Skip to contentSkip to conversation

usb: host: max3421-hcd: Add missing spi_device_id table

PATCH New Sasha Levin 1 participant 13 messages Apr 7, 2025

Subsystems (9)
lore
Share & tools
lore

Sasha Levin

author maintainer
#0 lore raw

PATCH 1/13 usb: host: max3421-hcd: Add missing spi_device_id table

From: Alexander Stein <[email protected]>

[ Upstream commit 41d5e3806cf589f658f92c75195095df0b66f66a ]

"maxim,max3421" DT compatible is missing its SPI device ID entry, not allowing module autoloading and leading to the following message:
"SPI driver max3421-hcd has no spi_device_id for maxim,max3421"

Fix this by adding the spi_device_id table.

---
 drivers/usb/host/max3421-hcd.c | 7 +++++++
 1 file changed, 7 insertions(+)
drivers/usb/host/max3421-hcd.c +7 -0
index ab12d76b01fbe..8aaafba058aa9 100644
@@ -1955,6 +1955,12 @@ max3421_remove(struct spi_device *spi)
 	usb_put_hcd(hcd);
 }
 
+static const struct spi_device_id max3421_spi_ids[] = {
+	{ "max3421" },
+	{ },
+};
+MODULE_DEVICE_TABLE(spi, max3421_spi_ids);
+
 static const struct of_device_id max3421_of_match_table[] = {
 	{ .compatible = "maxim,max3421", },
 	{},
@@ -1964,6 +1970,7 @@ MODULE_DEVICE_TABLE(of, max3421_of_match_table);
 static struct spi_driver max3421_driver = {
 	.probe		= max3421_probe,
 	.remove		= max3421_remove,
+	.id_table	= max3421_spi_ids,
 	.driver		= {
 		.name	= "max3421-hcd",
 		.of_match_table = of_match_ptr(max3421_of_match_table),
Signature · 1 line
2.39.5
author maintainer
#1 lore raw

PATCH 2/13 fs/ntfs3: Fix WARNING in ntfs_extend_initialized_size

From: Edward Adam Davis <[email protected]>

[ Upstream commit ff355926445897cc9fdea3b00611e514232c213c ]

Syzbot reported a WARNING in ntfs_extend_initialized_size. The data type of in->i_valid and to is u64 in ntfs_file_mmap(). If their values are greater than LLONG_MAX, overflow will occur because the data types of the parameters valid and new_valid corresponding to the function ntfs_extend_initialized_size() are loff_t.

Before calling ntfs_extend_initialized_size() in the ntfs_file_mmap(), the "ni->i_valid < to" has been determined, so the same WARN_ON determination is not required in ntfs_extend_initialized_size().
Just execute the ntfs_extend_initialized_size() in ntfs_extend() to make a WARN_ON check.

---
 fs/ntfs3/file.c | 1 +
 1 file changed, 1 insertion(+)
fs/ntfs3/file.c +1 -0
index 72e25842f5dc9..46eec986ec9ca 100644
@@ -354,6 +354,7 @@ static int ntfs_extend(struct inode *inode, loff_t pos, size_t count,
 	}
 
 	if (extend_init && !is_compressed(ni)) {
+		WARN_ON(ni->i_valid >= pos);
 		err = ntfs_extend_initialized_size(file, ni, ni->i_valid, pos);
 		if (err)
 			goto out;
Signature · 1 line
2.39.5
author maintainer
#2 lore raw

PATCH 3/13 usb: dwc3: gadget: Refactor loop to avoid NULL endpoints

In reply to Sasha Levin

From: Andy Shevchenko <[email protected]>

[ Upstream commit eafba0205426091354f050381c32ad1567c35844 ]

Prepare the gadget driver to handle the reserved endpoints that will be not allocated at the initialisation time.

While at it, add a warning where the NULL endpoint should never happen.

