- Linux-stable-mirror - lists.linaro.org

[PATCH printk v2 0/2] Fix reported suspend failures

by John Ogness

This is v2 of a series to address multiple reports [0][1] (+ 2 offlist) of suspend failing when NBCON console drivers are in use. With the help of NXP and NVIDIA we were able to isolate the problem and verify the fix. v1 is here [2]. The first NBCON drivers appeared in 6.13, so currently there is no LTS kernel that requires this series. But it should go into 6.17.x and 6.18. The changes since v1: - For printk_trigger_flush() add support for all flush types that are available. This will prevent printk_trigger_flush() from trying to inappropriately queue irq_work after this series is applied. - Add WARN_ON_ONCE() to the printk irq_work queueing functions in case they are called when irq_work is blocked. There should never be (and currently are no) such callers, but these functions are externally available. John Ogness [0] https://lore.kernel.org/lkml/80b020fc-c18a-4da4-b222-16da1cab2f4c@nvidia.com [1] https://lore.kernel.org/lkml/DB9PR04MB8429E7DDF2D93C2695DE401D92C4A@DB9PR04… [2] https://lore.kernel.org/lkml/20251111144328.887159-1-john.ogness@linutronix… John Ogness (2): printk: Allow printk_trigger_flush() to flush all types printk: Avoid scheduling irq_work on suspend kernel/printk/internal.h | 8 ++-- kernel/printk/nbcon.c | 9 ++++- kernel/printk/printk.c | 81 ++++++++++++++++++++++++++++++++-------- 3 files changed, 78 insertions(+), 20 deletions(-) base-commit: e9a6fb0bcdd7609be6969112f3fbfcce3b1d4a7c -- 2.47.3

2 weeks, 1 day

6
16
0 0

[PATCH] media: ipu6: isys: csi2: guard remote pad/subdev and zero link freq

by Alexei Safin

media_pad_remote_pad_first() may return NULL when the media link is absent or disabled. The code dereferenced remote_pad->entity unconditionally in ipu6_isys_csi2_enable_streams() and ipu6_isys_csi2_disable_streams(), leading to a possible NULL dereference. Guard the remote pad/subdev: in enable path return -ENOLINK/-ENODEV, in disable path always shut down the local stream first and return 0 if the remote side is missing. This keeps local shutdown semantics intact and prevents a crash when the graph is partially torn down. Also, ipu6_isys_csi2_calc_timing() passes link_freq into calc_timing() where it is used as a divisor. v4l2_get_link_freq() may yield 0; add an explicit check and return -EINVAL to avoid division by zero. Found by Linux Verification Center (linuxtesting.org) with SVACE. Fixes: 3a5c59ad926b ("media: ipu6: Rework CSI-2 sub-device streaming control") Cc: stable(a)vger.kernel.org Signed-off-by: Alexei Safin <a.safin(a)rosa.ru> --- drivers/media/pci/intel/ipu6/ipu6-isys-csi2.c | 18 ++++++++++++++++++ 1 file changed, 18 insertions(+) diff --git a/drivers/media/pci/intel/ipu6/ipu6-isys-csi2.c b/drivers/media/pci/intel/ipu6/ipu6-isys-csi2.c index d1fece6210ab..58944918c344 100644 --- a/drivers/media/pci/intel/ipu6/ipu6-isys-csi2.c +++ b/drivers/media/pci/intel/ipu6/ipu6-isys-csi2.c @@ -171,6 +171,10 @@ ipu6_isys_csi2_calc_timing(struct ipu6_isys_csi2 *csi2, if (link_freq < 0) return link_freq; + /* Avoid division by zero in calc_timing() if link frequency is zero */ + if (!link_freq) + return -EINVAL; + timing->ctermen = calc_timing(CSI2_CSI_RX_DLY_CNT_TERMEN_CLANE_A, CSI2_CSI_RX_DLY_CNT_TERMEN_CLANE_B, link_freq, accinv); @@ -352,7 +356,12 @@ static int ipu6_isys_csi2_enable_streams(struct v4l2_subdev *sd, int ret; remote_pad = media_pad_remote_pad_first(&sd->entity.pads[CSI2_PAD_SINK]); + if (!remote_pad) + return -ENOLINK; + remote_sd = media_entity_to_v4l2_subdev(remote_pad->entity); + if (!remote_sd) + return -ENODEV; sink_streams = v4l2_subdev_state_xlate_streams(state, pad, CSI2_PAD_SINK, @@ -389,7 +398,16 @@ static int ipu6_isys_csi2_disable_streams(struct v4l2_subdev *sd, &streams_mask); remote_pad = media_pad_remote_pad_first(&sd->entity.pads[CSI2_PAD_SINK]); + if (!remote_pad) { + ipu6_isys_csi2_set_stream(sd, NULL, 0, false); + return 0; + } + remote_sd = media_entity_to_v4l2_subdev(remote_pad->entity); + if (!remote_sd) { + ipu6_isys_csi2_set_stream(sd, NULL, 0, false); + return 0; + } ipu6_isys_csi2_set_stream(sd, NULL, 0, false); -- 2.50.1 (Apple Git-155)

