Skip to contentSkip to conversation

net: tundra: tsi108: use spin_lock_irqsave instead of spin_lock_irq in IRQ context

PATCH New Sasha Levin 1 participant 27 messages Aug 29, 2019

Subsystems (12)
lore
Share & tools
lore

Sasha Levin

author maintainer
#0 lore raw

PATCH 1/27 net: tundra: tsi108: use spin_lock_irqsave instead of spin_lock_irq in IRQ context

From: Fuqian Huang <[email protected]>

[ Upstream commit 8c25d0887a8bd0e1ca2074ac0c6dff173787a83b ]

As spin_unlock_irq will enable interrupts.
Function tsi108_stat_carry is called from interrupt handler tsi108_irq. Interrupts are enabled in interrupt handler.
Use spin_lock_irqsave/spin_unlock_irqrestore instead of spin_(un)lock_irq in IRQ context to avoid this.

---
 drivers/net/ethernet/tundra/tsi108_eth.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
drivers/net/ethernet/tundra/tsi108_eth.c +3 -2
index c2d15d9c0c33b..455979e47424c 100644
@@ -381,9 +381,10 @@ tsi108_stat_carry_one(int carry, int carry_bit, int carry_shift,
 static void tsi108_stat_carry(struct net_device *dev)
 {
 	struct tsi108_prv_data *data = netdev_priv(dev);
+	unsigned long flags;
 	u32 carry1, carry2;
 
-	spin_lock_irq(&data->misclock);
+	spin_lock_irqsave(&data->misclock, flags);
 
 	carry1 = TSI_READ(TSI108_STAT_CARRY1);
 	carry2 = TSI_READ(TSI108_STAT_CARRY2);
@@ -451,7 +452,7 @@ static void tsi108_stat_carry(struct net_device *dev)
 			      TSI108_STAT_TXPAUSEDROP_CARRY,
 			      &data->tx_pause_drop);
 
-	spin_unlock_irq(&data->misclock);
+	spin_unlock_irqrestore(&data->misclock, flags);
 }
 
 /* Read a stat counter atomically with respect to carries.
Signature · 1 line
2.20.1
author maintainer
#1 lore raw

PATCH 2/27 hv_netvsc: Fix a warning of suspicious RCU usage

From: Dexuan Cui <[email protected]>

[ Upstream commit 6d0d779dca73cd5acb649c54f81401f93098b298 ]

This fixes a warning of "suspicious rcu_dereference_check() usage" when nload runs.

Fixes 776e726bfb34 ("netvsc: fix RCU warning in get_stats")
---
 drivers/net/hyperv/netvsc_drv.c | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)
drivers/net/hyperv/netvsc_drv.c +7 -2
index eb92720dd1c4a..33c1f6548fb79 100644
@@ -1170,12 +1170,15 @@ static void netvsc_get_stats64(struct net_device *net,
 			       struct rtnl_link_stats64 *t)
 {
 	struct net_device_context *ndev_ctx = netdev_priv(net);
-	struct netvsc_device *nvdev = rcu_dereference_rtnl(ndev_ctx->nvdev);
+	struct netvsc_device *nvdev;
 	struct netvsc_vf_pcpu_stats vf_tot;
 	int i;
 
+	rcu_read_lock();
+
+	nvdev = rcu_dereference(ndev_ctx->nvdev);
 	if (!nvdev)
-		return;
+		goto out;
 
 	netdev_stats_to_stats64(t, &net->stats);
 
@@ -1214,6 +1217,8 @@ static void netvsc_get_stats64(struct net_device *net,
 		t->rx_packets	+= packets;
 		t->multicast	+= multicast;
 	}
+out:
+	rcu_read_unlock();
 }
 
 static int netvsc_set_mac_addr(struct net_device *ndev, void *p)
Signature · 1 line
2.20.1
author maintainer
#2 lore raw

PATCH 3/27 net: tc35815: Explicitly check NET_IP_ALIGN is not zero in tc35815_rx

In reply to Sasha Levin

From: Nathan Chancellor <[email protected]>

[ Upstream commit 125b7e0949d4e72b15c2b1a1590f8cece985a918 ]

clang warns:

drivers/net/ethernet/toshiba/tc35815.c:1507:30: warning: use of logical '&&' with constant operand [-Wconstant-logical-operand]

                        if (!HAVE_DMA_RXALIGN(lp) && NET_IP_ALIGN)
                                                  ^  ~~~~~~~~~~~~
drivers/net/ethernet/toshiba/tc35815.c:1507:30: note: use '&' for a
bitwise operation
                        if (!HAVE_DMA_RXALIGN(lp) && NET_IP_ALIGN)
                                                  ^~
                                                  &
drivers/net/ethernet/toshiba/tc35815.c:1507:30: note: remove constant to
silence this warning
                        if (!HAVE_DMA_RXALIGN(lp) && NET_IP_ALIGN)
                                                 ~^~~~~~~~~~~~~~~

1 warning generated.

Explicitly check that NET_IP_ALIGN is not zero, which matches how this is checked in other parts of the tree. Because NET_IP_ALIGN is a build time constant, this check will be constant folded away during optimization.

---
 drivers/net/ethernet/toshiba/tc35815.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/ethernet/toshiba/tc35815.c +1 -1
index cce9c9ed46aa9..9146068979d2c 100644
@@ -1497,7 +1497,7 @@ tc35815_rx(struct net_device *dev, int limit)
 			pci_unmap_single(lp->pci_dev,
 					 lp->rx_skbs[cur_bd].skb_dma,
 					 RX_BUF_SIZE, PCI_DMA_FROMDEVICE);
-			if (!HAVE_DMA_RXALIGN(lp) && NET_IP_ALIGN)
+			if (!HAVE_DMA_RXALIGN(lp) && NET_IP_ALIGN != 0)
 				memmove(skb->data, skb->data - NET_IP_ALIGN,
 					pkt_len);
 			data = skb_put(skb, pkt_len);
Signature · 1 line
2.20.1
author maintainer
#3 lore raw

PATCH 4/27 Bluetooth: btqca: Add a short delay before downloading the NVM

In reply to Sasha Levin

From: Matthias Kaehlcke <[email protected]>

[ Upstream commit 8059ba0bd0e4694e51c2ee6438a77b325f06c0d5 ]

On WCN3990 downloading the NVM sometimes fails with a "TLV response size mismatch" error:

[  174.949955] Bluetooth: btqca.c:qca_download_firmware() hci0: QCA Downloading qca/crnv21.bin
[  174.958718] Bluetooth: btqca.c:qca_tlv_send_segment() hci0: QCA TLV response size mismatch

It seems the controller needs a short time after downloading the firmware before it is ready for the NVM. A delay as short as 1 ms seems sufficient, make it 10 ms just in case. No event is received during the delay, hence we don't just silently drop an extra event.

---
 drivers/bluetooth/btqca.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/bluetooth/btqca.c +3 -0
index 0bbdfcef2aa84..a48a61f22f823 100644
@@ -363,6 +363,9 @@ int qca_uart_setup_rome(struct hci_dev *hdev, uint8_t baudrate)
 		return err;
 	}
 
+	/* Give the controller some time to get ready to receive the NVM */
+	msleep(10);
+
 	/* Download NVM configuration */
 	config.type = TLV_TYPE_NVM;
 	snprintf(config.fwname, sizeof(config.fwname), "qca/nvm_%08x.bin",
Signature · 1 line
2.20.1
author maintainer
#4 lore raw

PATCH 5/27 Bluetooth: hidp: Let hidp_send_message return number of queued bytes

In reply to Sasha Levin

From: Fabian Henneke <[email protected]>

[ Upstream commit 48d9cc9d85dde37c87abb7ac9bbec6598ba44b56 ]

Let hidp_send_message return the number of successfully queued bytes instead of an unconditional 0.

With the return value fixed to 0, other drivers relying on hidp, such as hidraw, can not return meaningful values from their respective implementations of write(). In particular, with the current behavior, a hidraw device's write() will have different return values depending on whether the device is connected via USB or Bluetooth, which makes it harder to abstract away the transport layer.

