
 		 (unsigned int __force)ndev->mem_start, lp->base_addr, ndev->irq);
 	return 0;
 
+put_node:
+	of_node_put(lp->phy_node);
 error:
 	free_netdev(ndev);
 	return rc;

2.34.1

From: Jiasheng Jiang <[email protected]>

[ Upstream commit 6babfc6e6fab068018c36e8f6605184b8c0b349d ]

As the potential failure of the clk_enable(),
it should be better to check it and return error
if fails.

---
 drivers/net/ethernet/ti/cpts.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

index 10b301e79086..01cc92f6a1f8 100644
@@ -445,7 +445,9 @@ int cpts_register(struct cpts *cpts)
 	for (i = 0; i < CPTS_MAX_EVENTS; i++)
 		list_add(&cpts->pool_data[i].list, &cpts->pool);
 
-	clk_enable(cpts->refclk);
+	err = clk_enable(cpts->refclk);
+	if (err)
+		return err;
 
 	cpts_write32(cpts, CPTS_EN, control);
 	cpts_write32(cpts, TS_PEND_EN, int_enable);

2.34.1

From: Jiasheng Jiang <[email protected]>

[ Upstream commit 2169b79258c8be803d2595d6456b1e77129fe154 ]

As the potential failure of the clk_enable(), it should be better to check it and return error if fails.

---
 drivers/net/ethernet/nxp/lpc_eth.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

index 415ac33341c5..27c22f0e9d25 100644
@@ -1512,6 +1512,7 @@ static int lpc_eth_drv_resume(struct platform_device *pdev)
 {
 	struct net_device *ndev = platform_get_drvdata(pdev);
 	struct netdata_local *pldat;
+	int ret;
 
 	if (device_may_wakeup(&pdev->dev))
 		disable_irq_wake(ndev->irq);

@@ -1521,7 +1522,9 @@ static int lpc_eth_drv_resume(struct platform_device *pdev)
 			pldat = netdev_priv(ndev);
 
 			/* Enable interface clock */
-			clk_enable(pldat->clk);
+			ret = clk_enable(pldat->clk);
+			if (ret)
+				return ret;
 
 			/* Reset and initialize */
 			__lpc_eth_reset(pldat);

2.34.1

From: Duoming Zhou <[email protected]>

[ Upstream commit 71171ac8eb34ce7fe6b3267dce27c313ab3cb3ac ]

When two ax25 devices attempted to establish connection, the requester use ax25_create(), ax25_bind() and ax25_connect() to initiate connection. The receiver use ax25_rcv() to accept connection and use ax25_create_cb() in ax25_rcv() to create ax25_cb, but the ax25_cb->sk is NULL. When the receiver is detaching, a NULL pointer dereference bug caused by sock_hold(sk) in ax25_kill_by_device() will happen. The corresponding fail log is shown below:

===============================================================

BUG: KASAN: null-ptr-deref in ax25_device_event+0xfd/0x290
Call Trace:
...
ax25_device_event+0xfd/0x290
raw_notifier_call_chain+0x5e/0x70
dev_close_many+0x174/0x220
unregister_netdevice_many+0x1f7/0xa60
unregister_netdevice_queue+0x12f/0x170
unregister_netdev+0x13/0x20
mkiss_close+0xcd/0x140
tty_ldisc_release+0xc0/0x220
tty_release_struct+0x17/0xa0
tty_release+0x62d/0x670

...

This patch add condition check in ax25_kill_by_device(). If s->sk is NULL, it will goto if branch to kill device.

---
 net/ax25/af_ax25.c | 7 +++++++
 1 file changed, 7 insertions(+)

index a2bf5e4e9fbe..3170b43b9f89 100644
@@ -90,6 +90,13 @@ static void ax25_kill_by_device(struct net_device *dev)
 	ax25_for_each(s, &ax25_list) {
 		if (s->ax25_dev == ax25_dev) {
 			sk = s->sk;
+			if (!sk) {
+				spin_unlock_bh(&ax25_list_lock);
+				s->ax25_dev = NULL;
+				ax25_disconnect(s, ENETUNREACH);
+				spin_lock_bh(&ax25_list_lock);
+				goto again;
+			}
 			sock_hold(sk);
 			spin_unlock_bh(&ax25_list_lock);
 			lock_sock(sk);

2.34.1

From: Mohammad Kabat <[email protected]>

[ Upstream commit ac77998b7ac3044f0509b097da9637184598980d ]

According to HW spec the field "size" should be 16 bits in bufferx register.

---
 include/linux/mlx5/mlx5_ifc.h | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

index 5f711b2983db..af040fcd2c5a 100644
@@ -8287,8 +8287,8 @@ struct mlx5_ifc_bufferx_reg_bits {
 	u8         reserved_at_0[0x6];
 	u8         lossy[0x1];
 	u8         epsb[0x1];
-	u8         reserved_at_8[0xc];
-	u8         size[0xc];
+	u8         reserved_at_8[0x8];
+	u8         size[0x10];
 
 	u8         xoff_threshold[0x10];
 	u8         xon_threshold[0x10];

2.34.1

From: Pavel Skripkin <[email protected]>

[ Upstream commit f80cfe2f26581f188429c12bd937eb905ad3ac7b ]

Syzbot reported UAF in port100_send_complete(). The root case is in missing usb_kill_urb() calls on error handling path of ->probe function.

port100_send_complete() accesses devm allocated memory which will be freed on probe failure. We should kill this urbs before returning an error from probe function to prevent reported use-after-free

Fail log:

BUG: KASAN: use-after-free in port100_send_complete+0x16e/0x1a0 drivers/nfc/port100.c:935
Read of size 1 at addr ffff88801bb59540 by task ksoftirqd/2/26
...
Call Trace:
 <TASK>
 __dump_stack lib/dump_stack.c:88 [inline]
 dump_stack_lvl+0xcd/0x134 lib/dump_stack.c:106
 print_address_description.constprop.0.cold+0x8d/0x303 mm/kasan/report.c:255
 __kasan_report mm/kasan/report.c:442 [inline]
 kasan_report.cold+0x83/0xdf mm/kasan/report.c:459
 port100_send_complete+0x16e/0x1a0 drivers/nfc/port100.c:935
 __usb_hcd_giveback_urb+0x2b0/0x5c0 drivers/usb/core/hcd.c:1670

...

Allocated by task 1255:
 kasan_save_stack+0x1e/0x40 mm/kasan/common.c:38
 kasan_set_track mm/kasan/common.c:45 [inline]
 set_alloc_info mm/kasan/common.c:436 [inline]
 ____kasan_kmalloc mm/kasan/common.c:515 [inline]
 ____kasan_kmalloc mm/kasan/common.c:474 [inline]
 __kasan_kmalloc+0xa6/0xd0 mm/kasan/common.c:524
 alloc_dr drivers/base/devres.c:116 [inline]
 devm_kmalloc+0x96/0x1d0 drivers/base/devres.c:823
 devm_kzalloc include/linux/device.h:209 [inline]
 port100_probe+0x8a/0x1320 drivers/nfc/port100.c:1502

Freed by task 1255:
 kasan_save_stack+0x1e/0x40 mm/kasan/common.c:38
 kasan_set_track+0x21/0x30 mm/kasan/common.c:45
 kasan_set_free_info+0x20/0x30 mm/kasan/generic.c:370
 ____kasan_slab_free mm/kasan/common.c:366 [inline]
 ____kasan_slab_free+0xff/0x140 mm/kasan/common.c:328
 kasan_slab_free include/linux/kasan.h:236 [inline]
 __cache_free mm/slab.c:3437 [inline]
 kfree+0xf8/0x2b0 mm/slab.c:3794
 release_nodes+0x112/0x1a0 drivers/base/devres.c:501
 devres_release_all+0x114/0x190 drivers/base/devres.c:530
 really_probe+0x626/0xcc0 drivers/base/dd.c:670

---
 drivers/nfc/port100.c | 2 ++
 1 file changed, 2 insertions(+)

index 0f37acec98ab..bc680b8be133 100644
@@ -1618,7 +1618,9 @@ static int port100_probe(struct usb_interface *interface,
 	nfc_digital_free_device(dev->nfc_digital_dev);
 
 error:
+	usb_kill_urb(dev->in_urb);
 	usb_free_urb(dev->in_urb);
+	usb_kill_urb(dev->out_urb);
 	usb_free_urb(dev->out_urb);
 	usb_put_dev(dev->udev);
 

2.34.1

From: Mark Featherston <[email protected]>

[ Upstream commit 03fe003547975680fdb9ff5ab0e41cb68276c4f2 ]

This works around an issue with the hardware where both OE and DAT are exposed in the same register. If both are updated simultaneously, the harware makes no guarantees that OE or DAT will actually change in any given order and may result in a glitch of a few ns on a GPIO pin when changing direction and value in a single write.