2 weeks, 1 day

1
0
0 0

[PATCH] libfs: Fix NULL pointer access in simple_recursive_removal

by rom.wang

From: Yufeng Wang <wangyufeng(a)kylinos.cn> There is an issue in the kernel: if inode is NULL pointer. the function "inode_lock_nested" (or function "inode_lock" before) a crash will happen at code "&inode->i_rwsem". [292618.520532] BUG: unable to handle kernel NULL pointer dereference at 00000000000000a0 [...] [292618.560398] RIP: 0010:down_write+0x12/0x30 [292618.565580] Code: 83 f8 01 74 08 48 c7 47 20 01 00 00 00 f3 c3 66 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 48 ba 01 00 00 00 ff ff ff ff 48 89 f8 <f0> 48 0f c1 10 85 d2 74 05 e8 00 43 ff ff 65 48 8b 04 25 80 5c 01 [292618.587219] RSP: 0018:ffffb898dc86fc20 EFLAGS: 00010246 [292618.593666] RAX: 00000000000000a0 RBX: ffff94c84f363950 RCX: ffffff8000000000 [292618.602255] RDX: ffffffff00000001 RSI: 0000000000000063 RDI: 00000000000000a0 [292618.610844] RBP: ffffb898dc86fc78 R08: 0000000000000000 R09: 0000000000000000 [292618.619434] R10: ffffb898dc86fca8 R11: 0000000000000000 R12: 0000000000000000 [292618.628022] R13: ffff94c84f362a20 R14: ffff954d3f2fb4a0 R15: ffff954c3afa5010 [292618.636612] FS: 0000555555989cc0(0000) GS:ffff956dbf900000(0000) knlGS:0000000000000000 [292618.646271] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [292618.653300] CR2: 00000000000000a0 CR3: 000000fc7f25a000 CR4: 00000000003406e0 [292618.661888] Call Trace: [292618.665225] simple_recursive_removal+0x4f/0x230 [292618.670994] ? debug_fill_super+0xe0/0xe0 [292618.676079] debugfs_remove+0x40/0x60 [292618.680799] kvm_vcpu_release+0x19/0x30 [kvm] Fixes: a3d1e7eb5abe ("simple_recursive_removal(): kernel-side rm -rf for ramfs-style filesystems") Cc: stable(a)vger.kernel.org Signed-off-by: Yufeng Wang <wangyufeng(a)kylinos.cn> --- fs/libfs.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/fs/libfs.c b/fs/libfs.c index 1661dcb7d983..9090cd1f97c4 100644 --- a/fs/libfs.c +++ b/fs/libfs.c @@ -611,6 +611,8 @@ static void __simple_recursive_removal(struct dentry *dentry, struct dentry *victim = NULL, *child; struct inode *inode = this->d_inode; + if (!inode) + return; inode_lock_nested(inode, I_MUTEX_CHILD); if (d_is_dir(this)) inode->i_flags |= S_DEAD; -- 2.34.1

2 weeks, 1 day

3
3
0 0

[PATCH REGRESSION] drm/panel: simple: restore connector_type fallback

by Ludovic Desroches

The switch from devm_kzalloc() + drm_panel_init() to devm_drm_panel_alloc() introduced a regression. Several panel descriptors do not set connector_type. For those panels, panel_simple_probe() used to compute a connector type (currently DPI as a fallback) and pass that value to drm_panel_init(). After the conversion to devm_drm_panel_alloc(), the call unconditionally used desc->connector_type instead, ignoring the computed fallback and potentially passing DRM_MODE_CONNECTOR_Unknown, which drm_panel_bridge_add() does not allow. Move the connector_type validation / fallback logic before the devm_drm_panel_alloc() call and pass the computed connector_type to devm_drm_panel_alloc(), so panels without an explicit connector_type once again get the DPI default. Signed-off-by: Ludovic Desroches <ludovic.desroches(a)microchip.com> Fixes: de04bb0089a9 ("drm/panel/panel-simple: Use the new allocation in place of devm_kzalloc()") --- Hi, I am not sure whether this regression has already been reported or addressed. If it has, please feel free to drop this patch. --- drivers/gpu/drm/panel/panel-simple.c | 86 ++++++++++++++++++------------------ 1 file changed, 43 insertions(+), 43 deletions(-) diff --git a/drivers/gpu/drm/panel/panel-simple.c b/drivers/gpu/drm/panel/panel-simple.c index da6b71b70a463400fcc45006788f87e97b0c148c..dc41789f6a53c78b928ff39291ab7219a2d835dd 100644 --- a/drivers/gpu/drm/panel/panel-simple.c +++ b/drivers/gpu/drm/panel/panel-simple.c @@ -623,49 +623,6 @@ static struct panel_simple *panel_simple_probe(struct device *dev) if (IS_ERR(desc)) return ERR_CAST(desc); - panel = devm_drm_panel_alloc(dev, struct panel_simple, base, - &panel_simple_funcs, desc->connector_type); - if (IS_ERR(panel)) - return ERR_CAST(panel); - - panel->desc = desc; - - panel->supply = devm_regulator_get(dev, "power"); - if (IS_ERR(panel->supply)) - return ERR_CAST(panel->supply); - - panel->enable_gpio = devm_gpiod_get_optional(dev, "enable", - GPIOD_OUT_LOW); - if (IS_ERR(panel->enable_gpio)) - return dev_err_cast_probe(dev, panel->enable_gpio, - "failed to request GPIO\n"); - - err = of_drm_get_panel_orientation(dev->of_node, &panel->orientation); - if (err) { - dev_err(dev, "%pOF: failed to get orientation %d\n", dev->of_node, err); - return ERR_PTR(err); - } - - ddc = of_parse_phandle(dev->of_node, "ddc-i2c-bus", 0); - if (ddc) { - panel->ddc = of_find_i2c_adapter_by_node(ddc); - of_node_put(ddc); - - if (!panel->ddc) - return ERR_PTR(-EPROBE_DEFER); - } - - if (!of_device_is_compatible(dev->of_node, "panel-dpi") && - !of_get_display_timing(dev->of_node, "panel-timing", &dt)) - panel_simple_parse_panel_timing_node(dev, panel, &dt); - - if (desc->connector_type == DRM_MODE_CONNECTOR_LVDS) { - /* Optional data-mapping property for overriding bus format */ - err = panel_simple_override_nondefault_lvds_datamapping(dev, panel); - if (err) - goto free_ddc; - } - connector_type = desc->connector_type; /* Catch common mistakes for panels. */ switch (connector_type) { @@ -720,6 +677,49 @@ static struct panel_simple *panel_simple_probe(struct device *dev) break; } + panel = devm_drm_panel_alloc(dev, struct panel_simple, base, + &panel_simple_funcs, connector_type); + if (IS_ERR(panel)) + return ERR_CAST(panel); + + panel->desc = desc; + + panel->supply = devm_regulator_get(dev, "power"); + if (IS_ERR(panel->supply)) + return ERR_CAST(panel->supply); + + panel->enable_gpio = devm_gpiod_get_optional(dev, "enable", + GPIOD_OUT_LOW); + if (IS_ERR(panel->enable_gpio)) + return dev_err_cast_probe(dev, panel->enable_gpio, + "failed to request GPIO\n"); + + err = of_drm_get_panel_orientation(dev->of_node, &panel->orientation); + if (err) { + dev_err(dev, "%pOF: failed to get orientation %d\n", dev->of_node, err); + return ERR_PTR(err); + } + + ddc = of_parse_phandle(dev->of_node, "ddc-i2c-bus", 0); + if (ddc) { + panel->ddc = of_find_i2c_adapter_by_node(ddc); + of_node_put(ddc); + + if (!panel->ddc) + return ERR_PTR(-EPROBE_DEFER); + } + + if (!of_device_is_compatible(dev->of_node, "panel-dpi") && + !of_get_display_timing(dev->of_node, "panel-timing", &dt)) + panel_simple_parse_panel_timing_node(dev, panel, &dt); + + if (desc->connector_type == DRM_MODE_CONNECTOR_LVDS) { + /* Optional data-mapping property for overriding bus format */ + err = panel_simple_override_nondefault_lvds_datamapping(dev, panel); + if (err) + goto free_ddc; + } + dev_set_drvdata(dev, panel); /* --- base-commit: 88cbd8ac379cf5ce68b7efcfd4d1484a6871ee0b change-id: 20251121-lcd_panel_connector_type_fix-f00fe3766a42 Best regards, -- Ludovic Desroches <ludovic.desroches(a)microchip.com>