---
 drivers/usb/dwc3/gadget.c | 22 ++++++++++++++++++----
 1 file changed, 18 insertions(+), 4 deletions(-)
drivers/usb/dwc3/gadget.c +18 -4
index f227f6c7c6d7e..d9b3c797077b5 100644
@@ -555,6 +555,7 @@ static int dwc3_gadget_set_xfer_resource(struct dwc3_ep *dep)
 int dwc3_gadget_start_config(struct dwc3 *dwc, unsigned int resource_index)
 {
 	struct dwc3_gadget_ep_cmd_params params;
+	struct dwc3_ep		*dep;
 	u32			cmd;
 	int			i;
 	int			ret;
@@ -571,8 +572,13 @@ int dwc3_gadget_start_config(struct dwc3 *dwc, unsigned int resource_index)
 		return ret;
 
 	/* Reset resource allocation flags */
-	for (i = resource_index; i < dwc->num_eps && dwc->eps[i]; i++)
-		dwc->eps[i]->flags &= ~DWC3_EP_RESOURCE_ALLOCATED;
+	for (i = resource_index; i < dwc->num_eps; i++) {
+		dep = dwc->eps[i];
+		if (!dep)
+			continue;
+
+		dep->flags &= ~DWC3_EP_RESOURCE_ALLOCATED;
+	}
 
 	return 0;
 }
@@ -721,9 +727,11 @@ void dwc3_gadget_clear_tx_fifos(struct dwc3 *dwc)
 
 	dwc->last_fifo_depth = fifo_depth;
 	/* Clear existing TXFIFO for all IN eps except ep0 */
-	for (num = 3; num < min_t(int, dwc->num_eps, DWC3_ENDPOINTS_NUM);
-	     num += 2) {
+	for (num = 3; num < min_t(int, dwc->num_eps, DWC3_ENDPOINTS_NUM); num += 2) {
 		dep = dwc->eps[num];
+		if (!dep)
+			continue;
+
 		/* Don't change TXFRAMNUM on usb31 version */
 		size = DWC3_IP_IS(DWC3) ? 0 :
 			dwc3_readl(dwc->regs, DWC3_GTXFIFOSIZ(num >> 1)) &
@@ -3525,6 +3533,8 @@ static bool dwc3_gadget_endpoint_trbs_complete(struct dwc3_ep *dep,
 
 		for (i = 0; i < DWC3_ENDPOINTS_NUM; i++) {
 			dep = dwc->eps[i];
+			if (!dep)
+				continue;
 
 			if (!(dep->flags & DWC3_EP_ENABLED))
 				continue;
@@ -3713,6 +3723,10 @@ static void dwc3_endpoint_interrupt(struct dwc3 *dwc,
 	u8			epnum = event->endpoint_number;
 
 	dep = dwc->eps[epnum];
+	if (!dep) {
+		dev_warn(dwc->dev, "spurious event, endpoint %u is not allocated\n", epnum);
+		return;
+	}
 
 	if (!(dep->flags & DWC3_EP_ENABLED)) {
 		if ((epnum > 1) && !(dep->flags & DWC3_EP_TRANSFER_STARTED))
Signature · 1 line
2.39.5
author maintainer
#3 lore raw

PATCH 4/13 usb: dwc3: gadget: Avoid using reserved endpoints on Intel Merrifield

In reply to Sasha Levin

From: Andy Shevchenko <[email protected]>

[ Upstream commit 461f24bff86808ee5fbfe74751a825f8a7ab24e0 ]

Intel Merrifield SoC uses these endpoints for tracing and they cannot be re-allocated if being used because the side band flow control signals are hard wired to certain endpoints:

• 1 High BW Bulk IN (IN#1) (RTIT)
• 1 1KB BW Bulk IN (IN#8) + 1 1KB BW Bulk OUT (Run Control) (OUT#8)

In device mode, since RTIT (EP#1) and EXI/RunControl (EP#8) uses External Buffer Control (EBC) mode, these endpoints are to be mapped to EBC mode (to be done by EXI target driver). Additionally TRB for RTIT and EXI are maintained in STM (System Trace Module) unit and the EXI target driver will as well configure the TRB location for EP #1 IN and EP#8 (IN and OUT). Since STM/PTI and EXI hardware blocks manage these endpoints and interface to OTG3 controller through EBC interface, there is no need to enable any events (such as XferComplete etc) for these end points.