Setting direction to input now only affects OE bit. Setting direction to output updates DAT first, then OE.

---
 drivers/gpio/gpio-ts4900.c | 24 +++++++++++++++++++-----
 1 file changed, 19 insertions(+), 5 deletions(-)

index 1da8d0586329..410452306bf7 100644
@@ -1,7 +1,7 @@
 /*
  * Digital I/O driver for Technologic Systems I2C FPGA Core
  *
- * Copyright (C) 2015 Technologic Systems
+ * Copyright (C) 2015, 2018 Technologic Systems
  * Copyright (C) 2016 Savoir-Faire Linux
  *
  * This program is free software; you can redistribute it and/or

@@ -52,19 +52,33 @@ static int ts4900_gpio_direction_input(struct gpio_chip *chip,
 {
 	struct ts4900_gpio_priv *priv = gpiochip_get_data(chip);
 
-	/*
-	 * This will clear the output enable bit, the other bits are
-	 * dontcare when this is cleared
+	/* Only clear the OE bit here, requires a RMW. Prevents potential issue
+	 * with OE and data getting to the physical pin at different times.
 	 */
-	return regmap_write(priv->regmap, offset, 0);
+	return regmap_update_bits(priv->regmap, offset, TS4900_GPIO_OE, 0);
 }
 
 static int ts4900_gpio_direction_output(struct gpio_chip *chip,
 					unsigned int offset, int value)
 {
 	struct ts4900_gpio_priv *priv = gpiochip_get_data(chip);
+	unsigned int reg;
 	int ret;
 
+	/* If changing from an input to an output, we need to first set the
+	 * proper data bit to what is requested and then set OE bit. This
+	 * prevents a glitch that can occur on the IO line
+	 */
+	regmap_read(priv->regmap, offset, &reg);
+	if (!(reg & TS4900_GPIO_OE)) {
+		if (value)
+			reg = TS4900_GPIO_OUT;
+		else
+			reg &= ~TS4900_GPIO_OUT;
+
+		regmap_write(priv->regmap, offset, reg);
+	}
+
 	if (value)
 		ret = regmap_write(priv->regmap, offset, TS4900_GPIO_OE |
 							 TS4900_GPIO_OUT);

2.34.1

From: Miaoqian Lin <[email protected]>

[ Upstream commit 2ac5b58e645c66932438bb021cb5b52097ce70b0 ]

The of_find_compatible_node() function returns a node pointer with refcount incremented, We should use of_node_put() on it when done Add the missing of_node_put() to release the refcount.

---
 drivers/net/ethernet/freescale/gianfar_ethtool.c | 1 +
 1 file changed, 1 insertion(+)

index 395a5266ea30..0cddaaaf48aa 100644
@@ -1528,6 +1528,7 @@ static int gfar_get_ts_info(struct net_device *dev,
 	ptp_node = of_find_compatible_node(NULL, NULL, "fsl,etsec-ptp");
 	if (ptp_node) {
 		ptp_dev = of_find_device_by_node(ptp_node);
+		of_node_put(ptp_node);
 		if (ptp_dev)
 			ptp = platform_get_drvdata(ptp_dev);
 	}

2.34.1

From: Clément Léger <[email protected]>

[ Upstream commit 37c9d66c95564c85a001d8a035354f0220a1e1c3 ]

MISR1 was cleared twice but the original author intention was probably to clear MISR1 & MISR2 to completely disable interrupts. Fix it to clear MISR2.

---
 drivers/net/phy/dp83822.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

index 6e8a2a4f3a6e..9e2ed98f7df2 100644
@@ -244,7 +244,7 @@ static int dp83822_config_intr(struct phy_device *phydev)
 		if (err < 0)
 			return err;
 
-		err = phy_write(phydev, MII_DP83822_MISR1, 0);
+		err = phy_write(phydev, MII_DP83822_MISR2, 0);
 		if (err < 0)
 			return err;
 

2.34.1

From: Eric Dumazet <[email protected]>

[ Upstream commit 633593a808980f82d251d0ca89730d8bb8b0220c ]

syzbot reported a kernel infoleak [1] of 4 bytes.

After analysis, it turned out r->idiag_expires is not initialized if inet_sctp_diag_fill() calls inet_diag_msg_common_fill()

Make sure to clear idiag_timer/idiag_retrans/idiag_expires and let inet_diag_msg_sctpasoc_fill() fill them again if needed.

[1]

BUG: KMSAN: kernel-infoleak in instrument_copy_to_user include/linux/instrumented.h:121 [inline]
BUG: KMSAN: kernel-infoleak in copyout lib/iov_iter.c:154 [inline]
BUG: KMSAN: kernel-infoleak in _copy_to_iter+0x6ef/0x25a0 lib/iov_iter.c:668
 instrument_copy_to_user include/linux/instrumented.h:121 [inline]
 copyout lib/iov_iter.c:154 [inline]
 _copy_to_iter+0x6ef/0x25a0 lib/iov_iter.c:668
 copy_to_iter include/linux/uio.h:162 [inline]
 simple_copy_to_iter+0xf3/0x140 net/core/datagram.c:519
 __skb_datagram_iter+0x2d5/0x11b0 net/core/datagram.c:425
 skb_copy_datagram_iter+0xdc/0x270 net/core/datagram.c:533
 skb_copy_datagram_msg include/linux/skbuff.h:3696 [inline]
 netlink_recvmsg+0x669/0x1c80 net/netlink/af_netlink.c:1977
 sock_recvmsg_nosec net/socket.c:948 [inline]
 sock_recvmsg net/socket.c:966 [inline]
 __sys_recvfrom+0x795/0xa10 net/socket.c:2097
 __do_sys_recvfrom net/socket.c:2115 [inline]
 __se_sys_recvfrom net/socket.c:2111 [inline]
 __x64_sys_recvfrom+0x19d/0x210 net/socket.c:2111
 do_syscall_x64 arch/x86/entry/common.c:51 [inline]
 do_syscall_64+0x54/0xd0 arch/x86/entry/common.c:82
 entry_SYSCALL_64_after_hwframe+0x44/0xae