---
 net/bluetooth/hidp/core.c | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)
net/bluetooth/hidp/core.c +7 -2
index b21fcc838784d..f6bffb3a95116 100644
@@ -101,6 +101,7 @@ static int hidp_send_message(struct hidp_session *session, struct socket *sock,
 {
 	struct sk_buff *skb;
 	struct sock *sk = sock->sk;
+	int ret;
 
 	BT_DBG("session %p data %p size %d", session, data, size);
 
@@ -114,13 +115,17 @@ static int hidp_send_message(struct hidp_session *session, struct socket *sock,
 	}
 
 	skb_put_u8(skb, hdr);
-	if (data && size > 0)
+	if (data && size > 0) {
 		skb_put_data(skb, data, size);
+		ret = size;
+	} else {
+		ret = 0;
+	}
 
 	skb_queue_tail(transmit, skb);
 	wake_up_interruptible(sk_sleep(sk));
 
-	return 0;
+	return ret;
 }
 
 static int hidp_send_ctrl_message(struct hidp_session *session,
Signature · 1 line
2.20.1
author maintainer
#5 lore raw

PATCH 6/27 ibmveth: Convert multicast list size for little-endian system

In reply to Sasha Levin

From: Thomas Falcon <[email protected]>

[ Upstream commit 66cf4710b23ab2adda11155684a2c8826f4fe732 ]

The ibm,mac-address-filters property defines the maximum number of addresses the hypervisor's multicast filter list can support. It is encoded as a big-endian integer in the OF device tree, but the virtual ethernet driver does not convert it for use by little-endian systems. As a result, the driver is not behaving as it should on affected systems when a large number of multicast addresses are assigned to the device.

---
 drivers/net/ethernet/ibm/ibmveth.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)
drivers/net/ethernet/ibm/ibmveth.c +5 -4
index 754dff4c1771e..880d925438c17 100644
@@ -1618,7 +1618,7 @@ static int ibmveth_probe(struct vio_dev *dev, const struct vio_device_id *id)
 	struct net_device *netdev;
 	struct ibmveth_adapter *adapter;
 	unsigned char *mac_addr_p;
-	unsigned int *mcastFilterSize_p;
+	__be32 *mcastFilterSize_p;
 	long ret;
 	unsigned long ret_attr;
 
@@ -1640,8 +1640,9 @@ static int ibmveth_probe(struct vio_dev *dev, const struct vio_device_id *id)
 		return -EINVAL;
 	}
 
-	mcastFilterSize_p = (unsigned int *)vio_get_attribute(dev,
-						VETH_MCAST_FILTER_SIZE, NULL);
+	mcastFilterSize_p = (__be32 *)vio_get_attribute(dev,
+							VETH_MCAST_FILTER_SIZE,
+							NULL);
 	if (!mcastFilterSize_p) {
 		dev_err(&dev->dev, "Can't find VETH_MCAST_FILTER_SIZE "
 			"attribute\n");
@@ -1658,7 +1659,7 @@ static int ibmveth_probe(struct vio_dev *dev, const struct vio_device_id *id)
 
 	adapter->vdev = dev;
 	adapter->netdev = netdev;
-	adapter->mcastFilterSize = *mcastFilterSize_p;
+	adapter->mcastFilterSize = be32_to_cpu(*mcastFilterSize_p);
 	adapter->pool_config = 0;
 
 	netif_napi_add(netdev, &adapter->napi, ibmveth_poll, 16);
Signature · 1 line
2.20.1
author maintainer
#6 lore raw

PATCH 7/27 gpio: Fix build error of function redefinition

In reply to Sasha Levin

From: YueHaibing <[email protected]>

[ Upstream commit 68e03b85474a51ec1921b4d13204782594ef7223 ]

when do randbuilding, I got this error:

In file included from drivers/hwmon/pmbus/ucd9000.c:19:0:
./include/linux/gpio/driver.h:576:1: error: redefinition of gpiochip_add_pin_range
gpiochip_add_pin_range(struct gpio_chip *chip, const char *pinctl_name, ^~~~~~~~~~~~~~~~~~~~~~
In file included from drivers/hwmon/pmbus/ucd9000.c:18:0:
./include/linux/gpio.h:245:1: note: previous definition of gpiochip_add_pin_range was here
gpiochip_add_pin_range(struct gpio_chip *chip, const char *pinctl_name, ^~~~~~~~~~~~~~~~~~~~~~

Fixes 964cb341882f ("gpio: move pincontrol calls to <linux/gpio/driver.h>")
---
 include/linux/gpio.h | 24 ------------------------
 1 file changed, 24 deletions(-)
include/linux/gpio.h +0 -24
index 8ef7fc0ce0f0c..b2f103b170a97 100644
@@ -230,30 +230,6 @@ static inline int irq_to_gpio(unsigned irq)
 	return -EINVAL;
 }
 
-static inline int
-gpiochip_add_pin_range(struct gpio_chip *chip, const char *pinctl_name,
-		       unsigned int gpio_offset, unsigned int pin_offset,
-		       unsigned int npins)
-{
-	WARN_ON(1);
-	return -EINVAL;
-}
-
-static inline int
-gpiochip_add_pingroup_range(struct gpio_chip *chip,
-			struct pinctrl_dev *pctldev,
-			unsigned int gpio_offset, const char *pin_group)
-{
-	WARN_ON(1);
-	return -EINVAL;
-}
-
-static inline void
-gpiochip_remove_pin_ranges(struct gpio_chip *chip)
-{
-	WARN_ON(1);
-}
-
 static inline int devm_gpio_request(struct device *dev, unsigned gpio,
 				    const char *label)
 {
Signature · 1 line
2.20.1
author maintainer
#7 lore raw

PATCH 8/27 drm/mediatek: use correct device to import PRIME buffers

In reply to Sasha Levin

From: Alexandre Courbot <[email protected]>

[ Upstream commit 4c6f3196e6ea111c456c6086dc3f57d4706b0b2d ]

PRIME buffers should be imported using the DMA device. To this end, use a custom import function that mimics drm_gem_prime_import_dev(), but passes the correct device.

Fixes 119f5173628aa ("drm/mediatek: Add DRM Driver for Mediatek SoC MT8173.")
---
 drivers/gpu/drm/mediatek/mtk_drm_drv.c | 14 +++++++++++++-
 1 file changed, 13 insertions(+), 1 deletion(-)
drivers/gpu/drm/mediatek/mtk_drm_drv.c +13 -1
index cada1c75c41cd..4a89cd2e4f1c5 100644
@@ -287,6 +287,18 @@ static const struct file_operations mtk_drm_fops = {
 	.compat_ioctl = drm_compat_ioctl,
 };
 
+/*
+ * We need to override this because the device used to import the memory is
+ * not dev->dev, as drm_gem_prime_import() expects.
+ */
+struct drm_gem_object *mtk_drm_gem_prime_import(struct drm_device *dev,
+						struct dma_buf *dma_buf)
+{
+	struct mtk_drm_private *private = dev->dev_private;
+
+	return drm_gem_prime_import_dev(dev, dma_buf, private->dma_dev);
+}
+
 static struct drm_driver mtk_drm_driver = {
 	.driver_features = DRIVER_MODESET | DRIVER_GEM | DRIVER_PRIME |
 			   DRIVER_ATOMIC,
@@ -298,7 +310,7 @@ static struct drm_driver mtk_drm_driver = {
 	.prime_handle_to_fd = drm_gem_prime_handle_to_fd,
 	.prime_fd_to_handle = drm_gem_prime_fd_to_handle,
 	.gem_prime_export = drm_gem_prime_export,
-	.gem_prime_import = drm_gem_prime_import,
+	.gem_prime_import = mtk_drm_gem_prime_import,
 	.gem_prime_get_sg_table = mtk_gem_prime_get_sg_table,
 	.gem_prime_import_sg_table = mtk_gem_prime_import_sg_table,
 	.gem_prime_mmap = mtk_drm_gem_mmap_buf,
Signature · 6 lines
2.20.1

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

PATCH 9/27 drm/mediatek: set DMA max segment size

In reply to Sasha Levin

From: Alexandre Courbot <[email protected]>

[ Upstream commit 070955558e820b9a89c570b91b1f21762f62b288 ]

This driver requires imported PRIME buffers to appear contiguously in its IO address space. Make sure this is the case by setting the maximum DMA segment size to a more suitable value than the default 64KB.

