Skip to contentSkip to conversation

virtio_mmio: Add missing PM calls to freeze/restore

PATCH New Sasha Levin 1 participant 11 messages Jul 6, 2022

Subsystems (9)
lore
Share & tools
lore

Sasha Levin

author maintainer
#0 lore raw

PATCH 1/11 virtio_mmio: Add missing PM calls to freeze/restore

From: Stephan Gerhold <[email protected]>

[ Upstream commit ed7ac37fde33ccd84e4bd2b9363c191f925364c7 ]

Most virtio drivers provide freeze/restore callbacks to finish up device usage before suspend and to reinitialize the virtio device after resume. However, these callbacks are currently only called when using virtio_pci. virtio_mmio does not have any PM ops defined.

This causes problems for example after suspend to disk (hibernation), since the virtio devices might lose their state after the VMM is restarted. Calling virtio_device_freeze()/restore() ensures that the virtio devices are re-initialized correctly.

Fix this by implementing the dev_pm_ops for virtio_mmio, similar to virtio_pci_common.

Message-Id: <[email protected]>

---
 drivers/virtio/virtio_mmio.c | 23 +++++++++++++++++++++++
 1 file changed, 23 insertions(+)
drivers/virtio/virtio_mmio.c +23 -0
index 5c970e6f664c..7dec1418bf7c 100644
@@ -62,6 +62,7 @@
 #include <linux/list.h>
 #include <linux/module.h>
 #include <linux/platform_device.h>
+#include <linux/pm.h>
 #include <linux/slab.h>
 #include <linux/spinlock.h>
 #include <linux/virtio.h>
@@ -543,6 +544,25 @@ static const struct virtio_config_ops virtio_mmio_config_ops = {
 	.get_shm_region = vm_get_shm_region,
 };
 
+#ifdef CONFIG_PM_SLEEP
+static int virtio_mmio_freeze(struct device *dev)
+{
+	struct virtio_mmio_device *vm_dev = dev_get_drvdata(dev);
+
+	return virtio_device_freeze(&vm_dev->vdev);
+}
+
+static int virtio_mmio_restore(struct device *dev)
+{
+	struct virtio_mmio_device *vm_dev = dev_get_drvdata(dev);
+
+	return virtio_device_restore(&vm_dev->vdev);
+}
+
+static const struct dev_pm_ops virtio_mmio_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(virtio_mmio_freeze, virtio_mmio_restore)
+};
+#endif
 
 static void virtio_mmio_release_dev(struct device *_d)
 {
@@ -787,6 +807,9 @@ static struct platform_driver virtio_mmio_driver = {
 		.name	= "virtio-mmio",
 		.of_match_table	= virtio_mmio_match,
 		.acpi_match_table = ACPI_PTR(virtio_mmio_acpi_match),
+#ifdef CONFIG_PM_SLEEP
+		.pm	= &virtio_mmio_pm_ops,
+#endif
 	},
 };
 
Signature · 6 lines
2.35.1

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

PATCH 2/11 virtio_mmio: Restore guest page size on resume

From: Stephan Gerhold <[email protected]>

[ Upstream commit e0c2ce8217955537dd5434baeba061f209797119 ]

Virtio devices might lose their state when the VMM is restarted after a suspend to disk (hibernation) cycle. This means that the guest page size register must be restored for the virtio_mmio legacy interface, since otherwise the virtio queues are not functional.

This is particularly problematic for QEMU that currently still defaults to using the legacy interface for virtio_mmio. Write the guest page size register again in virtio_mmio_restore() to make legacy virtio_mmio devices work correctly after hibernation.