Uninit was created at:

 slab_post_alloc_hook mm/slab.h:737 [inline]
 slab_alloc_node mm/slub.c:3247 [inline]
 __kmalloc_node_track_caller+0xe0c/0x1510 mm/slub.c:4975
 kmalloc_reserve net/core/skbuff.c:354 [inline]
 __alloc_skb+0x545/0xf90 net/core/skbuff.c:426
 alloc_skb include/linux/skbuff.h:1158 [inline]
 netlink_dump+0x3e5/0x16c0 net/netlink/af_netlink.c:2248
 __netlink_dump_start+0xcf8/0xe90 net/netlink/af_netlink.c:2373
 netlink_dump_start include/linux/netlink.h:254 [inline]
 inet_diag_handler_cmd+0x2e7/0x400 net/ipv4/inet_diag.c:1341
 sock_diag_rcv_msg+0x24a/0x620
 netlink_rcv_skb+0x40c/0x7e0 net/netlink/af_netlink.c:2494
 sock_diag_rcv+0x63/0x80 net/core/sock_diag.c:277
 netlink_unicast_kernel net/netlink/af_netlink.c:1317 [inline]
 netlink_unicast+0x1093/0x1360 net/netlink/af_netlink.c:1343
 netlink_sendmsg+0x14d9/0x1720 net/netlink/af_netlink.c:1919
 sock_sendmsg_nosec net/socket.c:705 [inline]
 sock_sendmsg net/socket.c:725 [inline]
 sock_write_iter+0x594/0x690 net/socket.c:1061
 do_iter_readv_writev+0xa7f/0xc70
 do_iter_write+0x52c/0x1500 fs/read_write.c:851
 vfs_writev fs/read_write.c:924 [inline]
 do_writev+0x645/0xe00 fs/read_write.c:967
 __do_sys_writev fs/read_write.c:1040 [inline]
 __se_sys_writev fs/read_write.c:1037 [inline]
 __x64_sys_writev+0xe5/0x120 fs/read_write.c:1037
 do_syscall_x64 arch/x86/entry/common.c:51 [inline]
 do_syscall_64+0x54/0xd0 arch/x86/entry/common.c:82
 entry_SYSCALL_64_after_hwframe+0x44/0xae

Bytes 68-71 of 2508 are uninitialized
Memory access of size 2508 starts at ffff888114f9b000 Data copied to user address 00007f7fe09ff2e0

CPU: 1 PID: 3478 Comm: syz-executor306 Not tainted 5.17.0-rc4-syzkaller #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011

---
 net/sctp/diag.c | 9 +++------
 1 file changed, 3 insertions(+), 6 deletions(-)

index 0a9db0a7f423..5f10984bf0f5 100644
@@ -76,10 +76,6 @@ static void inet_diag_msg_sctpasoc_fill(struct inet_diag_msg *r,
 		r->idiag_timer = SCTP_EVENT_TIMEOUT_T3_RTX;
 		r->idiag_retrans = asoc->rtx_data_chunks;
 		r->idiag_expires = jiffies_to_msecs(t3_rtx->expires - jiffies);
-	} else {
-		r->idiag_timer = 0;
-		r->idiag_retrans = 0;
-		r->idiag_expires = 0;
 	}
 }
 

@@ -159,13 +155,14 @@ static int inet_sctp_diag_fill(struct sock *sk, struct sctp_association *asoc,
 	r = nlmsg_data(nlh);
 	BUG_ON(!sk_fullsock(sk));
 
+	r->idiag_timer = 0;
+	r->idiag_retrans = 0;
+	r->idiag_expires = 0;
 	if (asoc) {
 		inet_diag_msg_sctpasoc_fill(r, sk, asoc);
 	} else {
 		inet_diag_msg_common_fill(r, sk);
 		r->idiag_state = sk->sk_state;
-		r->idiag_timer = 0;
-		r->idiag_retrans = 0;
 	}
 
 	if (inet_diag_msg_attrs_fill(sk, skb, r, ext, user_ns, net_admin))

2.34.1

From: suresh kumar <[email protected]>

[ Upstream commit 4224cfd7fb6523f7a9d1c8bb91bb5df1e38eb624 ]

When bringing down the netdevice or system shutdown, a panic can be triggered while accessing the sysfs path because the device is already removed.

    [  755.549084] mlx5_core 0000:12:00.1: Shutdown was called
    [  756.404455] mlx5_core 0000:12:00.0: Shutdown was called
    ...
    [  757.937260] BUG: unable to handle kernel NULL pointer dereference at           (null)
    [  758.031397] IP: [<ffffffff8ee11acb>] dma_pool_alloc+0x1ab/0x280

    crash> bt
    ...
    PID: 12649  TASK: ffff8924108f2100  CPU: 1   COMMAND: "amsd"
    ...
     #9 [ffff89240e1a38b0] page_fault at ffffffff8f38c778
        [exception RIP: dma_pool_alloc+0x1ab]
        RIP: ffffffff8ee11acb  RSP: ffff89240e1a3968  RFLAGS: 00010046
        RAX: 0000000000000246  RBX: ffff89243d874100  RCX: 0000000000001000
        RDX: 0000000000000000  RSI: 0000000000000246  RDI: ffff89243d874090
        RBP: ffff89240e1a39c0   R8: 000000000001f080   R9: ffff8905ffc03c00
        R10: ffffffffc04680d4  R11: ffffffff8edde9fd  R12: 00000000000080d0
        R13: ffff89243d874090  R14: ffff89243d874080  R15: 0000000000000000
        ORIG_RAX: ffffffffffffffff  CS: 0010  SS: 0018
    #10 [ffff89240e1a39c8] mlx5_alloc_cmd_msg at ffffffffc04680f3 [mlx5_core]
    #11 [ffff89240e1a3a18] cmd_exec at ffffffffc046ad62 [mlx5_core]
    #12 [ffff89240e1a3ab8] mlx5_cmd_exec at ffffffffc046b4fb [mlx5_core]
    #13 [ffff89240e1a3ae8] mlx5_core_access_reg at ffffffffc0475434 [mlx5_core]
    #14 [ffff89240e1a3b40] mlx5e_get_fec_caps at ffffffffc04a7348 [mlx5_core]
    #15 [ffff89240e1a3bb0] get_fec_supported_advertised at ffffffffc04992bf [mlx5_core]
    #16 [ffff89240e1a3c08] mlx5e_get_link_ksettings at ffffffffc049ab36 [mlx5_core]
    #17 [ffff89240e1a3ce8] __ethtool_get_link_ksettings at ffffffff8f25db46
    #18 [ffff89240e1a3d48] speed_show at ffffffff8f277208
    #19 [ffff89240e1a3dd8] dev_attr_show at ffffffff8f0b70e3
    #20 [ffff89240e1a3df8] sysfs_kf_seq_show at ffffffff8eedbedf
    #21 [ffff89240e1a3e18] kernfs_seq_show at ffffffff8eeda596
    #22 [ffff89240e1a3e28] seq_read at ffffffff8ee76d10
    #23 [ffff89240e1a3e98] kernfs_fop_read at ffffffff8eedaef5
    #24 [ffff89240e1a3ed8] vfs_read at ffffffff8ee4e3ff
    #25 [ffff89240e1a3f08] sys_read at ffffffff8ee4f27f
    #26 [ffff89240e1a3f50] system_call_fastpath at ffffffff8f395f92

    crash> net_device.state ffff89443b0c0000
      state = 0x5  (__LINK_STATE_START| __LINK_STATE_NOCARRIER)

To prevent this scenario, we also make sure that the netdevice is present.

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

index e5dc04cb5599..7a11b2d90975 100644
@@ -203,7 +203,7 @@ static ssize_t speed_show(struct device *dev,
 	if (!rtnl_trylock())
 		return restart_syscall();
 
-	if (netif_running(netdev)) {
+	if (netif_running(netdev) && netif_device_present(netdev)) {
 		struct ethtool_link_ksettings cmd;
 
 		if (!__ethtool_get_link_ksettings(netdev, &cmd))

2.34.1

From: Marek Marczykowski-Górecki <[email protected]>

[ Upstream commit 0f4558ae91870692ce7f509c31c9d6ee721d8cdc ]

This reverts commit 1f2565780e9b7218cf92c7630130e82dcc0fe9c2.

