Skip to contentSkip to conversation

opp: remove WARN when no valid OPPs remain

PATCH New Sasha Levin 1 participant 10 messages Aug 24, 2021

Subsystems (6)
lore
Share & tools
lore

Sasha Levin

author maintainer
#0 lore raw

PATCH 1/10 opp: remove WARN when no valid OPPs remain

From: Michał Mirosław <[email protected]>

[ Upstream commit 335ffab3ef864539e814b9a2903b0ae420c1c067 ]

This WARN can be triggered per-core and the stack trace is not useful. Replace it with plain dev_err(). Fix a comment while at it.

---
 drivers/opp/of.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
drivers/opp/of.c +3 -2
index d64a13d7881b..a53123356697 100644
@@ -423,8 +423,9 @@ static int _of_add_opp_table_v2(struct device *dev, struct device_node *opp_np)
 		}
 	}
 
-	/* There should be one of more OPP defined */
-	if (WARN_ON(!count)) {
+	/* There should be one or more OPPs defined */
+	if (!count) {
+		dev_err(dev, "%s: no supported OPPs", __func__);
 		ret = -ENOENT;
 		goto put_opp_table;
 	}
Signature · 1 line
2.30.2
author maintainer
#1 lore raw

PATCH 2/10 virtio: Improve vq->broken access to avoid any compiler optimization

From: Parav Pandit <[email protected]>

[ Upstream commit 60f0779862e4ab943810187752c462e85f5fa371 ]

Currently vq->broken field is read by virtqueue_is_broken() in busy loop in one context by virtnet_send_command().

vq->broken is set to true in other process context by virtio_break_device(). Reader and writer are accessing it without any synchronization. This may lead to a compiler optimization which may result to optimize reading vq->broken only once.

Hence, force reading vq->broken on each invocation of virtqueue_is_broken() and also force writing it so that such update is visible to the readers.

It is a theoretical fix that isn't yet encountered in the field.

---
 drivers/virtio/virtio_ring.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/virtio/virtio_ring.c +4 -2
index df7980aef927..0cc0cfd3a3cb 100644
@@ -1197,7 +1197,7 @@ bool virtqueue_is_broken(struct virtqueue *_vq)
 {
 	struct vring_virtqueue *vq = to_vvq(_vq);
 
-	return vq->broken;
+	return READ_ONCE(vq->broken);
 }
 EXPORT_SYMBOL_GPL(virtqueue_is_broken);
 
@@ -1211,7 +1211,9 @@ void virtio_break_device(struct virtio_device *dev)
 
 	list_for_each_entry(_vq, &dev->vqs, list) {
 		struct vring_virtqueue *vq = to_vvq(_vq);
-		vq->broken = true;
+
+		/* Pairs with READ_ONCE() in virtqueue_is_broken(). */
+		WRITE_ONCE(vq->broken, true);
 	}
 }
 EXPORT_SYMBOL_GPL(virtio_break_device);
Signature · 6 lines
2.30.2

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

PATCH 3/10 virtio_pci: Support surprise removal of virtio pci device

In reply to Sasha Levin

From: Parav Pandit <[email protected]>

[ Upstream commit 43bb40c5b92659966bdf4bfe584fde0a3575a049 ]

When a virtio pci device undergo surprise removal (aka async removal in PCIe spec), mark the device as broken so that any upper layer drivers can abort any outstanding operation.

When a virtio net pci device undergo surprise removal which is used by a NetworkManager, a below call trace was observed.

kernel:watchdog: BUG: soft lockup - CPU#1 stuck for 26s! [kworker/1:1:27059]