---
 drivers/gpu/drm/mediatek/mtk_drm_drv.c | 35 ++++++++++++++++++++++++--
 drivers/gpu/drm/mediatek/mtk_drm_drv.h |  2 ++
 2 files changed, 35 insertions(+), 2 deletions(-)
drivers/gpu/drm/mediatek/mtk_drm_drv.c +33 -2
index 4a89cd2e4f1c5..034b50080304f 100644
@@ -185,6 +185,7 @@ static int mtk_drm_kms_init(struct drm_device *drm)
 	struct mtk_drm_private *private = drm->dev_private;
 	struct platform_device *pdev;
 	struct device_node *np;
+	struct device *dma_dev;
 	int ret;
 
 	if (!iommu_present(&platform_bus_type))
@@ -242,7 +243,29 @@ static int mtk_drm_kms_init(struct drm_device *drm)
 		goto err_component_unbind;
 	}
 
-	private->dma_dev = &pdev->dev;
+	dma_dev = &pdev->dev;
+	private->dma_dev = dma_dev;
+
+	/*
+	 * Configure the DMA segment size to make sure we get contiguous IOVA
+	 * when importing PRIME buffers.
+	 */
+	if (!dma_dev->dma_parms) {
+		private->dma_parms_allocated = true;
+		dma_dev->dma_parms =
+			devm_kzalloc(drm->dev, sizeof(*dma_dev->dma_parms),
+				     GFP_KERNEL);
+	}
+	if (!dma_dev->dma_parms) {
+		ret = -ENOMEM;
+		goto err_component_unbind;
+	}
+
+	ret = dma_set_max_seg_size(dma_dev, (unsigned int)DMA_BIT_MASK(32));
+	if (ret) {
+		dev_err(dma_dev, "Failed to set DMA segment size\n");
+		goto err_unset_dma_parms;
+	}
 
 	/*
 	 * We don't use the drm_irq_install() helpers provided by the DRM
@@ -252,13 +275,16 @@ static int mtk_drm_kms_init(struct drm_device *drm)
 	drm->irq_enabled = true;
 	ret = drm_vblank_init(drm, MAX_CRTC);
 	if (ret < 0)
-		goto err_component_unbind;
+		goto err_unset_dma_parms;
 
 	drm_kms_helper_poll_init(drm);
 	drm_mode_config_reset(drm);
 
 	return 0;
 
+err_unset_dma_parms:
+	if (private->dma_parms_allocated)
+		dma_dev->dma_parms = NULL;
 err_component_unbind:
 	component_unbind_all(drm->dev, drm);
 err_config_cleanup:
@@ -269,9 +295,14 @@ static int mtk_drm_kms_init(struct drm_device *drm)
 
 static void mtk_drm_kms_deinit(struct drm_device *drm)
 {
+	struct mtk_drm_private *private = drm->dev_private;
+
 	drm_kms_helper_poll_fini(drm);
 	drm_atomic_helper_shutdown(drm);
 
+	if (private->dma_parms_allocated)
+		private->dma_dev->dma_parms = NULL;
+
 	component_unbind_all(drm->dev, drm);
 	drm_mode_config_cleanup(drm);
 }
drivers/gpu/drm/mediatek/mtk_drm_drv.h +2 -0
index c3378c452c0a0..445dd45e65ebc 100644
@@ -56,6 +56,8 @@ struct mtk_drm_private {
 	} commit;
 
 	struct drm_atomic_state *suspend_state;
+
+	bool dma_parms_allocated;
 };
 
 extern struct platform_driver mtk_ddp_driver;
Signature · 1 line
2.20.1
author maintainer
#9 lore raw

PATCH 10/27 cxgb4: fix a memory leak bug

In reply to Sasha Levin

From: Wenwen Wang <[email protected]>

[ Upstream commit c554336efa9bbc28d6ec14efbee3c7d63c61a34f ]

In blocked_fl_write(), 't' is not deallocated if bitmap_parse_user() fails, leading to a memory leak bug. To fix this issue, free t before returning the error.

---
 drivers/net/ethernet/chelsio/cxgb4/cxgb4_debugfs.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/net/ethernet/chelsio/cxgb4/cxgb4_debugfs.c +3 -1
index 76540b0e082d3..9e5cd18e7358c 100644
@@ -2777,8 +2777,10 @@ static ssize_t blocked_fl_write(struct file *filp, const char __user *ubuf,
 		return -ENOMEM;
 
 	err = bitmap_parse_user(ubuf, count, t, adap->sge.egr_sz);
-	if (err)
+	if (err) {
+		kvfree(t);
 		return err;
+	}
 
 	bitmap_copy(adap->sge.blocked_fl, t, adap->sge.egr_sz);
 	kvfree(t);
Signature · 1 line
2.20.1
author maintainer
#10 lore raw

PATCH 11/27 liquidio: add cleanup in octeon_setup_iq()

In reply to Sasha Levin

From: Wenwen Wang <[email protected]>

[ Upstream commit 6f967f8b1be7001b31c46429f2ee7d275af2190f ]

If oct->fn_list.enable_io_queues() fails, no cleanup is executed, leading to memory/resource leaks. To fix this issue, invoke octeon_delete_instr_queue() before returning from the function.

---
 drivers/net/ethernet/cavium/liquidio/request_manager.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/net/ethernet/cavium/liquidio/request_manager.c +3 -1
index 1e0fbce86d608..55e8731264634 100644
@@ -232,8 +232,10 @@ int octeon_setup_iq(struct octeon_device *oct,
 	}
 
 	oct->num_iqs++;
-	if (oct->fn_list.enable_io_queues(oct))
+	if (oct->fn_list.enable_io_queues(oct)) {
+		octeon_delete_instr_queue(oct, iq_no);
 		return 1;
+	}
 
 	return 0;
 }
Signature · 1 line
2.20.1
author maintainer
#11 lore raw

PATCH 12/27 net: myri10ge: fix memory leaks

In reply to Sasha Levin

From: Wenwen Wang <[email protected]>

[ Upstream commit 20fb7c7a39b5c719e2e619673b5f5729ee7d2306 ]

In myri10ge_probe(), myri10ge_alloc_slices() is invoked to allocate slices related structures. Later on, myri10ge_request_irq() is used to get an irq. However, if this process fails, the allocated slices related structures are not deallocated, leading to memory leaks. To fix this issue, revise the target label of the goto statement to 'abort_with_slices'.

---
 drivers/net/ethernet/myricom/myri10ge/myri10ge.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/ethernet/myricom/myri10ge/myri10ge.c +1 -1
index b171ed2015fe4..a0a555052d8ca 100644
@@ -3922,7 +3922,7 @@ static int myri10ge_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 	 * setup (if available). */
 	status = myri10ge_request_irq(mgp);
 	if (status != 0)
-		goto abort_with_firmware;
+		goto abort_with_slices;
 	myri10ge_free_irq(mgp);
 
 	/* Save configuration space to be restored if the
Signature · 1 line
2.20.1
author maintainer
#12 lore raw

PATCH 13/27 lan78xx: Fix memory leaks

In reply to Sasha Levin

From: Wenwen Wang <[email protected]>

[ Upstream commit b9cbf8a64865b50fd0f4a3915fa00ac7365cdf8f ]

In lan78xx_probe(), a new urb is allocated through usb_alloc_urb() and saved to 'dev->urb_intr'. However, in the following execution, if an error occurs, 'dev->urb_intr' is not deallocated, leading to memory leaks. To fix this issue, invoke usb_free_urb() to free the allocated urb before returning from the function.

---
 drivers/net/usb/lan78xx.c | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)
drivers/net/usb/lan78xx.c +5 -3
index b62c41114e34e..24b994c68bccd 100644
@@ -3645,7 +3645,7 @@ static int lan78xx_probe(struct usb_interface *intf,
 	ret = register_netdev(netdev);
 	if (ret != 0) {
 		netif_err(dev, probe, netdev, "couldn't register the device\n");
-		goto out3;
+		goto out4;
 	}
 
 	usb_set_intfdata(intf, dev);
@@ -3660,12 +3660,14 @@ static int lan78xx_probe(struct usb_interface *intf,
 
 	ret = lan78xx_phy_init(dev);
 	if (ret < 0)
-		goto out4;
+		goto out5;
 
 	return 0;
 
-out4:
+out5:
 	unregister_netdev(netdev);
+out4:
+	usb_free_urb(dev->urb_intr);
 out3:
 	lan78xx_unbind(dev, intf);
 out2:
Signature · 1 line
2.20.1
author maintainer
#13 lore raw

PATCH 14/27 vfs: fix page locking deadlocks when deduping files

In reply to Sasha Levin

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

[ Upstream commit edc58dd0123b552453a74369bd0c8d890b497b4b ]

When dedupe wants to use the page cache to compare parts of two files for dedupe, we must be very careful to handle locking correctly. The current code doesn't do this. It must lock and unlock the page only once if the two pages are the same, since the overlapping range check doesn't catch this when blocksize < pagesize. If the pages are distinct but from the same file, we must observe page locking order and lock them in order of increasing offset to avoid clashing with writeback locking.