The 'hotplug-status' node should not be removed as long as the vif device remains configured. Otherwise the xen-netback would wait for re-running the network script even if it was already called (in case of the frontent re-connecting). But also, it _should_ be removed when the vif device is destroyed (for example when unbinding the driver) - otherwise hotplug script would not configure the device whenever it re-appear.

Moving removal of the 'hotplug-status' node was a workaround for nothing calling network script after xen-netback module is reloaded. But when vif interface is re-created (on xen-netback unbind/bind for example), the script should be called, regardless of who does that - currently this case is not handled by the toolstack, and requires manual script call. Keeping hotplug-status=connected to skip the call is wrong and leads to not configured interface.

More discussion at
https://lore.kernel.org/xen-devel/[email protected]/T/#u

---
 drivers/net/xen-netback/xenbus.c | 1 +
 1 file changed, 1 insertion(+)

index 78c56149559c..6b678ab0a31f 100644
@@ -499,6 +499,7 @@ static void backend_disconnect(struct backend_info *be)
 		unsigned int queue_index;
 
 		xen_unregister_watchers(vif);
+		xenbus_rm(XBT_NIL, be->dev->nodename, "hotplug-status");
 #ifdef CONFIG_DEBUG_FS
 		xenvif_debugfs_delif(vif);
 #endif /* CONFIG_DEBUG_FS */

2.34.1

From: Marek Marczykowski-Górecki <[email protected]>

[ Upstream commit e8240addd0a3919e0fd7436416afe9aa6429c484 ]

This reverts commit 2afeec08ab5c86ae21952151f726bfe184f6b23d.

The reasoning in the commit was wrong - the code expected to setup the watch even if 'hotplug-status' didn't exist. In fact, it relied on the watch being fired the first time - to check if maybe 'hotplug-status' is already set to 'connected'. Not registering a watch for non-existing path (which is the case if hotplug script hasn't been executed yet), made the backend not waiting for the hotplug script to execute. This in turns, made the netfront think the interface is fully operational, while in fact it was not (the vif interface on xen-netback side might not be configured yet).

This was a workaround for 'hotplug-status' erroneously being removed. But since that is reverted now, the workaround is not necessary either.

More discussion at
https://lore.kernel.org/xen-devel/[email protected]/T/#u

---
 drivers/net/xen-netback/xenbus.c | 12 ++++--------
 1 file changed, 4 insertions(+), 8 deletions(-)

index 6b678ab0a31f..6d7fb0a95645 100644
@@ -1044,15 +1044,11 @@ static void connect(struct backend_info *be)
 	xenvif_carrier_on(be->vif);
 
 	unregister_hotplug_status_watch(be);
-	if (xenbus_exists(XBT_NIL, dev->nodename, "hotplug-status")) {
-		err = xenbus_watch_pathfmt(dev, &be->hotplug_status_watch,
-					   NULL, hotplug_status_changed,
-					   "%s/%s", dev->nodename,
-					   "hotplug-status");
-		if (err)
-			goto err;
+	err = xenbus_watch_pathfmt(dev, &be->hotplug_status_watch, NULL,
+				   hotplug_status_changed,
+				   "%s/%s", dev->nodename, "hotplug-status");
+	if (!err)
 		be->have_hotplug_status_watch = 1;
-	}
 
 	netif_tx_wake_all_queues(be->vif->dev);
 

2.34.1

From: Sven Schnelle <[email protected]>

[ Upstream commit 7acf3a127bb7c65ff39099afd78960e77b2ca5de ]

Booting the kernel with 'trace_buf_size=1' give a warning at boot during the ftrace selftests:

[    0.892809] Running postponed tracer tests:
[    0.892893] Testing tracer function:
[    0.901899] Callback from call_rcu_tasks_trace() invoked.
[    0.983829] Callback from call_rcu_tasks_rude() invoked.
[    1.072003] .. bad ring buffer .. corrupted trace buffer ..
[    1.091944] Callback from call_rcu_tasks() invoked.
[    1.097695] PASSED
[    1.097701] Testing dynamic ftrace: .. filter failed count=0 ..FAILED!
[    1.353474] ------------[ cut here ]------------
[    1.353478] WARNING: CPU: 0 PID: 1 at kernel/trace/trace.c:1951 run_tracer_selftest+0x13c/0x1b0

Therefore enforce a minimum of 4096 bytes to make the selftest pass.

---
 kernel/trace/trace.c | 10 ++++++----
 1 file changed, 6 insertions(+), 4 deletions(-)

index 19a6b088f1e7..063b434c89d2 100644
@@ -1118,10 +1118,12 @@ static int __init set_buf_size(char *str)
 	if (!str)
 		return 0;
 	buf_size = memparse(str, &str);
-	/* nr_entries can not be zero */
-	if (buf_size == 0)
-		return 0;
-	trace_buf_size = buf_size;
+	/*
+	 * nr_entries can not be zero and the startup
+	 * tests require some buffer space. Therefore
+	 * ensure we have at least 4096 bytes of buffer.
+	 */
+	trace_buf_size = max(4096UL, buf_size);
 	return 1;
 }
 __setup("trace_buf_size=", set_buf_size);

2.34.1

From: Mike Kravetz <[email protected]>

[ Upstream commit fda153c89af344d21df281009a9d046cf587ea0f ]

Running the memfd script ./run_hugetlbfs_test.sh will often end in error as follows:

memfd-hugetlb: CREATE
memfd-hugetlb: BASIC
memfd-hugetlb: SEAL-WRITE
memfd-hugetlb: SEAL-FUTURE-WRITE
memfd-hugetlb: SEAL-SHRINK fallocate(ALLOC) failed: No space left on device
./run_hugetlbfs_test.sh: line 60: 166855 Aborted (core dumped) ./memfd_test hugetlbfs
opening: ./mnt/memfd
fuse: DONE

If no hugetlb pages have been preallocated, run_hugetlbfs_test.sh will allocate 'just enough' pages to run the test. In the SEAL-FUTURE-WRITE test the mfd_fail_write routine maps the file, but does not unmap. As a result, two hugetlb pages remain reserved for the mapping. When the fallocate call in the SEAL-SHRINK test attempts allocate all hugetlb pages, it is short by the two reserved pages.

Fix by making sure to unmap in mfd_fail_write.

---
 tools/testing/selftests/memfd/memfd_test.c | 1 +
 1 file changed, 1 insertion(+)

index 10baa1652fc2..a4e520b94e43 100644
@@ -386,6 +386,7 @@ static void mfd_fail_write(int fd)
 			printf("mmap()+mprotect() didn't fail as expected\n");
 			abort();
 		}
+		munmap(p, mfd_def_size);
 	}
 
 	/* verify PUNCH_HOLE fails */

2.34.1

From: Randy Dunlap <[email protected]>

commit 68453767131a5deec1e8f9ac92a9042f929e585d upstream.

When CONFIG_GENERIC_CPU_VULNERABILITIES is not set, references to spectre_v2_update_state() cause a build error, so provide an empty stub for that function when the Kconfig option is not set.