watchdog: BUG: soft lockup - CPU#1 stuck for 52s! [kworker/1:1:27059]
CPU: 1 PID: 27059 Comm: kworker/1:1 Tainted: G S      W I  L    5.13.0-hotplug+ #8
Hardware name: Dell Inc. PowerEdge R640/0H28RR, BIOS 2.9.4 11/06/2020
Workqueue: events linkwatch_event
RIP: 0010:virtnet_send_command+0xfc/0x150 [virtio_net]
Call Trace:
 virtnet_set_rx_mode+0xcf/0x2a7 [virtio_net]
 ? __hw_addr_create_ex+0x85/0xc0
 __dev_mc_add+0x72/0x80
 igmp6_group_added+0xa7/0xd0
 ipv6_mc_up+0x3c/0x60
 ipv6_find_idev+0x36/0x80
 addrconf_add_dev+0x1e/0xa0
 addrconf_dev_config+0x71/0x130
 addrconf_notify+0x1f5/0xb40
 ? rtnl_is_locked+0x11/0x20
 ? __switch_to_asm+0x42/0x70
 ? finish_task_switch+0xaf/0x2c0
 ? raw_notifier_call_chain+0x3e/0x50
 raw_notifier_call_chain+0x3e/0x50
 netdev_state_change+0x67/0x90
 linkwatch_do_dev+0x3c/0x50
 __linkwatch_run_queue+0xd2/0x220
 linkwatch_event+0x21/0x30
 process_one_work+0x1c8/0x370
 worker_thread+0x30/0x380
 ? process_one_work+0x370/0x370
 kthread+0x118/0x140
 ? set_kthread_struct+0x40/0x40
 ret_from_fork+0x1f/0x30

Hence, add the ability to abort the command on surprise removal which prevents infinite loop and system lockup.

---
 drivers/virtio/virtio_pci_common.c | 7 +++++++
 1 file changed, 7 insertions(+)
drivers/virtio/virtio_pci_common.c +7 -0
index 45b04bc91f24..b7cc63f556ee 100644
@@ -579,6 +579,13 @@ static void virtio_pci_remove(struct pci_dev *pci_dev)
 	struct virtio_pci_device *vp_dev = pci_get_drvdata(pci_dev);
 	struct device *dev = get_device(&vp_dev->vdev.dev);
 
+	/*
+	 * Device is marked broken on surprise removal so that virtio upper
+	 * layers can abort any ongoing operation.
+	 */
+	if (!pci_device_is_present(pci_dev))
+		virtio_break_device(&vp_dev->vdev);
+
 	pci_disable_sriov(pci_dev);
 
 	unregister_virtio_device(&vp_dev->vdev);
Signature · 6 lines
2.30.2

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

PATCH 4/10 vringh: Use wiov->used to check for read/write desc order

In reply to Sasha Levin

From: Neeraj Upadhyay <[email protected]>

[ Upstream commit e74cfa91f42c50f7f649b0eca46aa049754ccdbd ]

As __vringh_iov() traverses a descriptor chain, it populates each descriptor entry into either read or write vring iov and increments that iov's ->used member. So, as we iterate over a descriptor chain, at any point, (riov/wriov)->used value gives the number of descriptor enteries available, which are to be read or written by the device. As all read iovs must precede the write iovs, wiov->used should be zero when we are traversing a read descriptor. Current code checks for wiov->i, to figure out whether any previous entry in the current descriptor chain was a write descriptor. However, iov->i is only incremented, when these vring iovs are consumed, at a later point, and remain 0 in __vringh_iov(). So, correct the check for read and write descriptor order, to use wiov->used.

---
 drivers/vhost/vringh.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/vhost/vringh.c +1 -1
index 59c61744dcc1..97aa9b87e572 100644
@@ -330,7 +330,7 @@ __vringh_iov(struct vringh *vrh, u16 i,
 			iov = wiov;
 		else {
 			iov = riov;
-			if (unlikely(wiov && wiov->i)) {
+			if (unlikely(wiov && wiov->used)) {
 				vringh_bad("Readable desc %p after writable",
 					   &descs[i]);
 				err = -EINVAL;
Signature · 6 lines
2.30.2

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

PATCH 5/10 qed: qed ll2 race condition fixes

In reply to Sasha Levin

From: Shai Malin <[email protected]>

[ Upstream commit 37110237f31105d679fc0aa7b11cdec867750ea7 ]

Avoiding qed ll2 race condition and NULL pointer dereference as part of the remove and recovery flows.

Changes form V1:
- Change (!p_rx->set_prod_addr).
- qed_ll2.c checkpatch fixes.

Change from V2:
- Revert "qed_ll2.c checkpatch fixes".