Fixes 876bec6f9bbfcb3 ("vfs: refactor clone/dedupe_file_range common functions")
---
 fs/read_write.c | 49 +++++++++++++++++++++++++++++++++++++++++--------
 1 file changed, 41 insertions(+), 8 deletions(-)
fs/read_write.c +41 -8
index d6f8bfb0f7942..38a8bcccf0dd0 100644
@@ -1882,10 +1882,7 @@ int vfs_clone_file_range(struct file *file_in, loff_t pos_in,
 }
 EXPORT_SYMBOL(vfs_clone_file_range);
 
-/*
- * Read a page's worth of file data into the page cache.  Return the page
- * locked.
- */
+/* Read a page's worth of file data into the page cache. */
 static struct page *vfs_dedupe_get_page(struct inode *inode, loff_t offset)
 {
 	struct address_space *mapping;
@@ -1901,10 +1898,32 @@ static struct page *vfs_dedupe_get_page(struct inode *inode, loff_t offset)
 		put_page(page);
 		return ERR_PTR(-EIO);
 	}
-	lock_page(page);
 	return page;
 }
 
+/*
+ * Lock two pages, ensuring that we lock in offset order if the pages are from
+ * the same file.
+ */
+static void vfs_lock_two_pages(struct page *page1, struct page *page2)
+{
+	/* Always lock in order of increasing index. */
+	if (page1->index > page2->index)
+		swap(page1, page2);
+
+	lock_page(page1);
+	if (page1 != page2)
+		lock_page(page2);
+}
+
+/* Unlock two pages, being careful not to unlock the same page twice. */
+static void vfs_unlock_two_pages(struct page *page1, struct page *page2)
+{
+	unlock_page(page1);
+	if (page1 != page2)
+		unlock_page(page2);
+}
+
 /*
  * Compare extents of two files to see if they are the same.
  * Caller must have locked both inodes to prevent write races.
@@ -1942,10 +1961,24 @@ int vfs_dedupe_file_range_compare(struct inode *src, loff_t srcoff,
 		dest_page = vfs_dedupe_get_page(dest, destoff);
 		if (IS_ERR(dest_page)) {
 			error = PTR_ERR(dest_page);
-			unlock_page(src_page);
 			put_page(src_page);
 			goto out_error;
 		}
+
+		vfs_lock_two_pages(src_page, dest_page);
+
+		/*
+		 * Now that we've locked both pages, make sure they're still
+		 * mapped to the file data we're interested in.  If not,
+		 * someone is invalidating pages on us and we lose.
+		 */
+		if (!PageUptodate(src_page) || !PageUptodate(dest_page) ||
+		    src_page->mapping != src->i_mapping ||
+		    dest_page->mapping != dest->i_mapping) {
+			same = false;
+			goto unlock;
+		}
+
 		src_addr = kmap_atomic(src_page);
 		dest_addr = kmap_atomic(dest_page);
 
@@ -1957,8 +1990,8 @@ int vfs_dedupe_file_range_compare(struct inode *src, loff_t srcoff,
 
 		kunmap_atomic(dest_addr);
 		kunmap_atomic(src_addr);
-		unlock_page(dest_page);
-		unlock_page(src_page);
+unlock:
+		vfs_unlock_two_pages(src_page, dest_page);
 		put_page(dest_page);
 		put_page(src_page);
 
Signature · 1 line
2.20.1
author maintainer
#14 lore raw

PATCH 15/27 cx82310_eth: fix a memory leak bug

In reply to Sasha Levin

From: Wenwen Wang <[email protected]>

[ Upstream commit 1eca92eef18719027d394bf1a2d276f43e7cf886 ]

In cx82310_bind(), 'dev->partial_data' is allocated through kmalloc(). Then, the execution waits for the firmware to become ready. If the firmware is not ready in time, the execution is terminated. However, the allocated 'dev->partial_data' is not deallocated on this path, leading to a memory leak bug. To fix this issue, free 'dev->partial_data' before returning the error.

---
 drivers/net/usb/cx82310_eth.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/net/usb/cx82310_eth.c +2 -1
index 947bea81d9241..dfbdea22fbad9 100644
@@ -175,7 +175,8 @@ static int cx82310_bind(struct usbnet *dev, struct usb_interface *intf)
 	}
 	if (!timeout) {
 		dev_err(&udev->dev, "firmware not ready in time\n");
-		return -ETIMEDOUT;
+		ret = -ETIMEDOUT;
+		goto err;
 	}
 
 	/* enable ethernet mode (?) */
Signature · 1 line
2.20.1
author maintainer
#15 lore raw

PATCH 16/27 net: kalmia: fix memory leaks

In reply to Sasha Levin

From: Wenwen Wang <[email protected]>

[ Upstream commit f1472cb09f11ddb41d4be84f0650835cb65a9073 ]

In kalmia_init_and_get_ethernet_addr(), 'usb_buf' is allocated through kmalloc(). In the following execution, if the 'status' returned by kalmia_send_init_packet() is not 0, 'usb_buf' is not deallocated, leading to memory leaks. To fix this issue, add the 'out' label to free 'usb_buf'.

---
 drivers/net/usb/kalmia.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
drivers/net/usb/kalmia.c +3 -3
index ce0b0b4e3a57c..c677ec2bae183 100644
@@ -117,16 +117,16 @@ kalmia_init_and_get_ethernet_addr(struct usbnet *dev, u8 *ethernet_addr)
 	status = kalmia_send_init_packet(dev, usb_buf, sizeof(init_msg_1)
 		/ sizeof(init_msg_1[0]), usb_buf, 24);
 	if (status != 0)
-		return status;
+		goto out;
 
 	memcpy(usb_buf, init_msg_2, 12);
 	status = kalmia_send_init_packet(dev, usb_buf, sizeof(init_msg_2)
 		/ sizeof(init_msg_2[0]), usb_buf, 28);
 	if (status != 0)
-		return status;
+		goto out;
 
 	memcpy(ethernet_addr, usb_buf + 10, ETH_ALEN);
-
+out:
 	kfree(usb_buf);
 	return status;
 }
Signature · 1 line
2.20.1
author maintainer
#16 lore raw

PATCH 17/27 wimax/i2400m: fix a memory leak bug

In reply to Sasha Levin

From: Wenwen Wang <[email protected]>

[ Upstream commit 44ef3a03252844a8753479b0cea7f29e4a804bdc ]

In i2400m_barker_db_init(), 'options_orig' is allocated through kstrdup() to hold the original command line options. Then, the options are parsed. However, if an error occurs during the parsing process, 'options_orig' is not deallocated, leading to a memory leak bug. To fix this issue, free 'options_orig' before returning the error.

---
 drivers/net/wimax/i2400m/fw.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/net/wimax/i2400m/fw.c +3 -1
index a89b5685e68b3..4577ee5bbddd6 100644
@@ -351,13 +351,15 @@ int i2400m_barker_db_init(const char *_options)
 			}
 			result = i2400m_barker_db_add(barker);
 			if (result < 0)
-				goto error_add;
+				goto error_parse_add;
 		}
 		kfree(options_orig);
 	}
 	return 0;
 
+error_parse_add:
 error_parse:
+	kfree(options_orig);
 error_add:
 	kfree(i2400m_barker_db);
 	return result;
Signature · 1 line
2.20.1
author maintainer
#17 lore raw

PATCH 18/27 ravb: Fix use-after-free ravb_tstamp_skb

In reply to Sasha Levin

From: Tho Vu <[email protected]>

[ Upstream commit cfef46d692efd852a0da6803f920cc756eea2855 ]

When a Tx timestamp is requested, a pointer to the skb is stored in the ravb_tstamp_skb struct. This was done without an skb_get. There exists the possibility that the skb could be freed by ravb_tx_free (when ravb_tx_free is called from ravb_start_xmit) before the timestamp was processed, leading to a use-after-free bug.

Use skb_get when filling a ravb_tstamp_skb struct, and add appropriate frees/consumes when a ravb_tstamp_skb struct is freed.