Fixes this build error:

arm-linux-gnueabi-ld: arch/arm/mm/proc-v7-bugs.o: in function `cpu_v7_bugs_init':
proc-v7-bugs.c:(.text+0x52): undefined reference to `spectre_v2_update_state'
arm-linux-gnueabi-ld: proc-v7-bugs.c:(.text+0x82): undefined reference to `spectre_v2_update_state'

---
 arch/arm/include/asm/spectre.h |    6 ++++++
 1 file changed, 6 insertions(+)

@@ -25,7 +25,13 @@ enum {
 	SPECTRE_V2_METHOD_LOOP8 = BIT(__SPECTRE_V2_METHOD_LOOP8),
 };
 
+#ifdef CONFIG_GENERIC_CPU_VULNERABILITIES
 void spectre_v2_update_state(unsigned int state, unsigned int methods);
+#else
+static inline void spectre_v2_update_state(unsigned int state,
+					   unsigned int methods)
+{}
+#endif
 
 int spectre_bhb_update_vectors(unsigned int method);
 

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

From: Dan Carpenter <[email protected]>

commit fc7f750dc9d102c1ed7bbe4591f991e770c99033 upstream.

The netif_rx_ni() function frees the skb so we can't dereference it to save the skb->len.

---
 drivers/staging/gdm724x/gdm_lte.c |    5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

@@ -76,14 +76,15 @@ static void tx_complete(void *arg)
 
 static int gdm_lte_rx(struct sk_buff *skb, struct nic *nic, int nic_type)
 {
-	int ret;
+	int ret, len;
 
+	len = skb->len + ETH_HLEN;
 	ret = netif_rx_ni(skb);
 	if (ret == NET_RX_DROP) {
 		nic->stats.rx_dropped++;
 	} else {
 		nic->stats.rx_packets++;
-		nic->stats.rx_bytes += skb->len + ETH_HLEN;
+		nic->stats.rx_bytes += len;
 	}
 
 	return 0;

From: Robert Hancock <[email protected]>

commit 0bf476fc3624e3a72af4ba7340d430a91c18cd67 upstream.

There is an oddity in the way the RSR register flags propagate to the ISR register (and the actual interrupt output) on this hardware: it appears that RSR register bits only result in ISR being asserted if the interrupt was actually enabled at the time, so enabling interrupts with RSR bits already set doesn't trigger an interrupt to be raised. There was already a partial fix for this race in the macb_poll function where it checked for RSR bits being set and re-triggered NAPI receive. However, there was a still a race window between checking RSR and actually enabling interrupts, where a lost wakeup could happen. It's necessary to check again after enabling interrupts to see if RSR was set just prior to the interrupt being enabled, and re-trigger receive in that case.

This issue was noticed in a point-to-point UDP request-response protocol which periodically saw timeouts or abnormally high response times due to received packets not being processed in a timely fashion. In many applications, more packets arriving, including TCP retransmissions, would cause the original packet to be processed, thus masking the issue.

---
 drivers/net/ethernet/cadence/macb_main.c |   25 ++++++++++++++++++++++++-
 1 file changed, 24 insertions(+), 1 deletion(-)

@@ -1269,7 +1269,14 @@ static int macb_poll(struct napi_struct
 	if (work_done < budget) {
 		napi_complete_done(napi, work_done);
 
-		/* Packets received while interrupts were disabled */
+		/* RSR bits only seem to propagate to raise interrupts when
+		 * interrupts are enabled at the time, so if bits are already
+		 * set due to packets received while interrupts were disabled,
+		 * they will not cause another interrupt to be generated when
+		 * interrupts are re-enabled.
+		 * Check for this case here. This has been seen to happen
+		 * around 30% of the time under heavy network load.
+		 */
 		status = macb_readl(bp, RSR);
 		if (status) {
 			if (bp->caps & MACB_CAPS_ISR_CLEAR_ON_WRITE)

@@ -1277,6 +1284,22 @@ static int macb_poll(struct napi_struct
 			napi_reschedule(napi);
 		} else {
 			queue_writel(queue, IER, bp->rx_intr_mask);
+
+			/* In rare cases, packets could have been received in
+			 * the window between the check above and re-enabling
+			 * interrupts. Therefore, a double-check is required
+			 * to avoid losing a wakeup. This can potentially race
+			 * with the interrupt handler doing the same actions
+			 * if an interrupt is raised just after enabling them,
+			 * but this should be harmless.
+			 */
+			status = macb_readl(bp, RSR);
+			if (unlikely(status)) {
+				queue_writel(queue, IDR, bp->rx_intr_mask);
+				if (bp->caps & MACB_CAPS_ISR_CLEAR_ON_WRITE)
+					queue_writel(queue, ISR, MACB_BIT(RCOMP));
+				napi_schedule(napi);
+			}
 		}
 	}
 

From: Emil Renner Berthing <[email protected]>

commit 0966d385830de3470b7131db8e86c0c5bc9c52dc upstream.

RISC-V can do PC-relative jumps with a 32bit range using the following two instructions:

	auipc	t0, imm20	; t0 = PC + imm20 * 2^12
	jalr	ra, t0, imm12	; ra = PC + 4, PC = t0 + imm12

Crucially both the 20bit immediate imm20 and the 12bit immediate imm12 are treated as two's-complement signed values. For this reason the immediates are usually calculated like this:

	imm20 = (offset + 0x800) >> 12
	imm12 = offset & 0xfff

..where offset is the signed offset from the auipc instruction. When the 11th bit of offset is 0 the addition of 0x800 doesn't change the top 20 bits and imm12 considered positive. When the 11th bit is 1 the carry of the addition by 0x800 means imm20 is one higher, but since imm12 is then considered negative the two's complement representation means it all cancels out nicely.

However, this addition by 0x800 (2^11) means an offset greater than or equal to 2^31 - 2^11 would overflow so imm20 is considered negative and result in a backwards jump. Similarly the lower range of offset is also moved down by 2^11 and hence the true 32bit range is

	[-2^31 - 2^11, 2^31 - 2^11)

---
 arch/riscv/kernel/module.c |   21 ++++++++++++++++-----
 1 file changed, 16 insertions(+), 5 deletions(-)

@@ -21,6 +21,19 @@
 #include <asm/pgtable.h>
 #include <asm/sections.h>
 
+/*
+ * The auipc+jalr instruction pair can reach any PC-relative offset
+ * in the range [-2^31 - 2^11, 2^31 - 2^11)
+ */
+static bool riscv_insn_valid_32bit_offset(ptrdiff_t val)
+{
+#ifdef CONFIG_32BIT
+	return true;
+#else
+	return (-(1L << 31) - (1L << 11)) <= val && val < ((1L << 31) - (1L << 11));
+#endif
+}
+
 static int apply_r_riscv_32_rela(struct module *me, u32 *location, Elf_Addr v)
 {
 	if (v != (u32)v) {

@@ -103,7 +116,7 @@ static int apply_r_riscv_pcrel_hi20_rela
 	ptrdiff_t offset = (void *)v - (void *)location;
 	s32 hi20;
 
-	if (offset != (s32)offset) {
+	if (!riscv_insn_valid_32bit_offset(offset)) {
 		pr_err(
 		  "%s: target %016llx can not be addressed by the 32-bit offset from PC = %p\n",
 		  me->name, (long long)v, location);

@@ -205,10 +218,9 @@ static int apply_r_riscv_call_plt_rela(s
 				       Elf_Addr v)
 {
 	ptrdiff_t offset = (void *)v - (void *)location;
-	s32 fill_v = offset;
 	u32 hi20, lo12;
 
-	if (offset != fill_v) {
+	if (!riscv_insn_valid_32bit_offset(offset)) {
 		/* Only emit the plt entry if offset over 32-bit range */
 		if (IS_ENABLED(CONFIG_MODULE_SECTIONS)) {
 			offset = module_emit_plt_entry(me, v);

@@ -232,10 +244,9 @@ static int apply_r_riscv_call_rela(struc
 				   Elf_Addr v)
 {
 	ptrdiff_t offset = (void *)v - (void *)location;
-	s32 fill_v = offset;
 	u32 hi20, lo12;
 
-	if (offset != fill_v) {
+	if (!riscv_insn_valid_32bit_offset(offset)) {
 		pr_err(
 		  "%s: target %016llx can not be addressed by the 32-bit offset from PC = %p\n",
 		  me->name, (long long)v, location);

From: James Morse <[email protected]>

The logic in commit 2a5f1b67ec57 "KVM: arm64: Don't access PMCR_EL0 when no PMU is available" relies on an empty reset handler being benign. This was not the case in earlier kernel versions, so the stable backport of this patch is causing problems.