---
 drivers/net/ethernet/qlogic/qed/qed_ll2.c | 20 ++++++++++++++++++++
 1 file changed, 20 insertions(+)
drivers/net/ethernet/qlogic/qed/qed_ll2.c +20 -0
index 2847509a183d..cb3569ac85f7 100644
@@ -354,6 +354,9 @@ static int qed_ll2_txq_completion(struct qed_hwfn *p_hwfn, void *p_cookie)
 	unsigned long flags;
 	int rc = -EINVAL;
 
+	if (!p_ll2_conn)
+		return rc;
+
 	spin_lock_irqsave(&p_tx->lock, flags);
 	if (p_tx->b_completing_packet) {
 		rc = -EBUSY;
@@ -527,7 +530,16 @@ static int qed_ll2_rxq_completion(struct qed_hwfn *p_hwfn, void *cookie)
 	unsigned long flags = 0;
 	int rc = 0;
 
+	if (!p_ll2_conn)
+		return rc;
+
 	spin_lock_irqsave(&p_rx->lock, flags);
+
+	if (!QED_LL2_RX_REGISTERED(p_ll2_conn)) {
+		spin_unlock_irqrestore(&p_rx->lock, flags);
+		return 0;
+	}
+
 	cq_new_idx = le16_to_cpu(*p_rx->p_fw_cons);
 	cq_old_idx = qed_chain_get_cons_idx(&p_rx->rcq_chain);
 
@@ -848,6 +860,9 @@ static int qed_ll2_lb_rxq_completion(struct qed_hwfn *p_hwfn, void *p_cookie)
 	struct qed_ll2_info *p_ll2_conn = (struct qed_ll2_info *)p_cookie;
 	int rc;
 
+	if (!p_ll2_conn)
+		return 0;
+
 	if (!QED_LL2_RX_REGISTERED(p_ll2_conn))
 		return 0;
 
@@ -871,6 +886,9 @@ static int qed_ll2_lb_txq_completion(struct qed_hwfn *p_hwfn, void *p_cookie)
 	u16 new_idx = 0, num_bds = 0;
 	int rc;
 
+	if (!p_ll2_conn)
+		return 0;
+
 	if (!QED_LL2_TX_REGISTERED(p_ll2_conn))
 		return 0;
 
@@ -1628,6 +1646,8 @@ int qed_ll2_post_rx_buffer(void *cxt,
 	if (!p_ll2_conn)
 		return -EINVAL;
 	p_rx = &p_ll2_conn->rx_queue;
+	if (!p_rx->set_prod_addr)
+		return -EIO;
 
 	spin_lock_irqsave(&p_rx->lock, flags);
 	if (!list_empty(&p_rx->free_descq))
Signature · 1 line
2.30.2
author maintainer
#5 lore raw

PATCH 6/10 qed: Fix null-pointer dereference in qed_rdma_create_qp()

In reply to Sasha Levin

From: Shai Malin <[email protected]>

[ Upstream commit d33d19d313d3466abdf8b0428be7837aff767802 ]

Fix a possible null-pointer dereference in qed_rdma_create_qp().

Changes from V2:
- Revert checkpatch fixes.

---
 drivers/net/ethernet/qlogic/qed/qed_rdma.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
drivers/net/ethernet/qlogic/qed/qed_rdma.c +1 -2
index 909422d93903..3392982ff374 100644
@@ -1244,8 +1244,7 @@ qed_rdma_create_qp(void *rdma_cxt,
 
 	if (!rdma_cxt || !in_params || !out_params ||
 	    !p_hwfn->p_rdma_info->active) {
-		DP_ERR(p_hwfn->cdev,
-		       "qed roce create qp failed due to NULL entry (rdma_cxt=%p, in=%p, out=%p, roce_info=?\n",
+		pr_err("qed roce create qp failed due to NULL entry (rdma_cxt=%p, in=%p, out=%p, roce_info=?\n",
 		       rdma_cxt, in_params, out_params);
 		return NULL;
 	}
Signature · 1 line
2.30.2
author maintainer
#6 lore raw

PATCH 7/10 drm: Copy drm_wait_vblank to user before returning

In reply to Sasha Levin

From: Mark Yacoub <[email protected]>

[ Upstream commit fa0b1ef5f7a694f48e00804a391245f3471aa155 ]

[Why]
Userspace should get back a copy of drm_wait_vblank that's been modified even when drm_wait_vblank_ioctl returns a failure.

Rationale:
drm_wait_vblank_ioctl modifies the request and expects the user to read it back. When the type is RELATIVE, it modifies it to ABSOLUTE and updates the sequence to become current_vblank_count + sequence (which was RELATIVE), but now it became ABSOLUTE.
drmWaitVBlank (in libdrm) expects this to be the case as it modifies the request to be Absolute so it expects the sequence to would have been updated.