Fixes c156633f1353 ("Renesas Ethernet AVB driver proper")
---
 drivers/net/ethernet/renesas/ravb_main.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)
drivers/net/ethernet/renesas/ravb_main.c +6 -2
index ce79af4a7f6fb..d73617cc3b159 100644
@@ -1,6 +1,6 @@
 /* Renesas Ethernet AVB device driver
  *
- * Copyright (C) 2014-2015 Renesas Electronics Corporation
+ * Copyright (C) 2014-2019 Renesas Electronics Corporation
  * Copyright (C) 2015 Renesas Solutions Corp.
  * Copyright (C) 2015-2016 Cogent Embedded, Inc. <[email protected]>
  *
@@ -513,7 +513,10 @@ static void ravb_get_tx_tstamp(struct net_device *ndev)
 			kfree(ts_skb);
 			if (tag == tfa_tag) {
 				skb_tstamp_tx(skb, &shhwtstamps);
+				dev_consume_skb_any(skb);
 				break;
+			} else {
+				dev_kfree_skb_any(skb);
 			}
 		}
 		ravb_modify(ndev, TCCR, TCCR_TFR, TCCR_TFR);
@@ -1576,7 +1579,7 @@ static netdev_tx_t ravb_start_xmit(struct sk_buff *skb, struct net_device *ndev)
 					 DMA_TO_DEVICE);
 			goto unmap;
 		}
-		ts_skb->skb = skb;
+		ts_skb->skb = skb_get(skb);
 		ts_skb->tag = priv->ts_skb_tag++;
 		priv->ts_skb_tag &= 0x3ff;
 		list_add_tail(&ts_skb->list, &priv->ts_skb_list);
@@ -1704,6 +1707,7 @@ static int ravb_close(struct net_device *ndev)
 	/* Clear the timestamp list */
 	list_for_each_entry_safe(ts_skb, ts_skb2, &priv->ts_skb_list, list) {
 		list_del(&ts_skb->list);
+		kfree_skb(ts_skb->skb);
 		kfree(ts_skb);
 	}
 
Signature · 1 line
2.20.1
author maintainer
#18 lore raw

PATCH 19/27 kprobes: Fix potential deadlock in kprobe_optimizer()

In reply to Sasha Levin

From: Andrea Righi <[email protected]>

[ Upstream commit f1c6ece23729257fb46562ff9224cf5f61b818da ]

lockdep reports the following deadlock scenario:

WARNING: possible circular locking dependency detected

kworker/1:1/48 is trying to acquire lock:
000000008d7a62b2 (text_mutex){+.+.}, at: kprobe_optimizer+0x163/0x290

but task is already holding lock:
00000000850b5e2d (module_mutex){+.+.}, at: kprobe_optimizer+0x31/0x290

which lock already depends on the new lock.

the existing dependency chain (in reverse order) is:

-> #1 (module_mutex){+.+.}:

        __mutex_lock+0xac/0x9f0
        mutex_lock_nested+0x1b/0x20
        set_all_modules_text_rw+0x22/0x90
        ftrace_arch_code_modify_prepare+0x1c/0x20
        ftrace_run_update_code+0xe/0x30
        ftrace_startup_enable+0x2e/0x50
        ftrace_startup+0xa7/0x100
        register_ftrace_function+0x27/0x70
        arm_kprobe+0xb3/0x130
        enable_kprobe+0x83/0xa0
        enable_trace_kprobe.part.0+0x2e/0x80
        kprobe_register+0x6f/0xc0
        perf_trace_event_init+0x16b/0x270
        perf_kprobe_init+0xa7/0xe0
        perf_kprobe_event_init+0x3e/0x70
        perf_try_init_event+0x4a/0x140
        perf_event_alloc+0x93a/0xde0
        __do_sys_perf_event_open+0x19f/0xf30
        __x64_sys_perf_event_open+0x20/0x30
        do_syscall_64+0x65/0x1d0
        entry_SYSCALL_64_after_hwframe+0x49/0xbe

-> #0 (text_mutex){+.+.}:

        __lock_acquire+0xfcb/0x1b60
        lock_acquire+0xca/0x1d0
        __mutex_lock+0xac/0x9f0
        mutex_lock_nested+0x1b/0x20
        kprobe_optimizer+0x163/0x290
        process_one_work+0x22b/0x560
        worker_thread+0x50/0x3c0
        kthread+0x112/0x150
        ret_from_fork+0x3a/0x50

other info that might help us debug this:

  Possible unsafe locking scenario:
        CPU0                    CPU1
        ----                    ----
   lock(module_mutex);
                                lock(text_mutex);
                                lock(module_mutex);

lock(text_mutex);

*** DEADLOCK ***

As a reproducer I've been using bcc's funccount.py
(https://github.com/iovisor/bcc/blob/master/tools/funccount.py),
for example:

 # ./funccount.py '*interrupt*'

That immediately triggers the lockdep splat.

Fix by acquiring text_mutex before module_mutex in kprobe_optimizer().

Cc Anil S Keshavamurthy <[email protected]>
Cc David S. Miller <[email protected]>
Cc Linus Torvalds <[email protected]>
Cc Naveen N. Rao <[email protected]>
Cc Peter Zijlstra <[email protected]>
Cc Thomas Gleixner <[email protected]>
Fixes d5b844a2cf50 ("ftrace/x86: Remove possible deadlock between register_kprobe() and ftrace_run_update_code()")
---
 kernel/kprobes.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)
kernel/kprobes.c +4 -4
index ec11bb986a8b4..c43bc2bc5b2ca 100644
@@ -483,6 +483,7 @@ static DECLARE_DELAYED_WORK(optimizing_work, kprobe_optimizer);
  */
 static void do_optimize_kprobes(void)
 {
+	lockdep_assert_held(&text_mutex);
 	/*
 	 * The optimization/unoptimization refers online_cpus via
 	 * stop_machine() and cpu-hotplug modifies online_cpus.
@@ -500,9 +501,7 @@ static void do_optimize_kprobes(void)
 	    list_empty(&optimizing_list))
 		return;
 
-	mutex_lock(&text_mutex);
 	arch_optimize_kprobes(&optimizing_list);
-	mutex_unlock(&text_mutex);
 }
 
 /*
@@ -513,6 +512,7 @@ static void do_unoptimize_kprobes(void)
 {
 	struct optimized_kprobe *op, *tmp;
 
+	lockdep_assert_held(&text_mutex);
 	/* See comment in do_optimize_kprobes() */
 	lockdep_assert_cpus_held();
 
@@ -520,7 +520,6 @@ static void do_unoptimize_kprobes(void)
 	if (list_empty(&unoptimizing_list))
 		return;
 
-	mutex_lock(&text_mutex);
 	arch_unoptimize_kprobes(&unoptimizing_list, &freeing_list);
 	/* Loop free_list for disarming */
 	list_for_each_entry_safe(op, tmp, &freeing_list, list) {
@@ -537,7 +536,6 @@ static void do_unoptimize_kprobes(void)
 		} else
 			list_del_init(&op->list);
 	}
-	mutex_unlock(&text_mutex);
 }
 
 /* Reclaim all kprobes on the free_list */
@@ -563,6 +561,7 @@ static void kprobe_optimizer(struct work_struct *work)
 {
 	mutex_lock(&kprobe_mutex);
 	cpus_read_lock();
+	mutex_lock(&text_mutex);
 	/* Lock modules while optimizing kprobes */
 	mutex_lock(&module_mutex);
 
@@ -590,6 +589,7 @@ static void kprobe_optimizer(struct work_struct *work)
 	do_free_cleaned_kprobes();
 
 	mutex_unlock(&module_mutex);
+	mutex_unlock(&text_mutex);
 	cpus_read_unlock();
 	mutex_unlock(&kprobe_mutex);
 
Signature · 1 line
2.20.1
author maintainer
#19 lore raw

PATCH 20/27 HID: cp2112: prevent sleeping function called from invalid context

In reply to Sasha Levin

From: Benjamin Tissoires <[email protected]>

[ Upstream commit 2d05dba2b25ecb0f8fc3a0b4eb2232da6454a47b ]

When calling request_threaded_irq() with a CP2112, the function cp2112_gpio_irq_startup() is called in a IRQ context.

Therefore we can not sleep, and we can not call
cp2112_gpio_direction_input() there.

Move the call to cp2112_gpio_direction_input() earlier to have a working driver.