KVMs behaviour in this area changed over time. In particular, prior to commit 03fdfb269009 ("KVM: arm64: Don't write junk to sysregs on reset"), an empty reset handler will trigger a warning, as the guest registers have been poisoned.
Prior to commit 20589c8cc47d ("arm/arm64: KVM: Don't panic on failure to properly reset system registers"), this warning was a panic().

Instead of reverting the backport, make it write 0 to the sys_reg[] array. This keeps the reset logic happy, and the dodgy value can't be seen by the guest as it can't request the emulation.

The original bug was accessing the PMCR_EL0 register on CPUs that don't implement that feature. There is no known silicon that does this, but v4.9's ACPI support is unable to find the PMU, so triggers this code:

| Kernel panic - not syncing: Didn't reset vcpu_sys_reg(24)
| CPU: 1 PID: 3055 Comm: lkvm Not tainted 4.9.302-00032-g64e078a56789 #13476
| Hardware name: ARM LTD ARM Juno Development Platform/ARM Juno Development Platform, BIOS EDK II Jul 30 2018
| Call trace:
| [<ffff00000808b4b0>] dump_backtrace+0x0/0x1a0
| [<ffff00000808b664>] show_stack+0x14/0x20
| [<ffff0000088f0e18>] dump_stack+0x98/0xb8
| [<ffff0000088eef08>] panic+0x118/0x274
| [<ffff0000080b50e0>] access_actlr+0x0/0x20
| [<ffff0000080b2620>] kvm_reset_vcpu+0x5c/0xac
| [<ffff0000080ac688>] kvm_arch_vcpu_ioctl+0x3e4/0x490
| [<ffff0000080a382c>] kvm_vcpu_ioctl+0x5b8/0x720
| [<ffff000008201e44>] do_vfs_ioctl+0x2f4/0x884
| [<ffff00000820244c>] SyS_ioctl+0x78/0x9c
| [<ffff000008083a9c>] __sys_trace_return+0x0/0x4

---
 arch/arm64/kvm/sys_regs.c |    4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

@@ -620,8 +620,10 @@ static void reset_pmcr(struct kvm_vcpu *
 	u64 pmcr, val;
 
 	/* No PMU available, PMCR_EL0 may UNDEF... */
-	if (!kvm_arm_support_pmu_v3())
+	if (!kvm_arm_support_pmu_v3()) {
+		vcpu_sys_reg(vcpu, PMCR_EL0) = 0;
 		return;
+	}
 
 	pmcr = read_sysreg(pmcr_el0);
 	/*

From: Michael S. Tsirkin <[email protected]>

commit 838d6d3461db0fdbf33fc5f8a69c27b50b4a46da upstream.

virtio_finalize_features is only used internally within virtio. No reason to export it.

---
 drivers/virtio/virtio.c |    3 +--
 include/linux/virtio.h  |    1 -
 2 files changed, 1 insertion(+), 3 deletions(-)

@@ -165,7 +165,7 @@ void virtio_add_status(struct virtio_dev
 }
 EXPORT_SYMBOL_GPL(virtio_add_status);
 
-int virtio_finalize_features(struct virtio_device *dev)
+static int virtio_finalize_features(struct virtio_device *dev)
 {
 	int ret = dev->config->finalize_features(dev);
 	unsigned status;

@@ -185,7 +185,6 @@ int virtio_finalize_features(struct virt
 	}
 	return 0;
 }
-EXPORT_SYMBOL_GPL(virtio_finalize_features);
 
 static int virtio_dev_probe(struct device *_d)
 {

@@ -151,7 +151,6 @@ void virtio_break_device(struct virtio_d
 void virtio_config_changed(struct virtio_device *dev);
 void virtio_config_disable(struct virtio_device *dev);
 void virtio_config_enable(struct virtio_device *dev);
-int virtio_finalize_features(struct virtio_device *dev);
 #ifdef CONFIG_PM_SLEEP
 int virtio_device_freeze(struct virtio_device *dev);
 int virtio_device_restore(struct virtio_device *dev);

From: Michael S. Tsirkin <[email protected]>

commit 4fa59ede95195f267101a1b8916992cf3f245cdb upstream.

The feature negotiation was designed in a way that makes it possible for devices to know which config fields will be accessed by drivers.

This is broken since commit 404123c2db79 ("virtio: allow drivers to validate features") with fallout in at least block and net. We have a partial work-around in commit 2f9a174f918e ("virtio: write back F_VERSION_1 before validate") which at least lets devices find out which format should config space have, but this is a partial fix: guests should not access config space without acknowledging features since otherwise we'll never be able to change the config space format.

To fix, split finalize_features from virtio_finalize_features and call finalize_features with all feature bits before validation, and then - if validation changed any bits - once again after.

Since virtio_finalize_features no longer writes out features rename it to virtio_features_ok - since that is what it does:
checks that features are ok with the device.

As a side effect, this also reduces the amount of hypervisor accesses - we now only acknowledge features once unless we are clearing any features when validating (which is uncommon).

IRC I think that this was more or less always the intent in the spec but unfortunately the way the spec is worded does not say this explicitly, I plan to address this at the spec level, too.

---
 drivers/virtio/virtio.c       |   39 +++++++++++++++++++++------------------
 include/linux/virtio_config.h |    3 ++-
 2 files changed, 23 insertions(+), 19 deletions(-)

@@ -165,14 +165,11 @@ void virtio_add_status(struct virtio_dev
 }
 EXPORT_SYMBOL_GPL(virtio_add_status);
 
-static int virtio_finalize_features(struct virtio_device *dev)
+/* Do some validation, then set FEATURES_OK */
+static int virtio_features_ok(struct virtio_device *dev)
 {
-	int ret = dev->config->finalize_features(dev);
 	unsigned status;
 
-	if (ret)
-		return ret;
-
 	if (!virtio_has_feature(dev, VIRTIO_F_VERSION_1))
 		return 0;
 

@@ -221,17 +218,6 @@ static int virtio_dev_probe(struct devic
 		driver_features_legacy = driver_features;
 	}
 
-	/*
-	 * Some devices detect legacy solely via F_VERSION_1. Write
-	 * F_VERSION_1 to force LE config space accesses before FEATURES_OK for
-	 * these when needed.
-	 */
-	if (drv->validate && !virtio_legacy_is_little_endian()
-			  && device_features & BIT_ULL(VIRTIO_F_VERSION_1)) {
-		dev->features = BIT_ULL(VIRTIO_F_VERSION_1);
-		dev->config->finalize_features(dev);
-	}
-
 	if (device_features & (1ULL << VIRTIO_F_VERSION_1))
 		dev->features = driver_features & device_features;
 	else

@@ -242,13 +228,26 @@ static int virtio_dev_probe(struct devic
 		if (device_features & (1ULL << i))
 			__virtio_set_bit(dev, i);
 
+	err = dev->config->finalize_features(dev);
+	if (err)
+		goto err;
+
 	if (drv->validate) {
+		u64 features = dev->features;
+
 		err = drv->validate(dev);
 		if (err)
 			goto err;
+
+		/* Did validation change any features? Then write them again. */
+		if (features != dev->features) {
+			err = dev->config->finalize_features(dev);
+			if (err)
+				goto err;
+		}
 	}
 
