6.18.39-rc1 review

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Alexey Nepomnyashih <[email protected]>

commit 03866d130ed33ab68cc7faaf4bf2c4abef96d42e upstream.

The dqp->q_id == 0 check inside the XFS_DQTYPE_BIGTIME block is unreachable because root dquots return successfully earlier. Reject root dquots with XFS_DQTYPE_BIGTIME before that early return, preserving the intended validation and removing the unreachable condition.

Found by Linux Verification Center (linuxtesting.org) with SVACE.

---
 fs/xfs/xfs_dquot.c |   16 ++++++++--------
 1 file changed, 8 insertions(+), 8 deletions(-)

@@ -1241,6 +1241,14 @@ xfs_qm_dqflush_check(
 	    type != XFS_DQTYPE_PROJ)
 		return __this_address;
 
+	/* bigtime flag should never be set on root dquots */
+	if (dqp->q_type & XFS_DQTYPE_BIGTIME) {
+		if (!xfs_has_bigtime(dqp->q_mount))
+			return __this_address;
+		if (dqp->q_id == 0)
+			return __this_address;
+	}
+
 	if (dqp->q_id == 0)
 		return NULL;
 

@@ -1256,14 +1264,6 @@ xfs_qm_dqflush_check(
 	    !dqp->q_rtb.timer)
 		return __this_address;
 
-	/* bigtime flag should never be set on root dquots */
-	if (dqp->q_type & XFS_DQTYPE_BIGTIME) {
-		if (!xfs_has_bigtime(dqp->q_mount))
-			return __this_address;
-		if (dqp->q_id == 0)
-			return __this_address;
-	}
-
 	return NULL;
 }
 

6.18-stable review patch. If anyone has any objections, please let me know.

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

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

commit 84eec3f7fc73144d1a230c9e8ad92721e37dcaab upstream.

The translation of the old XFS_BMBT_KEY_ADDR macro into a static function is not correct on 32-bit systems because the sizeof() argument went from being a xfs_bmbt_key_t (i.e. a struct) to a (struct xfs_bmbt_key *) (i.e. a pointer to the same struct). On 64-bit systems this turns out ok because they are the same size, but on 32-bit systems this is catastrophic because they are not the same size. So far there have been no complaints, most likely because the xfs developers urge against running it on 32-bit systems. But this needs fixing asap.

---
 fs/xfs/libxfs/xfs_bmap_btree.h |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -89,7 +89,7 @@ xfs_bmbt_key_addr(
 {
 	return (struct xfs_bmbt_key *)
 		((char *)block + xfs_bmbt_block_len(mp) +
-		 (index - 1) * sizeof(struct xfs_bmbt_key *));
+		 (index - 1) * sizeof(struct xfs_bmbt_key));
 }
 
 static inline xfs_bmbt_ptr_t *

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Yingjie Gao <[email protected]>

commit 0a5213bbff62b51c7d4999ac8c7e11ea57d00d45 upstream.

xfs_exchmaps_estimate_overhead() adds the bmbt and rmapbt overhead to a local resblks variable, but the final UINT_MAX check still tests req->resblks. That is the reservation value from before the overhead was added.

The computed value is stored back in req->resblks and later passed to xfs_trans_alloc(), whose block reservation argument is unsigned int. Check the computed reservation so the existing limit applies to the value that will be used.

---
 fs/xfs/libxfs/xfs_exchmaps.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -711,7 +711,7 @@ xfs_exchmaps_estimate_overhead(
 		return -ENOSPC;
 
 	/* Can't actually reserve more than UINT_MAX blocks. */
-	if (req->resblks > UINT_MAX)
+	if (resblks > UINT_MAX)
 		return -ENOSPC;
 
 	req->resblks = resblks;

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Dawei Feng <[email protected]>

commit 45de375b25060edf46e20abb36521ba530336ceb upstream.

If xfs_metadir_create() fails in xfs_dqinode_metadir_create(), the current code returns directly, leaking the allocated update and transaction state. If the subsequent commit fails, the caller-owned inode reference is left behind.

Fix this memory leak by routing the create failure path through xfs_metadir_cancel(). For both create and commit failures, finish and release any inode returned to the caller, mirroring the unwind pattern in xfs_metadir_mkdir().

The bug was first flagged by an experimental analysis tool we are developing for kernel memory-management bugs while analyzing v6.13-rc1. The tool is still under development and is not yet publicly available. Manual inspection confirms that the bug is still present in v7.1.1.

An x86_64 allyesconfig build showed no new warnings. Runtime validation used kprobe fault injection during mount -o uquota on a metadir XFS image. Injecting xfs_metadir_create() reproduced the old active-update path that left mount stuck later in mount setup; after this change, the same injection reported cancel_hits=1 and irele_hits=1. Injecting xfs_metadir_commit() exercised the old inode-reference leak path; after this change, it reported irele_hits=1.

---
 fs/xfs/libxfs/xfs_dquot_buf.c |   14 ++++++++++++--
 1 file changed, 12 insertions(+), 2 deletions(-)

@@ -436,17 +436,27 @@ xfs_dqinode_metadir_create(
 
 	error = xfs_metadir_create(&upd, S_IFREG);
 	if (error)
-		return error;
+		goto out_cancel;
 
 	xfs_trans_log_inode(upd.tp, upd.ip, XFS_ILOG_CORE);
 
 	error = xfs_metadir_commit(&upd);
 	if (error)
-		return error;
+		goto out_irele;
 
 	xfs_finish_inode_setup(upd.ip);
 	*ipp = upd.ip;
 	return 0;
+
+out_cancel:
+	xfs_metadir_cancel(&upd, error);
+out_irele:
+	/* Have to finish setting up the inode to ensure it's deleted. */
+	if (upd.ip) {
+		xfs_finish_inode_setup(upd.ip);
+		xfs_irele(upd.ip);
+	}
+	return error;
 }
 
 #ifndef __KERNEL__

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Zhao Zhang <[email protected]>

commit aa496720618f1a6054f1c870bf10b4f6c99bf656 upstream.

Devmap broadcast redirects clone the packet for all but the last destination.

For native XDP, that clone path copies only the linear xdp_frame data, while fragmented frames keep skb_shared_info in tailroom outside the linear area. Cloning such a frame leaves XDP_FLAGS_HAS_FRAGS set but without valid frag metadata, and the later free path can interpret uninitialized tail data as skb_shared_info, leading to an out-of-bounds access during frame return.

Reject fragmented native XDP frames in dev_map_enqueue_clone().

Add the same restriction to the generic XDP clone path in dev_map_redirect_clone(). Generic XDP represents fragmented packets as nonlinear skbs, and rejecting them here keeps clone-based broadcast support aligned between native and generic XDP.

---
 kernel/bpf/devmap.c |    7 +++++++
 1 file changed, 7 insertions(+)

@@ -581,6 +581,10 @@ static int dev_map_enqueue_clone(struct
 {
 	struct xdp_frame *nxdpf;
 
+	/* Frags live outside the linear frame and cannot be cloned safely. */
+	if (unlikely(xdp_frame_has_frags(xdpf)))
+		return -EOPNOTSUPP;
+
 	nxdpf = xdpf_clone(xdpf);
 	if (!nxdpf)
 		return -ENOMEM;

@@ -726,6 +730,9 @@ static int dev_map_redirect_clone(struct
 	struct sk_buff *nskb;
 	int err;
 
+	if (unlikely(skb_is_nonlinear(skb)))
+		return -EOPNOTSUPP;
+
 	nskb = skb_clone(skb, GFP_ATOMIC);
 	if (!nskb)
 		return -ENOMEM;

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Dawei Feng <[email protected]>

commit 2566c3b24219c5b30e35205cba029ff34ff7c78b upstream.

Commit 4e63acdff864 ("bpf: Introduce bpf_sysctl_{get,set}_new_value helpers") changed the success return value to 0, but failed to update the corresponding check in __cgroup_bpf_run_filter_sysctl(). Since bpf_prog_run_array_cg() now returns 0 on success, the legacy ret == 1 condition is never satisfied. As a result, the modified value is ignored, and bpf_sysctl_set_new_value() fails to replace the write buffer.

Fix this by checking for a return value of 0 instead, so cgroup/sysctl programs can correctly replace the pending sysctl buffer.

This bug was discovered during a manual code review. Tested via a cgroup/sysctl BPF reproducer overriding writes to a target sysctl. Pre-fix, bpf_sysctl_set_new_value("foo") was silently ignored: the write returned 8192 and the value remained "600". Post-fix, the BPF replacement buffer properly propagates: the write returns 3 and the value updates to "foo".

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

@@ -1939,7 +1939,7 @@ int __cgroup_bpf_run_filter_sysctl(struc
 
 	kfree(ctx.cur_val);
 
-	if (ret == 1 && ctx.new_updated) {
+	if (!ret && ctx.new_updated) {
 		kvfree(*buf);
 		*buf = ctx.new_val;
 		*pcount = ctx.new_len;

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Paul Moses <[email protected]>

commit b9452b594fd3aecbfd4aa0a6a1f741330a37dab7 upstream.

btf_parse_struct_metas() walks user-supplied BTF during BPF_BTF_LOAD, and btf_repeat_fields() expands repeatable fields from array elements into the fixed BTF_FIELDS_MAX scratch array used by btf_parse_fields().

The remaining-capacity check performs the expanded field count calculation in u32. A malformed BTF can wrap that calculation, causing the check to pass even when the expanded field count exceeds the scratch array capacity. The following memcpy() can then write past the end of the array.

Use checked addition and multiplication before copying repeated fields and reject impossible counts.

---
 kernel/bpf/btf.c |    9 ++++-----
 1 file changed, 4 insertions(+), 5 deletions(-)

@@ -3550,7 +3550,7 @@ end:
 static int btf_repeat_fields(struct btf_field_info *info, int info_cnt,
 			     u32 field_cnt, u32 repeat_cnt, u32 elem_size)
 {
-	u32 i, j;
+	u32 i, j, total_cnt, total_repeats;
 	u32 cur;
 
 	/* Ensure not repeating fields that should not be repeated. */

@@ -3568,10 +3568,9 @@ static int btf_repeat_fields(struct btf_
 		}
 	}
 
-	/* The type of struct size or variable size is u32,
-	 * so the multiplication will not overflow.
-	 */
-	if (field_cnt * (repeat_cnt + 1) > info_cnt)
+	if (check_add_overflow(repeat_cnt, 1, &total_repeats) ||
+	    check_mul_overflow(field_cnt, total_repeats, &total_cnt) ||
+	    total_cnt > (u32)info_cnt)
 		return -E2BIG;
 
 	cur = field_cnt;

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Maoyi Xie <[email protected]>

commit 47b6bcef6e679593d2e86e04ee72c46a4e2f7139 upstream.

kalmia_rx_fixup() computes usb_packet_length = skb->len - (2 * KALMIA_HEADER_LENGTH) as a u16, guarded only by a pre-loop check that skb->len is at least KALMIA_HEADER_LENGTH, which is 6. A device can deliver a short bulk-IN frame with skb->len in the 6 to 11 range, or leave a short trailing remainder on a later loop iteration. Either case underflows usb_packet_length to about 65530.

That bypasses the usb_packet_length < ether_packet_length truncation path. The device-supplied ether_packet_length, a le16 up to 65535 read from header_start[2], then drives a memcmp() and the following skb_trim() and skb_pull() past the end of the rx buffer. The rx buffer is hard_mtu * 10, which is 14000 bytes. That is an out of bounds read.

Require both the start and end framing headers to be present before subtracting them, on every loop iteration.

---
 drivers/net/usb/kalmia.c |    8 ++++++++
 1 file changed, 8 insertions(+)

@@ -276,6 +276,14 @@ kalmia_rx_fixup(struct usbnet *dev, stru
 				"Received header: %6phC. Package length: %i\n",
 				header_start, skb->len - KALMIA_HEADER_LENGTH);
 
+		/* both framing headers must be present before we subtract
+		 * them, otherwise usb_packet_length underflows and the
+		 * device-supplied ether_packet_length drives an out of bounds
+		 * access below
+		 */
+		if (skb->len < 2 * KALMIA_HEADER_LENGTH)
+			return 0;
+
 		/* subtract start header and end header */
 		usb_packet_length = skb->len - (2 * KALMIA_HEADER_LENGTH);
 		ether_packet_length = get_unaligned_le16(&header_start[2]);

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Jared Baldridge <[email protected]>

commit 6eba58568f6cc3ff8515a00b05e258d8cfb72b72 upstream.

Uniden BC125AT radio scanner has a USB interface which fails to work with the cdc_acm driver:
usb 1-1: new full-speed USB device number 2 using uhci_hcd cdc_acm 1-1:1.0: Zero length descriptor references cdc_acm 1-1:1.0: probe with driver cdc_acm failed with error -22
usbcore: registered new interface driver cdc_acm

Adding the NO_UNION_NORMAL quirk for the device fixes the issue:

usb 1-1: new full-speed USB device number 2 using uhci_hcd cdc_acm 1-1:1.0: ttyACM0: USB ACM device
usbcore: registered new interface driver cdc_acm

lsusb -v of the device:

Bus 001 Device 002: ID 1965:0017 Uniden Corporation BC125AT Negotiated speed: Full Speed (12Mbps)
Device Descriptor:
bLength 18
bDescriptorType 1
bcdUSB 2.00

   bDeviceClass            2 Communications
   bDeviceSubClass         0 [unknown]
   bDeviceProtocol         0
   bMaxPacketSize0        64
   idVendor           0x1965 Uniden Corporation
   idProduct          0x0017 BC125AT
   bcdDevice            0.01
   iManufacturer           1 Uniden America Corp.
   iProduct                2 BC125AT
   iSerial                 3 0001
   bNumConfigurations      1

Configuration Descriptor:
bLength 9
bDescriptorType 2
wTotalLength 0x0030
bNumInterfaces 2
bConfigurationValue 1
iConfiguration 0
bmAttributes 0x80
(Bus Powered)
MaxPower 500mA
Interface Descriptor:
bLength 9
bDescriptorType 4
bInterfaceNumber 0
bAlternateSetting 0
bNumEndpoints 1
bInterfaceClass 2 Communications
bInterfaceSubClass 2 Abstract (modem)
bInterfaceProtocol 0
iInterface 0
Endpoint Descriptor:
bLength 7
bDescriptorType 5
bEndpointAddress 0x87 EP 7 IN
bmAttributes 3
Transfer Type Interrupt
Synch Type None
Usage Type Data
wMaxPacketSize 0x0008 1x 8 bytes
bInterval 10
Interface Descriptor:
bLength 9
bDescriptorType 4
bInterfaceNumber 1
bAlternateSetting 0
bNumEndpoints 2
bInterfaceClass 10 CDC Data
bInterfaceSubClass 0 [unknown]
bInterfaceProtocol 0
iInterface 0
Endpoint Descriptor:
bLength 7
bDescriptorType 5
bEndpointAddress 0x81 EP 1 IN
bmAttributes 2
Transfer Type Bulk
Synch Type None
Usage Type Data
wMaxPacketSize 0x0040 1x 64 bytes
bInterval 0
Endpoint Descriptor:
bLength 7
bDescriptorType 5
bEndpointAddress 0x02 EP 2 OUT
bmAttributes 2
Transfer Type Bulk
Synch Type None
Usage Type Data
wMaxPacketSize 0x0040 1x 64 bytes
bInterval 0
Device Status: 0x0000
(Bus Powered)

---
 drivers/usb/class/cdc-acm.c |    3 +++
 1 file changed, 3 insertions(+)

@@ -1810,6 +1810,9 @@ static const struct usb_device_id acm_id
 	{ USB_DEVICE(0x1901, 0x0006), /* GE Healthcare Patient Monitor UI Controller */
 	.driver_info = DISABLE_ECHO, /* DISABLE ECHO in termios flag */
 	},
+	{ USB_DEVICE(0x1965, 0x0017), /* Uniden BC125AT */
+	.driver_info = NO_UNION_NORMAL, /* has no union descriptor */
+	},
 	{ USB_DEVICE(0x1965, 0x0018), /* Uniden UBC125XLT */
 	.driver_info = NO_UNION_NORMAL, /* has no union descriptor */
 	},

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Haoxiang Li <[email protected]>

commit 3348f444a4ce43dd5c2d1aa41634cb6eff33aa64 upstream.

cdnsp_alloc_stream_info() allocates stream_info->stream_ctx_array with cdnsp_alloc_stream_ctx(). If a later stream ring allocation or stream mapping update fails, the error path frees the allocated stream rings and stream_rings array, but leaves stream_ctx_array allocated.

Free the stream context array before falling through to the stream_rings cleanup path.

---
 drivers/usb/cdns3/cdnsp-mem.c |    2 ++
 1 file changed, 2 insertions(+)

@@ -632,6 +632,8 @@ cleanup_rings:
 		}
 	}
 
+	cdnsp_free_stream_ctx(pdev, pep);
+
 cleanup_stream_rings:
 	kfree(pep->stream_info.stream_rings);
 

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Rodrigo Lugathe da Conceição Alves <[email protected]>

commit bd728c3d9b1cc0bb0fda6a7055c5c8b55d7477b2 upstream.

The VIA Labs, Inc. USB 2.0 hub controller (2109:2817), found in a KVM switch, fails to enumerate high-power devices during cold boot and system restart.

Applying the kernel parameter
usbcore.quirks=2109:2817:k
resolves the issue.

Enumeration failure log:
usb 1-1.2.3: device descriptor read/64, error -32 usb 1-1.2.3: Device not responding to setup address. usb 1-1.2.3: device not accepting address 11, error -71 usb 1-1.2-port3: unable to enumerate USB device

Add USB_QUIRK_NO_LPM for this device.