---
 drivers/hid/hid-cp2112.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)
drivers/hid/hid-cp2112.c +6 -2
index 4e940a096b2ac..abf1079457664 100644
@@ -1149,8 +1149,6 @@ static unsigned int cp2112_gpio_irq_startup(struct irq_data *d)
 
 	INIT_DELAYED_WORK(&dev->gpio_poll_worker, cp2112_gpio_poll_callback);
 
-	cp2112_gpio_direction_input(gc, d->hwirq);
-
 	if (!dev->gpio_poll) {
 		dev->gpio_poll = true;
 		schedule_delayed_work(&dev->gpio_poll_worker, 0);
@@ -1198,6 +1196,12 @@ static int __maybe_unused cp2112_allocate_irq(struct cp2112_device *dev,
 		return PTR_ERR(dev->desc[pin]);
 	}
 
+	ret = cp2112_gpio_direction_input(&dev->gc, pin);
+	if (ret < 0) {
+		dev_err(dev->gc.parent, "Failed to set GPIO to input dir\n");
+		goto err_desc;
+	}
+
 	ret = gpiochip_lock_as_irq(&dev->gc, pin);
 	if (ret) {
 		dev_err(dev->gc.parent, "Failed to lock GPIO as interrupt\n");
Signature · 1 line
2.20.1
author maintainer
#20 lore raw

PATCH 21/27 Input: hyperv-keyboard: Use in-place iterator API in the channel callback

In reply to Sasha Levin

From: Dexuan Cui <[email protected]>

[ Upstream commit d09bc83640d524b8467a660db7b1d15e6562a1de ]

Simplify the ring buffer handling with the in-place API.

Also avoid the dynamic allocation and the memory leak in the channel callback function.

---
 drivers/input/serio/hyperv-keyboard.c | 35 +++++----------------------
 1 file changed, 6 insertions(+), 29 deletions(-)
drivers/input/serio/hyperv-keyboard.c +6 -29
index 55288a026e4e2..c137ffa6fdec8 100644
@@ -245,40 +245,17 @@ static void hv_kbd_handle_received_packet(struct hv_device *hv_dev,
 
 static void hv_kbd_on_channel_callback(void *context)
 {
+	struct vmpacket_descriptor *desc;
 	struct hv_device *hv_dev = context;
-	void *buffer;
-	int bufferlen = 0x100; /* Start with sensible size */
 	u32 bytes_recvd;
 	u64 req_id;
-	int error;
 
-	buffer = kmalloc(bufferlen, GFP_ATOMIC);
-	if (!buffer)
-		return;
-
-	while (1) {
-		error = vmbus_recvpacket_raw(hv_dev->channel, buffer, bufferlen,
-					     &bytes_recvd, &req_id);
-		switch (error) {
-		case 0:
-			if (bytes_recvd == 0) {
-				kfree(buffer);
-				return;
-			}
-
-			hv_kbd_handle_received_packet(hv_dev, buffer,
-						      bytes_recvd, req_id);
-			break;
+	foreach_vmbus_pkt(desc, hv_dev->channel) {
+		bytes_recvd = desc->len8 * 8;
+		req_id = desc->trans_id;
 
-		case -ENOBUFS:
-			kfree(buffer);
-			/* Handle large packet */
-			bufferlen = bytes_recvd;
-			buffer = kmalloc(bytes_recvd, GFP_ATOMIC);
-			if (!buffer)
-				return;
-			break;
-		}
+		hv_kbd_handle_received_packet(hv_dev, desc, bytes_recvd,
+					      req_id);
 	}
 }
 
Signature · 1 line
2.20.1
author maintainer
#21 lore raw

PATCH 22/27 Tools: hv: kvp: eliminate 'may be used uninitialized' warning

In reply to Sasha Levin

From: Vitaly Kuznetsov <[email protected]>

[ Upstream commit 89eb4d8d25722a0a0194cf7fa47ba602e32a6da7 ]

When building hv_kvp_daemon GCC-8.3 complains:

hv_kvp_daemon.c: In function ‘kvp_get_ip_info.constprop’:
hv_kvp_daemon.c:812:30: warning: ‘ip_buffer’ may be used uninitialized in this function [-Wmaybe-uninitialized]
struct hv_kvp_ipaddr_value *ip_buffer;

this seems to be a false positive: we only use ip_buffer when op == KVP_OP_GET_IP_INFO and it is only unset when op == KVP_OP_ENUMERATE.

Silence the warning by initializing ip_buffer to NULL.

---
 tools/hv/hv_kvp_daemon.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
tools/hv/hv_kvp_daemon.c +1 -1
index 62c9a503ae052..bb245d4afc0cf 100644
@@ -867,7 +867,7 @@ kvp_get_ip_info(int family, char *if_name, int op,
 	int sn_offset = 0;
 	int error = 0;
 	char *buffer;
-	struct hv_kvp_ipaddr_value *ip_buffer;
+	struct hv_kvp_ipaddr_value *ip_buffer = NULL;
 	char cidr_mask[5]; /* /xyz */
 	int weight;
 	int i;
Signature · 1 line
2.20.1
author maintainer
#22 lore raw

PATCH 23/27 IB/mlx4: Fix memory leaks

In reply to Sasha Levin

From: Wenwen Wang <[email protected]>

[ Upstream commit 5c1baaa82cea2c815a5180ded402a7cd455d1810 ]

In mlx4_ib_alloc_pv_bufs(), 'tun_qp->tx_ring' is allocated through kcalloc(). However, it is not always deallocated in the following execution if an error occurs, leading to memory leaks. To fix this issue, free 'tun_qp->tx_ring' whenever an error occurs.

---
 drivers/infiniband/hw/mlx4/mad.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/infiniband/hw/mlx4/mad.c +2 -2
index d604b3d5aa3e4..c69158ccab822 100644
@@ -1680,8 +1680,6 @@ static int mlx4_ib_alloc_pv_bufs(struct mlx4_ib_demux_pv_ctx *ctx,
 				    tx_buf_size, DMA_TO_DEVICE);
 		kfree(tun_qp->tx_ring[i].buf.addr);
 	}
-	kfree(tun_qp->tx_ring);
-	tun_qp->tx_ring = NULL;
 	i = MLX4_NUM_TUNNEL_BUFS;
 err:
 	while (i > 0) {
@@ -1690,6 +1688,8 @@ static int mlx4_ib_alloc_pv_bufs(struct mlx4_ib_demux_pv_ctx *ctx,
 				    rx_buf_size, DMA_FROM_DEVICE);
 		kfree(tun_qp->ring[i].addr);
 	}
+	kfree(tun_qp->tx_ring);
+	tun_qp->tx_ring = NULL;
 	kfree(tun_qp->ring);
 	tun_qp->ring = NULL;
 	return -ENOMEM;
Signature · 1 line
2.20.1
author maintainer
#23 lore raw

PATCH 24/27 ceph: fix buffer free while holding i_ceph_lock in __ceph_setxattr()

In reply to Sasha Levin

From: Luis Henriques <[email protected]>

[ Upstream commit 86968ef21596515958d5f0a40233d02be78ecec0 ]

Calling ceph_buffer_put() in __ceph_setxattr() may end up freeing the i_xattrs.prealloc_blob buffer while holding the i_ceph_lock. This can be fixed by postponing the call until later, when the lock is released.

The following backtrace was triggered by fstests generic/117.

  BUG: sleeping function called from invalid context at mm/vmalloc.c:2283
  in_atomic(): 1, irqs_disabled(): 0, pid: 650, name: fsstress
  3 locks held by fsstress/650:
   #0: 00000000870a0fe8 (sb_writers#8){.+.+}, at: mnt_want_write+0x20/0x50
   #1: 00000000ba0c4c74 (&type->i_mutex_dir_key#6){++++}, at: vfs_setxattr+0x55/0xa0
   #2: 000000008dfbb3f2 (&(&ci->i_ceph_lock)->rlock){+.+.}, at: __ceph_setxattr+0x297/0x810
  CPU: 1 PID: 650 Comm: fsstress Not tainted 5.2.0+ #437
  Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.12.1-0-ga5cab58-prebuilt.qemu.org 04/01/2014
  Call Trace:
   dump_stack+0x67/0x90
   ___might_sleep.cold+0x9f/0xb1
   vfree+0x4b/0x60
   ceph_buffer_release+0x1b/0x60
   __ceph_setxattr+0x2b4/0x810
   __vfs_setxattr+0x66/0x80
   __vfs_setxattr_noperm+0x59/0xf0
   vfs_setxattr+0x81/0xa0
   setxattr+0x115/0x230
   ? filename_lookup+0xc9/0x140
   ? rcu_read_lock_sched_held+0x74/0x80
   ? rcu_sync_lockdep_assert+0x2e/0x60
   ? __sb_start_write+0x142/0x1a0
   ? mnt_want_write+0x20/0x50
   path_setxattr+0xba/0xd0
   __x64_sys_lsetxattr+0x24/0x30
   do_syscall_64+0x50/0x1c0
   entry_SYSCALL_64_after_hwframe+0x49/0xbe
  RIP: 0033:0x7ff23514359a
---
 fs/ceph/xattr.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)
fs/ceph/xattr.c +6 -2
index 0376db8a74f85..932c6cdc22d66 100644
@@ -955,6 +955,7 @@ int __ceph_setxattr(struct inode *inode, const char *name,
 	struct ceph_inode_info *ci = ceph_inode(inode);
 	struct ceph_mds_client *mdsc = ceph_sb_to_client(inode->i_sb)->mdsc;
 	struct ceph_cap_flush *prealloc_cf = NULL;
+	struct ceph_buffer *old_blob = NULL;
 	int issued;
 	int err;
 	int dirty = 0;
@@ -1023,13 +1024,15 @@ int __ceph_setxattr(struct inode *inode, const char *name,
 		struct ceph_buffer *blob;
 