-	err = virtio_finalize_features(dev);
+	err = virtio_features_ok(dev);
 	if (err)
 		goto err;
 

@@ -412,7 +411,11 @@ int virtio_device_restore(struct virtio_
 	/* We have a driver! */
 	virtio_add_status(dev, VIRTIO_CONFIG_S_DRIVER);
 
-	ret = virtio_finalize_features(dev);
+	ret = dev->config->finalize_features(dev);
+	if (ret)
+		goto err;
+
+	ret = virtio_features_ok(dev);
 	if (ret)
 		goto err;
 

@@ -51,8 +51,9 @@ struct irq_affinity;
  *	Returns the first 32 feature bits (all we currently need).
  * @finalize_features: confirm what device features we'll be using.
  *	vdev: the virtio_device
- *	This gives the final feature bits for the device: it can change
+ *	This sends the driver feature bits to the device: it can change
  *	the dev->feature bits if it wants.
+ * Note: despite the name this can be called any number of times.
  *	Returns 0 on success or error status
  * @bus_name: return the bus name associated with the device
  *	vdev: the virtio_device

From: Russell King (Oracle) <[email protected]>

commit 6c7cb60bff7aec24b834343ff433125f469886a3 upstream.

When building for Thumb2, the vectors make use of a local label. Sadly, the Spectre BHB code also uses a local label with the same number which results in the Thumb2 reference pointing at the wrong place. Fix this by changing the number used for the Spectre BHB local label.

---
 arch/arm/kernel/entry-armv.S |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

@@ -1067,9 +1067,9 @@ vector_bhb_loop8_\name:
 
 	@ bhb workaround
 	mov	r0, #8
-1:	b	. + 4
+3:	b	. + 4
 	subs	r0, r0, #1
-	bne	1b
+	bne	3b
 	dsb
 	isb
 	b	2b

From: Josh Triplett <[email protected]>

commit b1489186cc8391e0c1e342f9fbc3eedf6b944c61 upstream.

The in-kernel ext4 resize code doesn't support filesystem with the
sparse_super2 feature. It fails with errors like this and doesn't finish
the resize:
EXT4-fs (loop0): resizing filesystem from 16640 to 7864320 blocks
EXT4-fs warning (device loop0): verify_reserved_gdb:760: reserved GDT 2 missing grp 1 (32770) EXT4-fs warning (device loop0): ext4_resize_fs:2111: error (-22) occurred during file system resize EXT4-fs (loop0): resized filesystem to 2097152

To reproduce:
mkfs.ext4 -b 4096 -I 256 -J size=32 -E resize=$((256*1024*1024)) -O sparse_super2 ext4.img 65M
truncate -s 30G ext4.img
mount ext4.img /mnt
python3 -c 'import fcntl, os, struct ; fd = os.open("/mnt", os.O_RDONLY | os.O_DIRECTORY) ; fcntl.ioctl(fd, 0x40086610, struct.pack("Q", 30 * 1024 * 1024 * 1024 // 4096), False) ; os.close(fd)' dmesg | tail
e2fsck ext4.img

The userspace resize2fs tool has a check for this case: it checks if the filesystem has sparse_super2 set and if the kernel provides /sys/fs/ext4/features/sparse_super2. However, the former check requires manually reading and parsing the filesystem superblock.

Detect this case in ext4_resize_begin and error out early with a clear error message.

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

@@ -74,6 +74,11 @@ int ext4_resize_begin(struct super_block
 		return -EPERM;
 	}
 
+	if (ext4_has_feature_sparse_super2(sb)) {
+		ext4_msg(sb, KERN_ERR, "Online resizing not supported with sparse_super2");
+		return -EOPNOTSUPP;
+	}
+
 	if (test_and_set_bit_lock(EXT4_FLAGS_RESIZING,
 				  &EXT4_SB(sb)->s_ext4_flags))
 		ret = -EBUSY;

From: Qu Wenruo <[email protected]>

commit 19ea40dddf1833db868533958ca066f368862211 upstream.

[BUG]
There is a bug report that injected ENOMEM error could leave a tree block locked while we return to user-space:

BTRFS info (device loop0): enabling ssd optimizations
FAULT_INJECTION: forcing a failure.
name failslab, interval 1, probability 0, space 0, times 0

  CPU: 0 PID: 7579 Comm: syz-executor Not tainted 5.15.0-rc1 #16
  Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS
  rel-1.12.0-59-gc9ba5276e321-prebuilt.qemu.org 04/01/2014
  Call Trace:
   __dump_stack lib/dump_stack.c:88 [inline]
   dump_stack_lvl+0x8d/0xcf lib/dump_stack.c:106
   fail_dump lib/fault-inject.c:52 [inline]
   should_fail+0x13c/0x160 lib/fault-inject.c:146
   should_failslab+0x5/0x10 mm/slab_common.c:1328
   slab_pre_alloc_hook.constprop.99+0x4e/0xc0 mm/slab.h:494
   slab_alloc_node mm/slub.c:3120 [inline]
   slab_alloc mm/slub.c:3214 [inline]
   kmem_cache_alloc+0x44/0x280 mm/slub.c:3219
   btrfs_alloc_delayed_extent_op fs/btrfs/delayed-ref.h:299 [inline]
   btrfs_alloc_tree_block+0x38c/0x670 fs/btrfs/extent-tree.c:4833
   __btrfs_cow_block+0x16f/0x7d0 fs/btrfs/ctree.c:415
   btrfs_cow_block+0x12a/0x300 fs/btrfs/ctree.c:570
   btrfs_search_slot+0x6b0/0xee0 fs/btrfs/ctree.c:1768
   btrfs_insert_empty_items+0x80/0xf0 fs/btrfs/ctree.c:3905
   btrfs_new_inode+0x311/0xa60 fs/btrfs/inode.c:6530
   btrfs_create+0x12b/0x270 fs/btrfs/inode.c:6783
   lookup_open+0x660/0x780 fs/namei.c:3282
   open_last_lookups fs/namei.c:3352 [inline]
   path_openat+0x465/0xe20 fs/namei.c:3557
   do_filp_open+0xe3/0x170 fs/namei.c:3588
   do_sys_openat2+0x357/0x4a0 fs/open.c:1200
   do_sys_open+0x87/0xd0 fs/open.c:1216
   do_syscall_x64 arch/x86/entry/common.c:50 [inline]
   do_syscall_64+0x34/0xb0 arch/x86/entry/common.c:80
   entry_SYSCALL_64_after_hwframe+0x44/0xae
  RIP: 0033:0x46ae99
  Code: f7 d8 64 89 02 b8 ff ff ff ff c3 66 0f 1f 44 00 00 48 89 f8 48
  89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d
  01 f0 ff ff 73 01 c3 48 c7 c1 bc ff ff ff f7 d8 64 89 01 48
  RSP: 002b:00007f46711b9c48 EFLAGS: 00000246 ORIG_RAX: 0000000000000055
  RAX: ffffffffffffffda RBX: 000000000078c0a0 RCX: 000000000046ae99
  RDX: 0000000000000000 RSI: 00000000000000a1 RDI: 0000000020005800
  RBP: 00007f46711b9c80 R08: 0000000000000000 R09: 0000000000000000
  R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000017
  R13: 0000000000000000 R14: 000000000078c0a0 R15: 00007ffc129da6e0

================================================
WARNING: lock held when returning to user space! 5.15.0-rc1 #16 Not tainted
------------------------------------------------
syz-executor/7579 is leaving the kernel with locks still held! 1 lock held by syz-executor/7579:
#0: ffff888104b73da8 (btrfs-tree-01/1){+.+.}-{3:3}, at:
__btrfs_tree_lock+0x2e/0x1a0 fs/btrfs/locking.c:112

[CAUSE]
In btrfs_alloc_tree_block(), after btrfs_init_new_buffer(), the new extent buffer @buf is locked, but if later operations like adding delayed tree ref fail, we just free @buf without unlocking it, resulting above warning.