---
 drivers/usb/core/quirks.c |    3 +++
 1 file changed, 3 insertions(+)

@@ -577,6 +577,9 @@ static const struct usb_device_id usb_qu
 	/* VLI disk */
 	{ USB_DEVICE(0x2109, 0x0711), .driver_info = USB_QUIRK_NO_LPM },
 
+	/* VIA Labs, Inc. USB2.0 Hub */
+	{ USB_DEVICE(0x2109, 0x2817), .driver_info = USB_QUIRK_NO_LPM },
+
 	/* Raydium Touchscreen */
 	{ USB_DEVICE(0x2386, 0x3114), .driver_info = USB_QUIRK_NO_LPM },
 

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: WenTao Liang <[email protected]>

commit 692c354bef03b77b30e57e61934da502c8a12d45 upstream.

If dwc3_meson_g12a_resume() succeeds in calling reset_control_reset(), an internal triggered_count reference is acquired. If any later step fails (usb_init, phy_init, phy_power_on, regulator_enable, or usb_post_init), the function returns the error without rearming the reset control. This leaks the reference and leaves the reset control in a triggered state, causing future reset_control_reset() calls to incorrectly return early as if already reset.

Add an error path that calls reset_control_rearm() to balance the reference before returning the error.

---
 drivers/usb/dwc3/dwc3-meson-g12a.c |   14 +++++++++-----
 1 file changed, 9 insertions(+), 5 deletions(-)

@@ -907,35 +907,39 @@ static int __maybe_unused dwc3_meson_g12
 
 	ret = priv->drvdata->usb_init(priv);
 	if (ret)
-		return ret;
+		goto err_rearm;
 
 	/* Init PHYs */
 	for (i = 0 ; i < PHY_COUNT ; ++i) {
 		ret = phy_init(priv->phys[i]);
 		if (ret)
-			return ret;
+			goto err_rearm;
 	}
 
 	/* Set PHY Power */
 	for (i = 0 ; i < PHY_COUNT ; ++i) {
 		ret = phy_power_on(priv->phys[i]);
 		if (ret)
-			return ret;
+			goto err_rearm;
 	}
 
 	if (priv->vbus && priv->otg_phy_mode == PHY_MODE_USB_HOST) {
 		ret = regulator_enable(priv->vbus);
 		if (ret)
-			return ret;
+			goto err_rearm;
 	}
 
 	if (priv->drvdata->usb_post_init) {
 		ret = priv->drvdata->usb_post_init(priv);
 		if (ret)
-			return ret;
+			goto err_rearm;
 	}
 
 	return 0;
+
+err_rearm:
+	reset_control_rearm(priv->reset);
+	return ret;
 }
 
 static const struct dev_pm_ops dwc3_meson_g12a_dev_pm_ops = {

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Dawei Feng <[email protected]>

commit b9399d25fbb34a05bbe76eeedd730f62ff2670e9 upstream.

EHCI and FOTG210 isochronous submits build an ehci_iso_sched before linking the URB to the endpoint queue, and keep the staged schedule in urb->hcpriv until iso_stream_schedule() and the link helpers consume it. If the controller is no longer accessible, or usb_hcd_link_urb_to_ep() fails, submit jumps to done_not_linked before that handoff happens and leaks the staged schedule still attached to urb->hcpriv.

Free the staged schedule from done_not_linked when submit fails before the URB is linked and clear urb->hcpriv after the free.

The bug was first flagged by an experimental analysis tool we are developing for kernel memory-management bugs while analyzing v6.13-rc1. The tool is still under development and is not yet publicly available. Manual inspection confirms that the bug is still present in v7.1.1.

An x86_64 allyesconfig build showed no new warnings. As we do not have an EHCI host controller with a USB isochronous device to test with, no runtime testing was able to be performed.

---
 drivers/usb/fotg210/fotg210-hcd.c |    6 ++++--
 drivers/usb/host/ehci-sched.c     |   11 +++++++++--
 2 files changed, 13 insertions(+), 4 deletions(-)

@@ -4267,8 +4267,6 @@ static int iso_stream_schedule(struct fo
 	return 0;
 
 fail:
-	iso_sched_free(stream, sched);
-	urb->hcpriv = NULL;
 	return status;
 }
 

@@ -4562,6 +4560,10 @@ static int itd_submit(struct fotg210_hcd
 	else
 		usb_hcd_unlink_urb_from_ep(fotg210_to_hcd(fotg210), urb);
 done_not_linked:
+	if (status < 0) {
+		iso_sched_free(stream, urb->hcpriv);
+		urb->hcpriv = NULL;
+	}
 	spin_unlock_irqrestore(&fotg210->lock, flags);
 done:
 	return status;

@@ -1624,6 +1624,7 @@ iso_stream_schedule(
 			status = 1;	/* and give it back immediately */
 			iso_sched_free(stream, sched);
 			sched = NULL;
+			urb->hcpriv = NULL;
 		}
 	}
 	urb->error_count = skip / period;

@@ -1654,8 +1655,6 @@ iso_stream_schedule(
 	return status;
 
  fail:
-	iso_sched_free(stream, sched);
-	urb->hcpriv = NULL;
 	return status;
 }
 

@@ -1967,6 +1966,10 @@ static int itd_submit(struct ehci_hcd *e
 		usb_hcd_unlink_urb_from_ep(ehci_to_hcd(ehci), urb);
 	}
  done_not_linked:
+	if (status < 0) {
+		iso_sched_free(stream, urb->hcpriv);
+		urb->hcpriv = NULL;
+	}
 	spin_unlock_irqrestore(&ehci->lock, flags);
  done:
 	return status;

@@ -2344,6 +2347,10 @@ static int sitd_submit(struct ehci_hcd *
 		usb_hcd_unlink_urb_from_ep(ehci_to_hcd(ehci), urb);
 	}
  done_not_linked:
+	if (status < 0) {
+		iso_sched_free(stream, urb->hcpriv);
+		urb->hcpriv = NULL;
+	}
 	spin_unlock_irqrestore(&ehci->lock, flags);
  done:
 	return status;

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Maoyi Xie <[email protected]>

commit f8f680609c2b3ab795ffcd6f21585b6dfc46d395 upstream.

The OTG branch of composite_setup() falls back to the first configuration when none is selected:

	if (cdev->config)
		config = cdev->config;
	else
		config = list_first_entry(&cdev->configs,
					  struct usb_configuration, list);
	if (!config)
		goto done;
	...
	memcpy(req->buf, config->descriptors[0], value);

list_first_entry() never returns NULL. On an empty list it returns container_of() of the list head. So the "if (!config)" check is dead.

When cdev->configs is empty, config points at the head inside struct usb_composite_dev. config->descriptors[0] reads whatever sits at that offset. The memcpy copies up to w_length bytes of it into the response buffer.

cdev->configs can be empty in two cases. One is a teardown race on gadget unbind with a control transfer in flight. The other is a driver that sets is_otg before it adds a config. A reproducer that holds cdev->configs empty triggers a KASAN fault in this branch.

Use list_first_entry_or_null() so the existing check does its job.

---
 drivers/usb/gadget/composite.c |    5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

@@ -1863,9 +1863,10 @@ composite_setup(struct usb_gadget *gadge
 				if (cdev->config)
 					config = cdev->config;
 				else
-					config = list_first_entry(
+					config = list_first_entry_or_null(
 							&cdev->configs,
-						struct usb_configuration, list);
+							struct usb_configuration,
+							list);
 				if (!config)
 					goto done;
 

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Jimmy Hu <[email protected]>

commit 67e511d2989eb1c8c588b599ce2fcc6bb8e6f7ea upstream.

The udc structure acts as the management structure for the gadget, but their lifecycles are decoupled. A race condition exists where usb_del_gadget() frees the udc memory (e.g., via mode-switch work) while gadget_match_driver() concurrently accesses the freed udc memory (e.g., via configfs), causing a Use-After-Free (UAF) that triggers a NULL pointer dereference when the freed memory is zeroed:

[39430.908615][ T1171] Unable to handle kernel NULL pointer dereference at virtual address 0000000000000000
[39430.911397][ T1171] pc : __pi_strcmp+0x20/0x140
[39430.911441][ T1171] lr : gadget_match_driver+0x34/0x60
...
[39430.911890][ T1171]  usb_gadget_register_driver_owner+0x50/0xf8
[39430.911910][ T1171]  gadget_dev_desc_UDC_store+0xf4/0x140
[39430.931308][ T1171]  configfs_write_iter+0xec/0x134

[39430.957058][ T1171] Workqueue: events_freezable __dwc3_set_mode
[39430.957287][ T1171]  dwc3_gadget_exit+0x34/0x8c
[39430.957304][ T1171]  __dwc3_set_mode+0xc0/0x664

Fix this by ensuring the udc structure remains allocated until the gadget is released. To achieve this, introduce a new usb_gadget_release() routine to the core. When the gadget is added, usb_add_gadget() stores the gadget's release routine in the udc structure and takes a reference to the udc. When the gadget is released, usb_gadget_release() drops the reference to the udc and then calls the gadget's release routine.

---
 drivers/usb/gadget/udc/core.c |   29 ++++++++++++++++++++++++++++-
 1 file changed, 28 insertions(+), 1 deletion(-)

@@ -31,8 +31,9 @@ static const struct bus_type gadget_bus_
 /**
  * struct usb_udc - describes one usb device controller
  * @driver: the gadget driver pointer. For use by the class code
- * @dev: the child device to the actual controller
  * @gadget: the gadget. For use by the class code
+ * @gadget_release: the gadget's release routine
+ * @dev: the child device to the actual controller
  * @list: for use by the udc class driver
  * @vbus: for udcs who care about vbus status, this value is real vbus status;
  * for udcs who do not care about vbus status, this value is always true

@@ -53,6 +54,7 @@ static const struct bus_type gadget_bus_
 struct usb_udc {
 	struct usb_gadget_driver	*driver;
 	struct usb_gadget		*gadget;
+	void				(*gadget_release)(struct device *dev);
 	struct device			dev;
 	struct list_head		list;
 	bool				vbus;

@@ -1356,6 +1358,17 @@ static void usb_udc_nop_release(struct d
 	dev_vdbg(dev, "%s\n", __func__);
 }
 
+static void usb_gadget_release(struct device *dev)
+{
+	struct usb_gadget *gadget = dev_to_usb_gadget(dev);
+	struct usb_udc *udc = gadget->udc;
+	/* Cache the gadget's release routine to prevent UAF */
+	void (*release)(struct device *dev) = udc->gadget_release;
+
+	put_device(&udc->dev);
+	release(dev);
+}
+
 /**
  * usb_initialize_gadget - initialize a gadget and its embedded struct device
  * @parent: the parent device to this udc. Usually the controller driver's

@@ -1412,6 +1425,14 @@ int usb_add_gadget(struct usb_gadget *ga
 	mutex_init(&udc->connect_lock);
 
 	udc->started = false;
+	/*
+	 * Align decoupled lifecycles: take a UDC reference to ensure it
+	 * remains allocated until the gadget is released, requiring an
+	 * override of the gadget's release routine to drop it.
+	 */
+	udc->gadget_release = gadget->dev.release;
+	gadget->dev.release = usb_gadget_release;
+	get_device(&udc->dev);
 
 	mutex_lock(&udc_lock);
 	list_add_tail(&udc->list, &udc_list);

@@ -1456,6 +1477,12 @@ int usb_add_gadget(struct usb_gadget *ga
 	mutex_lock(&udc_lock);
 	list_del(&udc->list);
 	mutex_unlock(&udc_lock);
+	/*
+	 * Revert the override and drop the UDC reference to prevent
+	 * leaking the UDC if the gadget was statically allocated.
+	 */
+	gadget->dev.release = udc->gadget_release;
+	put_device(&udc->dev);
 
  err_put_udc:
 	put_device(&udc->dev);

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Xu Rao <[email protected]>

commit 30adce93d5c4a5a1ec29d9249e3fdfcc391d406b upstream.

printer_open() returns -EBUSY when the character device is already open, but it increments dev->kref regardless of the return value. VFS does not call ->release() for a failed open, so every rejected second open permanently leaks one reference.

Move kref_get() into the successful-open branch.

---
 drivers/usb/gadget/function/f_printer.c |    3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

@@ -363,12 +363,11 @@ printer_open(struct inode *inode, struct
 		ret = 0;
 		/* Change the printer status to show that it's on-line. */
 		dev->printer_status |= PRINTER_SELECTED;
+		kref_get(&dev->kref);
 	}
 
 	spin_unlock_irqrestore(&dev->lock, flags);
 
-	kref_get(&dev->kref);
-
 	return ret;
 }
 

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Johan Hovold <[email protected]>

commit ff002c153f9722caece3983cc23dc4d9d4652cb4 upstream.

mutex_unlock() may access the mutex structure after releasing the lock and therefore cannot be used to manage lifetime of objects directly (unlike spinlocks and refcounts). [1][2]

Use a kref to release the driver data to avoid use-after-free in mutex_unlock() when release() races with disconnect().

[1] a51749ab34d9 ("locking/mutex: Document that mutex_unlock() is non-atomic")
[2] 2b9d9e0a9ba0 ("locking/mutex: Clarify that mutex_unlock(), and most other sleeping locks, can still use the lock object after it's unlocked")

---
 drivers/usb/misc/idmouse.c |   45 ++++++++++++++++++++++-----------------------
 1 file changed, 22 insertions(+), 23 deletions(-)

@@ -63,6 +63,7 @@ MODULE_DEVICE_TABLE(usb, idmouse_table);
 
 /* structure to hold all of our device specific stuff */
 struct usb_idmouse {
+	struct kref kref;
 
 	struct usb_device *udev; /* save off the usb device pointer */
 	struct usb_interface *interface; /* the interface for this device */

@@ -209,8 +210,10 @@ static int idmouse_resume(struct usb_int
 	return 0;
 }
 
-static inline void idmouse_delete(struct usb_idmouse *dev)
+static inline void idmouse_delete(struct kref *kref)
 {
+	struct usb_idmouse *dev = container_of(kref, struct usb_idmouse, kref);
+
 	kfree(dev->bulk_in_buffer);
 	kfree(dev);
 }

@@ -254,6 +257,8 @@ static int idmouse_open(struct inode *in
 		/* increment our usage count for the driver */
 		++dev->open;
 
+		kref_get(&dev->kref);
+
 		/* save our object in the file's private structure */
 		file->private_data = dev;
 

@@ -277,16 +282,11 @@ static int idmouse_release(struct inode
 
 	/* lock our device */
 	mutex_lock(&dev->lock);
-
 	--dev->open;
+	mutex_unlock(&dev->lock);
+
+	kref_put(&dev->kref, idmouse_delete);
 
-	if (!dev->present) {
-		/* the device was unplugged before the file was released */
-		mutex_unlock(&dev->lock);
-		idmouse_delete(dev);
-	} else {
-		mutex_unlock(&dev->lock);
-	}
 	return 0;
 }
 

@@ -334,6 +334,7 @@ static int idmouse_probe(struct usb_inte
 	if (dev == NULL)
 		return -ENOMEM;
 
+	kref_init(&dev->kref);
 	mutex_init(&dev->lock);
 	dev->udev = udev;
 	dev->interface = interface;

@@ -342,8 +343,7 @@ static int idmouse_probe(struct usb_inte
 	result = usb_find_bulk_in_endpoint(iface_desc, &endpoint);
 	if (result) {
 		dev_err(&interface->dev, "Unable to find bulk-in endpoint.\n");
-		idmouse_delete(dev);
-		return result;
+		goto err_put_kref;
 	}
 
 	dev->orig_bi_size = usb_endpoint_maxp(endpoint);

@@ -351,8 +351,8 @@ static int idmouse_probe(struct usb_inte
 	dev->bulk_in_endpointAddr = endpoint->bEndpointAddress;
 	dev->bulk_in_buffer = kmalloc(IMGSIZE + dev->bulk_in_size, GFP_KERNEL);
 	if (!dev->bulk_in_buffer) {
-		idmouse_delete(dev);
-		return -ENOMEM;
+		result = -ENOMEM;
+		goto err_put_kref;
 	}
 
 	/* allow device read, write and ioctl */

@@ -364,14 +364,18 @@ static int idmouse_probe(struct usb_inte
 	if (result) {
 		/* something prevented us from registering this device */
 		dev_err(&interface->dev, "Unable to allocate minor number.\n");
-		idmouse_delete(dev);
-		return result;
+		goto err_put_kref;
 	}
 
 	/* be noisy */
 	dev_info(&interface->dev,"%s now attached\n",DRIVER_DESC);
 
 	return 0;
+
+err_put_kref:
+	kref_put(&dev->kref, idmouse_delete);
+
+	return result;
 }
 
 static void idmouse_disconnect(struct usb_interface *interface)

@@ -387,14 +391,9 @@ static void idmouse_disconnect(struct us
 	/* prevent device read, write and ioctl */
 	dev->present = 0;
 
-	/* if the device is opened, idmouse_release will clean this up */
-	if (!dev->open) {
-		mutex_unlock(&dev->lock);
-		idmouse_delete(dev);
-	} else {
-		/* unlock */
-		mutex_unlock(&dev->lock);
-	}
+	mutex_unlock(&dev->lock);
+
+	kref_put(&dev->kref, idmouse_delete);
 
 	dev_info(&interface->dev, "disconnected\n");
 }

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Johan Hovold <[email protected]>

commit 19bdfc7b3c179331eafa423d87e1336f43bbfeb8 upstream.

mutex_unlock() may access the mutex structure after releasing the lock and therefore cannot be used to manage lifetime of objects directly (unlike spinlocks and refcounts). [1][2]

Use a kref to release the driver data to avoid use-after-free in mutex_unlock() when release() races with disconnect().