 		spin_unlock(&ci->i_ceph_lock);
-		dout(" preaallocating new blob size=%d\n", required_blob_size);
+		ceph_buffer_put(old_blob); /* Shouldn't be required */
+		dout(" pre-allocating new blob size=%d\n", required_blob_size);
 		blob = ceph_buffer_new(required_blob_size, GFP_NOFS);
 		if (!blob)
 			goto do_sync_unlocked;
 		spin_lock(&ci->i_ceph_lock);
+		/* prealloc_blob can't be released while holding i_ceph_lock */
 		if (ci->i_xattrs.prealloc_blob)
-			ceph_buffer_put(ci->i_xattrs.prealloc_blob);
+			old_blob = ci->i_xattrs.prealloc_blob;
 		ci->i_xattrs.prealloc_blob = blob;
 		goto retry;
 	}
@@ -1045,6 +1048,7 @@ int __ceph_setxattr(struct inode *inode, const char *name,
 	}
 
 	spin_unlock(&ci->i_ceph_lock);
+	ceph_buffer_put(old_blob);
 	if (lock_snap_rwsem)
 		up_read(&mdsc->snap_rwsem);
 	if (dirty)
Signature · 1 line
2.20.1
author maintainer
#24 lore raw

PATCH 25/27 ceph: fix buffer free while holding i_ceph_lock in __ceph_build_xattrs_blob()

In reply to Sasha Levin

From: Luis Henriques <[email protected]>

[ Upstream commit 12fe3dda7ed89c95cc0ef7abc001ad1ad3e092f8 ]

Calling ceph_buffer_put() in __ceph_build_xattrs_blob() may result in freeing the i_xattrs.blob buffer while holding the i_ceph_lock. This can be fixed by having this function returning the old blob buffer and have the callers of this function freeing it when the lock is released.

The following backtrace was triggered by fstests generic/117.

  BUG: sleeping function called from invalid context at mm/vmalloc.c:2283
  in_atomic(): 1, irqs_disabled(): 0, pid: 649, name: fsstress
  4 locks held by fsstress/649:
   #0: 00000000a7478e7e (&type->s_umount_key#19){++++}, at: iterate_supers+0x77/0xf0
   #1: 00000000f8de1423 (&(&ci->i_ceph_lock)->rlock){+.+.}, at: ceph_check_caps+0x7b/0xc60
   #2: 00000000562f2b27 (&s->s_mutex){+.+.}, at: ceph_check_caps+0x3bd/0xc60
   #3: 00000000f83ce16a (&mdsc->snap_rwsem){++++}, at: ceph_check_caps+0x3ed/0xc60
  CPU: 1 PID: 649 Comm: fsstress Not tainted 5.2.0+ #439
  Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.12.1-0-ga5cab58-prebuilt.qemu.org 04/01/2014
  Call Trace:
   dump_stack+0x67/0x90
   ___might_sleep.cold+0x9f/0xb1
   vfree+0x4b/0x60
   ceph_buffer_release+0x1b/0x60
   __ceph_build_xattrs_blob+0x12b/0x170
   __send_cap+0x302/0x540
   ? __lock_acquire+0x23c/0x1e40
   ? __mark_caps_flushing+0x15c/0x280
   ? _raw_spin_unlock+0x24/0x30
   ceph_check_caps+0x5f0/0xc60
   ceph_flush_dirty_caps+0x7c/0x150
   ? __ia32_sys_fdatasync+0x20/0x20
   ceph_sync_fs+0x5a/0x130
   iterate_supers+0x8f/0xf0
   ksys_sync+0x4f/0xb0
   __ia32_sys_sync+0xa/0x10
   do_syscall_64+0x50/0x1c0
   entry_SYSCALL_64_after_hwframe+0x49/0xbe
  RIP: 0033:0x7fc6409ab617
---
 fs/ceph/caps.c  |  5 ++++-
 fs/ceph/snap.c  |  4 +++-
 fs/ceph/super.h |  2 +-
 fs/ceph/xattr.c | 11 ++++++++---
 4 files changed, 16 insertions(+), 6 deletions(-)
fs/ceph/caps.c +4 -1
index 238d24348a98a..df95e39ccd45b 100644
@@ -1162,6 +1162,7 @@ static int __send_cap(struct ceph_mds_client *mdsc, struct ceph_cap *cap,
 {
 	struct ceph_inode_info *ci = cap->ci;
 	struct inode *inode = &ci->vfs_inode;
+	struct ceph_buffer *old_blob = NULL;
 	struct cap_msg_args arg;
 	int held, revoking, dropping;
 	int wake = 0;
@@ -1227,7 +1228,7 @@ static int __send_cap(struct ceph_mds_client *mdsc, struct ceph_cap *cap,
 	ci->i_requested_max_size = arg.max_size;
 
 	if (flushing & CEPH_CAP_XATTR_EXCL) {
-		__ceph_build_xattrs_blob(ci);
+		old_blob = __ceph_build_xattrs_blob(ci);
 		arg.xattr_version = ci->i_xattrs.version;
 		arg.xattr_buf = ci->i_xattrs.blob;
 	} else {
@@ -1262,6 +1263,8 @@ static int __send_cap(struct ceph_mds_client *mdsc, struct ceph_cap *cap,
 
 	spin_unlock(&ci->i_ceph_lock);
 
+	ceph_buffer_put(old_blob);
+
 	ret = send_cap_msg(&arg);
 	if (ret < 0) {
 		dout("error sending cap msg, must requeue %p\n", inode);
fs/ceph/snap.c +3 -1
index a7e763dac0385..29ed1688a1d3a 100644
@@ -460,6 +460,7 @@ void ceph_queue_cap_snap(struct ceph_inode_info *ci)
 	struct inode *inode = &ci->vfs_inode;
 	struct ceph_cap_snap *capsnap;
 	struct ceph_snap_context *old_snapc, *new_snapc;
+	struct ceph_buffer *old_blob = NULL;
 	int used, dirty;
 
 	capsnap = kzalloc(sizeof(*capsnap), GFP_NOFS);
@@ -536,7 +537,7 @@ void ceph_queue_cap_snap(struct ceph_inode_info *ci)
 	capsnap->gid = inode->i_gid;
 
 	if (dirty & CEPH_CAP_XATTR_EXCL) {
-		__ceph_build_xattrs_blob(ci);
+		old_blob = __ceph_build_xattrs_blob(ci);
 		capsnap->xattr_blob =
 			ceph_buffer_get(ci->i_xattrs.blob);
 		capsnap->xattr_version = ci->i_xattrs.version;
@@ -579,6 +580,7 @@ void ceph_queue_cap_snap(struct ceph_inode_info *ci)
 	}
 	spin_unlock(&ci->i_ceph_lock);
 
+	ceph_buffer_put(old_blob);
 	kfree(capsnap);
 	ceph_put_snap_context(old_snapc);
 }
fs/ceph/super.h +1 -1
index 60b70f0985f67..46f600107cb5b 100644
@@ -835,7 +835,7 @@ extern int ceph_getattr(const struct path *path, struct kstat *stat,
 int __ceph_setxattr(struct inode *, const char *, const void *, size_t, int);
 ssize_t __ceph_getxattr(struct inode *, const char *, void *, size_t);
 extern ssize_t ceph_listxattr(struct dentry *, char *, size_t);
-extern void __ceph_build_xattrs_blob(struct ceph_inode_info *ci);
+extern struct ceph_buffer *__ceph_build_xattrs_blob(struct ceph_inode_info *ci);
 extern void __ceph_destroy_xattrs(struct ceph_inode_info *ci);
 extern void __init ceph_xattr_init(void);
 extern void ceph_xattr_exit(void);
fs/ceph/xattr.c +8 -3
index 932c6cdc22d66..3a166f860b6cf 100644
@@ -681,12 +681,15 @@ static int __get_required_blob_size(struct ceph_inode_info *ci, int name_size,
 