---
 drivers/usb/dwc3/dwc3-pci.c | 10 ++++++++++
 1 file changed, 10 insertions(+)
drivers/usb/dwc3/dwc3-pci.c +10 -0
index 6110ab1f91318..e2401cc4f1556 100644
@@ -143,11 +143,21 @@ static const struct property_entry dwc3_pci_intel_byt_properties[] = {
 	{}
 };
 
+/*
+ * Intel Merrifield SoC uses these endpoints for tracing and they cannot
+ * be re-allocated if being used because the side band flow control signals
+ * are hard wired to certain endpoints:
+ * - 1 High BW Bulk IN (IN#1) (RTIT)
+ * - 1 1KB BW Bulk IN (IN#8) + 1 1KB BW Bulk OUT (Run Control) (OUT#8)
+ */
+static const u8 dwc3_pci_mrfld_reserved_endpoints[] = { 3, 16, 17 };
+
 static const struct property_entry dwc3_pci_mrfld_properties[] = {
 	PROPERTY_ENTRY_STRING("dr_mode", "otg"),
 	PROPERTY_ENTRY_STRING("linux,extcon-name", "mrfld_bcove_pwrsrc"),
 	PROPERTY_ENTRY_BOOL("snps,dis_u3_susphy_quirk"),
 	PROPERTY_ENTRY_BOOL("snps,dis_u2_susphy_quirk"),
+	PROPERTY_ENTRY_U8_ARRAY("snps,reserved-endpoints", dwc3_pci_mrfld_reserved_endpoints),
 	PROPERTY_ENTRY_BOOL("snps,usb2-gadget-lpm-disable"),
 	PROPERTY_ENTRY_BOOL("linux,sysdev_is_parent"),
 	{}
Signature · 1 line
2.39.5
author maintainer
#4 lore raw

PATCH 5/13 sound/virtio: Fix cancel_sync warnings on uninitialized work_structs

In reply to Sasha Levin

From: John Stultz <[email protected]>

[ Upstream commit 3c7df2e27346eb40a0e86230db1ccab195c97cfe ]

Betty reported hitting the following warning:

[    8.709131][  T221] WARNING: CPU: 2 PID: 221 at kernel/workqueue.c:4182
...
[    8.713282][  T221] Call trace:
[    8.713365][  T221]  __flush_work+0x8d0/0x914
[    8.713468][  T221]  __cancel_work_sync+0xac/0xfc
[    8.713570][  T221]  cancel_work_sync+0x24/0x34
[    8.713667][  T221]  virtsnd_remove+0xa8/0xf8 [virtio_snd ab15f34d0dd772f6d11327e08a81d46dc9c36276]
[    8.713868][  T221]  virtsnd_probe+0x48c/0x664 [virtio_snd ab15f34d0dd772f6d11327e08a81d46dc9c36276]
[    8.714035][  T221]  virtio_dev_probe+0x28c/0x390
[    8.714139][  T221]  really_probe+0x1bc/0x4c8

...

It seems we're hitting the error path in virtsnd_probe(), which triggers a virtsnd_remove() which iterates over the substreams calling cancel_work_sync() on the elapsed_period work_struct.

Looking at the code, from earlier in:
virtsnd_probe()->virtsnd_build_devs()->virtsnd_pcm_parse_cfg()

We set snd->nsubstreams, allocate the snd->substreams, and if we then hit an error on the info allocation or something in virtsnd_ctl_query_info() fails, we will exit without having initialized the elapsed_period work_struct.

When that error path unwinds we then call virtsnd_remove() which as long as the substreams array is allocated, will iterate through calling cancel_work_sync() on the uninitialized work struct hitting this warning.

Takashi Iwai suggested this fix, which initializes the substreams structure right after allocation, so that if we hit the error paths we avoid trying to cleanup uninitialized data.

Note: I have not yet managed to reproduce the issue myself, so this patch has had limited testing.

Feedback or thoughts would be appreciated!

Message-Id: <[email protected]>

---
 sound/virtio/virtio_pcm.c | 21 +++++++++++++++------
 1 file changed, 15 insertions(+), 6 deletions(-)
sound/virtio/virtio_pcm.c +15 -6
index c10d91fff2fb0..1ddec1f4f05d5 100644
@@ -337,6 +337,21 @@ int virtsnd_pcm_parse_cfg(struct virtio_snd *snd)
 	if (!snd->substreams)
 		return -ENOMEM;
 