Message-Id: <[email protected]>

---
 drivers/virtio/virtio_mmio.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/virtio/virtio_mmio.c +3 -0
index 7dec1418bf7c..e8ef0c66e558 100644
@@ -556,6 +556,9 @@ static int virtio_mmio_restore(struct device *dev)
 {
 	struct virtio_mmio_device *vm_dev = dev_get_drvdata(dev);
 
+	if (vm_dev->version == 1)
+		writel(PAGE_SIZE, vm_dev->base + VIRTIO_MMIO_GUEST_PAGE_SIZE);
+
 	return virtio_device_restore(&vm_dev->vdev);
 }
 
Signature · 6 lines
2.35.1

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

PATCH 3/11 netfilter: br_netfilter: do not skip all hooks with 0 priority

In reply to Sasha Levin

From: Florian Westphal <[email protected]>

[ Upstream commit c2577862eeb0be94f151f2f1fff662b028061b00 ]

When br_netfilter module is loaded, skbs may be diverted to the ipv4/ipv6 hooks, just like as if we were routing.

Unfortunately, bridge filter hooks with priority 0 may be skipped in this case.

Example:
1. an nftables bridge ruleset is loaded, with a prerouting hook that has priority 0.
2. interface is added to the bridge.
3. no tcp packet is ever seen by the bridge prerouting hook.
4. flush the ruleset
5. load the bridge ruleset again.
6. tcp packets are processed as expected.

After 1) the only registered hook is the bridge prerouting hook, but its not called yet because the bridge hasn't been brought up yet.

After 2), hook order is:
0 br_nf_pre_routing // br_netfilter internal hook 0 chain bridge f prerouting // nftables bridge ruleset

The packet is diverted to br_nf_pre_routing.
If call-iptables is off, the nftables bridge ruleset is called as expected.

But if its enabled, br_nf_hook_thresh() will skip it because it assumes that all 0-priority hooks had been called previously in bridge context.

To avoid this, check for the br_nf_pre_routing hook itself, we need to resume directly after it, even if this hook has a priority of 0.

Unfortunately, this still results in different packet flow. With this fix, the eval order after in 3) is:
1. br_nf_pre_routing
2. ip(6)tables (if enabled)
3. nftables bridge

but after 5 its the much saner:
1. nftables bridge
2. br_nf_pre_routing
3. ip(6)tables (if enabled)

Unfortunately I don't see a solution here:
It would be possible to move br_nf_pre_routing to a higher priority so that it will be called later in the pipeline, but this also impacts ebtables evaluation order, and would still result in this very ordering problem for all nftables-bridge hooks with the same priority as the br_nf_pre_routing one.

Searching back through the git history I don't think this has ever behaved in any other way, hence, no fixes-tag.

---
 net/bridge/br_netfilter_hooks.c | 21 ++++++++++++++++++---
 1 file changed, 18 insertions(+), 3 deletions(-)
net/bridge/br_netfilter_hooks.c +18 -3
index 68c0d0f92890..10a2c7bca719 100644
@@ -1012,9 +1012,24 @@ int br_nf_hook_thresh(unsigned int hook, struct net *net,
 		return okfn(net, sk, skb);
 
 	ops = nf_hook_entries_get_hook_ops(e);
-	for (i = 0; i < e->num_hook_entries &&
-	      ops[i]->priority <= NF_BR_PRI_BRNF; i++)
-		;
+	for (i = 0; i < e->num_hook_entries; i++) {
+		/* These hooks have already been called */
+		if (ops[i]->priority < NF_BR_PRI_BRNF)
+			continue;
+
+		/* These hooks have not been called yet, run them. */
+		if (ops[i]->priority > NF_BR_PRI_BRNF)
+			break;
+
+		/* take a closer look at NF_BR_PRI_BRNF. */
+		if (ops[i]->hook == br_nf_pre_routing) {
+			/* This hook diverted the skb to this function,
+			 * hooks after this have not been run yet.
+			 */
+			i++;
+			break;
+		}
+	}
 
 	nf_hook_state_init(&state, hook, NFPROTO_BRIDGE, indev, outdev,
 			   sk, net, okfn);
Signature · 1 line
2.35.1
author maintainer
#3 lore raw

PATCH 4/11 scsi: hisi_sas: Limit max hw sectors for v3 HW

In reply to Sasha Levin

From: John Garry <[email protected]>

[ Upstream commit fce54ed027577517df1e74b7d54dc2b1bd536887 ]

If the controller is behind an IOMMU then the IOMMU IOVA caching range can affect performance, as discussed in [0].