The change is in compat_drm_wait_vblank, which is called by drm_compat_ioctl. This change of copying the data back regardless of the return number makes it en par with drm_ioctl, which always copies the data before returning.

[How]
Return from the function after everything has been copied to user.

Fixes IGT:kms_flip::modeset-vs-vblank-race-interruptible Tested on ChromeOS Trogdor(msm)

---
 drivers/gpu/drm/drm_ioc32.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)
drivers/gpu/drm/drm_ioc32.c +1 -3
index ab8847c7dd96..87e13bcd7a67 100644
@@ -855,8 +855,6 @@ static int compat_drm_wait_vblank(struct file *file, unsigned int cmd,
 	req.request.sequence = req32.request.sequence;
 	req.request.signal = req32.request.signal;
 	err = drm_ioctl_kernel(file, drm_wait_vblank_ioctl, &req, DRM_UNLOCKED);
-	if (err)
-		return err;
 
 	req32.reply.type = req.reply.type;
 	req32.reply.sequence = req.reply.sequence;
@@ -865,7 +863,7 @@ static int compat_drm_wait_vblank(struct file *file, unsigned int cmd,
 	if (copy_to_user(argp, &req32, sizeof(req32)))
 		return -EFAULT;
 
-	return 0;
+	return err;
 }
 
 #if defined(CONFIG_X86)
Signature · 1 line
2.30.2
author maintainer
#7 lore raw

PATCH 8/10 drm/nouveau/disp: power down unused DP links during init

In reply to Sasha Levin

From: Ben Skeggs <[email protected]>

[ Upstream commit 6eaa1f3c59a707332e921e32782ffcad49915c5e ]

When booted with multiple displays attached, the EFI GOP driver on (at least) Ampere, can leave DP links powered up that aren't being used to display anything. This confuses our tracking of SOR routing, with the likely result being a failed modeset and display engine hang.

Fix this by (ab?)using the DisableLT IED script to power-down the link, restoring HW to a state the driver expects.

---
 drivers/gpu/drm/nouveau/nvkm/engine/disp/dp.c   | 2 +-
 drivers/gpu/drm/nouveau/nvkm/engine/disp/dp.h   | 1 +
 drivers/gpu/drm/nouveau/nvkm/engine/disp/outp.c | 9 +++++++++
 3 files changed, 11 insertions(+), 1 deletion(-)
drivers/gpu/drm/nouveau/nvkm/engine/disp/dp.c +1 -1
index 818d21bd28d3..1d2837c5a8f2 100644
@@ -419,7 +419,7 @@ nvkm_dp_train(struct nvkm_dp *dp, u32 dataKBps)
 	return ret;
 }
 
-static void
+void
 nvkm_dp_disable(struct nvkm_outp *outp, struct nvkm_ior *ior)
 {
 	struct nvkm_dp *dp = nvkm_dp(outp);
drivers/gpu/drm/nouveau/nvkm/engine/disp/dp.h +1 -0
index 495f665a0ee6..12d6ff4cfa95 100644
@@ -32,6 +32,7 @@ struct nvkm_dp {
 
 int nvkm_dp_new(struct nvkm_disp *, int index, struct dcb_output *,
 		struct nvkm_outp **);
+void nvkm_dp_disable(struct nvkm_outp *, struct nvkm_ior *);
 
 /* DPCD Receiver Capabilities */
 #define DPCD_RC00_DPCD_REV                                              0x00000
drivers/gpu/drm/nouveau/nvkm/engine/disp/outp.c +9 -0
index c62030c96fba..4b1c72fd8f03 100644
@@ -22,6 +22,7 @@
  * Authors: Ben Skeggs
  */
 #include "outp.h"
+#include "dp.h"
 #include "ior.h"
 
 #include <subdev/bios.h>
@@ -216,6 +217,14 @@ nvkm_outp_init_route(struct nvkm_outp *outp)
 	if (!ior->arm.head || ior->arm.proto != proto) {
 		OUTP_DBG(outp, "no heads (%x %d %d)", ior->arm.head,
 			 ior->arm.proto, proto);
+
+		/* The EFI GOP driver on Ampere can leave unused DP links routed,
+		 * which we don't expect.  The DisableLT IED script *should* get
+		 * us back to where we need to be.
+		 */
+		if (ior->func->route.get && !ior->arm.head && outp->info.type == DCB_OUTPUT_DP)
+			nvkm_dp_disable(outp, ior);
+
 		return;
 	}
 
Signature · 1 line
2.30.2
author maintainer
#8 lore raw

PATCH 9/10 drm/nouveau: block a bunch of classes from userspace

In reply to Sasha Levin

From: Ben Skeggs <[email protected]>

[ Upstream commit 148a8653789c01f159764ffcc3f370008966b42f ]

Long ago, there had been plans for making use of a bunch of these APIs from userspace and there's various checks in place to stop misbehaving.