[1] a51749ab34d9 ("locking/mutex: Document that mutex_unlock() is non-atomic")
[2] 2b9d9e0a9ba0 ("locking/mutex: Clarify that mutex_unlock(), and most other sleeping locks, can still use the lock object after it's unlocked")

---
 drivers/usb/misc/ldusb.c |   38 ++++++++++++++++++--------------------
 1 file changed, 18 insertions(+), 20 deletions(-)

@@ -150,6 +150,7 @@ MODULE_PARM_DESC(min_interrupt_out_inter
 
 /* Structure to hold all of our device specific stuff */
 struct ld_usb {
+	struct kref		kref;
 	struct mutex		mutex;		/* locks this structure */
 	struct usb_interface	*intf;		/* save off the usb interface pointer */
 	unsigned long		disconnected:1;

@@ -201,8 +202,10 @@ static void ld_usb_abort_transfers(struc
 /*
  *	ld_usb_delete
  */
-static void ld_usb_delete(struct ld_usb *dev)
+static void ld_usb_delete(struct kref *kref)
 {
+	struct ld_usb *dev = container_of(kref, struct ld_usb, kref);
+
 	/* free data structures */
 	usb_free_urb(dev->interrupt_in_urb);
 	usb_free_urb(dev->interrupt_out_urb);

@@ -355,6 +358,8 @@ static int ld_usb_open(struct inode *ino
 		goto unlock_exit;
 	}
 
+	kref_get(&dev->kref);
+
 	/* save device in the file's private structure */
 	file->private_data = dev;
 

@@ -381,17 +386,8 @@ static int ld_usb_release(struct inode *
 
 	mutex_lock(&dev->mutex);
 
-	if (dev->open_count != 1) {
-		retval = -ENODEV;
+	if (dev->disconnected)
 		goto unlock_exit;
-	}
-	if (dev->disconnected) {
-		/* the device was unplugged before the file was released */
-		mutex_unlock(&dev->mutex);
-		/* unlock here as ld_usb_delete frees dev */
-		ld_usb_delete(dev);
-		goto exit;
-	}
 
 	/* wait until write transfer is finished */
 	if (dev->interrupt_out_busy)

@@ -401,7 +397,7 @@ static int ld_usb_release(struct inode *
 
 unlock_exit:
 	mutex_unlock(&dev->mutex);
-
+	kref_put(&dev->kref, ld_usb_delete);
 exit:
 	return retval;
 }

@@ -659,6 +655,8 @@ static int ld_usb_probe(struct usb_inter
 	dev = kzalloc(sizeof(*dev), GFP_KERNEL);
 	if (!dev)
 		goto exit;
+
+	kref_init(&dev->kref);
 	mutex_init(&dev->mutex);
 	spin_lock_init(&dev->rbsl);
 	dev->intf = intf;

@@ -740,7 +738,7 @@ exit:
 	return retval;
 
 error:
-	ld_usb_delete(dev);
+	kref_put(&dev->kref, ld_usb_delete);
 
 	return retval;
 }

@@ -768,18 +766,18 @@ static void ld_usb_disconnect(struct usb
 
 	mutex_lock(&dev->mutex);
 
-	/* if the device is not opened, then we clean up right now */
-	if (!dev->open_count) {
-		mutex_unlock(&dev->mutex);
-		ld_usb_delete(dev);
-	} else {
-		dev->disconnected = 1;
+	dev->disconnected = 1;
+
+	if (dev->open_count) {
 		/* wake up pollers */
 		wake_up_interruptible_all(&dev->read_wait);
 		wake_up_interruptible_all(&dev->write_wait);
-		mutex_unlock(&dev->mutex);
 	}
 
+	mutex_unlock(&dev->mutex);
+
+	kref_put(&dev->kref, ld_usb_delete);
+
 	dev_info(&intf->dev, "LD USB Device #%d now disconnected\n",
 		 (minor - USB_LD_MINOR_BASE));
 }

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Johan Hovold <[email protected]>

commit bc0e4f16c44e50daa0b1ea729934baa3b4815dee upstream.

Submitted write URBs are not stopped on close() and therefore need to be stopped unconditionally on disconnect() to avoid use-after-free in the completion handler.

---
 drivers/usb/misc/iowarrior.c |    4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

@@ -919,13 +919,15 @@ static void iowarrior_disconnect(struct
 	/* prevent device read, write and ioctl */
 	dev->present = 0;
 
+	/* write urbs are not stopped on close() so kill unconditionally */
+	usb_kill_anchored_urbs(&dev->submitted);
+
 	if (dev->opened) {
 		/* There is a process that holds a filedescriptor to the device ,
 		   so we only shutdown read-/write-ops going on.
 		   Deleting the device is postponed until close() was called.
 		 */
 		usb_kill_urb(dev->int_in_urb);
-		usb_kill_anchored_urbs(&dev->submitted);
 		wake_up_interruptible(&dev->read_wait);
 		wake_up_interruptible(&dev->write_wait);
 		mutex_unlock(&dev->mutex);

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Erich E. Hoover <[email protected]>

commit fc591787785b9709a0bb65a7df3ba2537d611c47 upstream.

The Samsung T5 EVO Portable SSD (04e8:6200) exhibit two forms of link instability when USB Link Power Management is enabled:

1. The units fail to initialize properly on first detection, resulting in a lockup in the drive where it must be power cycled or the kernel will not recognize the presence of the device.

2. If used for sustained operations (small amounts of continuous data are transferred to the unit) then the unit will "hiccup" after roughly 8 hours of use and will disconnect and reconnect. This has a certain probability of triggering the first issue, but also causes mount points to become invalid since the device gets issued a new letter.

---
 drivers/usb/core/quirks.c |    3 +++
 1 file changed, 3 insertions(+)

@@ -297,6 +297,9 @@ static const struct usb_device_id usb_qu
 	/* CarrolTouch 4500U */
 	{ USB_DEVICE(0x04e7, 0x0030), .driver_info = USB_QUIRK_RESET_RESUME },
 
+	/* Samsung T5 EVO Portable SSD */
+	{ USB_DEVICE(0x04e8, 0x6200), .driver_info = USB_QUIRK_NO_LPM },
+
 	/* Samsung Android phone modem - ID conflict with SPH-I500 */
 	{ USB_DEVICE(0x04e8, 0x6601), .driver_info =
 			USB_QUIRK_CONFIG_INTF_STRINGS },

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Johan Hovold <[email protected]>

commit 62fc8eb1b1481051f7bab4aa93d79809053dd09f upstream.

mutex_unlock() may access the mutex structure after releasing the lock and therefore cannot be used to manage lifetime of objects directly (unlike spinlocks and refcounts). [1][2]

Use a kref to release the driver data to avoid use-after-free in mutex_unlock() when release() races with disconnect().

[1] a51749ab34d9 ("locking/mutex: Document that mutex_unlock() is non-atomic")
[2] 2b9d9e0a9ba0 ("locking/mutex: Clarify that mutex_unlock(), and most other sleeping locks, can still use the lock object after it's unlocked")

---
 drivers/usb/misc/legousbtower.c |   37 +++++++++++++++++++------------------
 1 file changed, 19 insertions(+), 18 deletions(-)

@@ -185,6 +185,7 @@ MODULE_DEVICE_TABLE(usb, tower_table);
 
 /* Structure to hold all of our device specific stuff */
 struct lego_usb_tower {
+	struct kref		kref;
 	struct mutex		lock;		/* locks this structure */
 	struct usb_device	*udev;		/* save off the usb device pointer */
 	unsigned char		minor;		/* the starting minor number for this device */

@@ -220,7 +221,6 @@ struct lego_usb_tower {
 /* local function prototypes */
 static ssize_t tower_read(struct file *file, char __user *buffer, size_t count, loff_t *ppos);
 static ssize_t tower_write(struct file *file, const char __user *buffer, size_t count, loff_t *ppos);
-static inline void tower_delete(struct lego_usb_tower *dev);
 static int tower_open(struct inode *inode, struct file *file);
 static int tower_release(struct inode *inode, struct file *file);
 static __poll_t tower_poll(struct file *file, poll_table *wait);

@@ -286,8 +286,10 @@ static inline void lego_usb_tower_debug_
 /*
  *	tower_delete
  */
-static inline void tower_delete(struct lego_usb_tower *dev)
+static inline void tower_delete(struct kref *kref)
 {
+	struct lego_usb_tower *dev = container_of(kref, struct lego_usb_tower, kref);
+
 	/* free data structures */
 	usb_free_urb(dev->interrupt_in_urb);
 	usb_free_urb(dev->interrupt_out_urb);

@@ -381,6 +383,8 @@ static int tower_open(struct inode *inod
 
 	dev->open_count = 1;
 
+	kref_get(&dev->kref);
+
 unlock_exit:
 	mutex_unlock(&dev->lock);
 

@@ -404,14 +408,8 @@ static int tower_release(struct inode *i
 
 	mutex_lock(&dev->lock);
 
-	if (dev->disconnected) {
-		/* the device was unplugged before the file was released */
-
-		/* unlock here as tower_delete frees dev */
-		mutex_unlock(&dev->lock);
-		tower_delete(dev);
-		goto exit;
-	}
+	if (dev->disconnected)
+		goto out_unlock;
 
 	/* wait until write transfer is finished */
 	if (dev->interrupt_out_busy) {

@@ -425,7 +423,9 @@ static int tower_release(struct inode *i
 
 	dev->open_count = 0;
 
+out_unlock:
 	mutex_unlock(&dev->lock);
+	kref_put(&dev->kref, tower_delete);
 exit:
 	return retval;
 }

@@ -752,6 +752,7 @@ static int tower_probe(struct usb_interf
 	if (!dev)
 		goto exit;
 
+	kref_init(&dev->kref);
 	mutex_init(&dev->lock);
 	dev->udev = usb_get_dev(udev);
 	spin_lock_init(&dev->read_buffer_lock);

@@ -828,7 +829,7 @@ exit:
 	return retval;
 
 error:
-	tower_delete(dev);
+	kref_put(&dev->kref, tower_delete);
 	return retval;
 }
 

@@ -856,18 +857,18 @@ static void tower_disconnect(struct usb_
 
 	mutex_lock(&dev->lock);
 
-	/* if the device is not opened, then we clean up right now */
-	if (!dev->open_count) {
-		mutex_unlock(&dev->lock);
-		tower_delete(dev);
-	} else {
-		dev->disconnected = 1;
+	dev->disconnected = 1;
+
+	if (dev->open_count) {
 		/* wake up pollers */
 		wake_up_interruptible_all(&dev->read_wait);
 		wake_up_interruptible_all(&dev->write_wait);
-		mutex_unlock(&dev->lock);
 	}
 
+	mutex_unlock(&dev->lock);
+
+	kref_put(&dev->kref, tower_delete);
+
 	dev_info(&interface->dev, "LEGO USB Tower #%d now disconnected\n",
 		 (minor - LEGO_USB_TOWER_MINOR_BASE));
 }

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Myeonghun Pak <[email protected]>

commit 4e8ba83ac4d311992e6a4c21de5dd705010df06e upstream.

sl811h_probe() enables the HCD controller device as a wakeup source after usb_add_hcd() succeeds, but sl811h_remove() removes the HCD and releases the driver resources without disabling that wakeup source.

Disable controller wakeup after usb_remove_hcd() and before usb_put_hcd() so the wakeup source object is detached while the controller device pointer is still available.

This issue was identified during our ongoing static-analysis research while reviewing kernel code.

---
 drivers/usb/host/sl811-hcd.c |    1 +
 1 file changed, 1 insertion(+)

@@ -1591,6 +1591,7 @@ sl811h_remove(struct platform_device *de
 
 	remove_debug_file(sl811);
 	usb_remove_hcd(hcd);
+	device_wakeup_disable(hcd->self.controller);
 
 	/* some platforms may use IORESOURCE_IO */
 	res = platform_get_resource(dev, IORESOURCE_MEM, 1);

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Xu Rao <[email protected]>

commit 2c00e09e3f9f06f8434f5ea2ee6179ce46692ee6 upstream.

usb_stor_adjust_quirks() parses the usb-storage.quirks module parameter into a new flag set and then applies it with the quirk mask to override built-in flags.

The mask is meant to cover the flags that can be overridden by the module parameter. The 'k' quirk character sets US_FL_NO_SAME, but US_FL_NO_SAME is not included in the mask.

As a result, the module parameter can set US_FL_NO_SAME, but it cannot clear a built-in US_FL_NO_SAME flag by providing an override entry that omits 'k'.

Add US_FL_NO_SAME to the mask so that the module parameter can override it in the same way as the other supported flags.

---
 drivers/usb/storage/usb.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -570,7 +570,7 @@ void usb_stor_adjust_quirks(struct usb_d
 			US_FL_INITIAL_READ10 | US_FL_WRITE_CACHE |
 			US_FL_NO_ATA_1X | US_FL_NO_REPORT_OPCODES |
 			US_FL_MAX_SECTORS_240 | US_FL_NO_REPORT_LUNS |
-			US_FL_ALWAYS_SYNC);
+			US_FL_ALWAYS_SYNC | US_FL_NO_SAME);
 
 	p = quirks;
 	while (*p) {

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: HE WEI (ギカク) <[email protected]>

commit 8c6314489550fa81d41723a0ff33f655b5b6c7b6 upstream.

usbio_bulk_msg() copies bpkt_len = le16_to_cpu(bpkt->len) bytes out of the bulk IN buffer (usbio->rxbuf, allocated with size usbio->rxbuf_len) into the caller's buffer. bpkt_len is fully controlled by the device and is only checked against ibuf_len; ibuf_len in turn is checked against usbio->txbuf_len, not against rxbuf_len:

	if ((obuf_len > (usbio->txbuf_len - sizeof(*bpkt))) ||
	    (ibuf_len > (usbio->txbuf_len - sizeof(*bpkt))))
		return -EMSGSIZE;

txbuf_len and rxbuf_len are taken independently from the bulk OUT and bulk IN endpoint wMaxPacketSize in usbio_probe(). A malicious or malfunctioning device that advertises a large bulk OUT endpoint and a small bulk IN endpoint (e.g. by claiming one of the quirk-free IDs such as the Lattice NX33U, 0x2ac1:0x20cb) therefore makes ibuf_len, and hence the device-supplied bpkt_len, exceed rxbuf_len. memcpy() then reads up to txbuf_len - rxbuf_len bytes past the end of the rxbuf slab object. The over-read bytes are handed back to the i2c layer and on to user space through i2c-dev, disclosing adjacent slab memory; with KASAN this is reported as a slab-out-of-bounds read.

The number of bytes actually received is already known: act equals the URB actual_length and is bounded by rxbuf_len. Reject any response that claims more payload than was received, mirroring the existing "act < sizeof(*bpkt)" check just above.

The control path (usbio_ctrl_msg()) is not affected: it uses a single buffer (ctrlbuf) for both directions, so its analogous copy can never leave the allocation.

Found by code review. The out-of-bounds read was confirmed under AddressSanitizer with a faithful userspace model of usbio_bulk_msg()'s receive path (an rxbuf_len-sized buffer, the same act/ibuf_len/bpkt_len checks and the memcpy). A USB raw-gadget + dummy_hcd reproducer is also available.

---
 drivers/usb/misc/usbio.c |    4 ++++
 1 file changed, 4 insertions(+)

@@ -344,6 +344,10 @@ read:
 	if (ibuf_len < bpkt_len)
 		return -ENOSPC;
 
+	/* The device must not claim more payload than it actually sent. */
+	if (bpkt_len > act - sizeof(*bpkt))
+		return -EPROTO;
+
 	memcpy(ibuf, bpkt->data, bpkt_len);
 
 	return bpkt_len;

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Myeonghun Pak <[email protected]>

commit b4ecbdc4f8830f5586c4a5cfc384c00f20f8f8b3 upstream.

uss720_probe() registers a parport before reading the 1284 register used to detect unsupported Belkin F5U002 adapters. If get_1284_register() fails, the error path drops the driver private data and the USB device reference, but leaves the parport device registered.

Leaving the port registered is more than a private allocation leak:
parport_register_port() has already reserved a parport number and registered the parport bus device, while pp->private_data still points at the private data that the common error path is about to release.

Undo the pre-announce registration in the get_1284_register() failure branch before jumping to the common private-data cleanup path. Clear priv->pp first, matching the disconnect path and avoiding a stale pointer in the private data.

This issue was identified during our ongoing static-analysis research while reviewing kernel code.

---
 drivers/usb/misc/uss720.c |    5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

@@ -735,8 +735,11 @@ static int uss720_probe(struct usb_inter
 	 * here. */
 	ret = get_1284_register(pp, 0, &reg, GFP_KERNEL);
 	dev_dbg(&intf->dev, "reg: %7ph\n", priv->reg);
-	if (ret < 0)
+	if (ret < 0) {
+		priv->pp = NULL;
+		parport_del_port(pp);
 		goto probe_abort;
+	}
 
 	ret = usb_find_last_int_in_endpoint(interface, &epd);
 	if (!ret) {

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Haoxiang Li <[email protected]>

commit 0bddda5a11665c210339de76d27ebbd1a2e0b43c upstream.

mtu3_gadget_queue() maps the request before checking whether the QMU GPD ring can accept another transfer. the request is returned with -EAGAIN before it is linked on the endpoint request list if mtu3_prepare_transfer() fails.

Normal completion and dequeue paths unmap requests from mtu3_req_complete(), but this error path never reaches that helper, so the DMA mapping is left active. Unmap the request before returning from the failed queue path.

---
 drivers/usb/mtu3/mtu3_gadget.c |    1 +
 1 file changed, 1 insertion(+)

@@ -305,6 +305,7 @@ static int mtu3_gadget_queue(struct usb_
 
 	if (mtu3_prepare_transfer(mep)) {
 		ret = -EAGAIN;
+		usb_gadget_unmap_request(&mtu->g, req, mep->is_in);
 		goto error;
 	}
 

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Johan Hovold <[email protected]>

commit 6bfc8d01ac4068eced509f8fc74d0cd205e4dcec upstream.