Limit the max HW sectors to not exceed this limit. We need to hardcode the value until a proper DMA mapping API is available.

[0] https://lore.kernel.org/linux-iommu/[email protected]/

---
 drivers/scsi/hisi_sas/hisi_sas_v3_hw.c | 7 +++++++
 1 file changed, 7 insertions(+)
drivers/scsi/hisi_sas/hisi_sas_v3_hw.c +7 -0
index cd41dc061d87..dfe7e6370d84 100644
@@ -2738,6 +2738,7 @@ static int slave_configure_v3_hw(struct scsi_device *sdev)
 	struct hisi_hba *hisi_hba = shost_priv(shost);
 	struct device *dev = hisi_hba->dev;
 	int ret = sas_slave_configure(sdev);
+	unsigned int max_sectors;
 
 	if (ret)
 		return ret;
@@ -2755,6 +2756,12 @@ static int slave_configure_v3_hw(struct scsi_device *sdev)
 		}
 	}
 
+	/* Set according to IOMMU IOVA caching limit */
+	max_sectors = min_t(size_t, queue_max_hw_sectors(sdev->request_queue),
+			    (PAGE_SIZE * 32) >> SECTOR_SHIFT);
+
+	blk_queue_max_hw_sectors(sdev->request_queue, max_sectors);
+
 	return 0;
 }
 
Signature · 1 line
2.35.1
author maintainer
#4 lore raw

PATCH 5/11 cpufreq: pmac32-cpufreq: Fix refcount leak bug

In reply to Sasha Levin

From: Liang He <[email protected]>

[ Upstream commit ccd7567d4b6cf187fdfa55f003a9e461ee629e36 ]

In pmac_cpufreq_init_MacRISC3(), we need to add corresponding of_node_put() for the three node pointers whose refcount have been incremented by of_find_node_by_name().

---
 drivers/cpufreq/pmac32-cpufreq.c | 4 ++++
 1 file changed, 4 insertions(+)
drivers/cpufreq/pmac32-cpufreq.c +4 -0
index 73621bc11976..3704476bb83a 100644
@@ -471,6 +471,10 @@ static int pmac_cpufreq_init_MacRISC3(struct device_node *cpunode)
 	if (slew_done_gpio_np)
 		slew_done_gpio = read_gpio(slew_done_gpio_np);
 
+	of_node_put(volt_gpio_np);
+	of_node_put(freq_gpio_np);
+	of_node_put(slew_done_gpio_np);
+
 	/* If we use the frequency GPIOs, calculate the min/max speeds based
 	 * on the bus frequencies
 	 */
Signature · 1 line
2.35.1
author maintainer
#5 lore raw

PATCH 6/11 platform/x86: hp-wmi: Ignore Sanitization Mode event

In reply to Sasha Levin

From: Kai-Heng Feng <[email protected]>

[ Upstream commit 9ab762a84b8094540c18a170e5ddd6488632c456 ]

After system resume the hp-wmi driver may complain:

[ 702.620180] hp_wmi: Unknown event_id - 23 - 0x0

According to HP it means 'Sanitization Mode' and it's harmless to just ignore the event.