Countless other projects have occurred in the meantime, and the pieces didn't finish falling into place for that to happen.

They will (hopefully) in the not-too-distant future, but it won't look quite as insane. The super checks are causing problems right now, and are going to be removed.

---
 drivers/gpu/drm/nouveau/include/nvif/cl0080.h |  3 +-
 drivers/gpu/drm/nouveau/nouveau_drm.c         |  1 +
 drivers/gpu/drm/nouveau/nouveau_usif.c        | 57 ++++++++++++++-----
 .../gpu/drm/nouveau/nvkm/engine/device/user.c |  2 +-
 4 files changed, 48 insertions(+), 15 deletions(-)
drivers/gpu/drm/nouveau/include/nvif/cl0080.h +2 -1
index 4f5233107f5f..82daa2bf8415 100644
@@ -4,7 +4,8 @@
 
 struct nv_device_v0 {
 	__u8  version;
-	__u8  pad01[7];
+	__u8  priv;
+	__u8  pad02[6];
 	__u64 device;	/* device identifier, ~0 for client default */
 };
 
drivers/gpu/drm/nouveau/nouveau_drm.c +1 -0
index 81999bed1e4a..f4f0a1772774 100644
@@ -237,6 +237,7 @@ nouveau_cli_init(struct nouveau_drm *drm, const char *sname,
 	ret = nvif_device_init(&cli->base.object, 0, NV_DEVICE,
 			       &(struct nv_device_v0) {
 					.device = ~0,
+					.priv = true,
 			       }, sizeof(struct nv_device_v0),
 			       &cli->device);
 	if (ret) {
drivers/gpu/drm/nouveau/nouveau_usif.c +44 -13

97 lines not on this page · raw message

drivers/gpu/drm/nouveau/nvkm/engine/device/user.c +1 -1
index dde6bbafa709..f8658894e753 100644
@@ -426,7 +426,7 @@ nvkm_udevice_new(const struct nvkm_oclass *oclass, void *data, u32 size,
 		return ret;
 
 	/* give priviledged clients register access */
-	if (client->super)
+	if (args->v0.priv)
 		func = &nvkm_udevice_super;
 	else
 		func = &nvkm_udevice;
Signature · 1 line
2.30.2
author maintainer
#9 lore raw

PATCH 10/10 net/rds: dma_map_sg is entitled to merge entries

In reply to Sasha Levin

From: Gerd Rausch <[email protected]>

[ Upstream commit fb4b1373dcab086d0619c29310f0466a0b2ceb8a ]

Function "dma_map_sg" is entitled to merge adjacent entries and return a value smaller than what was passed as "nents".

Subsequently "ib_map_mr_sg" needs to work with this value ("sg_dma_len") rather than the original "nents" parameter ("sg_len").

This old RDS bug was exposed and reliably causes kernel panics (using RDMA operations "rds-stress -D") on x86_64 starting with:
commit c588072bba6b ("iommu/vt-d: Convert intel iommu driver to the iommu ops")

Simply put: Linux 5.11 and later.

---
 net/rds/ib_frmr.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
net/rds/ib_frmr.c +2 -2
index 6431a023ac89..46988c009a3e 100644
@@ -111,9 +111,9 @@ static int rds_ib_post_reg_frmr(struct rds_ib_mr *ibmr)
 		cpu_relax();
 	}
 
-	ret = ib_map_mr_sg_zbva(frmr->mr, ibmr->sg, ibmr->sg_len,
+	ret = ib_map_mr_sg_zbva(frmr->mr, ibmr->sg, ibmr->sg_dma_len,
 				&off, PAGE_SIZE);
-	if (unlikely(ret != ibmr->sg_len))
+	if (unlikely(ret != ibmr->sg_dma_len))
 		return ret < 0 ? ret : -EINVAL;
 
 	/* Perform a WR for the fast_reg_mr. Each individual page
Signature · 1 line
2.30.2