The write() callback is supposed to return the number of characters accepted or a negative errno. Since the addition of write fifo support the keyspan_pda implementation will however return the number characters submitted to the device if the write urb is not already in use. If this number is larger than the number of characters passed to write(), the line discipline continues writing data from beyond the tty write buffer.

Fix the information leak by making sure that keyspan_pda_write_start() returns zero on success as intended.

---
 drivers/usb/serial/keyspan_pda.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -518,7 +518,7 @@ static int keyspan_pda_write_start(struc
 	if (count == room)
 		schedule_work(&priv->unthrottle_work);
 
-	return count;
+	return 0;
 }
 
 static void keyspan_pda_write_bulk_callback(struct urb *urb)

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Fabio Porcedda <[email protected]>

commit b33ab1dd80f5c1742f49eb6ec7b337c5ffcf3d32 upstream.

Add support for Telit Cinterion FE990D50 compositions:

0x990: RNDIS + tty (AT/NMEA) + tty (AT) + tty (AT) + tty (AT) +
tty (diag) + ADPL + adb

T:  Bus=01 Lev=01 Prnt=01 Port=06 Cnt=03 Dev#=  3 Spd=480  MxCh= 0
D:  Ver= 2.10 Cls=00(>ifc ) Sub=00 Prot=00 MxPS=64 #Cfgs=  1
P:  Vendor=1bc7 ProdID=0990 Rev=06.06
S:  Manufacturer=Telit Cinterion
S:  Product=FE990
S:  SerialNumber=90b6a3ed
C:  #Ifs=10 Cfg#= 1 Atr=e0 MxPwr=500mA
I:  If#= 0 Alt= 0 #EPs= 1 Cls=ef(misc ) Sub=04 Prot=01 Driver=rndis_host
E:  Ad=82(I) Atr=03(Int.) MxPS=   8 Ivl=32ms
I:  If#= 1 Alt= 0 #EPs= 2 Cls=0a(data ) Sub=00 Prot=00 Driver=rndis_host
E:  Ad=01(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=81(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
I:  If#= 2 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=ff Prot=60 Driver=option
E:  Ad=02(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=83(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=84(I) Atr=03(Int.) MxPS=  10 Ivl=32ms
I:  If#= 3 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=ff Prot=40 Driver=option
E:  Ad=03(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=85(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=86(I) Atr=03(Int.) MxPS=  10 Ivl=32ms
I:  If#= 4 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=ff Prot=40 Driver=option
E:  Ad=04(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=87(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=88(I) Atr=03(Int.) MxPS=  10 Ivl=32ms
I:  If#= 5 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=ff Prot=40 Driver=option
E:  Ad=05(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=89(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=8a(I) Atr=03(Int.) MxPS=  10 Ivl=32ms
I:  If#= 6 Alt= 0 #EPs= 2 Cls=ff(vend.) Sub=ff Prot=30 Driver=option
E:  Ad=06(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=8b(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
I:  If#= 7 Alt= 0 #EPs= 1 Cls=ff(vend.) Sub=ff Prot=80 Driver=(none)
E:  Ad=8c(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
I:  If#= 8 Alt= 0 #EPs= 1 Cls=ff(vend.) Sub=ff Prot=70 Driver=(none)
E:  Ad=8d(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
I:  If#= 9 Alt= 0 #EPs= 2 Cls=ff(vend.) Sub=42 Prot=01 Driver=(none)
E:  Ad=07(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=8e(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms

0x991: rmnet + tty (AT/NMEA) + tty (AT) + tty (AT) + tty (AT) +
tty (diag) + ADPL + adb

T:  Bus=01 Lev=01 Prnt=01 Port=06 Cnt=03 Dev#=  9 Spd=480  MxCh= 0
D:  Ver= 2.10 Cls=00(>ifc ) Sub=00 Prot=00 MxPS=64 #Cfgs=  1
P:  Vendor=1bc7 ProdID=0991 Rev=06.06
S:  Manufacturer=Telit Cinterion
S:  Product=FE990
S:  SerialNumber=90b6a3ed
C:  #Ifs= 9 Cfg#= 1 Atr=e0 MxPwr=500mA
I:  If#= 0 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=ff Prot=50 Driver=(none)
E:  Ad=01(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=81(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=82(I) Atr=03(Int.) MxPS=   8 Ivl=32ms
I:  If#= 1 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=ff Prot=60 Driver=option
E:  Ad=02(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=83(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=84(I) Atr=03(Int.) MxPS=  10 Ivl=32ms
I:  If#= 2 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=ff Prot=40 Driver=option
E:  Ad=03(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=85(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=86(I) Atr=03(Int.) MxPS=  10 Ivl=32ms
I:  If#= 3 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=ff Prot=40 Driver=option
E:  Ad=04(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=87(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=88(I) Atr=03(Int.) MxPS=  10 Ivl=32ms
I:  If#= 4 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=ff Prot=40 Driver=option
E:  Ad=05(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=89(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=8a(I) Atr=03(Int.) MxPS=  10 Ivl=32ms
I:  If#= 5 Alt= 0 #EPs= 2 Cls=ff(vend.) Sub=ff Prot=30 Driver=option
E:  Ad=06(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=8b(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
I:  If#= 6 Alt= 0 #EPs= 1 Cls=ff(vend.) Sub=ff Prot=80 Driver=(none)
E:  Ad=8c(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
I:  If#= 7 Alt= 0 #EPs= 1 Cls=ff(vend.) Sub=ff Prot=70 Driver=(none)
E:  Ad=8d(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
I:  If#= 8 Alt= 0 #EPs= 2 Cls=ff(vend.) Sub=42 Prot=01 Driver=(none)
E:  Ad=07(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=8e(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms

0x992: MBIM + tty (AT/NMEA) + tty (AT) + tty (AT) + tty (AT) +
tty (diag) + ADPL + adb

T:  Bus=01 Lev=01 Prnt=01 Port=06 Cnt=03 Dev#= 12 Spd=480  MxCh= 0
D:  Ver= 2.10 Cls=ef(misc ) Sub=02 Prot=01 MxPS=64 #Cfgs=  1
P:  Vendor=1bc7 ProdID=0992 Rev=06.06
S:  Manufacturer=Telit Cinterion
S:  Product=FE990
S:  SerialNumber=90b6a3ed
C:  #Ifs=10 Cfg#= 1 Atr=e0 MxPwr=500mA
I:  If#= 0 Alt= 0 #EPs= 1 Cls=02(commc) Sub=0e Prot=00 Driver=cdc_mbim
E:  Ad=82(I) Atr=03(Int.) MxPS=  64 Ivl=32ms
I:  If#= 1 Alt= 1 #EPs= 2 Cls=0a(data ) Sub=00 Prot=02 Driver=cdc_mbim
E:  Ad=01(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=81(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
I:  If#= 2 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=ff Prot=60 Driver=option
E:  Ad=02(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=83(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=84(I) Atr=03(Int.) MxPS=  10 Ivl=32ms
I:  If#= 3 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=ff Prot=40 Driver=option
E:  Ad=03(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=85(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=86(I) Atr=03(Int.) MxPS=  10 Ivl=32ms
I:  If#= 4 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=ff Prot=40 Driver=option
E:  Ad=04(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=87(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=88(I) Atr=03(Int.) MxPS=  10 Ivl=32ms
I:  If#= 5 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=ff Prot=40 Driver=option
E:  Ad=05(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=89(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=8a(I) Atr=03(Int.) MxPS=  10 Ivl=32ms
I:  If#= 6 Alt= 0 #EPs= 2 Cls=ff(vend.) Sub=ff Prot=30 Driver=option
E:  Ad=06(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=8b(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
I:  If#= 7 Alt= 0 #EPs= 1 Cls=ff(vend.) Sub=ff Prot=80 Driver=(none)
E:  Ad=8c(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
I:  If#= 8 Alt= 0 #EPs= 1 Cls=ff(vend.) Sub=ff Prot=70 Driver=(none)
E:  Ad=8d(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
I:  If#= 9 Alt= 0 #EPs= 2 Cls=ff(vend.) Sub=42 Prot=01 Driver=(none)
E:  Ad=07(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=8e(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms

0x993: ECM + tty (AT/NMEA) + tty (AT) + tty (AT) + tty (AT) +
tty (diag) + ADPL + adb

T:  Bus=01 Lev=01 Prnt=01 Port=06 Cnt=03 Dev#= 15 Spd=480  MxCh= 0
D:  Ver= 2.10 Cls=00(>ifc ) Sub=00 Prot=00 MxPS=64 #Cfgs=  1
P:  Vendor=1bc7 ProdID=0993 Rev=06.06
S:  Manufacturer=Telit Cinterion
S:  Product=FE990
S:  SerialNumber=90b6a3ed
C:  #Ifs=10 Cfg#= 1 Atr=e0 MxPwr=500mA
I:  If#= 0 Alt= 0 #EPs= 1 Cls=02(commc) Sub=06 Prot=00 Driver=cdc_ether
E:  Ad=82(I) Atr=03(Int.) MxPS=  16 Ivl=32ms
I:  If#= 1 Alt= 1 #EPs= 2 Cls=0a(data ) Sub=00 Prot=00 Driver=cdc_ether
E:  Ad=01(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=81(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
I:  If#= 2 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=ff Prot=60 Driver=option
E:  Ad=02(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=83(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=84(I) Atr=03(Int.) MxPS=  10 Ivl=32ms
I:  If#= 3 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=ff Prot=40 Driver=option
E:  Ad=03(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=85(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=86(I) Atr=03(Int.) MxPS=  10 Ivl=32ms
I:  If#= 4 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=ff Prot=40 Driver=option
E:  Ad=04(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=87(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=88(I) Atr=03(Int.) MxPS=  10 Ivl=32ms
I:  If#= 5 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=ff Prot=40 Driver=option
E:  Ad=05(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=89(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=8a(I) Atr=03(Int.) MxPS=  10 Ivl=32ms
I:  If#= 6 Alt= 0 #EPs= 2 Cls=ff(vend.) Sub=ff Prot=30 Driver=option
E:  Ad=06(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=8b(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
I:  If#= 7 Alt= 0 #EPs= 1 Cls=ff(vend.) Sub=ff Prot=80 Driver=(none)
E:  Ad=8c(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
I:  If#= 8 Alt= 0 #EPs= 1 Cls=ff(vend.) Sub=ff Prot=70 Driver=(none)
E:  Ad=8d(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
I:  If#= 9 Alt= 0 #EPs= 2 Cls=ff(vend.) Sub=42 Prot=01 Driver=(none)
E:  Ad=07(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=8e(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms

---
 drivers/usb/serial/option.c |   16 ++++++++++++++++
 1 file changed, 16 insertions(+)

@@ -1326,6 +1326,22 @@ static const struct usb_device_id option
 	{ USB_DEVICE(TELIT_VENDOR_ID, TELIT_PRODUCT_CC864_SINGLE) },
 	{ USB_DEVICE(TELIT_VENDOR_ID, TELIT_PRODUCT_DE910_DUAL) },
 	{ USB_DEVICE(TELIT_VENDOR_ID, TELIT_PRODUCT_UE910_V2) },
+	{ USB_DEVICE_AND_INTERFACE_INFO(TELIT_VENDOR_ID, 0x0990, 0xff, 0xff, 0x30),	/* Telit FE990D50 (RNDIS) */
+	  .driver_info = NCTRL(6) },
+	{ USB_DEVICE_AND_INTERFACE_INFO(TELIT_VENDOR_ID, 0x0990, 0xff, 0xff, 0x40) },
+	{ USB_DEVICE_AND_INTERFACE_INFO(TELIT_VENDOR_ID, 0x0990, 0xff, 0xff, 0x60) },
+	{ USB_DEVICE_AND_INTERFACE_INFO(TELIT_VENDOR_ID, 0x0991, 0xff, 0xff, 0x30),	/* Telit FE990D50 (rmnet) */
+	  .driver_info = NCTRL(5) },
+	{ USB_DEVICE_AND_INTERFACE_INFO(TELIT_VENDOR_ID, 0x0991, 0xff, 0xff, 0x40) },
+	{ USB_DEVICE_AND_INTERFACE_INFO(TELIT_VENDOR_ID, 0x0991, 0xff, 0xff, 0x60) },
+	{ USB_DEVICE_AND_INTERFACE_INFO(TELIT_VENDOR_ID, 0x0992, 0xff, 0xff, 0x30),	/* Telit FE990D50 (MBIM) */
+	  .driver_info = NCTRL(6) },
+	{ USB_DEVICE_AND_INTERFACE_INFO(TELIT_VENDOR_ID, 0x0992, 0xff, 0xff, 0x40) },
+	{ USB_DEVICE_AND_INTERFACE_INFO(TELIT_VENDOR_ID, 0x0992, 0xff, 0xff, 0x60) },
+	{ USB_DEVICE_AND_INTERFACE_INFO(TELIT_VENDOR_ID, 0x0993, 0xff, 0xff, 0x30),	/* Telit FE990D50 (ECM) */
+	  .driver_info = NCTRL(6) },
+	{ USB_DEVICE_AND_INTERFACE_INFO(TELIT_VENDOR_ID, 0x0993, 0xff, 0xff, 0x40) },
+	{ USB_DEVICE_AND_INTERFACE_INFO(TELIT_VENDOR_ID, 0x0993, 0xff, 0xff, 0x60) },
 	{ USB_DEVICE_INTERFACE_CLASS(TELIT_VENDOR_ID, 0x1031, 0xff),	/* Telit LE910C1-EUX */
 	 .driver_info = NCTRL(0) | RSVD(3) },
 	{ USB_DEVICE_INTERFACE_CLASS(TELIT_VENDOR_ID, 0x1033, 0xff),	/* Telit LE910C1-EUX (ECM) */

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Johan Hovold <[email protected]>

commit 83a3dfc018943b05b6daf3a6f891833e1aabfa1f upstream.

If the port is closed while throttled, the read urb is never resubmitted and the port will not receive any further data until the device is reconnected (or the driver is rebound).

Clear the throttle flags and submit the urb if needed when opening the port.

---
 drivers/usb/serial/digi_acceleport.c |   16 ++++++++++++++++
 1 file changed, 16 insertions(+)

@@ -1071,6 +1071,7 @@ static int digi_open(struct tty_struct *
 	unsigned char buf[32];
 	struct digi_port *priv = usb_get_serial_port_data(port);
 	struct ktermios not_termios;
+	int throttled;
 
 	/* be sure the device is started up */
 	if (digi_startup_device(port->serial) != 0)

@@ -1098,6 +1099,21 @@ static int digi_open(struct tty_struct *
 		not_termios.c_iflag = ~tty->termios.c_iflag;
 		digi_set_termios(tty, port, &not_termios);
 	}
+
+	spin_lock_irq(&priv->dp_port_lock);
+	throttled = priv->dp_throttle_restart;
+	priv->dp_throttled = 0;
+	priv->dp_throttle_restart = 0;
+	spin_unlock_irq(&priv->dp_port_lock);
+
+	if (throttled) {
+		ret = usb_submit_urb(port->read_urb, GFP_KERNEL);
+		if (ret) {
+			dev_err(&port->dev, "failed to submit read urb: %d\n", ret);
+			return ret;
+		}
+	}
+
 	return 0;
 }
 

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Johan Hovold <[email protected]>

commit 5c1ea24b53bf3bfb859f0a05573997487975da23 upstream.

If submitting the OOB write urb fails persistently (e.g if the device is being disconnected) the driver would loop indefinitely with interrupts disabled.

Check for urb submission errors when sending OOB commands to avoid hanging if, for example, open(), set_termios() or close() races with a physical disconnect.

This is issue was flagged by Sashiko when reviewing an unrelated change to the driver.

---
 drivers/usb/serial/digi_acceleport.c |   12 +++++++-----
 1 file changed, 7 insertions(+), 5 deletions(-)

@@ -394,12 +394,14 @@ static int digi_write_oob_command(struct
 			len &= ~3;
 		memcpy(oob_port->write_urb->transfer_buffer, buf, len);
 		oob_port->write_urb->transfer_buffer_length = len;
+
 		ret = usb_submit_urb(oob_port->write_urb, GFP_ATOMIC);
-		if (ret == 0) {
-			oob_priv->dp_write_urb_in_use = 1;
-			count -= len;
-			buf += len;
-		}
+		if (ret)
+			break;
+
+		oob_priv->dp_write_urb_in_use = 1;
+		count -= len;
+		buf += len;
 	}
 	spin_unlock_irqrestore(&oob_priv->dp_port_lock, flags);
 	if (ret)

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Johan Hovold <[email protected]>

commit 24ca1fea8f2753bf33e1d458ec1ae5d9b7796a65 upstream.

The digi_write_inb_command() is supposed to wait for the write urb to become available or return an error, but instead it updates the transfer buffer and tries to resubmit the urb on timeout.

To make things worse, for commands like break control where no timeout is used, the driver would corrupt the urb immediately due to a broken jiffies comparison (on 32-bit machines this takes five minutes of uptime to trigger due to INITIAL_JIFFIES).

Fix this by adding the missing return on timeout and waiting indefinitely when no timeout has been specified as intended.

This issue was (sort of) flagged by Sashiko when reviewing an unrelated change to the driver.

---
 drivers/usb/serial/digi_acceleport.c |   15 ++++++++++-----
 1 file changed, 10 insertions(+), 5 deletions(-)

@@ -431,20 +431,22 @@ static int digi_write_inb_command(struct
 	int len;
 	struct digi_port *priv = usb_get_serial_port_data(port);
 	unsigned char *data = port->write_urb->transfer_buffer;
+	unsigned long expire;
 	unsigned long flags;
 
 	dev_dbg(&port->dev, "digi_write_inb_command: TOP: port=%d, count=%d\n",
 		priv->dp_port_num, count);
 
 	if (timeout)
-		timeout += jiffies;
-	else
-		timeout = ULONG_MAX;
+		expire = jiffies + timeout;
 
 	spin_lock_irqsave(&priv->dp_port_lock, flags);
 	while (count > 0 && ret == 0) {
-		while (priv->dp_write_urb_in_use &&
-		       time_before(jiffies, timeout)) {
+		while (priv->dp_write_urb_in_use) {
+			if (timeout && time_after(jiffies, expire)) {
+				ret = -ETIMEDOUT;
+				break;
+			}
 			cond_wait_interruptible_timeout_irqrestore(
 				&priv->write_wait, DIGI_RETRY_TIMEOUT,
 				&priv->dp_port_lock, flags);

@@ -453,6 +455,9 @@ static int digi_write_inb_command(struct
 			spin_lock_irqsave(&priv->dp_port_lock, flags);
 		}
 
+		if (ret)
+			break;
+
 		/* len must be a multiple of 4 and small enough to */
 		/* guarantee the write will send buffered data first, */
 		/* so commands are in order with data and not split */

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Johan Hovold <[email protected]>

commit 8af6812795869a66e9b26044f455b13deecdb69c upstream.

The allocated device name is never freed on early ULPI device registration failures.

Fix this by initialising the device structure earlier and releasing the initial reference whenever registration fails.

---
 drivers/usb/common/ulpi.c |   23 ++++++++++-------------
 1 file changed, 10 insertions(+), 13 deletions(-)

@@ -281,28 +281,24 @@ static int ulpi_register(struct device *
 	ulpi->dev.parent = dev; /* needed early for ops */
 	ulpi->dev.bus = &ulpi_bus;
 	ulpi->dev.type = &ulpi_dev_type;
+
+	device_initialize(&ulpi->dev);
+
 	dev_set_name(&ulpi->dev, "%s.ulpi", dev_name(dev));
 
 	ACPI_COMPANION_SET(&ulpi->dev, ACPI_COMPANION(dev));
 
 	ret = ulpi_of_register(ulpi);
-	if (ret) {
-		kfree(ulpi);
+	if (ret)
 		return ret;
-	}
 
 	ret = ulpi_read_id(ulpi);
-	if (ret) {
-		of_node_put(ulpi->dev.of_node);
-		kfree(ulpi);
+	if (ret)
 		return ret;
-	}
 
-	ret = device_register(&ulpi->dev);
-	if (ret) {
-		put_device(&ulpi->dev);
+	ret = device_add(&ulpi->dev);
+	if (ret)
 		return ret;
-	}
 
 	root = debugfs_create_dir(dev_name(&ulpi->dev), ulpi_root);
 	debugfs_create_file("regs", 0444, root, ulpi, &ulpi_regs_fops);

@@ -334,9 +330,10 @@ struct ulpi *ulpi_register_interface(str
 	ulpi->ops = ops;
 
 	ret = ulpi_register(dev, ulpi);
-	if (ret)
+	if (ret) {
+		put_device(&ulpi->dev);
 		return ERR_PTR(ret);
-
+	}
 
 	return ulpi;
 }

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Xu Rao <[email protected]>

commit 5fc3f333c001f1e308bbcdeecdec0d054d24338b upstream.

Commit 1892bf90677a ("USB: usb-storage: Fix use of bitfields for hardware data in ene_ub6250.c") converted the media status fields from bitfields to bit masks.

The original ene_transport() test called ene_init() only when neither media type was ready:

        !(sd_ready || ms_ready)

The converted test became:

        !sd_ready || ms_ready

This is not equivalent. Restore the original semantics by testing that both ready bits are clear before calling ene_init().

---
 drivers/usb/storage/ene_ub6250.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

@@ -2306,7 +2306,8 @@ static int ene_transport(struct scsi_cmn
 
 	/*US_DEBUG(usb_stor_show_command(us, srb)); */
 	scsi_set_resid(srb, 0);
-	if (unlikely(!(info->SD_Status & SD_Ready) || (info->MS_Status & MS_Ready)))
+	if (unlikely(!(info->SD_Status & SD_Ready) &&
+		     !(info->MS_Status & MS_Ready)))
 		result = ene_init(us);
 	if (result == USB_STOR_XFER_GOOD) {
 		result = USB_STOR_TRANSPORT_ERROR;

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Yichong Chen <[email protected]>

commit 195e667c8719480c320cd48ac0fbf1cb81d6ffe0 upstream.

USB devices running at SuperSpeedPlus report "10000" or "20000" in their sysfs speed attribute. usbip currently maps only "5000" to USB_SPEED_SUPER, so a SuperSpeedPlus device is imported as USB_SPEED_UNKNOWN.

The attach request is then rejected by vhci_hcd:

vhci_hcd: Failed attach request for unsupported USB speed: UNKNOWN

Map the SuperSpeedPlus sysfs speed values to USB_SPEED_SUPER_PLUS, use the SuperSpeed VHCI hub for SuperSpeedPlus devices, and recognize the gadget current_speed string used by the kernel.

---
 tools/usb/usbip/libsrc/usbip_common.c        |    2 ++
 tools/usb/usbip/libsrc/usbip_device_driver.c |    4 ++++
 tools/usb/usbip/libsrc/vhci_driver.c         |    1 +
 3 files changed, 7 insertions(+)

@@ -29,6 +29,8 @@ static const struct speed_string speed_s
 	{ USB_SPEED_HIGH, "480", "High Speed(480Mbps)" },
 	{ USB_SPEED_WIRELESS, "53.3-480", "Wireless"},
 	{ USB_SPEED_SUPER, "5000", "Super Speed(5000Mbps)" },
+	{ USB_SPEED_SUPER_PLUS, "10000", "Super Speed Plus(10000Mbps)" },
+	{ USB_SPEED_SUPER_PLUS, "20000", "Super Speed Plus(20000Mbps)" },
 	{ 0, NULL, NULL }
 };
 

@@ -57,6 +57,10 @@ static struct {
 		.speed = USB_SPEED_SUPER,
 		.name = "super-speed",
 	},
+	{
+		.speed = USB_SPEED_SUPER_PLUS,
+		.name = "super-speed-plus",
+	},
 };
 
 static

@@ -338,6 +338,7 @@ int usbip_vhci_get_free_port(uint32_t sp
 
 		switch (speed) {
 		case	USB_SPEED_SUPER:
+		case	USB_SPEED_SUPER_PLUS:
 			if (vhci_driver->idev[i].hub != HUB_SPEED_SUPER)
 				continue;
 		break;

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Sam Day <[email protected]>

commit c5371e0b91b24159a3ebaa61e70b0980bcf03c0a upstream.

vep_alloc_request() wasn't initializing vrequest->udc, so cancellations on the FunctionFS AIO path were arriving in vep_dequeue without a valid UDC reference.

Since vrequest->udc is never actually properly used anywhere, we opt to remove it, and update vep_dequeue to obtain a reference to the udc with ep_to_vudc(), consistent with the other vep_ ops.

AFAICT this bug has existed for ~10 years. Seems that nobody has really stressed the FunctionFS AIO path on usbip's vudc.

I tested this fix in a QEMU aarch64 guest driving FunctionFS endpoints via AIO. Before the fix, running usbip attach from the host would cause the guest to oops with the following backtrace:

Call trace:
 vep_dequeue+0x1c/0xe4 (P)
 usb_ep_dequeue+0x14/0x20
 ffs_aio_cancel+0x24/0x34
 __arm64_sys_io_cancel+0xb0/0x124
 do_el0_svc+0x68/0x100
 el0_svc+0x18/0x5c
 el0t_64_sync_handler+0x98/0xdc
 el0t_64_sync+0x154/0x158

---
 drivers/usb/usbip/vudc.h     |    1 -
 drivers/usb/usbip/vudc_dev.c |    4 +---
 2 files changed, 1 insertion(+), 4 deletions(-)

@@ -38,7 +38,6 @@ struct vep {
 
 struct vrequest {
 	struct usb_request req;
-	struct vudc *udc;
 	struct list_head req_entry; /* Request queue */
 };
 

@@ -333,7 +333,6 @@ static int vep_queue(struct usb_ep *_ep,
 static int vep_dequeue(struct usb_ep *_ep, struct usb_request *_req)
 {
 	struct vep *ep;
-	struct vrequest *req;
 	struct vudc *udc;
 	struct vrequest *lst;
 	unsigned long flags;

@@ -343,8 +342,7 @@ static int vep_dequeue(struct usb_ep *_e
 		return ret;
 
 	ep = to_vep(_ep);
-	req = to_vrequest(_req);
-	udc = req->udc;
+	udc = ep_to_vudc(ep);
 
 	if (!udc->driver)
 		return -ESHUTDOWN;

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Myeonghun Pak <[email protected]>

commit 0ef7cc27da8b9e315a4a5a665c68c44206f5e559 upstream.

anx7411_i2c_probe() enables runtime PM before returning successfully, but anx7411_i2c_remove() tears down the Type-C partner state, workqueue, dummy I2C device, mux, switch and port without disabling runtime PM.

Use devm_pm_runtime_enable() so runtime PM is disabled automatically on driver detach. Since devres action registration can fail, route that failure through the existing probe unwind path.

This issue was identified during our ongoing static-analysis research while reviewing kernel code.

---
 drivers/usb/typec/anx7411.c |    4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

@@ -1538,7 +1538,9 @@ static int anx7411_i2c_probe(struct i2c_
 	if (anx7411_typec_check_connection(plat))
 		dev_err(dev, "check status\n");
 
-	pm_runtime_enable(dev);
+	ret = devm_pm_runtime_enable(dev);
+	if (ret)
+		goto free_wq;
 
 	return 0;
 

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Shuangpeng Bai <[email protected]>

commit 43ae2f90b70cda374c487c1639a01d0f14e5d583 upstream.

usb_power_delivery_find() wraps class_find_device_by_name(). That helper returns a device reference that must be released by the caller.

select_usb_power_delivery_store() only needs this reference while calling the pd_set callback. Drop it once the callback returns. Otherwise the sysfs write can pin the selected USB Power Delivery object and prevent it from being released on unregister.

---
 drivers/usb/typec/class.c |    1 +
 1 file changed, 1 insertion(+)

@@ -1525,6 +1525,7 @@ static ssize_t select_usb_power_delivery
 		return -EINVAL;
 
 	ret = port->ops->pd_set(port, pd);
+	put_device(&pd->dev);
 	if (ret)
 		return ret;
 

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Andy Yan <[email protected]>

commit 9cff680e47632b7723cb19f9c5e63669063c3417 upstream.

VDO() second parameter is VDM type (bit 15): 1 for SVDM, 0 for UVDM. Using 'vdo ? 2 : 1' corrupts SVID low bit when vdo is non-NULL (2 << 15 = BIT(16)). Enter Mode is always SVDM, hardcode to 1.

---
 drivers/usb/typec/tcpm/tcpm.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

@@ -2928,7 +2928,7 @@ static int tcpm_altmode_enter(struct typ
 	if (svdm_version < 0)
 		return svdm_version;
 
-	header = VDO(altmode->svid, vdo ? 2 : 1, svdm_version, CMD_ENTER_MODE);
+	header = VDO(altmode->svid, 1, svdm_version, CMD_ENTER_MODE);
 	header |= VDO_OPOS(altmode->mode);
 
 	return tcpm_queue_vdm_unlocked(port, header, vdo, vdo ? 1 : 0, TCPC_TX_SOP);

@@ -2976,7 +2976,7 @@ static int tcpm_cable_altmode_enter(stru
 	if (svdm_version < 0)
 		return svdm_version;
 
-	header = VDO(altmode->svid, vdo ? 2 : 1, svdm_version, CMD_ENTER_MODE);
+	header = VDO(altmode->svid, 1, svdm_version, CMD_ENTER_MODE);
 	header |= VDO_OPOS(altmode->mode);
 
 	return tcpm_queue_vdm_unlocked(port, header, vdo, vdo ? 1 : 0, TCPC_TX_SOP_PRIME);

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Badhri Jagan Sridharan <[email protected]>

commit 7b681dd5fbf60b24a13c14661e5b7735759fb491 upstream.

In svdm_consume_modes(), the SVID value is read from pmdata->svids using pmdata->svid_index as an array index without bounds validation:

    paltmode->svid = pmdata->svids[pmdata->svid_index];

If pmdata->svid_index is driven beyond SVID_DISCOVERY_MAX (16), it results in an out-of-bounds read of the pmdata->svids array. Because pd_mode_data is embedded inside struct tcpm_port, indexing past svids reads into adjacent fields. In particular:
- At index 16, it reads the altmodes count.
- At index 18 and beyond, it reads into altmode_desc[], which contains partner-supplied SVDM Discovery Modes VDOs.

By injecting a chosen SVID into altmode_desc[0].vdo and driving svid_index to 20, the partner can force paltmode->svid to be loaded with an arbitrary, partner- chosen SVID, which is then registered via typec_partner_register_altmode().

Fix this by validating that pmdata->svid_index is non-negative and strictly less than pmdata->nsvids before accessing the pmdata->svids array inside svdm_consume_modes().

---
 drivers/usb/typec/tcpm/tcpm.c |    5 +++++
 1 file changed, 5 insertions(+)

@@ -1840,6 +1840,11 @@ static void svdm_consume_modes(struct tc
 		return;
 	}
 
+	if (pmdata->svid_index < 0 || pmdata->svid_index >= pmdata->nsvids) {
+		tcpm_log(port, "Invalid SVID index %d", pmdata->svid_index);
+		return;
+	}
+
 	for (i = 1; i < cnt; i++) {
 		if (pmdata->altmodes >= ALTMODE_DISCOVERY_MAX) {
 			/* Already logged in svdm_consume_svids() */

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Andrei Kuchynski <[email protected]>

commit d092d7edf8faefa3e27b9fc7f0e7904b06c833a2 upstream.

The existing implementation assigned these flags backwards, configuring the partner's DisplayPort role to match the port's role instead of complementing it.
This prevents proper configuration during DP altmode activation, often causing pin_assignment to remain 0 in dp_altmode_configure() and resulting in VDM negotiation failures:

    [  583.328246] typec port1.1: VDM 0xff01a150 failed

Additionally, the fix ensures that the pin_assignment sysfs attribute displays the correct values.

---
 drivers/usb/typec/ucsi/displayport.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

@@ -166,12 +166,12 @@ static int ucsi_displayport_status_updat
 	 * that Multi-function is preferred.
 	 */
 	if (DP_CAP_CAPABILITY(cap) & DP_CAP_UFP_D) {
-		dp->data.status |= DP_STATUS_CON_UFP_D;
+		dp->data.status |= DP_STATUS_CON_DFP_D;
 
 		if (DP_CAP_UFP_D_PIN_ASSIGN(cap) & BIT(DP_PIN_ASSIGN_D))
 			dp->data.status |= DP_STATUS_PREFER_MULTI_FUNC;
 	} else {
-		dp->data.status |= DP_STATUS_CON_DFP_D;
+		dp->data.status |= DP_STATUS_CON_UFP_D;
 
 		if (DP_CAP_DFP_D_PIN_ASSIGN(cap) & BIT(DP_PIN_ASSIGN_D))
 			dp->data.status |= DP_STATUS_PREFER_MULTI_FUNC;

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Madhu M <[email protected]>

commit 735a461d060d4eeb2f9732aba1295bc32a3982e2 upstream.

In the UCSI Specification Revision 3.1 RC1, bits 32-63 of the SET_NEW_CAM command hold the 32-bit Alternate Mode Specific (AMSpecific) field.

For DisplayPort Alternate Mode, this field must contain the full 32-bit DisplayPort configuration VDO payload that the OPM wants the connector to operate in, rather than just the pin assignment value. This AMSpecific value follows the DisplayPort Configurations defined in the DisplayPort Alt Mode on USB Type-C Specification v2.1a, Table 5-13: SOP DisplayPort Configurations.

---
 drivers/usb/typec/ucsi/displayport.c |    3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

@@ -185,13 +185,12 @@ static int ucsi_displayport_status_updat
 
 static int ucsi_displayport_configure(struct ucsi_dp *dp)
 {
-	u32 pins = DP_CONF_GET_PIN_ASSIGN(dp->data.conf);
 	u64 command;
 
 	if (!dp->override)
 		return 0;
 
-	command = UCSI_CMD_SET_NEW_CAM(dp->con->num, 1, dp->offset, pins);
+	command = UCSI_CMD_SET_NEW_CAM(dp->con->num, 1, dp->offset, dp->data.conf);
 
 	return ucsi_send_command(dp->con->ucsi, command, NULL, 0);
 }

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Fan Wu <[email protected]>

commit 1f0bdc2884b67de337215079bba166df0cdf4ac5 upstream.

The threaded IRQ handler ccg_irq_handler() calls ucsi_notify_common(), which on a connector-change event calls ucsi_connector_change() and schedules connector work. In ucsi_ccg_remove(), ucsi_destroy() frees uc->ucsi (kfree) before free_irq() is called, so a handler invocation already in flight may access the freed object after ucsi_destroy().

  CPU 0 (remove)            | CPU 1 (threaded IRQ)
    ucsi_destroy(uc->ucsi)  |   ccg_irq_handler()
      kfree(ucsi) // FREE   |     ucsi_notify_common(uc->ucsi) // USE

Move free_irq() before ucsi_destroy() in the remove path. It is kept after ucsi_unregister(): ucsi_unregister() cancels connector work whose handler issues GET_CONNECTOR_STATUS through ucsi_send_command_common(), which waits for a completion that is signalled from the IRQ handler, so the IRQ must stay active until that work has been cancelled.

The probe error path already orders free_irq() before ucsi_destroy().

This bug was found by static analysis.

---
 drivers/usb/typec/ucsi/ucsi_ccg.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -1518,8 +1518,8 @@ static void ucsi_ccg_remove(struct i2c_c
 	cancel_work_sync(&uc->work);
 	pm_runtime_disable(uc->dev);
 	ucsi_unregister(uc->ucsi);
-	ucsi_destroy(uc->ucsi);
 	free_irq(uc->irq, uc);
+	ucsi_destroy(uc->ucsi);
 }
 
 static const struct of_device_id ucsi_ccg_of_match_table[] = {

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Paul Menzel <[email protected]>

commit 7c4a234bd31a64a8dbd0140dc812da592c5e0787 upstream.

On a Dell XPS 13 9360 (BIOS 2.21.0), entering system suspend (deep/S3) races a pending UCSI connector-change worker against the ACPI EC teardown. The worker evaluates the UCSI _DSM (GET_CONNECTOR_STATUS), whose AML accesses the Embedded Controller. By that point the ACPI EC has already been stopped for suspend, so the EC address space handler rejects the access with AE_BAD_PARAMETER, aborting the AML and failing the connector query:

    [22314.689495] ACPI: EC: interrupt blocked
    [22314.711981] ACPI: PM: Preparing to enter system sleep state S3
    [22314.743260] ACPI: EC: event blocked
    [22314.743265] ACPI: EC: EC stopped
    [22314.743267] ACPI: PM: Saving platform NVS memory
    [22314.744241] ACPI Error: AE_BAD_PARAMETER, Returned by Handler for [EmbeddedControl] (20260408/evregion-303)
    [22314.744432] ACPI Error: Aborting method \_SB.PCI0.LPCB.ECDV.ECW1 due to previous error (AE_BAD_PARAMETER) (20260408/psparse-543)
    [22314.744673] ACPI Error: Aborting method \ECWB due to previous error (AE_BAD_PARAMETER) (20260408/psparse-543)
    [22314.745201] ACPI Error: Aborting method \_SB.UBTC._DSM due to previous error (AE_BAD_PARAMETER) (20260408/psparse-543)
    [22314.745394] ACPI: \_SB_.UBTC: failed to evaluate _DSM c298836f-a47c-e411-ad36-631042b5008f rev:1 func:1 (0x1001)
    [22314.745414] ucsi_acpi USBC000:00: ucsi_acpi_dsm: failed to evaluate _DSM 1
    [22314.745424] ucsi_acpi USBC000:00: ucsi_handle_connector_change: GET_CONNECTOR_STATUS failed (-5)

ucsi_acpi implements a resume callback but no suspend callback, so nothing cancels the connector-change work before the firmware/EC is torn down.

Add a ucsi_suspend() core helper that cancels the pending init and connector-change work, and wire it into ucsi_acpi's PM ops. The connector state is re-read on resume by ucsi_resume(), so cancelling the work loses nothing.

---
 drivers/usb/typec/ucsi/ucsi.c      |   20 ++++++++++++++++++++
 drivers/usb/typec/ucsi/ucsi.h      |    1 +
 drivers/usb/typec/ucsi/ucsi_acpi.c |   10 +++++++++-
 3 files changed, 30 insertions(+), 1 deletion(-)

@@ -1918,6 +1918,26 @@ static void ucsi_resume_work(struct work
 	}
 }
 
+int ucsi_suspend(struct ucsi *ucsi)
+{
+	int i;
+
+	/*
+	 * Cancel pending work so it cannot access the firmware after the ACPI
+	 * EC is stopped for suspend; state is re-read on resume.
+	 */
+	cancel_delayed_work_sync(&ucsi->work);
+
+	if (!ucsi->connector)
+		return 0;
+
+	for (i = 0; i < ucsi->cap.num_connectors; i++)
+		cancel_work_sync(&ucsi->connector[i].work);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ucsi_suspend);
+
 int ucsi_resume(struct ucsi *ucsi)
 {
 	if (ucsi->connector)

@@ -551,6 +551,7 @@ int ucsi_send_command(struct ucsi *ucsi,
 		      void *retval, size_t size);
 
 void ucsi_altmode_update_active(struct ucsi_connector *con);
+int ucsi_suspend(struct ucsi *ucsi);
 int ucsi_resume(struct ucsi *ucsi);
 
 void ucsi_notify_common(struct ucsi *ucsi, u32 cci);

@@ -245,6 +245,13 @@ static void ucsi_acpi_remove(struct plat
 				   ucsi_acpi_notify);
 }
 
+static int ucsi_acpi_suspend(struct device *dev)
+{
+	struct ucsi_acpi *ua = dev_get_drvdata(dev);
+
+	return ucsi_suspend(ua->ucsi);
+}
+
 static int ucsi_acpi_resume(struct device *dev)
 {
 	struct ucsi_acpi *ua = dev_get_drvdata(dev);

@@ -252,7 +259,8 @@ static int ucsi_acpi_resume(struct devic
 	return ucsi_resume(ua->ucsi);
 }
 
-static DEFINE_SIMPLE_DEV_PM_OPS(ucsi_acpi_pm_ops, NULL, ucsi_acpi_resume);
+static DEFINE_SIMPLE_DEV_PM_OPS(ucsi_acpi_pm_ops, ucsi_acpi_suspend,
+				ucsi_acpi_resume);
 
 static const struct acpi_device_id ucsi_acpi_match[] = {
 	{ "PNP0CA0", 0 },

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Paul Cercueil <[email protected]>

commit baa6b6068a3f2bf2ed525a1cb37975905dadc658 upstream.

In ffs_dmabuf_transfer(), a ffs_dma_fence object is kmalloc'd, with the underlying dma_fence later initialized by dma_fence_init(), which sets its kref counter to 1. Then, dma_resv_add_fence() gets a second reference, and a pointer to the ffs_dma_fence is passed as the usb_request's "context" field.

The dma-resv mechanism will manage the second reference, but the first reference is never properly released; the ffs_dmabuf_cleanup() function decreases the reference count, but only to balance with the reference grab in ffs_dmabuf_signal_done().

The code will then slowly leak memory as more ffs_dma_fence objects are created without being ever freed.

Address this issue by transferring ownership of the fence to the DMA reservation object, by calling dma_fence_put() right after dma_resv_add_fence(). The ffs_dma_fence then gets properly discarded after being signalled.

---
 drivers/usb/gadget/function/f_fs.c |    1 +
 1 file changed, 1 insertion(+)

@@ -1684,6 +1684,7 @@ static int ffs_dmabuf_transfer(struct fi
 	resv_dir = epfile->in ? DMA_RESV_USAGE_READ : DMA_RESV_USAGE_WRITE;
 
 	dma_resv_add_fence(dmabuf->resv, &fence->base, resv_dir);
+	dma_fence_put(&fence->base);
 	dma_resv_unlock(dmabuf->resv);
 
 	/* Now that the dma_fence is in place, queue the transfer. */

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Chao Shi <[email protected]>

commit 49f06cff50a4ccf3b7a1a662ceb892b3b21a527a upstream.

bdev_mark_dead()'s @surprise == true means the device is already gone. The filesystem callback fs_bdev_mark_dead() honours this and skips sync_filesystem(), but the bare block device path (no ->mark_dead op) lost its !surprise guard when the holder ->mark_dead callback was wired up (see Fixes), and now calls sync_blockdev() unconditionally, which can hang forever waiting on writeback that can no longer complete.

syzkaller hit this via nvme_reset_work()'s "I/O queues lost" path:
nvme_mark_namespaces_dead() -> blk_mark_disk_dead() -> bdev_mark_dead(bdev, true) -> sync_blockdev() blocks in folio_wait_writeback(), wedging the reset worker and every task waiting on it.

Skip the sync on surprise removal, matching fs_bdev_mark_dead(); invalidate_bdev() still runs. Orderly removal (surprise == false) is unchanged.

Found by FuzzNvme(Syzkaller with FEMU fuzzing framework).

---
 block/bdev.c |    8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

@@ -1237,7 +1237,13 @@ void bdev_mark_dead(struct block_device
 		bdev->bd_holder_ops->mark_dead(bdev, surprise);
 	else {
 		mutex_unlock(&bdev->bd_holder_lock);
-		sync_blockdev(bdev);
+		/*
+		 * On surprise removal the device is already gone; syncing is
+		 * futile and can hang forever waiting on I/O that will never
+		 * complete.  Match fs_bdev_mark_dead(), which also skips it.
+		 */
+		if (!surprise)
+			sync_blockdev(bdev);
 	}
 
 	invalidate_bdev(bdev);

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Baolin Wang <[email protected]>

When skipping swapcache for synchronous IO swap devices, swapcache_prepare() is used to prevent parallel swapin from proceeding with the swap cache flag. However, on PREEMPT kernels this can lead to a livelock, as reported by Chao[1]:

Thread A starts direct swapin of a shmem folio and calls swapcache_prepare() to set SWAP_HAS_CACHE. It may then be preempted inside workingset_refault(). Meanwhile, a higher priority thread B also attempts direct swapin of the same shmem swap entry. Since swapcache_prepare() already marks the entry, thread B repeatedly gets -EEXIST and busy-loops waiting for thread A to finish. But as thread B runs at higher priority, thread A cannot preempt it, resulting in starvation and a livelock.

Fix it by yielding the CPU with schedule_timeout_uninterruptible(1) when swapcache_prepare() fails, following the same approach used in commit 029c4628b2eb ("mm: swap: get rid of livelock in swapin readahead") and commit 13ddaf26be32 ("mm/swap: fix race when skipping swapcache").

However, commit 01626a1823 ("mm: avoid unconditional one-tick sleep when swapcache_prepare fails") found that the unconditional one-tick sleep can cause UI stuttering on latency-sensitive Android devices. So we can follow the same approach by adding a waitqueue to wake up tasks when needed, instead of always sleeping for a full tick.

Note that mainline does not have this potential issue, which has already been resolved by Kairui's swap refactoring work[2].

[1] https://lore.kernel.org/all/[email protected]/
[2] https://lore.kernel.org/all/[email protected]/

---
 mm/shmem.c | 14 +++++++++++++-
 1 file changed, 13 insertions(+), 1 deletion(-)

index 94c5b0d78ac31e..3c329b794ae4df 100644
@@ -2005,11 +2005,14 @@ static struct folio *shmem_alloc_and_add_folio(struct vm_fault *vmf,
 	return ERR_PTR(error);
 }
 
+static DECLARE_WAIT_QUEUE_HEAD(shmem_swapcache_wq);
+
 static struct folio *shmem_swap_alloc_folio(struct inode *inode,
 		struct vm_area_struct *vma, pgoff_t index,
 		swp_entry_t entry, int order, gfp_t gfp)
 {
 	struct shmem_inode_info *info = SHMEM_I(inode);
+	DECLARE_WAITQUEUE(wait, current);
 	int nr_pages = 1 << order;
 	struct folio *new;
 	gfp_t alloc_gfp;

@@ -2066,6 +2069,10 @@ static struct folio *shmem_swap_alloc_folio(struct inode *inode,
 	if (swapcache_prepare(entry, nr_pages)) {
 		folio_put(new);
 		new = ERR_PTR(-EEXIST);
+		/* Relax a bit to prevent rapid repeated page faults */
+		add_wait_queue(&shmem_swapcache_wq, &wait);
+		schedule_timeout_uninterruptible(1);
+		remove_wait_queue(&shmem_swapcache_wq, &wait);
 		/* Try smaller folio to avoid cache conflict */
 		goto fallback;
 	}

@@ -2423,6 +2430,8 @@ static int shmem_swapin_folio(struct inode *inode, pgoff_t index,
 	if (skip_swapcache) {
 		folio->swap.val = 0;
 		swapcache_clear(si, swap, nr_pages);
+		if (waitqueue_active(&shmem_swapcache_wq))
+			wake_up(&shmem_swapcache_wq);
 	} else {
 		swap_cache_del_folio(folio);
 	}

@@ -2442,8 +2451,11 @@ static int shmem_swapin_folio(struct inode *inode, pgoff_t index,
 	if (folio)
 		folio_unlock(folio);
 failed_nolock:
-	if (skip_swapcache)
+	if (skip_swapcache) {
 		swapcache_clear(si, folio->swap, folio_nr_pages(folio));
+		if (waitqueue_active(&shmem_swapcache_wq))
+			wake_up(&shmem_swapcache_wq);
+	}
 	if (folio)
 		folio_put(folio);
 	put_swap_device(si);

2.53.0

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Reinette Chatre <[email protected]>

commit fc16126cc11d9f507130bf84ab137ee0938c900e upstream.

The architecture updates the cpu_mask in a domain's header to track which online CPUs are associated with the domain. When this mask becomes empty the architecture initiates offline of the domain that includes calling on resctrl fs to offline the domain. If it is a monitoring domain in which LLC occupancy is tracked resctrl fs forces the limbo handler to clear all busy RMID state associated with the domain.

The limbo handler always reads the current event value associated with a busy RMID irrespective of it being checked as part of regular "is it still busy" check or whether it will be forced released anyway. When reading an RMID on a system with SNC enabled the "logical RMID" is converted to the "physical RMID" and this conversion requires the NUMA node ID of the resctrl monitoring domain that is in turn determined by querying the NUMA node ID of any CPU belonging to the monitoring domain.

When the monitoring domain is going offline its cpu_mask is empty causing the NUMA node ID query via cpu_to_node() to be done with "nr_cpu_ids" as argument resulting in an out-of-bounds access.

Refactor the limbo handler to skip reading the RMID when the RMID will just be forced to no longer be dirty in the domain anyway. Add a safety check to the architecture's RMID reader to protect against this scenario.

---
 arch/x86/kernel/cpu/resctrl/monitor.c |  8 +++++-
 fs/resctrl/monitor.c                  | 39 +++++++++++++++------------
 2 files changed, 29 insertions(+), 18 deletions(-)

index fe1a2aa53c16ab..4b2b539ddff953 100644
@@ -243,14 +243,20 @@ int resctrl_arch_rmid_read(struct rdt_resource *r, struct rdt_mon_domain *d,
 			   u64 *val, void *ignored)
 {
 	struct rdt_hw_mon_domain *hw_dom = resctrl_to_arch_mon_dom(d);
-	int cpu = cpumask_any(&d->hdr.cpu_mask);
 	struct arch_mbm_state *am;
 	u64 msr_val;
 	u32 prmid;
+	int cpu;
 	int ret;
 
 	resctrl_arch_rmid_read_context_check();
 
+	if (cpumask_empty(&d->hdr.cpu_mask)) {
+		pr_warn_once("Domain %d has no CPUs\n", d->hdr.id);
+		return -EINVAL;
+	}
+
+	cpu = cpumask_any(&d->hdr.cpu_mask);
 	prmid = logical_rmid_to_physical_rmid(cpu, rmid);
 	ret = __rmid_read_phys(prmid, eventid, &msr_val);
 

index 572a9925bd6ca0..7012a75814ac36 100644
@@ -135,9 +135,9 @@ void __check_limbo(struct rdt_mon_domain *d, bool force_free)
 	struct rdt_resource *r = resctrl_arch_get_resource(RDT_RESOURCE_L3);
 	u32 idx_limit = resctrl_arch_system_num_rmid_idx();
 	struct rmid_entry *entry;
+	bool rmid_dirty = true;
 	u32 idx, cur_idx = 1;
 	void *arch_mon_ctx;
-	bool rmid_dirty;
 	u64 val = 0;
 
 	arch_mon_ctx = resctrl_arch_mon_ctx_alloc(r, QOS_L3_OCCUP_EVENT_ID);

@@ -159,22 +159,27 @@ void __check_limbo(struct rdt_mon_domain *d, bool force_free)
 			break;
 
 		entry = __rmid_entry(idx);
-		if (resctrl_arch_rmid_read(r, d, entry->closid, entry->rmid,
-					   QOS_L3_OCCUP_EVENT_ID, &val,
-					   arch_mon_ctx)) {
-			rmid_dirty = true;
-		} else {
-			rmid_dirty = (val >= resctrl_rmid_realloc_threshold);
-
-			/*
-			 * x86's CLOSID and RMID are independent numbers, so the entry's
-			 * CLOSID is an empty CLOSID (X86_RESCTRL_EMPTY_CLOSID). On Arm the
-			 * RMID (PMG) extends the CLOSID (PARTID) space with bits that aren't
-			 * used to select the configuration. It is thus necessary to track both
-			 * CLOSID and RMID because there may be dependencies between them
-			 * on some architectures.
-			 */
-			trace_mon_llc_occupancy_limbo(entry->closid, entry->rmid, d->hdr.id, val);
+		if (!force_free) {
+			if (resctrl_arch_rmid_read(r, d, entry->closid,
+						   entry->rmid, QOS_L3_OCCUP_EVENT_ID,
+						   &val, arch_mon_ctx)) {
+				rmid_dirty = true;
+			} else {
+				rmid_dirty = (val >= resctrl_rmid_realloc_threshold);
+
+				/*
+				 * x86's CLOSID and RMID are independent numbers,
+				 * so the entry's CLOSID is an empty CLOSID
+				 * (X86_RESCTRL_EMPTY_CLOSID). On Arm the RMID
+				 * (PMG) extends the CLOSID (PARTID) space with
+				 * bits that aren't used to select the configuration.
+				 * It is thus necessary to track both CLOSID and
+				 * RMID because there may be dependencies between
+				 * them on some architectures.
+				 */
+				trace_mon_llc_occupancy_limbo(entry->closid, entry->rmid,
+							      d->hdr.id, val);
+			}
 		}
 
 		if (force_free || !rmid_dirty) {

2.53.0

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Alice Ryhl <[email protected]>

[ Upstream commit 5326a18e3e640061ca4b65c1b732feaeace61c39 ]

Rust Binder exposes information about transactions that are sent in various ways: printing to the kernel log, tracepoints, files in binderfs, and the upcoming netlink support. Currently all these mechanisms use disparate ways of obtaining the same information, so let's introduce a single Info struct that collects all the required information in a single place, so that all of these different mechanisms can operate in a more uniform way.

For now, the new info struct is only used to replace a few things:
* The BinderTransactionDataSg struct that is passed as an argument to several methods is removed as the information is moved into the new info struct and passed down that way.
* The oneway spam detection fields on Transaction and Allocation can be removed, as the information can be returned to the caller via the mutable info struct instead.
But several other uses of the info struct are planned in follow-up patches.

---
 drivers/android/binder/allocation.rs  |   3 -
 drivers/android/binder/error.rs       |  10 +-
 drivers/android/binder/process.rs     |  15 +-
 drivers/android/binder/thread.rs      | 190 +++++++++++++++-----------
 drivers/android/binder/transaction.rs |  83 ++++++-----
 rust/kernel/uaccess.rs                |   3 +
 6 files changed, 170 insertions(+), 134 deletions(-)

index a9475e136d76fe..3aaeefef6bdd99 100644
@@ -56,7 +56,6 @@ pub(crate) struct Allocation {
     pub(crate) process: Arc<Process>,
     allocation_info: Option<AllocationInfo>,
     free_on_drop: bool,
-    pub(crate) oneway_spam_detected: bool,
     #[allow(dead_code)]
     pub(crate) debug_id: usize,
 }

@@ -68,7 +67,6 @@ impl Allocation {
         offset: usize,
         size: usize,
         ptr: usize,
-        oneway_spam_detected: bool,
     ) -> Self {
         Self {
             process,

@@ -76,7 +74,6 @@ impl Allocation {
             size,
             ptr,
             debug_id,
-            oneway_spam_detected,
             allocation_info: None,
             free_on_drop: true,
         }

index 9921827267d0d6..c6a834071f8cef 100644
@@ -12,7 +12,7 @@ pub(crate) type BinderResult<T = ()> = core::result::Result<T, BinderError>;
 /// errno.
 pub(crate) struct BinderError {
     pub(crate) reply: u32,
-    source: Option<Error>,
+    pub(crate) source: Option<Error>,
 }
 
 impl BinderError {

@@ -40,14 +40,6 @@ impl BinderError {
     pub(crate) fn is_dead(&self) -> bool {
         self.reply == BR_DEAD_REPLY
     }
-
-    pub(crate) fn as_errno(&self) -> kernel::ffi::c_int {
-        self.source.unwrap_or(EINVAL).to_errno()
-    }
-
-    pub(crate) fn should_pr_warn(&self) -> bool {
-        self.source.is_some()
-    }
 }
 
 /// Convert an errno into a `BinderError` and store the errno used to construct it. The errno

index 8ccc9ffbcd90e3..d33af8a043676c 100644
@@ -47,6 +47,7 @@ use crate::{
     range_alloc::{RangeAllocator, ReserveNew, ReserveNewArgs},
     stats::BinderStats,
     thread::{PushWorkRes, Thread},
+    transaction::TransactionInfo,
     BinderfsProcFile, DArc, DLArc, DTRWrap, DeliverToRead,
 };
 

@@ -981,16 +982,15 @@ impl Process {
         self: &Arc<Self>,
         debug_id: usize,
         size: usize,
-        is_oneway: bool,
-        from_pid: i32,
+        info: &mut TransactionInfo,
     ) -> BinderResult<NewAllocation> {
         use kernel::page::PAGE_SIZE;
 
         let mut reserve_new_args = ReserveNewArgs {
             debug_id,
             size,
-            is_oneway,
-            pid: from_pid,
+            is_oneway: info.is_oneway(),
+            pid: info.from_pid,
             ..ReserveNewArgs::default()
         };
 

@@ -1006,13 +1006,13 @@ impl Process {
             reserve_new_args = alloc_request.make_alloc()?;
         };
 
+        info.oneway_spam_suspect = new_alloc.oneway_spam_detected;
         let res = Allocation::new(
             self.clone(),
             debug_id,
             new_alloc.offset,
             size,
             addr + new_alloc.offset,
-            new_alloc.oneway_spam_detected,
         );
 
         // This allocation will be marked as in use until the `Allocation` is used to free it.

@@ -1044,7 +1044,7 @@ impl Process {
         let mapping = inner.mapping.as_mut()?;
         let offset = ptr.checked_sub(mapping.address)?;
         let (size, debug_id, odata) = mapping.alloc.reserve_existing(offset).ok()?;
-        let mut alloc = Allocation::new(self.clone(), debug_id, offset, size, ptr, false);
+        let mut alloc = Allocation::new(self.clone(), debug_id, offset, size, ptr);
         if let Some(data) = odata {
             alloc.set_info(data);
         }

@@ -1397,8 +1397,7 @@ impl Process {
                 .alloc
                 .take_for_each(|offset, size, debug_id, odata| {
                     let ptr = offset + address;
-                    let mut alloc =
-                        Allocation::new(self.clone(), debug_id, offset, size, ptr, false);
+                    let mut alloc = Allocation::new(self.clone(), debug_id, offset, size, ptr);
                     if let Some(data) = odata {
                         alloc.set_info(data);
                     }

index a8fb4764185a1a..2035c2877a2b2f 100644
@@ -21,6 +21,9 @@ use core::mem::{size_of, MaybeUninit};
 #[derive(Copy, Clone)]
 pub struct UserPtr(*mut c_void);
 
+// SAFETY: Null pointer is a valid value.
+unsafe impl Zeroable for UserPtr {}
+
 impl UserPtr {
     /// Create a `UserPtr` from an integer representing the userspace address.
     #[inline]

2.53.0

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Alice Ryhl <[email protected]>

[ Upstream commit 77bfebf110773f5a0d6b5ff8110896adb2c9c335 ]

This code currently copies the ExtendedError struct to the stack, modifies the copy, and then doesn't modify the original. Thus, fix it.

Furthermore, errors when replying must be delivered directly to the remote thread, so update deliver_reply() to take an extended error argument.

---
 drivers/android/binder/error.rs       | 13 +++---
 drivers/android/binder/thread.rs      | 65 +++++++++++++++++++--------
 drivers/android/binder/transaction.rs | 15 +++----
 3 files changed, 58 insertions(+), 35 deletions(-)

index c6a834071f8cef..fd9e9d9e5658b1 100644
@@ -72,20 +72,17 @@ impl core::fmt::Debug for BinderError {
     fn fmt(&self, f: &mut core::fmt::Formatter<'_>) -> core::fmt::Result {
         match self.reply {
             BR_FAILED_REPLY => match self.source.as_ref() {
-                Some(source) => f
-                    .debug_struct("BR_FAILED_REPLY")
-                    .field("source", source)
-                    .finish(),
+                Some(source) => source.fmt(f),
                 None => f.pad("BR_FAILED_REPLY"),
             },
             BR_DEAD_REPLY => f.pad("BR_DEAD_REPLY"),
             BR_FROZEN_REPLY => f.pad("BR_FROZEN_REPLY"),
             BR_TRANSACTION_PENDING_FROZEN => f.pad("BR_TRANSACTION_PENDING_FROZEN"),
             BR_TRANSACTION_COMPLETE => f.pad("BR_TRANSACTION_COMPLETE"),
-            _ => f
-                .debug_struct("BinderError")
-                .field("reply", &self.reply)
-                .finish(),
+            _ => match self.source.as_ref() {
+                Some(source) => source.fmt(f),
+                None => self.reply.fmt(f),
+            },
         }
     }
 }

index 733f452d8f49dd..b5434a2ae8193c 100644
@@ -42,6 +42,7 @@ pub(crate) struct TransactionInfo {
     pub(crate) reply: u32,
     pub(crate) oneway_spam_suspect: bool,
     pub(crate) is_reply: bool,
+    pub(crate) debug_id: usize,
 }
 
 impl TransactionInfo {

@@ -82,7 +83,6 @@ impl Transaction {
         from: &Arc<Thread>,
         info: &mut TransactionInfo,
     ) -> BinderResult<DLArc<Self>> {
-        let debug_id = super::next_debug_id();
         let allow_fds = node_ref.node.flags & FLAT_BINDER_FLAG_ACCEPTS_FDS != 0;
         let txn_security_ctx = node_ref.node.flags & FLAT_BINDER_FLAG_TXN_SECURITY_CTX != 0;
         let mut txn_security_ctx_off = if txn_security_ctx { Some(0) } else { None };

@@ -90,7 +90,7 @@ impl Transaction {
         let mut alloc = match from.copy_transaction_data(
             to.clone(),
             info,
-            debug_id,
+            info.debug_id,
             allow_fds,
             txn_security_ctx_off.as_mut(),
         ) {

@@ -117,7 +117,7 @@ impl Transaction {
         let data_address = alloc.ptr;
 
         Ok(DTRWrap::arc_pin_init(pin_init!(Transaction {
-            debug_id,
+            debug_id: info.debug_id,
             target_node: Some(target_node),
             from_parent,
             sender_euid: from.process.task.euid(),

@@ -141,9 +141,8 @@ impl Transaction {
         info: &mut TransactionInfo,
         allow_fds: bool,
     ) -> BinderResult<DLArc<Self>> {
-        let debug_id = super::next_debug_id();
         let mut alloc =
-            match from.copy_transaction_data(to.clone(), info, debug_id, allow_fds, None) {
+            match from.copy_transaction_data(to.clone(), info, info.debug_id, allow_fds, None) {
                 Ok(alloc) => alloc,
                 Err(err) => {
                     pr_warn!("Failure in copy_transaction_data: {:?}", err);

@@ -154,7 +153,7 @@ impl Transaction {
             alloc.set_info_clear_on_drop();
         }
         Ok(DTRWrap::arc_pin_init(pin_init!(Transaction {
-            debug_id,
+            debug_id: info.debug_id,
             target_node: None,
             from_parent: None,
             sender_euid: from.process.task.euid(),

@@ -380,7 +379,7 @@ impl DeliverToRead for Transaction {
         let send_failed_reply = ScopeGuard::new(|| {
             if self.target_node.is_some() && self.flags & TF_ONE_WAY == 0 {
                 let reply = Err(BR_FAILED_REPLY);
-                self.from.deliver_reply(reply, &self);
+                self.from.deliver_reply(reply, &self, None);
             }
             self.drop_outstanding_txn();
         });

@@ -462,7 +461,7 @@ impl DeliverToRead for Transaction {
         // If this is not a reply or oneway transaction, then send a dead reply.
         if self.target_node.is_some() && self.flags & TF_ONE_WAY == 0 {
             let reply = Err(BR_DEAD_REPLY);
-            self.from.deliver_reply(reply, &self);
+            self.from.deliver_reply(reply, &self, None);
         }
 
         self.drop_outstanding_txn();

2.53.0

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Pawan Gupta <[email protected]>

commit 96cce16e26dd02a8678f1e87f88a4b5cdb63b995 upstream.

The BPF JIT allocator packs many small programs into larger executable allocations and reuses space within those allocations as programs are loaded and freed. When fresh code is written into space that a previous program occupied, an indirect jump into the new program can reuse a branch prediction left behind by the old one.

Flush the indirect branch predictors before reusing JIT memory so that indirect jumps into a newly written program don't reuse predictions from an old program that occupied the same space.

Introduce bpf_arch_pred_flush_enabled static key and bpf_arch_pred_flush static call for flushing the branch predictors on JIT memory reuse. Architectures that need a flush, can update it to a predictor flush function. By default, its a NOP and does not emit any CALL.

Allocations larger than a pack are not covered by this flush. That is safe because cBPF programs (the unprivileged attack surface) are bounded well below a pack size. Issue a warning if this assumption is ever violated while the flush is active.

---
 include/linux/filter.h | 10 ++++++++++
 kernel/bpf/core.c      | 19 +++++++++++++++++++
 2 files changed, 29 insertions(+)

index cf7a0bce1bb6d8..3c4dd718beced8 100644
@@ -22,6 +22,7 @@
 #include <linux/vmalloc.h>
 #include <linux/sockptr.h>
 #include <crypto/sha1.h>
+#include <linux/static_call.h>
 #include <linux/u64_stats_sync.h>
 
 #include <net/sch_generic.h>

@@ -1266,6 +1267,15 @@ extern long bpf_jit_limit_max;
 
 typedef void (*bpf_jit_fill_hole_t)(void *area, unsigned int size);
 
+/*
+ * Flush the indirect branch predictors before reusing JIT memory, so that
+ * indirect jumps into a newly written program don't reuse predictions left
+ * behind by an old program that occupied the same space.
+ */
+void bpf_arch_pred_flush(void);
+DECLARE_STATIC_CALL(bpf_arch_pred_flush, bpf_arch_pred_flush);
+DECLARE_STATIC_KEY_FALSE(bpf_pred_flush_enabled);
+
 void bpf_jit_fill_hole_with_zero(void *area, unsigned int size);
 
 struct bpf_binary_header *

index 931a4ddd8530ce..a043fccb917e86 100644
@@ -21,6 +21,7 @@
 #include <crypto/sha1.h>
 #include <linux/filter.h>
 #include <linux/skbuff.h>
+#include <linux/static_call.h>
 #include <linux/vmalloc.h>
 #include <linux/prandom.h>
 #include <linux/bpf.h>

@@ -864,6 +865,15 @@ void bpf_jit_fill_hole_with_zero(void *area, unsigned int size)
 	memset(area, 0, size);
 }
 
+DEFINE_STATIC_CALL_NULL(bpf_arch_pred_flush, bpf_arch_pred_flush);
+
+/*
+ * Enabled once bpf_arch_pred_flush points at a real flush routine. Lets the
+ * pack allocator test "is a predictor flush wired up at all" with a cheap
+ * static branch instead of repeatedly querying the static call target.
+ */
+DEFINE_STATIC_KEY_FALSE(bpf_pred_flush_enabled);
+
 #define BPF_PROG_SIZE_TO_NBITS(size)	(round_up(size, BPF_PROG_CHUNK_SIZE) / BPF_PROG_CHUNK_SIZE)
 
 static DEFINE_MUTEX(pack_mutex);

@@ -923,6 +933,14 @@ void *bpf_prog_pack_alloc(u32 size, bpf_jit_fill_hole_t bpf_fill_ill_insns)
 
 	mutex_lock(&pack_mutex);
 	if (size > BPF_PROG_PACK_SIZE) {
+		/*
+		 * Allocations larger than a pack get their own pages, and
+		 * predictors are not flushed for such allocation. This is only
+		 * safe because cBPF programs (the unprivileged attack surface)
+		 * are bounded well below a pack size.
+		 */
+		if (static_branch_unlikely(&bpf_pred_flush_enabled))
+			pr_warn_once("BPF: Predictors not flushed for allocations greater than BPF_PROG_PACK_SIZE\n");
 		size = round_up(size, PAGE_SIZE);
 		ptr = bpf_jit_alloc_exec(size);
 		if (ptr) {

@@ -953,6 +971,7 @@ void *bpf_prog_pack_alloc(u32 size, bpf_jit_fill_hole_t bpf_fill_ill_insns)
 	pos = 0;
 
 found_free_area:
+	static_call_cond(bpf_arch_pred_flush)();
 	bitmap_set(pack->bitmap, pos, nbits);
 	ptr = (void *)(pack->ptr) + (pos << BPF_PROG_CHUNK_SHIFT);
 

2.53.0

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Pawan Gupta <[email protected]>

commit a3af84b0fa00ead01fcd0e28b5d773ff25990a0d upstream.

Enable hardening against JIT spraying when Spectre-v2 mitigations are in use. Specifically, issue an IBPB flush on BPF JIT memory reuse. Skip enabling the IBPB flush if the BPF dispatcher is already using a retpoline sequence.

This hardening applies only when BPF-JIT is in use. Guard the enabling under CONFIG_BPF_JIT so that bugs.c still builds with CONFIG_BPF_JIT=n.

---
 arch/x86/include/asm/nospec-branch.h |  4 +++
 arch/x86/kernel/cpu/bugs.c           | 50 +++++++++++++++++++++++++---
 2 files changed, 49 insertions(+), 5 deletions(-)

index 08ed5a2e46a5fd..71e9861a16a896 100644
@@ -386,6 +386,10 @@ extern void srso_alias_return_thunk(void);
 extern void entry_untrain_ret(void);
 extern void write_ibpb(void);
 
+#ifdef CONFIG_BPF_JIT
+extern void bpf_arch_ibpb(void);
+#endif
+
 #ifdef CONFIG_X86_64
 extern void clear_bhb_loop(void);
 #endif

2.53.0

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Pawan Gupta <[email protected]>

commit 0bb99f2cfaae6822d734d69722de30af823efdf3 upstream.

Currently predictor flush on memory reuse is done for all BPF JIT allocations, but only cBPF programs can be loaded by an unprivileged user. eBPF is privileged by default, and flushing predictors for all CPUs on every eBPF reuse penalizes the common case for no security benefit.

eBPF allocations can be frequent on busy systems, only flush predictors for cBPF programs. Trampoline and dispatcher allocations also skip the flush as they are eBPF-only.