---
 drivers/platform/x86/hp-wmi.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/platform/x86/hp-wmi.c +3 -0
index e94e59283ecb..012639f6d335 100644
@@ -62,6 +62,7 @@ enum hp_wmi_event_ids {
 	HPWMI_BACKLIT_KB_BRIGHTNESS	= 0x0D,
 	HPWMI_PEAKSHIFT_PERIOD		= 0x0F,
 	HPWMI_BATTERY_CHARGE_PERIOD	= 0x10,
+	HPWMI_SANITIZATION_MODE		= 0x17,
 };
 
 struct bios_args {
@@ -629,6 +630,8 @@ static void hp_wmi_notify(u32 value, void *context)
 		break;
 	case HPWMI_BATTERY_CHARGE_PERIOD:
 		break;
+	case HPWMI_SANITIZATION_MODE:
+		break;
 	default:
 		pr_info("Unknown event_id - %d - 0x%x\n", event_id, event_data);
 		break;
Signature · 1 line
2.35.1
author maintainer
#6 lore raw

PATCH 7/11 net: tipc: fix possible refcount leak in tipc_sk_create()

In reply to Sasha Levin

From: Hangyu Hua <[email protected]>

[ Upstream commit 00aff3590fc0a73bddd3b743863c14e76fd35c0c ]

Free sk in case tipc_sk_insert() fails.

---
 net/tipc/socket.c | 1 +
 1 file changed, 1 insertion(+)
net/tipc/socket.c +1 -0
index 42283dc6c5b7..38256aabf4f1 100644
@@ -489,6 +489,7 @@ static int tipc_sk_create(struct net *net, struct socket *sock,
 	sock_init_data(sock, sk);
 	tipc_set_sk_state(sk, TIPC_OPEN);
 	if (tipc_sk_insert(tsk)) {
+		sk_free(sk);
 		pr_warn("Socket create failed; port number exhausted\n");
 		return -EINVAL;
 	}
Signature · 1 line
2.35.1
author maintainer
#7 lore raw

PATCH 8/11 NFC: nxp-nci: don't print header length mismatch on i2c error

In reply to Sasha Levin

From: Michael Walle <[email protected]>

[ Upstream commit 9577fc5fdc8b07b891709af6453545db405e24ad ]

Don't print a misleading header length mismatch error if the i2c call returns an error. Instead just return the error code without any error message.

---
 drivers/nfc/nxp-nci/i2c.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)
drivers/nfc/nxp-nci/i2c.c +6 -2
index 9f60e4dc5a90..a45064b3389c 100644
@@ -122,7 +122,9 @@ static int nxp_nci_i2c_fw_read(struct nxp_nci_i2c_phy *phy,
 	skb_put_data(*skb, &header, NXP_NCI_FW_HDR_LEN);
 
 	r = i2c_master_recv(client, skb_put(*skb, frame_len), frame_len);
-	if (r != frame_len) {
+	if (r < 0) {
+		goto fw_read_exit_free_skb;
+	} else if (r != frame_len) {
 		nfc_err(&client->dev,
 			"Invalid frame length: %u (expected %zu)\n",
 			r, frame_len);
@@ -163,7 +165,9 @@ static int nxp_nci_i2c_nci_read(struct nxp_nci_i2c_phy *phy,
 	skb_put_data(*skb, (void *)&header, NCI_CTRL_HDR_SIZE);
 
 	r = i2c_master_recv(client, skb_put(*skb, header.plen), header.plen);
-	if (r != header.plen) {
+	if (r < 0) {
+		goto nci_read_exit_free_skb;
+	} else if (r != header.plen) {
 		nfc_err(&client->dev,
 			"Invalid frame payload length: %u (expected %u)\n",
 			r, header.plen);
Signature · 1 line
2.35.1
author maintainer
#8 lore raw

PATCH 9/11 nvme-tcp: always fail a request when sending it failed

In reply to Sasha Levin

From: Sagi Grimberg <[email protected]>

[ Upstream commit 41d07df7de841bfbc32725ce21d933ad358f2844 ]

queue stoppage and inflight requests cancellation is fully fenced from io_work and thus failing a request from this context. Hence we don't need to try to guess from the socket retcode if this failure is because the queue is about to be torn down or not.

We are perfectly safe to just fail it, the request will not be cancelled later on.