+	/*
+	 * Initialize critical substream fields early in case we hit an
+	 * error path and end up trying to clean up uninitialized structures
+	 * elsewhere.
+	 */
+	for (i = 0; i < snd->nsubstreams; ++i) {
+		struct virtio_pcm_substream *vss = &snd->substreams[i];
+
+		vss->snd = snd;
+		vss->sid = i;
+		INIT_WORK(&vss->elapsed_period, virtsnd_pcm_period_elapsed);
+		init_waitqueue_head(&vss->msg_empty);
+		spin_lock_init(&vss->lock);
+	}
+
 	info = kcalloc(snd->nsubstreams, sizeof(*info), GFP_KERNEL);
 	if (!info)
 		return -ENOMEM;
@@ -350,12 +365,6 @@ int virtsnd_pcm_parse_cfg(struct virtio_snd *snd)
 		struct virtio_pcm_substream *vss = &snd->substreams[i];
 		struct virtio_pcm *vpcm;
 
-		vss->snd = snd;
-		vss->sid = i;
-		INIT_WORK(&vss->elapsed_period, virtsnd_pcm_period_elapsed);
-		init_waitqueue_head(&vss->msg_empty);
-		spin_lock_init(&vss->lock);
-
 		rc = virtsnd_pcm_build_hw(vss, &info[i]);
 		if (rc)
 			goto on_exit;
Signature · 1 line
2.39.5
author maintainer
#5 lore raw

PATCH 6/13 dmaengine: dmatest: Fix dmatest waiting less when interrupted

In reply to Sasha Levin

From: Vinicius Costa Gomes <[email protected]>

[ Upstream commit e87ca16e99118ab4e130a41bdf12abbf6a87656c ]

Change the "wait for operation finish" logic to take interrupts into account.

When using dmatest with idxd DMA engine, it's possible that during longer tests, the interrupt notifying the finish of an operation happens during wait_event_freezable_timeout(), which causes dmatest to cleanup all the resources, some of which might still be in use.

This fix ensures that the wait logic correctly handles interrupts, preventing premature cleanup of resources.

---
 drivers/dma/dmatest.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
drivers/dma/dmatest.c +3 -3
index ffe621695e472..78b8a97b23637 100644
@@ -827,9 +827,9 @@ static int dmatest_func(void *data)
 		} else {
 			dma_async_issue_pending(chan);
 
-			wait_event_freezable_timeout(thread->done_wait,
-					done->done,
-					msecs_to_jiffies(params->timeout));
+			wait_event_timeout(thread->done_wait,
+					   done->done,
+					   msecs_to_jiffies(params->timeout));
 
 			status = dma_async_is_tx_complete(chan, cookie, NULL,
 							  NULL);
Signature · 1 line
2.39.5
author maintainer
#6 lore raw

PATCH 7/13 usb: xhci: Avoid Stop Endpoint retry loop if the endpoint seems Running

In reply to Sasha Levin

From: Michal Pecio <[email protected]>

[ Upstream commit 28a76fcc4c85dd39633fb96edb643c91820133e3 ]

Nothing prevents a broken HC from claiming that an endpoint is Running and repeatedly rejecting Stop Endpoint with Context State Error.

Avoid infinite retries and give back cancelled TDs.

No such cases known so far, but HCs have bugs.

---
 drivers/usb/host/xhci-ring.c | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)
drivers/usb/host/xhci-ring.c +7 -4
index 85a2b3ca05075..0862fdd3e5682 100644
@@ -1175,16 +1175,19 @@ static void xhci_handle_cmd_stop_ep(struct xhci_hcd *xhci, int slot_id,
 			 * Stopped state, but it will soon change to Running.
 			 *
 			 * Assume this bug on unexpected Stop Endpoint failures.
-			 * Keep retrying until the EP starts and stops again, on
-			 * chips where this is known to help. Wait for 100ms.
+			 * Keep retrying until the EP starts and stops again.
 			 */
-			if (time_is_before_jiffies(ep->stop_time + msecs_to_jiffies(100)))
-				break;
 			fallthrough;
 		case EP_STATE_RUNNING:
 			/* Race, HW handled stop ep cmd before ep was running */
 			xhci_dbg(xhci, "Stop ep completion ctx error, ctx_state %d\n",
 					GET_EP_CTX_STATE(ep_ctx));
+			/*
+			 * Don't retry forever if we guessed wrong or a defective HC never starts
+			 * the EP or says 'Running' but fails the command. We must give back TDs.
+			 */
+			if (time_is_before_jiffies(ep->stop_time + msecs_to_jiffies(100)))
+				break;
 
 			command = xhci_alloc_command(xhci, false, GFP_ATOMIC);
 			if (!command) {
Signature · 1 line
2.39.5
author maintainer
#7 lore raw

PATCH 8/13 usb: gadget: aspeed: Add NULL pointer check in ast_vhub_init_dev()

In reply to Sasha Levin

From: Chenyuan Yang <[email protected]>

[ Upstream commit 8c75f3e6a433d92084ad4e78b029ae680865420f ]

The variable d->name, returned by devm_kasprintf(), could be NULL. A pointer check is added to prevent potential NULL pointer dereference. This is similar to the fix in commit 3027e7b15b02
("ice: Fix some null pointer dereference issues in ice_ptp.c").