[FIX]
Unlock @buf in out_free_buf: label.

---
 fs/btrfs/extent-tree.c |    1 +
 1 file changed, 1 insertion(+)

@@ -8327,6 +8327,7 @@ struct extent_buffer *btrfs_alloc_tree_b
 out_free_delayed:
 	btrfs_free_delayed_extent_op(extent_op);
 out_free_buf:
+	btrfs_tree_unlock(buf);
 	free_extent_buffer(buf);
 out_free_reserved:
 	btrfs_free_reserved_extent(fs_info, ins.objectid, ins.offset, 0);

From: Valentin Schneider <[email protected]>

commit 620a6dc40754dc218f5b6389b5d335e9a107fd29 upstream.

The deduplicating sort in sched_init_numa() assumes that the first line in the distance table contains all unique values in the entire table. I've been trying to pen what this exactly means for the topology, but it's not straightforward. For instance, topology.c uses this example:

node 0 1 2 3

    0:  10  20  20  30
    1:  20  10  20  20
    2:  20  20  10  20
    3:  30  20  20  10

0 ----- 1

  |     / |
  |   /   |
  | /     |

2 ----- 3

Which works out just fine. However, if we swap nodes 0 and 1:

1 ----- 0

  |     / |
  |   /   |
  | /     |

2 ----- 3

we get this distance table:

node 0 1 2 3

    0:  10 20 20 20
    1:  20 10 20 30
    2:  20 20 10 20
    3:  20 30 20 10

Which breaks the deduplicating sort (non-representative first line). In this case this would just be a renumbering exercise, but it so happens that we can have a deduplicating sort that goes through the whole table in O(n²) at the extra cost of a temporary memory allocation (i.e. any form of set).

The ACPI spec (SLIT) mentions distances are encoded on 8 bits. Following this, implement the set as a 256-bits bitmap. Should this not be satisfactory (i.e. we want to support 32-bit values), then we'll have to go for some other sparse set implementation.

This has the added benefit of letting us allocate just the right amount of memory for sched_domains_numa_distance[], rather than an arbitrary (nr_node_ids + 1).

Note: DT binding equivalent (distance-map) decodes distances as 32-bit values.

---
 include/linux/topology.h |    1 
 kernel/sched/topology.c  |   99 ++++++++++++++++++++++-------------------------
 2 files changed, 49 insertions(+), 51 deletions(-)

@@ -47,6 +47,7 @@ int arch_update_cpu_topology(void);
 /* Conform to ACPI 2.0 SLIT distance definitions */
 #define LOCAL_DISTANCE		10
 #define REMOTE_DISTANCE		20
+#define DISTANCE_BITS           8
 #ifndef node_distance
 #define node_distance(from,to)	((from) == (to) ? LOCAL_DISTANCE : REMOTE_DISTANCE)
 #endif

From: Dietmar Eggemann <[email protected]>

commit 71e5f6644fb2f3304fcb310145ded234a37e7cc1 upstream.

Commit "sched/topology: Make sched_init_numa() use a set for the deduplicating sort" allocates 'i + nr_levels (level)' instead of 'i + nr_levels + 1' sched_domain_topology_level.

This led to an Oops (on Arm64 juno with CONFIG_SCHED_DEBUG):

sched_init_domains
build_sched_domains()

    __free_domain_allocs()
      __sdt_free() {
	...
        for_each_sd_topology(tl)
	  ...
          sd = *per_cpu_ptr(sdd->sd, j); <--
	  ...
      }

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

@@ -1428,7 +1428,7 @@ void sched_init_numa(void)
 	/* Compute default topology size */
 	for (i = 0; sched_domain_topology[i].mask; i++);
 
-	tl = kzalloc((i + nr_levels) *
+	tl = kzalloc((i + nr_levels + 1) *
 			sizeof(struct sched_domain_topology_level), GFP_KERNEL);
 	if (!tl)
 		return;

From: Valentin Schneider <[email protected]>

commit b22a8f7b4bde4e4ab73b64908ffd5d90ecdcdbfd upstream.

John Paul reported a warning about bogus NUMA distance values spurred by commit:

620a6dc40754 ("sched/topology: Make sched_init_numa() use a set for the deduplicating sort")

In this case, the afflicted machine comes up with a reported 256 possible nodes, all of which are 0 distance away from one another. This was previously silently ignored, but is now caught by the aforementioned commit.

The culprit is ia64's node_possible_map which remains unchanged from its initialization value of NODE_MASK_ALL. In John's case, the machine doesn't have any SRAT nor SLIT table, but AIUI the possible map remains untouched regardless of what ACPI tables end up being parsed. Thus, !online && possible nodes remain with a bogus distance of 0 (distances \in [0, 9] are "reserved and have no meaning" as per the ACPI spec).

Follow x86 / drivers/base/arch_numa's example and set the possible map to the parsed map, which in this case seems to be the online map.

---
 arch/ia64/kernel/acpi.c |    7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

@@ -537,7 +537,8 @@ void __init acpi_numa_fixup(void)
 	if (srat_num_cpus == 0) {
 		node_set_online(0);
 		node_cpuid[0].phys_id = hard_smp_processor_id();
-		return;
+		slit_distance(0, 0) = LOCAL_DISTANCE;
+		goto out;
 	}
 
 	/*

@@ -580,7 +581,7 @@ void __init acpi_numa_fixup(void)
 			for (j = 0; j < MAX_NUMNODES; j++)
 				node_distance(i, j) = i == j ? LOCAL_DISTANCE :
 							REMOTE_DISTANCE;
-		return;
+		goto out;
 	}
 
 	memset(numa_slit, -1, sizeof(numa_slit));

@@ -605,6 +606,8 @@ void __init acpi_numa_fixup(void)
 		printk("\n");
 	}
 #endif
+out:
+	node_possible_map = node_online_map;
 }
 #endif				/* CONFIG_ACPI_NUMA */
 

Hi Greg,

...

The above is causing the following build error for ARM64 ...

arch/arm64/kvm/sys_regs.c: In function ‘reset_pmcr’:
arch/arm64/kvm/sys_regs.c:624:3: error: implicit declaration of function ‘vcpu_sys_reg’ [-Werror=implicit-function-declaration]
vcpu_sys_reg(vcpu, PMCR_EL0) = 0;
^~~~~~~~~~~~
arch/arm64/kvm/sys_regs.c:624:32: error: lvalue required as left operand of assignment
vcpu_sys_reg(vcpu, PMCR_EL0) = 0;

Cheers
Jon

nvpublic

Is this also broken in Linus's tree?

thanks,

greg k-h

No, Linus' tree is not broken. However, I don't see this change in Linus' tree (v5.17-rc8).

Jon

nvpublic

Ah, this is a "fix something broken in stable-only" type patch :(

James, I'm dropping this from the 4.19, 4.9, and 4.14 trees right now as it looks broken :(

thanks,

greg k-h

Hi Greg,

What would you prefer I do here:
1 post a revert for the original problematic backport.
2 post versions of this to fix each of the above 3 stable kernels. (instead of putting conditions in the stable tag).

Thanks,

James

I don't see what I did wrong with the "conditions" in the existing commit you sent. How did I get it wrong?

Best case, send a patch series for each kernel tree. That way I "know" I got the right thing here.

thanks,

greg k-h

nope.
It is also only on 4.19.

--
Linaro LKFT
https://lkft.linaro.org
_______________________________________________
kvmarm mailing list
[email protected]
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm

Thanks. I've pushed out -rc2 releases with this commit dropped.

greg k-h