This solves possible very long shutdown delays when the users issues a 'nvme disconnect-all'

---
 drivers/nvme/host/tcp.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
drivers/nvme/host/tcp.c +1 -2
index 7e3932033707..d5e162f2c23a 100644
@@ -1149,8 +1149,7 @@ static int nvme_tcp_try_send(struct nvme_tcp_queue *queue)
 	} else if (ret < 0) {
 		dev_err(queue->ctrl->ctrl.device,
 			"failed to send request %d\n", ret);
-		if (ret != -EPIPE && ret != -ECONNRESET)
-			nvme_tcp_fail_request(queue->request);
+		nvme_tcp_fail_request(queue->request);
 		nvme_tcp_done_send_req(queue);
 	}
 	return ret;
Signature · 1 line
2.35.1
author maintainer
#9 lore raw

PATCH 10/11 nvme: fix regression when disconnect a recovering ctrl

In reply to Sasha Levin

From: Ruozhu Li <[email protected]>

[ Upstream commit f7f70f4aa09dc43d7455c060143e86a017c30548 ]

We encountered a problem that the disconnect command hangs. After analyzing the log and stack, we found that the triggering process is as follows:
CPU0 CPU1
nvme_rdma_error_recovery_work
nvme_rdma_teardown_io_queues
nvme_do_delete_ctrl nvme_stop_queues
nvme_remove_namespaces
--clear ctrl->namespaces
nvme_start_queues --no ns in ctrl->namespaces
nvme_ns_remove return(because ctrl is deleting)
blk_freeze_queue
blk_mq_freeze_queue_wait
--wait for ns to unquiesce to clean infligt IO, hang forever

This problem was not found in older kernels because we will flush err work in nvme_stop_ctrl before nvme_remove_namespaces.It does not seem to be modified for functional reasons, the patch can be revert to solve the problem.

Revert commit 794a4cb3d2f7 ("nvme: remove the .stop_ctrl callout")

---
 drivers/nvme/host/core.c |  2 ++
 drivers/nvme/host/nvme.h |  1 +
 drivers/nvme/host/rdma.c | 12 +++++++++---
 drivers/nvme/host/tcp.c  | 10 +++++++---
 4 files changed, 19 insertions(+), 6 deletions(-)
drivers/nvme/host/core.c +2 -0
index af2902d70b19..ab060b4911ff 100644
@@ -4460,6 +4460,8 @@ void nvme_stop_ctrl(struct nvme_ctrl *ctrl)
 	nvme_stop_keep_alive(ctrl);
 	flush_work(&ctrl->async_event_work);
 	cancel_work_sync(&ctrl->fw_act_work);
+	if (ctrl->ops->stop_ctrl)
+		ctrl->ops->stop_ctrl(ctrl);
 }
 EXPORT_SYMBOL_GPL(nvme_stop_ctrl);
 
drivers/nvme/host/nvme.h +1 -0
index 8e40a6306e53..58cf9e39d613 100644
@@ -478,6 +478,7 @@ struct nvme_ctrl_ops {
 	void (*free_ctrl)(struct nvme_ctrl *ctrl);
 	void (*submit_async_event)(struct nvme_ctrl *ctrl);
 	void (*delete_ctrl)(struct nvme_ctrl *ctrl);
+	void (*stop_ctrl)(struct nvme_ctrl *ctrl);
 	int (*get_address)(struct nvme_ctrl *ctrl, char *buf, int size);
 };
 
drivers/nvme/host/rdma.c +9 -3
index 8eacc9bd58f5..b61924394032 100644
@@ -1057,6 +1057,14 @@ static void nvme_rdma_teardown_io_queues(struct nvme_rdma_ctrl *ctrl,
 	}
 }
 