This issue is found by our static analysis tool

---
 drivers/usb/gadget/udc/aspeed-vhub/dev.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/usb/gadget/udc/aspeed-vhub/dev.c +3 -0
index 4f3bc27c1c628..73664a123c7a0 100644
@@ -549,6 +549,9 @@ int ast_vhub_init_dev(struct ast_vhub *vhub, unsigned int idx)
 	d->vhub = vhub;
 	d->index = idx;
 	d->name = devm_kasprintf(parent, GFP_KERNEL, "port%d", idx+1);
+	if (!d->name)
+		return -ENOMEM;
+
 	d->regs = vhub->regs + 0x100 + 0x10 * idx;
 
 	ast_vhub_init_ep0(vhub, &d->ep0, d);
Signature · 1 line
2.39.5
author maintainer
#8 lore raw

PATCH 9/13 usb: host: xhci-plat: mvebu: use ->quirks instead of ->init_quirk() func

In reply to Sasha Levin

From: Théo Lebrun <[email protected]>

[ Upstream commit 64eb182d5f7a5ec30227bce4f6922ff663432f44 ]

Compatible "marvell,armada3700-xhci" match data uses the struct xhci_plat_priv::init_quirk() function pointer to add XHCI_RESET_ON_RESUME as quirk on XHCI.

Instead, use the struct xhci_plat_priv::quirks field.

---
 drivers/usb/host/xhci-mvebu.c | 10 ----------
 drivers/usb/host/xhci-mvebu.h |  6 ------
 drivers/usb/host/xhci-plat.c  |  2 +-
 3 files changed, 1 insertion(+), 17 deletions(-)
drivers/usb/host/xhci-mvebu.c +0 -10
index 87f1597a0e5ab..257e4d79971fd 100644
@@ -73,13 +73,3 @@ int xhci_mvebu_mbus_init_quirk(struct usb_hcd *hcd)
 
 	return 0;
 }
-
-int xhci_mvebu_a3700_init_quirk(struct usb_hcd *hcd)
-{
-	struct xhci_hcd	*xhci = hcd_to_xhci(hcd);
-
-	/* Without reset on resume, the HC won't work at all */
-	xhci->quirks |= XHCI_RESET_ON_RESUME;
-
-	return 0;
-}
drivers/usb/host/xhci-mvebu.h +0 -6
index 3be021793cc8b..9d26e22c48422 100644
@@ -12,16 +12,10 @@ struct usb_hcd;
 
 #if IS_ENABLED(CONFIG_USB_XHCI_MVEBU)
 int xhci_mvebu_mbus_init_quirk(struct usb_hcd *hcd);
-int xhci_mvebu_a3700_init_quirk(struct usb_hcd *hcd);
 #else
 static inline int xhci_mvebu_mbus_init_quirk(struct usb_hcd *hcd)
 {
 	return 0;
 }
-
-static inline int xhci_mvebu_a3700_init_quirk(struct usb_hcd *hcd)
-{
-	return 0;
-}
 #endif
 #endif /* __LINUX_XHCI_MVEBU_H */
drivers/usb/host/xhci-plat.c +1 -1
index b387d39bfb81d..59f1cb68e6579 100644
@@ -111,7 +111,7 @@ static const struct xhci_plat_priv xhci_plat_marvell_armada = {
 };
 
 static const struct xhci_plat_priv xhci_plat_marvell_armada3700 = {
-	.init_quirk = xhci_mvebu_a3700_init_quirk,
+	.quirks = XHCI_RESET_ON_RESUME,
 };
 
 static const struct xhci_plat_priv xhci_plat_renesas_rcar_gen2 = {
Signature · 1 line
2.39.5
author maintainer
#9 lore raw

PATCH 10/13 thunderbolt: Scan retimers after device router has been enumerated

In reply to Sasha Levin

From: Mika Westerberg <[email protected]>

[ Upstream commit 75749d2c1d8cef439f8b69fa1f4f36d0fc3193e6 ]

Thomas reported connection issues on AMD system with Pluggable UD-4VPD dock. After some experiments it looks like the device has some sort of internal timeout that triggers reconnect. This is completely against the USB4 spec, as there is no requirement for the host to enumerate the device right away or even at all.