 /*
  * If there are dirty xattrs, reencode xattrs into the prealloc_blob
- * and swap into place.
+ * and swap into place.  It returns the old i_xattrs.blob (or NULL) so
+ * that it can be freed by the caller as the i_ceph_lock is likely to be
+ * held.
  */
-void __ceph_build_xattrs_blob(struct ceph_inode_info *ci)
+struct ceph_buffer *__ceph_build_xattrs_blob(struct ceph_inode_info *ci)
 {
 	struct rb_node *p;
 	struct ceph_inode_xattr *xattr = NULL;
+	struct ceph_buffer *old_blob = NULL;
 	void *dest;
 
 	dout("__build_xattrs_blob %p\n", &ci->vfs_inode);
@@ -717,12 +720,14 @@ void __ceph_build_xattrs_blob(struct ceph_inode_info *ci)
 			dest - ci->i_xattrs.prealloc_blob->vec.iov_base;
 
 		if (ci->i_xattrs.blob)
-			ceph_buffer_put(ci->i_xattrs.blob);
+			old_blob = ci->i_xattrs.blob;
 		ci->i_xattrs.blob = ci->i_xattrs.prealloc_blob;
 		ci->i_xattrs.prealloc_blob = NULL;
 		ci->i_xattrs.dirty = false;
 		ci->i_xattrs.version++;
 	}
+
+	return old_blob;
 }
 
 static inline int __get_request_mask(struct inode *in) {
Signature · 1 line
2.20.1
author maintainer
#25 lore raw

PATCH 26/27 ceph: fix buffer free while holding i_ceph_lock in fill_inode()

In reply to Sasha Levin

From: Luis Henriques <[email protected]>

[ Upstream commit af8a85a41734f37b67ba8ce69d56b685bee4ac48 ]

Calling ceph_buffer_put() in fill_inode() may result in freeing the i_xattrs.blob buffer while holding the i_ceph_lock. This can be fixed by postponing the call until later, when the lock is released.

The following backtrace was triggered by fstests generic/070.

  BUG: sleeping function called from invalid context at mm/vmalloc.c:2283
  in_atomic(): 1, irqs_disabled(): 0, pid: 3852, name: kworker/0:4
  6 locks held by kworker/0:4/3852:
   #0: 000000004270f6bb ((wq_completion)ceph-msgr){+.+.}, at: process_one_work+0x1b8/0x5f0
   #1: 00000000eb420803 ((work_completion)(&(&con->work)->work)){+.+.}, at: process_one_work+0x1b8/0x5f0
   #2: 00000000be1c53a4 (&s->s_mutex){+.+.}, at: dispatch+0x288/0x1476
   #3: 00000000559cb958 (&mdsc->snap_rwsem){++++}, at: dispatch+0x2eb/0x1476
   #4: 000000000d5ebbae (&req->r_fill_mutex){+.+.}, at: dispatch+0x2fc/0x1476
   #5: 00000000a83d0514 (&(&ci->i_ceph_lock)->rlock){+.+.}, at: fill_inode.isra.0+0xf8/0xf70
  CPU: 0 PID: 3852 Comm: kworker/0:4 Not tainted 5.2.0+ #441
  Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.12.1-0-ga5cab58-prebuilt.qemu.org 04/01/2014
  Workqueue: ceph-msgr ceph_con_workfn
  Call Trace:
   dump_stack+0x67/0x90
   ___might_sleep.cold+0x9f/0xb1
   vfree+0x4b/0x60
   ceph_buffer_release+0x1b/0x60
   fill_inode.isra.0+0xa9b/0xf70
   ceph_fill_trace+0x13b/0xc70
   ? dispatch+0x2eb/0x1476
   dispatch+0x320/0x1476
   ? __mutex_unlock_slowpath+0x4d/0x2a0
   ceph_con_workfn+0xc97/0x2ec0
   ? process_one_work+0x1b8/0x5f0
   process_one_work+0x244/0x5f0
   worker_thread+0x4d/0x3e0
   kthread+0x105/0x140
   ? process_one_work+0x5f0/0x5f0
   ? kthread_park+0x90/0x90
   ret_from_fork+0x3a/0x50
---
 fs/ceph/inode.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)
fs/ceph/inode.c +4 -3
index f2b722f0df5d0..9bda8c7a80a05 100644
@@ -730,6 +730,7 @@ static int fill_inode(struct inode *inode, struct page *locked_page,
 	int issued = 0, implemented, new_issued;
 	struct timespec mtime, atime, ctime;
 	struct ceph_buffer *xattr_blob = NULL;
+	struct ceph_buffer *old_blob = NULL;
 	struct ceph_string *pool_ns = NULL;
 	struct ceph_cap *new_cap = NULL;
 	int err = 0;
@@ -847,7 +848,7 @@ static int fill_inode(struct inode *inode, struct page *locked_page,
 	if ((ci->i_xattrs.version == 0 || !(issued & CEPH_CAP_XATTR_EXCL))  &&
 	    le64_to_cpu(info->xattr_version) > ci->i_xattrs.version) {
 		if (ci->i_xattrs.blob)
-			ceph_buffer_put(ci->i_xattrs.blob);
+			old_blob = ci->i_xattrs.blob;
 		ci->i_xattrs.blob = xattr_blob;
 		if (xattr_blob)
 			memcpy(ci->i_xattrs.blob->vec.iov_base,
@@ -993,8 +994,8 @@ static int fill_inode(struct inode *inode, struct page *locked_page,
 out:
 	if (new_cap)
 		ceph_put_cap(mdsc, new_cap);
-	if (xattr_blob)
-		ceph_buffer_put(xattr_blob);
+	ceph_buffer_put(old_blob);
+	ceph_buffer_put(xattr_blob);
 	ceph_put_string(pool_ns);
 	return err;
 }
Signature · 1 line
2.20.1
author maintainer
#26 lore raw

PATCH 27/27 KVM: arm/arm64: Only skip MMIO insn once

In reply to Sasha Levin

From: Andrew Jones <[email protected]>

[ Upstream commit 2113c5f62b7423e4a72b890bd479704aa85c81ba ]

If after an MMIO exit to userspace a VCPU is immediately run with an immediate_exit request, such as when a signal is delivered or an MMIO emulation completion is needed, then the VCPU completes the MMIO emulation and immediately returns to userspace. As the exit_reason does not get changed from KVM_EXIT_MMIO in these cases we have to be careful not to complete the MMIO emulation again, when the VCPU is eventually run again, because the emulation does an instruction skip (and doing too many skips would be a waste of guest code :-) We need to use additional VCPU state to track if the emulation is complete. As luck would have it, we already have 'mmio_needed', which even appears to be used in this way by other architectures already.

Fixes 0d640732dbeb ("arm64: KVM: Skip MMIO insn after emulation")
---
 virt/kvm/arm/mmio.c | 7 +++++++
 1 file changed, 7 insertions(+)
virt/kvm/arm/mmio.c +7 -0
index 08443a15e6be8..3caee91bca089 100644
@@ -98,6 +98,12 @@ int kvm_handle_mmio_return(struct kvm_vcpu *vcpu, struct kvm_run *run)
 	unsigned int len;
 	int mask;
 
+	/* Detect an already handled MMIO return */
+	if (unlikely(!vcpu->mmio_needed))
+		return 0;
+
+	vcpu->mmio_needed = 0;
+
 	if (!run->mmio.is_write) {
 		len = run->mmio.len;
 		if (len > sizeof(unsigned long))
@@ -200,6 +206,7 @@ int io_mem_abort(struct kvm_vcpu *vcpu, struct kvm_run *run,
 	run->mmio.is_write	= is_write;
 	run->mmio.phys_addr	= fault_ipa;
 	run->mmio.len		= len;
+	vcpu->mmio_needed	= 1;
 
 	if (!ret) {
 		/* We handled the access successfully in the kernel. */
Signature · 6 lines
2.20.1

_______________________________________________
kvmarm mailing list
[email protected]
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm