Skip to contentSkip to conversation

5.4.118-rc1 review

PATCH In review Greg Kroah-Hartman 10 participants 202 messages May 10, 2021 – Jun 18, 2021 Page 2 of 2

Subsystems (4)
lore
Share & tools
lore
author maintainer
#173 lore raw

PATCH 163/184 ext4: fix check to prevent false positive report of incorrect used inodes

In reply to Greg Kroah-Hartman

From: Zhang Yi <[email protected]>

commit a149d2a5cabbf6507a7832a1c4fd2593c55fd450 upstream.

Commit <50122847007> ("ext4: fix check to prevent initializing reserved inodes") check the block group zero and prevent initializing reserved inodes. But in some special cases, the reserved inode may not all belong to the group zero, it may exist into the second group if we format filesystem below.

mkfs.ext4 -b 4096 -g 8192 -N 1024 -I 4096 /dev/sda

So, it will end up triggering a false positive report of a corrupted file system. This patch fix it by avoid check reserved inodes if no free inode blocks will be zeroed.

Fixes 50122847007 ("ext4: fix check to prevent initializing reserved inodes")
---
 fs/ext4/ialloc.c |   48 ++++++++++++++++++++++++++++++++----------------
 1 file changed, 32 insertions(+), 16 deletions(-)
fs/ext4/ialloc.c +32 -16
@@ -1353,6 +1353,7 @@ int ext4_init_inode_table(struct super_b
 	handle_t *handle;
 	ext4_fsblk_t blk;
 	int num, ret = 0, used_blks = 0;
+	unsigned long used_inos = 0;
 
 	/* This should not happen, but just to be sure check this */
 	if (sb_rdonly(sb)) {
@@ -1383,22 +1384,37 @@ int ext4_init_inode_table(struct super_b
 	 * used inodes so we need to skip blocks with used inodes in
 	 * inode table.
 	 */
-	if (!(gdp->bg_flags & cpu_to_le16(EXT4_BG_INODE_UNINIT)))
-		used_blks = DIV_ROUND_UP((EXT4_INODES_PER_GROUP(sb) -
-			    ext4_itable_unused_count(sb, gdp)),
-			    sbi->s_inodes_per_block);
-
-	if ((used_blks < 0) || (used_blks > sbi->s_itb_per_group) ||
-	    ((group == 0) && ((EXT4_INODES_PER_GROUP(sb) -
-			       ext4_itable_unused_count(sb, gdp)) <
-			      EXT4_FIRST_INO(sb)))) {
-		ext4_error(sb, "Something is wrong with group %u: "
-			   "used itable blocks: %d; "
-			   "itable unused count: %u",
-			   group, used_blks,
-			   ext4_itable_unused_count(sb, gdp));
-		ret = 1;
-		goto err_out;
+	if (!(gdp->bg_flags & cpu_to_le16(EXT4_BG_INODE_UNINIT))) {
+		used_inos = EXT4_INODES_PER_GROUP(sb) -
+			    ext4_itable_unused_count(sb, gdp);
+		used_blks = DIV_ROUND_UP(used_inos, sbi->s_inodes_per_block);
+
+		/* Bogus inode unused count? */
+		if (used_blks < 0 || used_blks > sbi->s_itb_per_group) {
+			ext4_error(sb, "Something is wrong with group %u: "
+				   "used itable blocks: %d; "
+				   "itable unused count: %u",
+				   group, used_blks,
+				   ext4_itable_unused_count(sb, gdp));
+			ret = 1;
+			goto err_out;
+		}
+
+		used_inos += group * EXT4_INODES_PER_GROUP(sb);
+		/*
+		 * Are there some uninitialized inodes in the inode table
+		 * before the first normal inode?
+		 */
+		if ((used_blks != sbi->s_itb_per_group) &&
+		     (used_inos < EXT4_FIRST_INO(sb))) {
+			ext4_error(sb, "Something is wrong with group %u: "
+				   "itable unused count: %u; "
+				   "itables initialized count: %ld",
+				   group, ext4_itable_unused_count(sb, gdp),
+				   used_inos);
+			ret = 1;
+			goto err_out;
+		}
 	}
 
 	blk = ext4_inode_table(sb, gdp) + used_blks;
author maintainer
#174 lore raw

PATCH 164/184 ext4: do not set SB_ACTIVE in ext4_orphan_cleanup()

In reply to Greg Kroah-Hartman

From: Zhang Yi <[email protected]>

commit 72ffb49a7b623c92a37657eda7cc46a06d3e8398 upstream.

When CONFIG_QUOTA is enabled, if we failed to mount the filesystem due to some error happens behind ext4_orphan_cleanup(), it will end up triggering a after free issue of super_block. The problem is that ext4_orphan_cleanup() will set SB_ACTIVE flag if CONFIG_QUOTA is enabled, after we cleanup the truncated inodes, the last iput() will put them into the lru list, and these inodes' pages may probably dirty and will be write back by the writeback thread, so it could be raced by freeing super_block in the error path of mount_bdev().

After check the setting of SB_ACTIVE flag in ext4_orphan_cleanup(), it was used to ensure updating the quota file properly, but evict inode and trash data immediately in the last iput does not affect the quotafile, so setting the SB_ACTIVE flag seems not required[1]. Fix this issue by just remove the SB_ACTIVE setting.

[1] https://lore.kernel.org/linux-ext4/[email protected]/T/#m04990cfbc4f44592421736b504afcc346b2a7c00

---
 fs/ext4/super.c |    3 ---
 1 file changed, 3 deletions(-)
fs/ext4/super.c +0 -3
@@ -2675,9 +2675,6 @@ static void ext4_orphan_cleanup(struct s
 		sb->s_flags &= ~SB_RDONLY;
 	}
 #ifdef CONFIG_QUOTA
-	/* Needed for iput() to work correctly and not trash data */
-	sb->s_flags |= SB_ACTIVE;
-
 	/*
 	 * Turn on quotas which were not enabled for read-only mounts if
 	 * filesystem has quota feature, so that they are updated correctly.
author maintainer
#175 lore raw

PATCH 165/184 ext4: fix error code in ext4_commit_super

In reply to Greg Kroah-Hartman

From: Fengnan Chang <[email protected]>

commit f88f1466e2a2e5ca17dfada436d3efa1b03a3972 upstream.

We should set the error code when ext4_commit_super check argument failed. Found in code review.

Fixes c4be0c1dc4cdc ("filesystem freeze: add error handling of write_super_lockfs/unlockfs").
---
 fs/ext4/super.c |    6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
fs/ext4/super.c +4 -2
@@ -5057,8 +5057,10 @@ static int ext4_commit_super(struct supe
 	struct buffer_head *sbh = EXT4_SB(sb)->s_sbh;
 	int error = 0;
 
-	if (!sbh || block_device_ejected(sb))
-		return error;
+	if (!sbh)
+		return -EINVAL;
+	if (block_device_ejected(sb))
+		return -ENODEV;
 
 	/*
 	 * If the file system is mounted read-only, don't update the
author maintainer
#176 lore raw

PATCH 166/184 media: dvbdev: Fix memory leak in dvb_media_device_free()

In reply to Greg Kroah-Hartman

From: Peilin Ye <[email protected]>

commit bf9a40ae8d722f281a2721779595d6df1c33a0bf upstream.

dvb_media_device_free() is leaking memory. Free dvbdev->adapter->conn before setting it to NULL, as documented in include/media/media-device.h:
"The media_entity instance itself must be freed explicitly by the driver if required."

---
 drivers/media/dvb-core/dvbdev.c |    1 +
 1 file changed, 1 insertion(+)
drivers/media/dvb-core/dvbdev.c +1 -0
@@ -241,6 +241,7 @@ static void dvb_media_device_free(struct
 
 	if (dvbdev->adapter->conn) {
 		media_device_unregister_entity(dvbdev->adapter->conn);
+		kfree(dvbdev->adapter->conn);
 		dvbdev->adapter->conn = NULL;
 		kfree(dvbdev->adapter->conn_pads);
 		dvbdev->adapter->conn_pads = NULL;
author maintainer
#177 lore raw

PATCH 167/184 media: dvb-usb: Fix use-after-free access

In reply to Greg Kroah-Hartman

From: Takashi Iwai <[email protected]>

commit c49206786ee252f28b7d4d155d1fff96f145a05d upstream.

dvb_usb_device_init() copies the properties to the own data, so that the callers can release the original properties later (as done in the commit 299c7007e936 ("media: dw2102: Fix memleak on sequence of probes")). However, it also stores dev->desc pointer that is a reference to the original properties data. Since dev->desc is referred later, it may result in use-after-free, in the worst case, leading to a kernel Oops as reported.

This patch addresses the problem by allocating and copying the properties at first, then get the desc from the copied properties.

---
 drivers/media/usb/dvb-usb/dvb-usb-init.c |   23 +++++++++++++----------
 1 file changed, 13 insertions(+), 10 deletions(-)
drivers/media/usb/dvb-usb/dvb-usb-init.c +13 -10
@@ -267,27 +267,30 @@ int dvb_usb_device_init(struct usb_inter
 	if (du != NULL)
 		*du = NULL;
 
-	if ((desc = dvb_usb_find_device(udev, props, &cold)) == NULL) {
+	d = kzalloc(sizeof(*d), GFP_KERNEL);
+	if (!d) {
+		err("no memory for 'struct dvb_usb_device'");
+		return -ENOMEM;
+	}
+
+	memcpy(&d->props, props, sizeof(struct dvb_usb_device_properties));
+
+	desc = dvb_usb_find_device(udev, &d->props, &cold);
+	if (!desc) {
 		deb_err("something went very wrong, device was not found in current device list - let's see what comes next.\n");
-		return -ENODEV;
+		ret = -ENODEV;
+		goto error;
 	}
 
 	if (cold) {
 		info("found a '%s' in cold state, will try to load a firmware", desc->name);
 		ret = dvb_usb_download_firmware(udev, props);
 		if (!props->no_reconnect || ret != 0)
-			return ret;
+			goto error;
 	}
 
 	info("found a '%s' in warm state.", desc->name);
-	d = kzalloc(sizeof(struct dvb_usb_device), GFP_KERNEL);
-	if (d == NULL) {
-		err("no memory for 'struct dvb_usb_device'");
-		return -ENOMEM;
-	}
-
 	d->udev = udev;
-	memcpy(&d->props, props, sizeof(struct dvb_usb_device_properties));
 	d->desc = desc;
 	d->owner = owner;
 
author maintainer
#178 lore raw

PATCH 168/184 media: dvb-usb: Fix memory leak at error in dvb_usb_device_init()

In reply to Greg Kroah-Hartman

From: Takashi Iwai <[email protected]>

commit 13a79f14ab285120bc4977e00a7c731e8143f548 upstream.

dvb_usb_device_init() allocates a dvb_usb_device object, but it doesn't release the object by itself even at errors. The object is released in the callee side (dvb_usb_init()) in some error cases via dvb_usb_exit() call, but it also missed the object free in other error paths. And, the caller (it's only dvb_usb_device_init()) doesn't seem caring the resource management as well, hence those memories are leaked.

This patch assures releasing the memory at the error path in dvb_usb_device_init(). Now dvb_usb_init() frees the resources it allocated but leaves the passed dvb_usb_device object intact. In turn, the dvb_usb_device object is released in dvb_usb_device_init() instead.
We could use dvb_usb_exit() function for releasing everything in the callee (as it was used for some error cases in the original code), but releasing the passed object in the callee is non-intuitive and error-prone. So I took this approach (which is more standard in Linus kernel code) although it ended with a bit more open codes.

Along with the change, the patch makes sure that USB intfdata is reset and don't return the bogus pointer to the caller of dvb_usb_device_init() at the error path, too.

---
 drivers/media/usb/dvb-usb/dvb-usb-init.c |   47 ++++++++++++++++++++-----------
 1 file changed, 31 insertions(+), 16 deletions(-)
drivers/media/usb/dvb-usb/dvb-usb-init.c +31 -16
@@ -170,22 +170,20 @@ static int dvb_usb_init(struct dvb_usb_d
 
 		if (d->props.priv_init != NULL) {
 			ret = d->props.priv_init(d);
-			if (ret != 0) {
-				kfree(d->priv);
-				d->priv = NULL;
-				return ret;
-			}
+			if (ret != 0)
+				goto err_priv_init;
 		}
 	}
 
 	/* check the capabilities and set appropriate variables */
 	dvb_usb_device_power_ctrl(d, 1);
 
-	if ((ret = dvb_usb_i2c_init(d)) ||
-		(ret = dvb_usb_adapter_init(d, adapter_nums))) {
-		dvb_usb_exit(d);
-		return ret;
-	}
+	ret = dvb_usb_i2c_init(d);
+	if (ret)
+		goto err_i2c_init;
+	ret = dvb_usb_adapter_init(d, adapter_nums);
+	if (ret)
+		goto err_adapter_init;
 
 	if ((ret = dvb_usb_remote_init(d)))
 		err("could not initialize remote control.");
@@ -193,6 +191,17 @@ static int dvb_usb_init(struct dvb_usb_d
 	dvb_usb_device_power_ctrl(d, 0);
 
 	return 0;
+
+err_adapter_init:
+	dvb_usb_adapter_exit(d);
+err_i2c_init:
+	dvb_usb_i2c_exit(d);
+	if (d->priv && d->props.priv_destroy)
+		d->props.priv_destroy(d);
+err_priv_init:
+	kfree(d->priv);
+	d->priv = NULL;
+	return ret;
 }
 
 /* determine the name and the state of the just found USB device */
@@ -296,15 +305,21 @@ int dvb_usb_device_init(struct usb_inter
 
 	usb_set_intfdata(intf, d);
 
-	if (du != NULL)
+	ret = dvb_usb_init(d, adapter_nums);
+	if (ret) {
+		info("%s error while loading driver (%d)", desc->name, ret);
+		goto error;
+	}
+
+	if (du)
 		*du = d;
 
-	ret = dvb_usb_init(d, adapter_nums);
+	info("%s successfully initialized and connected.", desc->name);
+	return 0;
 
-	if (ret == 0)
-		info("%s successfully initialized and connected.", desc->name);
-	else
-		info("%s error while loading driver (%d)", desc->name, ret);
+ error:
+	usb_set_intfdata(intf, NULL);
+	kfree(d);
 	return ret;
 }
 EXPORT_SYMBOL(dvb_usb_device_init);
author maintainer
#179 lore raw

PATCH 169/184 media: staging/intel-ipu3: Fix memory leak in imu_fmt

In reply to Greg Kroah-Hartman

From: Ricardo Ribalda <[email protected]>

commit 3630901933afba1d16c462b04d569b7576339223 upstream.

We are losing the reference to an allocated memory if try. Change the order of the check to avoid that.

Fixes 6d5f26f2e045 ("media: staging/intel-ipu3-v4l: reduce kernel stack usage")
---
 drivers/staging/media/ipu3/ipu3-v4l2.c |   11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)
drivers/staging/media/ipu3/ipu3-v4l2.c +7 -4
@@ -692,6 +692,13 @@ static int imgu_fmt(struct imgu_device *
 		if (inode == IMGU_NODE_STAT_3A || inode == IMGU_NODE_PARAMS)
 			continue;
 
+		/* CSS expects some format on OUT queue */
+		if (i != IPU3_CSS_QUEUE_OUT &&
+		    !imgu_pipe->nodes[inode].enabled) {
+			fmts[i] = NULL;
+			continue;
+		}
+
 		if (try) {
 			fmts[i] = kmemdup(&imgu_pipe->nodes[inode].vdev_fmt.fmt.pix_mp,
 					  sizeof(struct v4l2_pix_format_mplane),
@@ -704,10 +711,6 @@ static int imgu_fmt(struct imgu_device *
 			fmts[i] = &imgu_pipe->nodes[inode].vdev_fmt.fmt.pix_mp;
 		}
 
-		/* CSS expects some format on OUT queue */
-		if (i != IPU3_CSS_QUEUE_OUT &&
-		    !imgu_pipe->nodes[inode].enabled)
-			fmts[i] = NULL;
 	}
 
 	if (!try) {
author maintainer
#180 lore raw

PATCH 170/184 media: staging/intel-ipu3: Fix set_fmt error handling

In reply to Greg Kroah-Hartman

From: Ricardo Ribalda <[email protected]>

commit ad91849996f9dd79741a961fd03585a683b08356 upstream.

If there in an error during a set_fmt, do not overwrite the previous sizes with the invalid config.

Without this patch, v4l2-compliance ends up allocating 4GiB of RAM and causing the following OOPs

[   38.662975] ipu3-imgu 0000:00:05.0: swiotlb buffer is full (sz: 4096 bytes)
[   38.662980] DMA: Out of SW-IOMMU space for 4096 bytes at device 0000:00:05.0
[   38.663010] general protection fault: 0000 [#1] PREEMPT SMP
Fixes 6d5f26f2e045 ("media: staging/intel-ipu3-v4l: reduce kernel stack usage")
---
 drivers/staging/media/ipu3/ipu3-v4l2.c |    5 +++++
 1 file changed, 5 insertions(+)
drivers/staging/media/ipu3/ipu3-v4l2.c +5 -0
@@ -668,6 +668,7 @@ static int imgu_fmt(struct imgu_device *
 	struct imgu_css_pipe *css_pipe = &imgu->css.pipes[pipe];
 	struct imgu_media_pipe *imgu_pipe = &imgu->imgu_pipe[pipe];
 	struct imgu_v4l2_subdev *imgu_sd = &imgu_pipe->imgu_sd;
+	struct v4l2_pix_format_mplane fmt_backup;
 
 	dev_dbg(dev, "set fmt node [%u][%u](try = %u)", pipe, node, try);
 
@@ -736,6 +737,7 @@ static int imgu_fmt(struct imgu_device *
 		ret = -EINVAL;
 		goto out;
 	}
+	fmt_backup = *fmts[css_q];
 	*fmts[css_q] = f->fmt.pix_mp;
 
 	if (try)
@@ -743,6 +745,9 @@ static int imgu_fmt(struct imgu_device *
 	else
 		ret = imgu_css_fmt_set(&imgu->css, fmts, rects, pipe);
 
+	if (try || ret < 0)
+		*fmts[css_q] = fmt_backup;
+
 	/* ret is the binary number in the firmware blob */
 	if (ret < 0)
 		goto out;
author maintainer
#181 lore raw

PATCH 171/184 media: staging/intel-ipu3: Fix race condition during set_fmt

In reply to Greg Kroah-Hartman

From: Ricardo Ribalda <[email protected]>

commit dccfe2548746ca9cca3a20401ece4cf255d1f171 upstream.

Do not modify imgu_pipe->nodes[inode].vdev_fmt.fmt.pix_mp, until the format has been correctly validated.

Otherwise, even if we use a backup variable, there is a period of time where imgu_pipe->nodes[inode].vdev_fmt.fmt.pix_mp might have an invalid value that can be used by other functions.

Fixes ad91849996f9 ("media: staging/intel-ipu3: Fix set_fmt error handling")
---
 drivers/staging/media/ipu3/ipu3-v4l2.c |   30 ++++++++++++++----------------
 1 file changed, 14 insertions(+), 16 deletions(-)
drivers/staging/media/ipu3/ipu3-v4l2.c +14 -16
@@ -668,7 +668,6 @@ static int imgu_fmt(struct imgu_device *
 	struct imgu_css_pipe *css_pipe = &imgu->css.pipes[pipe];
 	struct imgu_media_pipe *imgu_pipe = &imgu->imgu_pipe[pipe];
 	struct imgu_v4l2_subdev *imgu_sd = &imgu_pipe->imgu_sd;
-	struct v4l2_pix_format_mplane fmt_backup;
 
 	dev_dbg(dev, "set fmt node [%u][%u](try = %u)", pipe, node, try);
 
@@ -686,6 +685,7 @@ static int imgu_fmt(struct imgu_device *
 
 	dev_dbg(dev, "IPU3 pipe %u pipe_id = %u", pipe, css_pipe->pipe_id);
 
+	css_q = imgu_node_to_queue(node);
 	for (i = 0; i < IPU3_CSS_QUEUES; i++) {
 		unsigned int inode = imgu_map_node(imgu, i);
 
@@ -700,6 +700,11 @@ static int imgu_fmt(struct imgu_device *
 			continue;
 		}
 
+		if (i == css_q) {
+			fmts[i] = &f->fmt.pix_mp;
+			continue;
+		}
+
 		if (try) {
 			fmts[i] = kmemdup(&imgu_pipe->nodes[inode].vdev_fmt.fmt.pix_mp,
 					  sizeof(struct v4l2_pix_format_mplane),
@@ -728,39 +733,32 @@ static int imgu_fmt(struct imgu_device *
 		rects[IPU3_CSS_RECT_GDC]->height = pad_fmt.height;
 	}
 
-	/*
-	 * imgu doesn't set the node to the value given by user
-	 * before we return success from this function, so set it here.
-	 */
-	css_q = imgu_node_to_queue(node);
 	if (!fmts[css_q]) {
 		ret = -EINVAL;
 		goto out;
 	}
-	fmt_backup = *fmts[css_q];
-	*fmts[css_q] = f->fmt.pix_mp;
 
 	if (try)
 		ret = imgu_css_fmt_try(&imgu->css, fmts, rects, pipe);
 	else
 		ret = imgu_css_fmt_set(&imgu->css, fmts, rects, pipe);
 
-	if (try || ret < 0)
-		*fmts[css_q] = fmt_backup;
-
 	/* ret is the binary number in the firmware blob */
 	if (ret < 0)
 		goto out;
 
-	if (try)
-		f->fmt.pix_mp = *fmts[css_q];
-	else
-		f->fmt = imgu_pipe->nodes[node].vdev_fmt.fmt;
+	/*
+	 * imgu doesn't set the node to the value given by user
+	 * before we return success from this function, so set it here.
+	 */
+	if (!try)
+		imgu_pipe->nodes[node].vdev_fmt.fmt.pix_mp = f->fmt.pix_mp;
 
 out:
 	if (try) {
 		for (i = 0; i < IPU3_CSS_QUEUES; i++)
-			kfree(fmts[i]);
+			if (i != css_q)
+				kfree(fmts[i]);
 	}
 
 	return ret;
author maintainer
#182 lore raw

PATCH 172/184 usb: gadget: dummy_hcd: fix gpf in gadget_setup

In reply to Greg Kroah-Hartman

From: Anirudh Rayabharam <[email protected]>

commit 4a5d797a9f9c4f18585544237216d7812686a71f upstream.

Fix a general protection fault reported by syzbot due to a race between gadget_setup() and gadget_unbind() in raw_gadget.

The gadget core is supposed to guarantee that there won't be any more callbacks to the gadget driver once the driver's unbind routine is called. That guarantee is enforced in usb_gadget_remove_driver as follows:

        usb_gadget_disconnect(udc->gadget);
        if (udc->gadget->irq)
                synchronize_irq(udc->gadget->irq);
        udc->driver->unbind(udc->gadget);
        usb_gadget_udc_stop(udc);

usb_gadget_disconnect turns off the pullup resistor, telling the host that the gadget is no longer connected and preventing the transmission of any more USB packets. Any packets that have already been received are sure to processed by the UDC driver's interrupt handler by the time synchronize_irq returns.

But this doesn't work with dummy_hcd, because dummy_hcd doesn't use interrupts; it uses a timer instead. It does have code to emulate the effect of synchronize_irq, but that code doesn't get invoked at the right time -- it currently runs in usb_gadget_udc_stop, after the unbind callback instead of before. Indeed, there's no way for usb_gadget_remove_driver to invoke this code before the unbind callback.

To fix this, move the synchronize_irq() emulation code to dummy_pullup so that it runs before unbind. Also, add a comment explaining why it is necessary to have it there.

---
 drivers/usb/gadget/udc/dummy_hcd.c |   23 +++++++++++++++--------
 1 file changed, 15 insertions(+), 8 deletions(-)
drivers/usb/gadget/udc/dummy_hcd.c +15 -8
@@ -900,6 +900,21 @@ static int dummy_pullup(struct usb_gadge
 	spin_lock_irqsave(&dum->lock, flags);
 	dum->pullup = (value != 0);
 	set_link_state(dum_hcd);
+	if (value == 0) {
+		/*
+		 * Emulate synchronize_irq(): wait for callbacks to finish.
+		 * This seems to be the best place to emulate the call to
+		 * synchronize_irq() that's in usb_gadget_remove_driver().
+		 * Doing it in dummy_udc_stop() would be too late since it
+		 * is called after the unbind callback and unbind shouldn't
+		 * be invoked until all the other callbacks are finished.
+		 */
+		while (dum->callback_usage > 0) {
+			spin_unlock_irqrestore(&dum->lock, flags);
+			usleep_range(1000, 2000);
+			spin_lock_irqsave(&dum->lock, flags);
+		}
+	}
 	spin_unlock_irqrestore(&dum->lock, flags);
 
 	usb_hcd_poll_rh_status(dummy_hcd_to_hcd(dum_hcd));
@@ -1001,14 +1016,6 @@ static int dummy_udc_stop(struct usb_gad
 	spin_lock_irq(&dum->lock);
 	dum->ints_enabled = 0;
 	stop_activity(dum);
-
-	/* emulate synchronize_irq(): wait for callbacks to finish */
-	while (dum->callback_usage > 0) {
-		spin_unlock_irq(&dum->lock);
-		usleep_range(1000, 2000);
-		spin_lock_irq(&dum->lock);
-	}
-
 	dum->driver = NULL;
 	spin_unlock_irq(&dum->lock);
 
author maintainer
#183 lore raw

PATCH 173/184 usb: gadget: Fix double free of device descriptor pointers

In reply to Greg Kroah-Hartman

From: Hemant Kumar <[email protected]>

commit 43c4cab006f55b6ca549dd1214e22f5965a8675f upstream.

Upon driver unbind usb_free_all_descriptors() function frees all speed descriptor pointers without setting them to NULL. In case gadget speed changes (i.e from super speed plus to super speed) after driver unbind only upto super speed descriptor pointers get populated. Super speed plus desc still holds the stale (already freed) pointer. Fix this issue by setting all descriptor pointers to NULL after freeing them in usb_free_all_descriptors().

---
 drivers/usb/gadget/config.c |    4 ++++
 1 file changed, 4 insertions(+)
drivers/usb/gadget/config.c +4 -0
@@ -194,9 +194,13 @@ EXPORT_SYMBOL_GPL(usb_assign_descriptors
 void usb_free_all_descriptors(struct usb_function *f)
 {
 	usb_free_descriptors(f->fs_descriptors);
+	f->fs_descriptors = NULL;
 	usb_free_descriptors(f->hs_descriptors);
+	f->hs_descriptors = NULL;
 	usb_free_descriptors(f->ss_descriptors);
+	f->ss_descriptors = NULL;
 	usb_free_descriptors(f->ssp_descriptors);
+	f->ssp_descriptors = NULL;
 }
 EXPORT_SYMBOL_GPL(usb_free_all_descriptors);
 
author maintainer
#184 lore raw

PATCH 174/184 usb: gadget/function/f_fs string table fix for multiple languages

In reply to Greg Kroah-Hartman

From: Dean Anderson <[email protected]>

commit 55b74ce7d2ce0b0058f3e08cab185a0afacfe39e upstream.

Fixes bug with the handling of more than one language in the string table in f_fs.c.
str_count was not reset for subsequent language codes. str_count-- "rolls under" and processes u32 max strings on the processing of the second language entry.
The existing bug can be reproduced by adding a second language table to the structure "strings" in tools/usb/ffs-test.c.

---
 drivers/usb/gadget/function/f_fs.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/usb/gadget/function/f_fs.c +2 -1
@@ -2658,6 +2658,7 @@ static int __ffs_data_got_strings(struct
 
 	do { /* lang_count > 0 so we can use do-while */
 		unsigned needed = needed_count;
+		u32 str_per_lang = str_count;
 
 		if (unlikely(len < 3))
 			goto error_free;
@@ -2693,7 +2694,7 @@ static int __ffs_data_got_strings(struct
 
 			data += length + 1;
 			len -= length + 1;
-		} while (--str_count);
+		} while (--str_per_lang);
 
 		s->id = 0;   /* terminator */
 		s->s = NULL;
author maintainer
#185 lore raw

PATCH 175/184 usb: dwc3: gadget: Fix START_TRANSFER link state check

In reply to Greg Kroah-Hartman

From: Thinh Nguyen <[email protected]>

commit c560e76319a94a3b9285bc426c609903408e4826 upstream.

The START_TRANSFER command needs to be executed while in ON/U0 link state (with an exception during register initialization). Don't use dwc->link_state to check this since the driver only tracks the link state when the link state change interrupt is enabled. Check the link state from DSTS register instead.

Note that often the host already brings the device out of low power before it sends/requests the next transfer. So, the user won't see any issue when the device starts transfer then. This issue is more noticeable in cases when the device delays starting transfer, which can happen during delayed control status after the host put the device in low power.

---
 drivers/usb/dwc3/gadget.c |   13 +++++++------
 1 file changed, 7 insertions(+), 6 deletions(-)
drivers/usb/dwc3/gadget.c +7 -6
@@ -304,13 +304,12 @@ int dwc3_send_gadget_ep_cmd(struct dwc3_
 	}
 
 	if (DWC3_DEPCMD_CMD(cmd) == DWC3_DEPCMD_STARTTRANSFER) {
-		int		needs_wakeup;
+		int link_state;
 
-		needs_wakeup = (dwc->link_state == DWC3_LINK_STATE_U1 ||
-				dwc->link_state == DWC3_LINK_STATE_U2 ||
-				dwc->link_state == DWC3_LINK_STATE_U3);
-
-		if (unlikely(needs_wakeup)) {
+		link_state = dwc3_gadget_get_link_state(dwc);
+		if (link_state == DWC3_LINK_STATE_U1 ||
+		    link_state == DWC3_LINK_STATE_U2 ||
+		    link_state == DWC3_LINK_STATE_U3) {
 			ret = __dwc3_gadget_wakeup(dwc);
 			dev_WARN_ONCE(dwc->dev, ret, "wakeup failed --> %d\n",
 					ret);
@@ -1862,6 +1861,8 @@ static int __dwc3_gadget_wakeup(struct d
 	case DWC3_LINK_STATE_RESET:
 	case DWC3_LINK_STATE_RX_DET:	/* in HS, means Early Suspend */
 	case DWC3_LINK_STATE_U3:	/* in HS, means SUSPEND */
+	case DWC3_LINK_STATE_U2:	/* in HS, means Sleep (L1) */
+	case DWC3_LINK_STATE_U1:
 	case DWC3_LINK_STATE_RESUME:
 		break;
 	default:
author maintainer
#186 lore raw

PATCH 176/184 usb: dwc2: Fix session request interrupt handler

In reply to Greg Kroah-Hartman

From: Artur Petrosyan <[email protected]>

commit 42b32b164acecd850edef010915a02418345a033 upstream.

According to programming guide in host mode, port
power must be turned on in session request
interrupt handlers.

---
 drivers/usb/dwc2/core_intr.c |    8 ++++++++
 1 file changed, 8 insertions(+)
drivers/usb/dwc2/core_intr.c +8 -0
@@ -312,6 +312,7 @@ static void dwc2_handle_conn_id_status_c
 static void dwc2_handle_session_req_intr(struct dwc2_hsotg *hsotg)
 {
 	int ret;
+	u32 hprt0;
 
 	/* Clear interrupt */
 	dwc2_writel(hsotg, GINTSTS_SESSREQINT, GINTSTS);
@@ -332,6 +333,13 @@ static void dwc2_handle_session_req_intr
 		 * established
 		 */
 		dwc2_hsotg_disconnect(hsotg);
+	} else {
+		/* Turn on the port power bit. */
+		hprt0 = dwc2_read_hprt0(hsotg);
+		hprt0 |= HPRT0_PWR;
+		dwc2_writel(hsotg, hprt0, HPRT0);
+		/* Connect hcd after port power is set. */
+		dwc2_hcd_connect(hsotg);
 	}
 }
 
author maintainer
#187 lore raw

PATCH 177/184 tty: fix memory leak in vc_deallocate

In reply to Greg Kroah-Hartman

From: Pavel Skripkin <[email protected]>

commit 211b4d42b70f1c1660feaa968dac0efc2a96ac4d upstream.

syzbot reported memory leak in tty/vt.
The problem was in VT_DISALLOCATE ioctl cmd. After allocating unimap with PIO_UNIMAP it wasn't freed via VT_DISALLOCATE, but vc_cons[currcons].d was zeroed.

---
 drivers/tty/vt/vt.c |    1 +
 1 file changed, 1 insertion(+)
drivers/tty/vt/vt.c +1 -0
@@ -1377,6 +1377,7 @@ struct vc_data *vc_deallocate(unsigned i
 		atomic_notifier_call_chain(&vt_notifier_list, VT_DEALLOCATE, &param);
 		vcs_remove_sysfs(currcons);
 		visual_deinit(vc);
+		con_free_unimap(vc);
 		put_pid(vc->vt_pid);
 		vc_uniscr_set(vc, NULL);
 		kfree(vc->vc_screenbuf);
author maintainer
#188 lore raw

PATCH 178/184 rsi: Use resume_noirq for SDIO

In reply to Greg Kroah-Hartman

From: Marek Vasut <[email protected]>

commit c434e5e48dc4e626364491455f97e2db0aa137b1 upstream.

The rsi_resume() does access the bus to enable interrupts on the RSI SDIO WiFi card, however when calling sdio_claim_host() in the resume path, it is possible the bus is already claimed and sdio_claim_host() spins indefinitelly. Enable the SDIO card interrupts in resume_noirq instead to prevent anything else from claiming the SDIO bus first.

Fixes 20db07332736 ("rsi: sdio suspend and resume support")
Cc Amitkumar Karwar <[email protected]>
Cc Angus Ainslie <[email protected]>
Cc David S. Miller <[email protected]>
Cc Jakub Kicinski <[email protected]>
Cc Kalle Valo <[email protected]>
Cc Karun Eagalapati <[email protected]>
Cc Martin Kepplinger <[email protected]>
Cc Sebastian Krzyszkowiak <[email protected]>
Cc Siva Rebbagondla <[email protected]>
---
 drivers/net/wireless/rsi/rsi_91x_sdio.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/wireless/rsi/rsi_91x_sdio.c +1 -1
@@ -1511,7 +1511,7 @@ static int rsi_restore(struct device *de
 }
 static const struct dev_pm_ops rsi_pm_ops = {
 	.suspend = rsi_suspend,
-	.resume = rsi_resume,
+	.resume_noirq = rsi_resume,
 	.freeze = rsi_freeze,
 	.thaw = rsi_thaw,
 	.restore = rsi_restore,
author maintainer
#189 lore raw

PATCH 179/184 tracing: Map all PIDs to command lines

In reply to Greg Kroah-Hartman

From: Steven Rostedt (VMware) <[email protected]>

commit 785e3c0a3a870e72dc530856136ab4c8dd207128 upstream.

The default max PID is set by PID_MAX_DEFAULT, and the tracing infrastructure uses this number to map PIDs to the comm names of the tasks, such output of the trace can show names from the recorded PIDs in the ring buffer. This mapping is also exported to user space via the "saved_cmdlines" file in the tracefs directory.

But currently the mapping expects the PIDs to be less than PID_MAX_DEFAULT, which is the default maximum and not the real maximum. Recently, systemd will increases the maximum value of a PID on the system, and when tasks are traced that have a PID higher than PID_MAX_DEFAULT, its comm is not recorded. This leads to the entire trace to have "<...>" as the comm name, which is pretty useless.

Instead, keep the array mapping the size of PID_MAX_DEFAULT, but instead of just mapping the index to the comm, map a mask of the PID (PID_MAX_DEFAULT - 1) to the comm, and find the full PID from the map_cmdline_to_pid array (that already exists).

This bug goes back to the beginning of ftrace, but hasn't been an issue until user space started increasing the maximum value of PIDs.

Fixes bc0c38d139ec7 ("ftrace: latency tracer infrastructure")
---
 kernel/trace/trace.c |   41 +++++++++++++++--------------------------
 1 file changed, 15 insertions(+), 26 deletions(-)
kernel/trace/trace.c +15 -26
@@ -2140,14 +2140,13 @@ static void tracing_stop_tr(struct trace
 
 static int trace_save_cmdline(struct task_struct *tsk)
 {
-	unsigned pid, idx;
+	unsigned tpid, idx;
 
 	/* treat recording of idle task as a success */
 	if (!tsk->pid)
 		return 1;
 
-	if (unlikely(tsk->pid > PID_MAX_DEFAULT))
-		return 0;
+	tpid = tsk->pid & (PID_MAX_DEFAULT - 1);
 
 	/*
 	 * It's not the end of the world if we don't get
@@ -2158,26 +2157,15 @@ static int trace_save_cmdline(struct tas
 	if (!arch_spin_trylock(&trace_cmdline_lock))
 		return 0;
 
-	idx = savedcmd->map_pid_to_cmdline[tsk->pid];
+	idx = savedcmd->map_pid_to_cmdline[tpid];
 	if (idx == NO_CMDLINE_MAP) {
 		idx = (savedcmd->cmdline_idx + 1) % savedcmd->cmdline_num;
 
-		/*
-		 * Check whether the cmdline buffer at idx has a pid
-		 * mapped. We are going to overwrite that entry so we
-		 * need to clear the map_pid_to_cmdline. Otherwise we
-		 * would read the new comm for the old pid.
-		 */
-		pid = savedcmd->map_cmdline_to_pid[idx];
-		if (pid != NO_CMDLINE_MAP)
-			savedcmd->map_pid_to_cmdline[pid] = NO_CMDLINE_MAP;
-
-		savedcmd->map_cmdline_to_pid[idx] = tsk->pid;
-		savedcmd->map_pid_to_cmdline[tsk->pid] = idx;
-
+		savedcmd->map_pid_to_cmdline[tpid] = idx;
 		savedcmd->cmdline_idx = idx;
 	}
 
+	savedcmd->map_cmdline_to_pid[idx] = tsk->pid;
 	set_cmdline(idx, tsk->comm);
 
 	arch_spin_unlock(&trace_cmdline_lock);
@@ -2188,6 +2176,7 @@ static int trace_save_cmdline(struct tas
 static void __trace_find_cmdline(int pid, char comm[])
 {
 	unsigned map;
+	int tpid;
 
 	if (!pid) {
 		strcpy(comm, "<idle>");
@@ -2199,16 +2188,16 @@ static void __trace_find_cmdline(int pid
 		return;
 	}
 
-	if (pid > PID_MAX_DEFAULT) {
-		strcpy(comm, "<...>");
-		return;
+	tpid = pid & (PID_MAX_DEFAULT - 1);
+	map = savedcmd->map_pid_to_cmdline[tpid];
+	if (map != NO_CMDLINE_MAP) {
+		tpid = savedcmd->map_cmdline_to_pid[map];
+		if (tpid == pid) {
+			strlcpy(comm, get_saved_cmdlines(map), TASK_COMM_LEN);
+			return;
+		}
 	}
-
-	map = savedcmd->map_pid_to_cmdline[pid];
-	if (map != NO_CMDLINE_MAP)
-		strlcpy(comm, get_saved_cmdlines(map), TASK_COMM_LEN);
-	else
-		strcpy(comm, "<...>");
+	strcpy(comm, "<...>");
 }
 
 void trace_find_cmdline(int pid, char comm[])
author maintainer
#190 lore raw

PATCH 180/184 tracing: Restructure trace_clock_global() to never block

In reply to Greg Kroah-Hartman

From: Steven Rostedt (VMware) <[email protected]>

commit aafe104aa9096827a429bc1358f8260ee565b7cc upstream.

It was reported that a fix to the ring buffer recursion detection would cause a hung machine when performing suspend / resume testing. The following backtrace was extracted from debugging that case:

Call Trace:
 trace_clock_global+0x91/0xa0
 __rb_reserve_next+0x237/0x460
 ring_buffer_lock_reserve+0x12a/0x3f0
 trace_buffer_lock_reserve+0x10/0x50
 __trace_graph_return+0x1f/0x80
 trace_graph_return+0xb7/0xf0
 ? trace_clock_global+0x91/0xa0
 ftrace_return_to_handler+0x8b/0xf0
 ? pv_hash+0xa0/0xa0
 return_to_handler+0x15/0x30
 ? ftrace_graph_caller+0xa0/0xa0
 ? trace_clock_global+0x91/0xa0
 ? __rb_reserve_next+0x237/0x460
 ? ring_buffer_lock_reserve+0x12a/0x3f0
 ? trace_event_buffer_lock_reserve+0x3c/0x120
 ? trace_event_buffer_reserve+0x6b/0xc0
 ? trace_event_raw_event_device_pm_callback_start+0x125/0x2d0
 ? dpm_run_callback+0x3b/0xc0
 ? pm_ops_is_empty+0x50/0x50
 ? platform_get_irq_byname_optional+0x90/0x90
 ? trace_device_pm_callback_start+0x82/0xd0
 ? dpm_run_callback+0x49/0xc0

With the following RIP:

RIP: 0010:native_queued_spin_lock_slowpath+0x69/0x200

Since the fix to the recursion detection would allow a single recursion to happen while tracing, this lead to the trace_clock_global() taking a spin lock and then trying to take it again:

ring_buffer_lock_reserve() {

  trace_clock_global() {
    arch_spin_lock() {
      queued_spin_lock_slowpath() {
        /* lock taken */
        (something else gets traced by function graph tracer)
          ring_buffer_lock_reserve() {
            trace_clock_global() {
              arch_spin_lock() {
                queued_spin_lock_slowpath() {
                /* DEAD LOCK! */

Tracing should *never* block, as it can lead to strange lockups like the above.

Restructure the trace_clock_global() code to instead of simply taking a lock to update the recorded "prev_time" simply use it, as two events happening on two different CPUs that calls this at the same time, really doesn't matter which one goes first. Use a trylock to grab the lock for updating the prev_time, and if it fails, simply try again the next time. If it failed to be taken, that means something else is already updating it.

Fixes b02414c8f045 ("ring-buffer: Fix recursion protection transitions between interrupt context") # started showing the problem
Fixes 14131f2f98ac3 ("tracing: implement trace_clock_*() APIs") # where the bug happened

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=212761

---
 kernel/trace/trace_clock.c |   48 ++++++++++++++++++++++++++++++---------------
 1 file changed, 32 insertions(+), 16 deletions(-)
kernel/trace/trace_clock.c +32 -16
@@ -95,33 +95,49 @@ u64 notrace trace_clock_global(void)
 {
 	unsigned long flags;
 	int this_cpu;
-	u64 now;
+	u64 now, prev_time;
 
 	raw_local_irq_save(flags);
 
 	this_cpu = raw_smp_processor_id();
-	now = sched_clock_cpu(this_cpu);
+
 	/*
-	 * If in an NMI context then dont risk lockups and return the
-	 * cpu_clock() time:
+	 * The global clock "guarantees" that the events are ordered
+	 * between CPUs. But if two events on two different CPUS call
+	 * trace_clock_global at roughly the same time, it really does
+	 * not matter which one gets the earlier time. Just make sure
+	 * that the same CPU will always show a monotonic clock.
+	 *
+	 * Use a read memory barrier to get the latest written
+	 * time that was recorded.
 	 */
-	if (unlikely(in_nmi()))
-		goto out;
+	smp_rmb();
+	prev_time = READ_ONCE(trace_clock_struct.prev_time);
+	now = sched_clock_cpu(this_cpu);
 
-	arch_spin_lock(&trace_clock_struct.lock);
+	/* Make sure that now is always greater than prev_time */
+	if ((s64)(now - prev_time) < 0)
+		now = prev_time + 1;
 
 	/*
-	 * TODO: if this happens often then maybe we should reset
-	 * my_scd->clock to prev_time+1, to make sure
-	 * we start ticking with the local clock from now on?
+	 * If in an NMI context then dont risk lockups and simply return
+	 * the current time.
 	 */
-	if ((s64)(now - trace_clock_struct.prev_time) < 0)
-		now = trace_clock_struct.prev_time + 1;
-
-	trace_clock_struct.prev_time = now;
-
-	arch_spin_unlock(&trace_clock_struct.lock);
+	if (unlikely(in_nmi()))
+		goto out;
 
+	/* Tracing can cause strange recursion, always use a try lock */
+	if (arch_spin_trylock(&trace_clock_struct.lock)) {
+		/* Reread prev_time in case it was already updated */
+		prev_time = READ_ONCE(trace_clock_struct.prev_time);
+		if ((s64)(now - prev_time) < 0)
+			now = prev_time + 1;
+
+		trace_clock_struct.prev_time = now;
+
+		/* The unlock acts as the wmb for the above rmb */
+		arch_spin_unlock(&trace_clock_struct.lock);
+	}
  out:
 	raw_local_irq_restore(flags);
 
author maintainer
#191 lore raw

PATCH 181/184 dm persistent data: packed struct should have an aligned() attribute too

In reply to Greg Kroah-Hartman

From: Joe Thornber <[email protected]>

commit a88b2358f1da2c9f9fcc432f2e0a79617fea397c upstream.

Otherwise most non-x86 architectures (e.g. riscv, arm) will resort to byte-by-byte access.

---
 drivers/md/persistent-data/dm-btree-internal.h   |    4 ++--
 drivers/md/persistent-data/dm-space-map-common.h |    8 ++++----
 2 files changed, 6 insertions(+), 6 deletions(-)
drivers/md/persistent-data/dm-btree-internal.h +2 -2
@@ -34,12 +34,12 @@ struct node_header {
 	__le32 max_entries;
 	__le32 value_size;
 	__le32 padding;
-} __packed;
+} __attribute__((packed, aligned(8)));
 
 struct btree_node {
 	struct node_header header;
 	__le64 keys[0];
-} __packed;
+} __attribute__((packed, aligned(8)));
 
 
 /*
drivers/md/persistent-data/dm-space-map-common.h +4 -4
@@ -33,7 +33,7 @@ struct disk_index_entry {
 	__le64 blocknr;
 	__le32 nr_free;
 	__le32 none_free_before;
-} __packed;
+} __attribute__ ((packed, aligned(8)));
 
 
 #define MAX_METADATA_BITMAPS 255
@@ -43,7 +43,7 @@ struct disk_metadata_index {
 	__le64 blocknr;
 
 	struct disk_index_entry index[MAX_METADATA_BITMAPS];
-} __packed;
+} __attribute__ ((packed, aligned(8)));
 
 struct ll_disk;
 
@@ -86,7 +86,7 @@ struct disk_sm_root {
 	__le64 nr_allocated;
 	__le64 bitmap_root;
 	__le64 ref_count_root;
-} __packed;
+} __attribute__ ((packed, aligned(8)));
 
 #define ENTRIES_PER_BYTE 4
 
@@ -94,7 +94,7 @@ struct disk_bitmap_header {
 	__le32 csum;
 	__le32 not_used;
 	__le64 blocknr;
-} __packed;
+} __attribute__ ((packed, aligned(8)));
 
 enum allocation_event {
 	SM_NONE,
author maintainer
#192 lore raw

PATCH 182/184 dm space map common: fix division bug in sm_ll_find_free_block()

In reply to Greg Kroah-Hartman

From: Joe Thornber <[email protected]>

commit 5208692e80a1f3c8ce2063a22b675dd5589d1d80 upstream.

This division bug meant the search for free metadata space could skip the final allocation bitmap's worth of entries. Fix affects DM thinp, cache and era targets.

---
 drivers/md/persistent-data/dm-space-map-common.c |    2 ++
 1 file changed, 2 insertions(+)
drivers/md/persistent-data/dm-space-map-common.c +2 -0
@@ -339,6 +339,8 @@ int sm_ll_find_free_block(struct ll_disk
 	 */
 	begin = do_div(index_begin, ll->entries_per_block);
 	end = do_div(end, ll->entries_per_block);
+	if (end == 0)
+		end = ll->entries_per_block;
 
 	for (i = index_begin; i < index_end; i++, begin = 0) {
 		struct dm_block *blk;
author maintainer
#193 lore raw

PATCH 183/184 dm integrity: fix missing goto in bitmap_flush_interval error handling

In reply to Greg Kroah-Hartman

From: Tian Tao <[email protected]>

commit 17e9e134a8efabbbf689a0904eee92bb5a868172 upstream.

Fixes 468dfca38b1a ("dm integrity: add a bitmap mode")
---
 drivers/md/dm-integrity.c |    1 +
 1 file changed, 1 insertion(+)
drivers/md/dm-integrity.c +1 -0
@@ -3762,6 +3762,7 @@ static int dm_integrity_ctr(struct dm_ta
 			if (val >= (uint64_t)UINT_MAX * 1000 / HZ) {
 				r = -EINVAL;
 				ti->error = "Invalid bitmap_flush_interval argument";
+				goto bad;
 			}
 			ic->bitmap_flush_interval = msecs_to_jiffies(val);
 		} else if (!strncmp(opt_string, "internal_hash:", strlen("internal_hash:"))) {
author maintainer
#194 lore raw

PATCH 184/184 dm rq: fix double free of blk_mq_tag_set in dev remove after table load fails

In reply to Greg Kroah-Hartman

From: Benjamin Block <[email protected]>

commit 8e947c8f4a5620df77e43c9c75310dc510250166 upstream.

When loading a device-mapper table for a request-based mapped device, and the allocation/initialization of the blk_mq_tag_set for the device fails, a following device remove will cause a double free.

E.g. (dmesg):
device-mapper: core: Cannot initialize queue for request-based dm-mq mapped device
device-mapper: ioctl: unable to set up device queue for new table.

  Unable to handle kernel pointer dereference in virtual kernel address space
  Failing address: 0305e098835de000 TEID: 0305e098835de803
  Fault in home space mode while using kernel ASCE.
  AS:000000025efe0007 R3:0000000000000024
  Oops: 0038 ilc:3 [#1] SMP
  Modules linked in: ... lots of modules ...
  Supported: Yes, External
  CPU: 0 PID: 7348 Comm: multipathd Kdump: loaded Tainted: G        W      X    5.3.18-53-default #1 SLE15-SP3
  Hardware name: IBM 8561 T01 7I2 (LPAR)
  Krnl PSW : 0704e00180000000 000000025e368eca (kfree+0x42/0x330)
             R:0 T:1 IO:1 EX:1 Key:0 M:1 W:0 P:0 AS:3 CC:2 PM:0 RI:0 EA:3
  Krnl GPRS: 000000000000004a 000000025efe5230 c1773200d779968d 0000000000000000
             000000025e520270 000000025e8d1b40 0000000000000003 00000007aae10000
             000000025e5202a2 0000000000000001 c1773200d779968d 0305e098835de640
             00000007a8170000 000003ff80138650 000000025e5202a2 000003e00396faa8
  Krnl Code: 000000025e368eb8: c4180041e100       lgrl    %r1,25eba50b8
             000000025e368ebe: ecba06b93a55       risbg   %r11,%r10,6,185,58
            #000000025e368ec4: e3b010000008       ag      %r11,0(%r1)
            >000000025e368eca: e310b0080004       lg      %r1,8(%r11)
             000000025e368ed0: a7110001           tmll    %r1,1
             000000025e368ed4: a7740129           brc     7,25e369126
             000000025e368ed8: e320b0080004       lg      %r2,8(%r11)
             000000025e368ede: b904001b           lgr     %r1,%r11
  Call Trace:
   [<000000025e368eca>] kfree+0x42/0x330
   [<000000025e5202a2>] blk_mq_free_tag_set+0x72/0xb8
   [<000003ff801316a8>] dm_mq_cleanup_mapped_device+0x38/0x50 [dm_mod]
   [<000003ff80120082>] free_dev+0x52/0xd0 [dm_mod]
   [<000003ff801233f0>] __dm_destroy+0x150/0x1d0 [dm_mod]
   [<000003ff8012bb9a>] dev_remove+0x162/0x1c0 [dm_mod]
   [<000003ff8012a988>] ctl_ioctl+0x198/0x478 [dm_mod]
   [<000003ff8012ac8a>] dm_ctl_ioctl+0x22/0x38 [dm_mod]
   [<000000025e3b11ee>] ksys_ioctl+0xbe/0xe0
   [<000000025e3b127a>] __s390x_sys_ioctl+0x2a/0x40
   [<000000025e8c15ac>] system_call+0xd8/0x2c8
  Last Breaking-Event-Address:
   [<000000025e52029c>] blk_mq_free_tag_set+0x6c/0xb8
  Kernel panic - not syncing: Fatal exception: panic_on_oops

When allocation/initialization of the blk_mq_tag_set fails in dm_mq_init_request_queue(), it is uninitialized/freed, but the pointer is not reset to NULL; so when dev_remove() later gets into dm_mq_cleanup_mapped_device() it sees the pointer and tries to uninitialize and free it again.

Fix this by setting the pointer to NULL in dm_mq_init_request_queue() error-handling. Also set it to NULL in dm_mq_cleanup_mapped_device().

Cc <[email protected]> # 4.6+
Fixes 1c357a1e86a4 ("dm: allocate blk_mq_tag_set rather than embed in mapped_device")
---
 drivers/md/dm-rq.c |    2 ++
 1 file changed, 2 insertions(+)
drivers/md/dm-rq.c +2 -0
@@ -572,6 +572,7 @@ out_tag_set:
 	blk_mq_free_tag_set(md->tag_set);
 out_kfree_tag_set:
 	kfree(md->tag_set);
+	md->tag_set = NULL;
 
 	return err;
 }
@@ -581,6 +582,7 @@ void dm_mq_cleanup_mapped_device(struct
 	if (md->tag_set) {
 		blk_mq_free_tag_set(md->tag_set);
 		kfree(md->tag_set);
+		md->tag_set = NULL;
 	}
 }
 
maintainer Tested-by on the v1 series
#195 lore raw
In reply to Greg Kroah-Hartman
Quoted 9 lines · Greg Kroah-Hartman

This is the start of the stable review cycle for the 5.4.118 release. There are 184 patches in this series, all will be posted as a response to this one. If anyone has any issues with these being applied, please let me know.

Responses should be made by Wed, 12 May 2021 10:19:23 +0000. Anything received after that time might be too late.

The whole patch series can be found in one patch at:

Quoted 8 lines
	https://www.kernel.org/pub/linux/kernel/v5.x/stable-review/patch-5.4.118-rc1.gz

or in the git tree and branch at:
git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable-rc.git linux-5.4.y
and the diffstat can be found below.

thanks,

greg k-h

All tests passing for Tegra ...

Test results for stable-v5.4:

    12 builds:	12 pass, 0 fail
    26 boots:	26 pass, 0 fail
    59 tests:	59 pass, 0 fail

Linux version: 5.4.118-rc1-geb078a943f97
Boards tested: tegra124-jetson-tk1, tegra186-p2771-0000,

                tegra194-p2972-0000, tegra20-ventana,
                tegra210-p2371-2180, tegra210-p3450-0000,
                tegra30-cardhu-a04

Jon

Tested-by on the v1 series
#196 lore raw
In reply to Greg Kroah-Hartman
Quoted 9 lines · Greg Kroah-Hartman

This is the start of the stable review cycle for the 5.4.118 release. There are 184 patches in this series, all will be posted as a response to this one. If anyone has any issues with these being applied, please let me know.

Responses should be made by Wed, 12 May 2021 10:19:23 +0000. Anything received after that time might be too late.

The whole patch series can be found in one patch at:

Quoted 8 lines
	https://www.kernel.org/pub/linux/kernel/v5.x/stable-review/patch-5.4.118-rc1.gz

or in the git tree and branch at:
git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable-rc.git linux-5.4.y
and the diffstat can be found below.

thanks,

greg k-h

On ARCH_BRCMSTB, using 32-bit and 64-bit ARM kernels:

Signature · 1 line
Florian
Tested-by on the v1 series
#197 lore raw
In reply to Greg Kroah-Hartman
Quoted 8 lines · Greg Kroah-Hartman

This is the start of the stable review cycle for the 5.4.118 release. There are 184 patches in this series, all will be posted as a response to this one. If anyone has any issues with these being applied, please let me know.

Responses should be made by Wed, 12 May 2021 10:19:23 +0000. Anything received after that time might be too late.