In Linux case the delay is caused by scanning of retimers on the link so we can work this around by doing the scanning after the device router has been enumerated.

---
 drivers/thunderbolt/tb.c | 16 ++++++++++++++--
 1 file changed, 14 insertions(+), 2 deletions(-)
drivers/thunderbolt/tb.c +14 -2
index c592032657a1e..0668e1645bc50 100644
@@ -640,11 +640,15 @@ static void tb_scan_port(struct tb_port *port)
 		goto out_rpm_put;
 	}
 
-	tb_retimer_scan(port, true);
-
 	sw = tb_switch_alloc(port->sw->tb, &port->sw->dev,
 			     tb_downstream_route(port));
 	if (IS_ERR(sw)) {
+		/*
+		 * Make the downstream retimers available even if there
+		 * is no router connected.
+		 */
+		tb_retimer_scan(port, true);
+
 		/*
 		 * If there is an error accessing the connected switch
 		 * it may be connected to another domain. Also we allow
@@ -704,6 +708,14 @@ static void tb_scan_port(struct tb_port *port)
 	tb_switch_lane_bonding_enable(sw);
 	/* Set the link configured */
 	tb_switch_configure_link(sw);
+	/*
+	 * Scan for downstream retimers. We only scan them after the
+	 * router has been enumerated to avoid issues with certain
+	 * Pluggable devices that expect the host to enumerate them
+	 * within certain timeout.
+	 */
+	tb_retimer_scan(port, true);
+
 	/*
 	 * CL0s and CL1 are enabled and supported together.
 	 * Silently ignore CLx enabling in case CLx is not supported.
Signature · 1 line
2.39.5
author maintainer
#10 lore raw

PATCH 11/13 objtool: Silence more KCOV warnings

In reply to Sasha Levin

From: Josh Poimboeuf <[email protected]>

[ Upstream commit 6b023c7842048c4bbeede802f3cf36b96c7a8b25 ]

In the past there were issues with KCOV triggering unreachable instruction warnings, which is why unreachable warnings are now disabled with CONFIG_KCOV.

Now some new KCOV warnings are showing up with GCC 14:

vmlinux.o: warning: objtool: cpuset_write_resmask() falls through to next function cpuset_update_active_cpus.cold() drivers/usb/core/driver.o: error: objtool: usb_deregister() falls through to next function usb_match_device() sound/soc/codecs/snd-soc-wcd934x.o: warning: objtool: .text.wcd934x_slim_irq_handler: unexpected end of section

All are caused by GCC KCOV not finishing an optimization, leaving behind a never-taken conditional branch to a basic block which falls through to the next function (or end of section).

At a high level this is similar to the unreachable warnings mentioned above, in that KCOV isn't fully removing dead code. Treat it the same way by adding these to the list of warnings to ignore with CONFIG_KCOV.

---
 tools/objtool/check.c | 6 ++++++
 1 file changed, 6 insertions(+)
tools/objtool/check.c +6 -0
index 6ea78612635ba..b6c91bb5ce3e3 100644
@@ -3368,6 +3368,9 @@ static int validate_branch(struct objtool_file *file, struct symbol *func,
 			if (!strncmp(func->name, "__cfi_", 6))
 				return 0;
 
+			if (file->ignore_unreachables)
+				return 0;
+
 			WARN("%s() falls through to next function %s()",
 			     func->name, insn->func->name);
 			return 1;
@@ -3582,6 +3585,9 @@ static int validate_branch(struct objtool_file *file, struct symbol *func,
 		if (!next_insn) {
 			if (state.cfi.cfa.base == CFI_UNDEFINED)
 				return 0;
+			if (file->ignore_unreachables)
+				return 0;
+
 			WARN("%s: unexpected end of section", sec->name);
 			return 1;
 		}
Signature · 1 line
2.39.5
author maintainer
#11 lore raw

PATCH 12/13 objtool, ASoC: codecs: wcd934x: Remove potential undefined behavior in wcd934x_slim_irq_handler()

In reply to Sasha Levin

From: Josh Poimboeuf <[email protected]>

[ Upstream commit 060aed9c0093b341480770457093449771cf1496 ]

If 'port_id' is negative, the shift counts in wcd934x_slim_irq_handler() also become negative, resulting in undefined behavior due to shift out of bounds.