+static void nvme_rdma_stop_ctrl(struct nvme_ctrl *nctrl)
+{
+	struct nvme_rdma_ctrl *ctrl = to_rdma_ctrl(nctrl);
+
+	cancel_work_sync(&ctrl->err_work);
+	cancel_delayed_work_sync(&ctrl->reconnect_work);
+}
+
 static void nvme_rdma_free_ctrl(struct nvme_ctrl *nctrl)
 {
 	struct nvme_rdma_ctrl *ctrl = to_rdma_ctrl(nctrl);
@@ -2236,9 +2244,6 @@ static const struct blk_mq_ops nvme_rdma_admin_mq_ops = {
 
 static void nvme_rdma_shutdown_ctrl(struct nvme_rdma_ctrl *ctrl, bool shutdown)
 {
-	cancel_work_sync(&ctrl->err_work);
-	cancel_delayed_work_sync(&ctrl->reconnect_work);
-
 	nvme_rdma_teardown_io_queues(ctrl, shutdown);
 	blk_mq_quiesce_queue(ctrl->ctrl.admin_q);
 	if (shutdown)
@@ -2288,6 +2293,7 @@ static const struct nvme_ctrl_ops nvme_rdma_ctrl_ops = {
 	.submit_async_event	= nvme_rdma_submit_async_event,
 	.delete_ctrl		= nvme_rdma_delete_ctrl,
 	.get_address		= nvmf_get_address,
+	.stop_ctrl		= nvme_rdma_stop_ctrl,
 };
 
 /*
drivers/nvme/host/tcp.c +7 -3
index d5e162f2c23a..fe8c27bbc3f2 100644
@@ -2135,9 +2135,6 @@ static void nvme_tcp_error_recovery_work(struct work_struct *work)
 
 static void nvme_tcp_teardown_ctrl(struct nvme_ctrl *ctrl, bool shutdown)
 {
-	cancel_work_sync(&to_tcp_ctrl(ctrl)->err_work);
-	cancel_delayed_work_sync(&to_tcp_ctrl(ctrl)->connect_work);
-
 	nvme_tcp_teardown_io_queues(ctrl, shutdown);
 	blk_mq_quiesce_queue(ctrl->admin_q);
 	if (shutdown)
@@ -2177,6 +2174,12 @@ static void nvme_reset_ctrl_work(struct work_struct *work)
 	nvme_tcp_reconnect_or_remove(ctrl);
 }
 
+static void nvme_tcp_stop_ctrl(struct nvme_ctrl *ctrl)
+{
+	cancel_work_sync(&to_tcp_ctrl(ctrl)->err_work);
+	cancel_delayed_work_sync(&to_tcp_ctrl(ctrl)->connect_work);
+}
+
 static void nvme_tcp_free_ctrl(struct nvme_ctrl *nctrl)
 {
 	struct nvme_tcp_ctrl *ctrl = to_tcp_ctrl(nctrl);
@@ -2499,6 +2502,7 @@ static const struct nvme_ctrl_ops nvme_tcp_ctrl_ops = {
 	.submit_async_event	= nvme_tcp_submit_async_event,
 	.delete_ctrl		= nvme_tcp_delete_ctrl,
 	.get_address		= nvmf_get_address,
+	.stop_ctrl		= nvme_tcp_stop_ctrl,
 };
 
 static bool
Signature · 1 line
2.35.1
author maintainer
#10 lore raw

PATCH 11/11 net: sfp: fix memory leak in sfp_probe()

In reply to Sasha Levin

From: Jianglei Nie <[email protected]>

[ Upstream commit 0a18d802d65cf662644fd1d369c86d84a5630652 ]

sfp_probe() allocates a memory chunk from sfp with sfp_alloc(). When devm_add_action() fails, sfp is not freed, which leads to a memory leak.

We should use devm_add_action_or_reset() instead of devm_add_action().

---
 drivers/net/phy/sfp.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/phy/sfp.c +1 -1
index 96068e0d841a..dcbe278086dc 100644
@@ -2427,7 +2427,7 @@ static int sfp_probe(struct platform_device *pdev)
 
 	platform_set_drvdata(pdev, sfp);
 
-	err = devm_add_action(sfp->dev, sfp_cleanup, sfp);
+	err = devm_add_action_or_reset(sfp->dev, sfp_cleanup, sfp);
 	if (err < 0)
 		return err;
 
Signature · 1 line
2.35.1