Build results:

	total: 157 pass: 157 fail: 0
Qemu test results:
	total: 428 pass: 428 fail: 0

Guenter

Tested-by on the v1 series
#198 lore raw
In reply to Greg Kroah-Hartman
Quoted 9 lines · Greg Kroah-Hartman

This is the start of the stable review cycle for the 5.4.118 release. There are 184 patches in this series, all will be posted as a response to this one. If anyone has any issues with these being applied, please let me know.

Responses should be made by Wed, 12 May 2021 10:19:23 +0000. Anything received after that time might be too late.

The whole patch series can be found in one patch at:

Quoted 8 lines
	https://www.kernel.org/pub/linux/kernel/v5.x/stable-review/patch-5.4.118-rc1.gz

or in the git tree and branch at:
git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable-rc.git linux-5.4.y
and the diffstat can be found below.

thanks,

greg k-h

Compiled and booted on my test system. No dmesg regressions.

thanks,
-- Shuah

Tested-by Linux Kernel Functional Testing on the v1 series
#199 lore raw
In reply to Greg Kroah-Hartman
Quoted 10 lines · Greg Kroah-Hartman

This is the start of the stable review cycle for the 5.4.118 release. There are 184 patches in this series, all will be posted as a response to this one. If anyone has any issues with these being applied, please let me know.

Responses should be made by Wed, 12 May 2021 10:19:23 +0000. Anything received after that time might be too late.

The whole patch series can be found in one patch at:

Quoted 8 lines
        https://www.kernel.org/pub/linux/kernel/v5.x/stable-review/patch-5.4.118-rc1.gz

or in the git tree and branch at:
git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable-rc.git linux-5.4.y
and the diffstat can be found below.

thanks,

greg k-h

Results from Linaro’s test farm.
No regressions on arm64, arm, x86_64, and i386.

## Build
* kernel: 5.4.118-rc1
* git: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable-rc.git
* git branch: linux-5.4.y
* git commit: eb078a943f9777461b67e3f61b4b6376593eb08c
* git describe: v5.4.117-185-geb078a943f97
* test details:
https://qa-reports.linaro.org/lkft/linux-stable-rc-linux-5.4.y/build/v5.4.117-185-geb078a943f97

## No regressions (compared to v5.4.117-135-g07e3ee575ce0)

## No fixes (compared to v5.4.117-135-g07e3ee575ce0)

## Test result summary
total: 78227, pass: 63083, fail: 2683, skip: 12195, xfail: 266,

## Build Summary
* arc: 10 total, 10 passed, 0 failed
* arm: 192 total, 192 passed, 0 failed
* arm64: 26 total, 26 passed, 0 failed
* dragonboard-410c: 1 total, 1 passed, 0 failed
* hi6220-hikey: 1 total, 1 passed, 0 failed
* i386: 14 total, 14 passed, 0 failed
* juno-r2: 1 total, 1 passed, 0 failed
* mips: 45 total, 45 passed, 0 failed
* parisc: 9 total, 9 passed, 0 failed
* powerpc: 27 total, 27 passed, 0 failed
* riscv: 21 total, 21 passed, 0 failed
* s390: 9 total, 9 passed, 0 failed
* sh: 18 total, 18 passed, 0 failed
* sparc: 9 total, 9 passed, 0 failed
* x15: 1 total, 1 passed, 0 failed
* x86: 1 total, 1 passed, 0 failed
* x86_64: 26 total, 26 passed, 0 failed

## Test suites summary
* fwts
* igt-gpu-tools
* install-android-platform-tools-r2600
* kselftest-
* kselftest-android
* kselftest-bpf
* kselftest-breakpoints
* kselftest-capabilities
* kselftest-cgroup
* kselftest-clone3
* kselftest-core
* kselftest-cpu-hotplug
* kselftest-cpufreq
* kselftest-drivers
* kselftest-efivarfs
* kselftest-filesystems
* kselftest-firmware
* kselftest-fpu
* kselftest-futex
* kselftest-gpio
* kselftest-intel_pstate
* kselftest-ipc
* kselftest-ir
* kselftest-kcmp
* kselftest-kexec
* kselftest-kvm
* kselftest-lib
* kselftest-livepatch
* kselftest-lkdtm
* kselftest-membarrier
* kselftest-memfd
* kselftest-memory-hotplug
* kselftest-mincore
* kselftest-mount
* kselftest-mqueue
* kselftest-net
* kselftest-netfilter
* kselftest-nsfs
* kselftest-openat2
* kselftest-pid_namespace
* kselftest-pidfd
* kselftest-proc
* kselftest-pstore
* kselftest-ptrace
* kselftest-rseq
* kselftest-rtc
* kselftest-seccomp
* kselftest-sigaltstack
* kselftest-size
* kselftest-splice
* kselftest-static_keys
* kselftest-sync
* kselftest-sysctl
* kselftest-tc-testing
* kselftest-timens
* kselftest-timers
* kselftest-tmpfs
* kselftest-tpm2
* kselftest-user
* kselftest-vm
* kselftest-x86
* kselftest-zram
* kvm-unit-tests
* libhugetlbfs
* linux-log-parser
* ltp-cap_bounds-tests
* ltp-commands-tests
* ltp-containers-tests
* ltp-controllers-tests
* ltp-cpuhotplug-tests
* ltp-crypto-tests
* ltp-cve-tests
* ltp-dio-tests
* ltp-fcntl-locktests-tests
* ltp-filecaps-tests
* ltp-fs-tests
* ltp-fs_bind-tests
* ltp-fs_perms_simple-tests
* ltp-fsx-tests
* ltp-hugetlb-tests
* ltp-io-tests
* ltp-ipc-tests
* ltp-math-tests
* ltp-mm-tests
* ltp-nptl-tests
* ltp-open-posix-tests
* ltp-pty-tests
* ltp-sched-tests
* ltp-securebits-tests
* ltp-syscalls-tests
* ltp-tracing-tests
* network-basic-tests
* packetdrill
* perf
* rcutorture
* ssuite
* v4l2-compliance

--
Linaro LKFT
https://lkft.linaro.org

Tested-by on the v1 series
#200 lore raw
In reply to Greg Kroah-Hartman

Hi Greg,

Quoted 7 lines · Greg Kroah-Hartman

This is the start of the stable review cycle for the 5.4.118 release. There are 184 patches in this series, all will be posted as a response to this one. If anyone has any issues with these being applied, please let me know.

Responses should be made by Wed, 12 May 2021 10:19:23 +0000. Anything received after that time might be too late.

Build test:
mips (gcc version 11.1.1 20210430): 65 configs -> no new failure arm (gcc version 11.1.1 20210430): 107 configs -> no new failure x86_64 (gcc version 10.2.1 20210110): 2 configs -> no failure

Boot test:
x86_64: Booted on my test laptop. No regression.
x86_64: Booted on qemu. No regression.

--
Regards
Sudip

Tested-by Hulk Robot on the v1 series
#201 lore raw
In reply to Greg Kroah-Hartman
Quoted 9 lines · Greg Kroah-Hartman

This is the start of the stable review cycle for the 5.4.118 release. There are 184 patches in this series, all will be posted as a response to this one. If anyone has any issues with these being applied, please let me know.

Responses should be made by Wed, 12 May 2021 10:19:23 +0000. Anything received after that time might be too late.

The whole patch series can be found in one patch at:

Quoted 9 lines
	https://www.kernel.org/pub/linux/kernel/v5.x/stable-review/patch-5.4.118-rc1.gz

or in the git tree and branch at:
git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable-rc.git linux-5.4.y
and the diffstat can be found below.

thanks,

greg k-h

Tested on arm64 and x86 for 5.4.118-rc1,

Kernel repo:
https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable-rc.git
Branch: linux-5.4.y
Version: 5.4.118-rc1
Commit: eb078a943f9777461b67e3f61b4b6376593eb08c
Compiler: gcc version 7.3.0 (GCC)

arm64:
--------------------------------------------------------------------
Testcase Result Summary:
total: 8476
passed: 8476
failed: 0
timeout: 0
--------------------------------------------------------------------

x86:
--------------------------------------------------------------------
Testcase Result Summary:
total: 8476
passed: 8476
failed: 0
timeout: 0
--------------------------------------------------------------------