If I'm reading the code correctly, that appears to be not possible, but with KCOV enabled, Clang's range analysis isn't always able to determine that and generates undefined behavior.

As a result the code generation isn't optimal, and undefined behavior should be avoided regardless. Improve code generation and remove the undefined behavior by converting the signed variables to unsigned.

Fixes the following warning with UBSAN:

sound/soc/codecs/snd-soc-wcd934x.o: warning: objtool: .text.wcd934x_slim_irq_handler: unexpected end of section

---
 sound/soc/codecs/wcd934x.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
sound/soc/codecs/wcd934x.c +1 -1
index 0b5999c819db9..04c50f9acda18 100644
@@ -2281,7 +2281,7 @@ static irqreturn_t wcd934x_slim_irq_handler(int irq, void *data)
 {
 	struct wcd934x_codec *wcd = data;
 	unsigned long status = 0;
-	int i, j, port_id;
+	unsigned int i, j, port_id;
 	unsigned int val, int_val = 0;
 	irqreturn_t ret = IRQ_NONE;
 	bool tx;
Signature · 1 line
2.39.5
author maintainer
#12 lore raw

PATCH 13/13 objtool, lkdtm: Obfuscate the do_nothing() pointer

In reply to Sasha Levin

From: Josh Poimboeuf <[email protected]>

[ Upstream commit 05026ea01e95ffdeb0e5ac8fb7fb1b551e3a8726 ]

If execute_location()'s memcpy of do_nothing() gets inlined and unrolled by the compiler, it copies one word at a time:

    mov    0x0(%rip),%rax    R_X86_64_PC32    .text+0x1374
    mov    %rax,0x38(%rbx)
    mov    0x0(%rip),%rax    R_X86_64_PC32    .text+0x136c
    mov    %rax,0x30(%rbx)
    ...

Those .text references point to the middle of the function, causing objtool to complain about their lack of ENDBR.

Prevent that by resolving the function pointer at runtime rather than build time. This fixes the following warning:

drivers/misc/lkdtm/lkdtm.o: warning: objtool: execute_location+0x23: relocation to !ENDBR: .text+0x1378

---
 drivers/misc/lkdtm/perms.c | 14 +++++++++++---
 1 file changed, 11 insertions(+), 3 deletions(-)
drivers/misc/lkdtm/perms.c +11 -3
index b93404d656509..e82d9543a0c5f 100644
@@ -28,6 +28,13 @@ static const unsigned long rodata = 0xAA55AA55;
 /* This is marked __ro_after_init, so it should ultimately be .rodata. */
 static unsigned long ro_after_init __ro_after_init = 0x55AA5500;
 
+/*
+ * This is a pointer to do_nothing() which is initialized at runtime rather
+ * than build time to avoid objtool IBT validation warnings caused by an
+ * inlined unrolled memcpy() in execute_location().
+ */
+static void __ro_after_init *do_nothing_ptr;
+
 /*
  * This just returns to the caller. It is designed to be copied into
  * non-executable memory regions.
@@ -65,13 +72,12 @@ static noinline void execute_location(void *dst, bool write)
 {
 	void (*func)(void);
 	func_desc_t fdesc;
-	void *do_nothing_text = dereference_function_descriptor(do_nothing);
 
-	pr_info("attempting ok execution at %px\n", do_nothing_text);
+	pr_info("attempting ok execution at %px\n", do_nothing_ptr);
 	do_nothing();
 
 	if (write == CODE_WRITE) {
-		memcpy(dst, do_nothing_text, EXEC_SIZE);
+		memcpy(dst, do_nothing_ptr, EXEC_SIZE);
 		flush_icache_range((unsigned long)dst,
 				   (unsigned long)dst + EXEC_SIZE);
 	}
@@ -267,6 +273,8 @@ static void lkdtm_ACCESS_NULL(void)
 
 void __init lkdtm_perms_init(void)
 {
+	do_nothing_ptr = dereference_function_descriptor(do_nothing);
+
 	/* Make sure we can write to __ro_after_init values during __init */
 	ro_after_init |= 0xAA;
 }
Signature · 1 line
2.39.5