[pawan: backport had various conflicts in arches bpf_int_jit_compile().
loongarch bpf_int_jit_compile() doesn't use pack allocator, dropped was_classic hunk. ]

---
 arch/arm64/net/bpf_jit_comp.c   |  4 ++--
 arch/loongarch/net/bpf_jit.c    |  2 +-
 arch/powerpc/net/bpf_jit_comp.c |  4 ++--
 arch/riscv/net/bpf_jit_comp64.c |  2 +-
 arch/riscv/net/bpf_jit_core.c   |  3 ++-
 arch/x86/net/bpf_jit_comp.c     |  5 +++--
 include/linux/filter.h          |  5 +++--
 kernel/bpf/core.c               | 13 ++++++++-----
 kernel/bpf/dispatcher.c         |  2 +-
 9 files changed, 23 insertions(+), 17 deletions(-)

index 873c1b784a872d..c563bae8fce1c2 100644
@@ -2117,7 +2117,7 @@ struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *prog)
 	image_size = extable_offset + extable_size;
 	ro_header = bpf_jit_binary_pack_alloc(image_size, &ro_image_ptr,
 					      sizeof(u64), &header, &image_ptr,
-					      jit_fill_hole);
+					      jit_fill_hole, was_classic);
 	if (!ro_header) {
 		prog = orig_prog;
 		goto out_off;

@@ -2754,7 +2754,7 @@ int arch_bpf_trampoline_size(const struct btf_func_model *m, u32 flags,
 
 void *arch_alloc_bpf_trampoline(unsigned int size)
 {
-	return bpf_prog_pack_alloc(size, jit_fill_hole);
+	return bpf_prog_pack_alloc(size, jit_fill_hole, false);
 }
 
 void arch_free_bpf_trampoline(void *image, unsigned int size)

index e9d666508ae28f..a000620ca99da7 100644
@@ -1474,7 +1474,7 @@ static void invoke_bpf_mod_ret(struct jit_ctx *ctx, struct bpf_tramp_links *tl,
 
 void *arch_alloc_bpf_trampoline(unsigned int size)
 {
-	return bpf_prog_pack_alloc(size, jit_fill_hole);
+	return bpf_prog_pack_alloc(size, jit_fill_hole, false);
 }
 
 void arch_free_bpf_trampoline(void *image, unsigned int size)

index 189ef7b72081d2..cedab25aa3a823 100644
@@ -246,7 +246,7 @@ struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *fp)
 	alloclen = proglen + FUNCTION_DESCR_SIZE + fixup_len + extable_len;
 
 	fhdr = bpf_jit_binary_pack_alloc(alloclen, &fimage, 4, &hdr, &image,
-					      bpf_jit_fill_ill_insns);
+					 bpf_jit_fill_ill_insns, bpf_prog_was_classic(fp));
 	if (!fhdr) {
 		fp = org_fp;
 		goto out_addrs;

@@ -468,7 +468,7 @@ bool bpf_jit_supports_insn(struct bpf_insn *insn, bool in_arena)
 
 void *arch_alloc_bpf_trampoline(unsigned int size)
 {
-	return bpf_prog_pack_alloc(size, bpf_jit_fill_ill_insns);
+	return bpf_prog_pack_alloc(size, bpf_jit_fill_ill_insns, false);
 }
 
 void arch_free_bpf_trampoline(void *image, unsigned int size)

index 45cbc7c6fe490c..9e9e6dcfc48254 100644
@@ -1268,7 +1268,7 @@ int arch_bpf_trampoline_size(const struct btf_func_model *m, u32 flags,
 
 void *arch_alloc_bpf_trampoline(unsigned int size)
 {
-	return bpf_prog_pack_alloc(size, bpf_fill_ill_insns);
+	return bpf_prog_pack_alloc(size, bpf_fill_ill_insns, false);
 }
 
 void arch_free_bpf_trampoline(void *image, unsigned int size)

index 191bf0e66c8243..e4ab5bb9c9f646 100644
@@ -125,7 +125,8 @@ struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *prog)
 				bpf_jit_binary_pack_alloc(prog_size + extable_size,
 							  &jit_data->ro_image, sizeof(u32),
 							  &jit_data->header, &jit_data->image,
-							  bpf_fill_ill_insns);
+							  bpf_fill_ill_insns,
+							  bpf_prog_was_classic(prog));
 			if (!jit_data->ro_header) {
 				prog = orig_prog;
 				goto out_offset;

index 788671a32d8ee0..c3798cab3b7db0 100644
@@ -3447,7 +3447,7 @@ static int __arch_prepare_bpf_trampoline(struct bpf_tramp_image *im, void *rw_im
 
 void *arch_alloc_bpf_trampoline(unsigned int size)
 {
-	return bpf_prog_pack_alloc(size, jit_fill_hole);
+	return bpf_prog_pack_alloc(size, jit_fill_hole, false);
 }
 
 void arch_free_bpf_trampoline(void *image, unsigned int size)

@@ -3780,7 +3780,8 @@ struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *prog)
 			/* allocate module memory for x86 insns and extable */
 			header = bpf_jit_binary_pack_alloc(roundup(proglen, align) + extable_size,
 							   &image, align, &rw_header, &rw_image,
-							   jit_fill_hole);
+							   jit_fill_hole,
+							   bpf_prog_was_classic(prog));
 			if (!header) {
 				prog = orig_prog;
 				goto out_addrs;

index 3c4dd718beced8..2469fd2e401578 100644
@@ -1290,7 +1290,7 @@ void bpf_jit_free(struct bpf_prog *fp);
 struct bpf_binary_header *
 bpf_jit_binary_pack_hdr(const struct bpf_prog *fp);
 
-void *bpf_prog_pack_alloc(u32 size, bpf_jit_fill_hole_t bpf_fill_ill_insns);
+void *bpf_prog_pack_alloc(u32 size, bpf_jit_fill_hole_t bpf_fill_ill_insns, bool was_classic);
 void bpf_prog_pack_free(void *ptr, u32 size);
 
 static inline bool bpf_prog_kallsyms_verify_off(const struct bpf_prog *fp)

@@ -1304,7 +1304,8 @@ bpf_jit_binary_pack_alloc(unsigned int proglen, u8 **ro_image,
 			  unsigned int alignment,
 			  struct bpf_binary_header **rw_hdr,
 			  u8 **rw_image,
-			  bpf_jit_fill_hole_t bpf_fill_ill_insns);
+			  bpf_jit_fill_hole_t bpf_fill_ill_insns,
+			  bool was_classic);
 int bpf_jit_binary_pack_finalize(struct bpf_binary_header *ro_header,
 				 struct bpf_binary_header *rw_header);
 void bpf_jit_binary_pack_free(struct bpf_binary_header *ro_header,

index a043fccb917e86..338a806e25c045 100644
@@ -924,7 +924,7 @@ static struct bpf_prog_pack *alloc_new_pack(bpf_jit_fill_hole_t bpf_fill_ill_ins
 	return NULL;
 }
 
-void *bpf_prog_pack_alloc(u32 size, bpf_jit_fill_hole_t bpf_fill_ill_insns)
+void *bpf_prog_pack_alloc(u32 size, bpf_jit_fill_hole_t bpf_fill_ill_insns, bool was_classic)
 {
 	unsigned int nbits = BPF_PROG_SIZE_TO_NBITS(size);
 	struct bpf_prog_pack *pack;

@@ -939,7 +939,7 @@ void *bpf_prog_pack_alloc(u32 size, bpf_jit_fill_hole_t bpf_fill_ill_insns)
 		 * safe because cBPF programs (the unprivileged attack surface)
 		 * are bounded well below a pack size.
 		 */
-		if (static_branch_unlikely(&bpf_pred_flush_enabled))
+		if (was_classic && static_branch_unlikely(&bpf_pred_flush_enabled))
 			pr_warn_once("BPF: Predictors not flushed for allocations greater than BPF_PROG_PACK_SIZE\n");
 		size = round_up(size, PAGE_SIZE);
 		ptr = bpf_jit_alloc_exec(size);

@@ -971,7 +971,9 @@ void *bpf_prog_pack_alloc(u32 size, bpf_jit_fill_hole_t bpf_fill_ill_insns)
 	pos = 0;
 
 found_free_area:
-	static_call_cond(bpf_arch_pred_flush)();
+	/* Flush only for cBPF as it may contain a crafted gadget */
+	if (static_branch_unlikely(&bpf_pred_flush_enabled) && was_classic)
+		static_call_cond(bpf_arch_pred_flush)();
 	bitmap_set(pack->bitmap, pos, nbits);
 	ptr = (void *)(pack->ptr) + (pos << BPF_PROG_CHUNK_SHIFT);
 

@@ -1131,7 +1133,8 @@ bpf_jit_binary_pack_alloc(unsigned int proglen, u8 **image_ptr,
 			  unsigned int alignment,
 			  struct bpf_binary_header **rw_header,
 			  u8 **rw_image,
-			  bpf_jit_fill_hole_t bpf_fill_ill_insns)
+			  bpf_jit_fill_hole_t bpf_fill_ill_insns,
+			  bool was_classic)
 {
 	struct bpf_binary_header *ro_header;
 	u32 size, hole, start;

@@ -1144,7 +1147,7 @@ bpf_jit_binary_pack_alloc(unsigned int proglen, u8 **image_ptr,
 
 	if (bpf_jit_charge_modmem(size))
 		return NULL;
-	ro_header = bpf_prog_pack_alloc(size, bpf_fill_ill_insns);
+	ro_header = bpf_prog_pack_alloc(size, bpf_fill_ill_insns, was_classic);
 	if (!ro_header) {
 		bpf_jit_uncharge_modmem(size);
 		return NULL;

index b77db7413f8c70..ea2d60dc1feeb7 100644
@@ -145,7 +145,7 @@ void bpf_dispatcher_change_prog(struct bpf_dispatcher *d, struct bpf_prog *from,
 
 	mutex_lock(&d->mutex);
 	if (!d->image) {
-		d->image = bpf_prog_pack_alloc(PAGE_SIZE, bpf_jit_fill_hole_with_zero);
+		d->image = bpf_prog_pack_alloc(PAGE_SIZE, bpf_jit_fill_hole_with_zero, false);
 		if (!d->image)
 			goto out;
 		d->rw_image = bpf_jit_alloc_exec(PAGE_SIZE);

2.53.0

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Pawan Gupta <[email protected]>

commit a23c1c5396a91680703360d1ee28a44657c503c4 upstream.

bpf_prog_pack_alloc() issues IBPB on all CPUs on every cBPF allocation, even when reusing chunks from an existing pack where no new memory was touched since the last IBPB.

Since IBPB on all CPUs is heavy, Dave Hansen suggested to track allocation since last IBPB, and only issue IBPB at reuse for the chunks that have not seen an IBPB since they were last freed.

Track per-pack whether an IBPB is needed via arch_flush_needed. Set it when allocating a chunk, reset on IBPB flush. On reuse, conditionally issue the flush. Since IBPB invalidates all BTB entries, clear the flag on all packs after flushing.

---
 kernel/bpf/core.c | 15 ++++++++++++++-
 1 file changed, 14 insertions(+), 1 deletion(-)

index 338a806e25c045..49ca543a41115a 100644
@@ -857,6 +857,7 @@ int bpf_jit_add_poke_descriptor(struct bpf_prog *prog,
 struct bpf_prog_pack {
 	struct list_head list;
 	void *ptr;
+	bool arch_flush_needed;
 	unsigned long bitmap[];
 };
 

@@ -910,6 +911,8 @@ static struct bpf_prog_pack *alloc_new_pack(bpf_jit_fill_hole_t bpf_fill_ill_ins
 	bpf_fill_ill_insns(pack->ptr, BPF_PROG_PACK_SIZE);
 	bitmap_zero(pack->bitmap, BPF_PROG_PACK_SIZE / BPF_PROG_CHUNK_SIZE);
 
+	if (static_branch_unlikely(&bpf_pred_flush_enabled))
+		pack->arch_flush_needed = true;
 	set_vm_flush_reset_perms(pack->ptr);
 	err = set_memory_rox((unsigned long)pack->ptr,
 			     BPF_PROG_PACK_SIZE / PAGE_SIZE);

@@ -972,8 +975,15 @@ void *bpf_prog_pack_alloc(u32 size, bpf_jit_fill_hole_t bpf_fill_ill_insns, bool
 
 found_free_area:
 	/* Flush only for cBPF as it may contain a crafted gadget */
-	if (static_branch_unlikely(&bpf_pred_flush_enabled) && was_classic)
+	if (static_branch_unlikely(&bpf_pred_flush_enabled) &&
+	    pack->arch_flush_needed &&
+	    was_classic) {
+		struct bpf_prog_pack *p;
+
 		static_call_cond(bpf_arch_pred_flush)();
+		list_for_each_entry(p, &pack_list, list)
+			p->arch_flush_needed = false;
+	}
 	bitmap_set(pack->bitmap, pos, nbits);
 	ptr = (void *)(pack->ptr) + (pos << BPF_PROG_CHUNK_SHIFT);
 

@@ -1011,6 +1021,9 @@ void bpf_prog_pack_free(void *ptr, u32 size)
 		  "bpf_prog_pack bug: missing bpf_arch_text_invalidate?\n");
 
 	bitmap_clear(pack->bitmap, pos, nbits);
+
+	if (static_branch_unlikely(&bpf_pred_flush_enabled))
+		pack->arch_flush_needed = true;
 	if (bitmap_find_next_zero_area(pack->bitmap, BPF_PROG_CHUNK_COUNT, 0,
 				       BPF_PROG_CHUNK_COUNT, 0) == 0) {
 		list_del(&pack->list);

2.53.0

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Pawan Gupta <[email protected]>

commit a9b1f19a6a673ba06820898d0f1ad02883ea1639 upstream.

Currently BPF pack allocator picks the chunks from the first available pack. While this is okay, it naturally leads to more frequent flushes when there are multiple packs in the system that weren't used since the last flush.

As an optimization prefer allocating the new programs from packs that are unused since last flush. When all packs are dirty, allocation forces a flush and marks all packs clean.

Below are some future optimizations ideas:

1. Currently, the "dirty" tracking is only done at the pack-level. Flush frequency can further be reduced with chunk-level tracking. This requires a new bitmap per-pack to track the dirty state.
2. IBPB flush is done on all CPUs, even if only a single CPU ran the BPF program. On a system with hundreds of CPUs this could be a major bottleneck forcing hundreds of IPIs to deliver the flush. The solution is to track the CPUs where a BPF program ran, and issue IBPB only on those CPUs.
3. Avoid IBPB when flush is already done at other sources (e.g. context switch).

---
 kernel/bpf/core.c | 27 ++++++++++++++++++++++++---
 1 file changed, 24 insertions(+), 3 deletions(-)

index 49ca543a41115a..e0ffdd985eb5d7 100644
@@ -930,8 +930,8 @@ static struct bpf_prog_pack *alloc_new_pack(bpf_jit_fill_hole_t bpf_fill_ill_ins
 void *bpf_prog_pack_alloc(u32 size, bpf_jit_fill_hole_t bpf_fill_ill_insns, bool was_classic)
 {
 	unsigned int nbits = BPF_PROG_SIZE_TO_NBITS(size);
-	struct bpf_prog_pack *pack;
-	unsigned long pos;
+	struct bpf_prog_pack *pack, *fallback_pack = NULL;
+	unsigned long pos, fallback_pos = 0;
 	void *ptr = NULL;
 
 	mutex_lock(&pack_mutex);

@@ -963,8 +963,29 @@ void *bpf_prog_pack_alloc(u32 size, bpf_jit_fill_hole_t bpf_fill_ill_insns, bool
 	list_for_each_entry(pack, &pack_list, list) {
 		pos = bitmap_find_next_zero_area(pack->bitmap, BPF_PROG_CHUNK_COUNT, 0,
 						 nbits, 0);
-		if (pos < BPF_PROG_CHUNK_COUNT)
+		if (pos >= BPF_PROG_CHUNK_COUNT)
+			continue;
+		/* Flush not enabled, use any pack */
+		if (!static_branch_unlikely(&bpf_pred_flush_enabled))
 			goto found_free_area;
+		/*
+		 * cBPF reuse of a dirty pack triggers a flush, so prefer a
+		 * clean pack for cBPF. eBPF never flushes, so pick the first
+		 * free pack, dirty or clean.
+		 */
+		if (!was_classic || !pack->arch_flush_needed)
+			goto found_free_area;
+		if (!fallback_pack) {
+			fallback_pack = pack;
+			fallback_pos = pos;
+		}
+	}
+
+	/* No preferred pack found */
+	if (fallback_pack) {
+		pack = fallback_pack;
+		pos = fallback_pos;
+		goto found_free_area;
 	}
 
 	pack = alloc_new_pack(bpf_fill_ill_insns);

2.53.0

6.18-stable review patch. If anyone has any objections, please let me know.

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

From: Pawan Gupta <[email protected]>

commit b72e29e0f7ee329d89f86db8700c8ea99b4a370a upstream.

The pack allocator only flushes predictors when reusing a dirty pack for cBPF, eBPF allocations never trigger a flush. Currently, eBPF picks the first free pack, which could be a clean pack. As an optimization, leaving a clean pack for cBPF can avoid flushes.

Prefer dirty packs for eBPF and keep clean packs free for cBPF. This mirrors the existing cBPF preference for clean packs: each program kind prefers the pack that avoids an extra flush, and falls back to the other kind only when no preferred pack has room. eBPF reuse of a dirty pack is harmless since eBPF being privileged does not flush.

---
 kernel/bpf/core.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

index e0ffdd985eb5d7..b102704f89bc77 100644
@@ -970,10 +970,10 @@ void *bpf_prog_pack_alloc(u32 size, bpf_jit_fill_hole_t bpf_fill_ill_insns, bool
 			goto found_free_area;
 		/*
 		 * cBPF reuse of a dirty pack triggers a flush, so prefer a
-		 * clean pack for cBPF. eBPF never flushes, so pick the first
-		 * free pack, dirty or clean.
+		 * clean pack for cBPF. eBPF never flushes, so steer it to a
+		 * dirty pack and keep clean packs free for cBPF.
 		 */
-		if (!was_classic || !pack->arch_flush_needed)
+		if (was_classic ^ pack->arch_flush_needed)
 			goto found_free_area;
 		if (!fallback_pack) {
 			fallback_pack = pack;