2 weeks, 1 day

6
6
0 0

[PATCH 5.15.y 1/2] Revert "block: Move checking GENHD_FL_NO_PART to bdev_add_partition()"

by Gulam Mohamed

This reverts commit 7777f47f2ea64efd1016262e7b59fab34adfb869. The commit 1a721de8489f ("block: don't add or resize partition on the disk with GENHD_FL_NO_PART") and the commit 7777f47f2ea6 ("block: Move checking GENHD_FL_NO_PART to bdev_add_partition()") used the flag GENHD_FL_NO_PART to prevent the add or resize of partitions in 5.15 stable kernels.But in these 5.15 kernels, this is giving an issue with the following error where the loop driver wants to create a partition when the partscan is disabled on the loop device: dd if=/dev/zero of=loopDisk.dsk bs=1M count=1 seek=10240; losetup -f loopDisk.dsk;parted -s /dev/loop0 -- mklabel gpt mkpart primary 2048s 4096s 1+0 records in 1+0 records out 1048576 bytes (1.0 MB, 1.0 MiB) copied, 0.0016293 s, 644 MB/s "" Error: Partition(s) 1 on /dev/loop0 have been written, but we have been unable to inform the kernel of the change, probably because it/they are in use. As a result, the old partition(s) will remain in use. You should reboot now before making further changes. "" If the partition scan is not enabled on the loop device, this flag GENHD_FL_NO_PART is getting set and when partition creation is tried, it returns an error EINVAL thereby preventing the creation of partitions. So, there is no such distinction between disabling of partition scan and partition creation. Later in 6.xxx kernels, the commit b9684a71fca7 ("block, loop: support partitions without scanning") a new flag GD_SUPPRESS_PART_SCAN was introduced that just disables the partition scan and uses GENHD_FL_NO_PART only to prevent creating partition scan. So, the partition creationg can proceed with even if partition scan is disabled. As the commit b9684a71fca7 ("block, loop: support partitions without scanning") is not available in 5.15 stable kernel, and since there is no distinction between disabling of "partition scan" and "partition creation", we need to revert the commits 1a721de8489f and 7777f47f2ea6 from 5.15 stable kernel to allow partition creation when partscan is disabled. Cc: stable(a)vger.kernel.org Signed-off-by: Gulam Mohamed <gulam.mohamed(a)oracle.com> --- block/ioctl.c | 2 ++ block/partitions/core.c | 5 ----- 2 files changed, 2 insertions(+), 5 deletions(-) diff --git a/block/ioctl.c b/block/ioctl.c index a260e39e56a4..d25b84441237 100644 --- a/block/ioctl.c +++ b/block/ioctl.c @@ -20,6 +20,8 @@ static int blkpg_do_ioctl(struct block_device *bdev, struct blkpg_partition p; sector_t start, length; + if (disk->flags & GENHD_FL_NO_PART) + return -EINVAL; if (!capable(CAP_SYS_ADMIN)) return -EACCES; if (copy_from_user(&p, upart, sizeof(struct blkpg_partition))) diff --git a/block/partitions/core.c b/block/partitions/core.c index 0d1fe2b42b85..7b5750db7eaf 100644 --- a/block/partitions/core.c +++ b/block/partitions/core.c @@ -463,11 +463,6 @@ int bdev_add_partition(struct gendisk *disk, int partno, sector_t start, goto out; } - if (disk->flags & GENHD_FL_NO_PART) { - ret = -EINVAL; - goto out; - } - if (partition_overlaps(disk, start, length, -1)) { ret = -EBUSY; goto out; -- 2.47.3

2 weeks, 1 day

2
4
0 0

[PATCH] LoongArch: Fix build errors for CONFIG_RANDSTRUCT

by Huacai Chen

When CONFIG_RANDSTRUCT enabled, members of task_struct are randomized. There is a chance that TASK_STACK_CANARY be out of 12bit immediate's range and causes build errors. TASK_STACK_CANARY is naturally aligned, so fix it by replacing ld.d/st.d with ldptr.d/stptr.d which have 14bit immediates. Cc: stable(a)vger.kernel.org Reported-by: kernel test robot <lkp(a)intel.com> Closes: https://lore.kernel.org/oe-kbuild-all/202511240656.0NaPcJs1-lkp@intel.com/ Suggested-by: Rui Wang <wangrui(a)loongson.cn> Signed-off-by: Huacai Chen <chenhuacai(a)loongson.cn> --- arch/loongarch/kernel/switch.S | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/arch/loongarch/kernel/switch.S b/arch/loongarch/kernel/switch.S index 9c23cb7e432f..3007e909e0d8 100644 --- a/arch/loongarch/kernel/switch.S +++ b/arch/loongarch/kernel/switch.S @@ -25,8 +25,8 @@ SYM_FUNC_START(__switch_to) stptr.d a4, a0, THREAD_SCHED_CFA #if defined(CONFIG_STACKPROTECTOR) && !defined(CONFIG_SMP) la t7, __stack_chk_guard - LONG_L t8, a1, TASK_STACK_CANARY - LONG_S t8, t7, 0 + ldptr.d t8, a1, TASK_STACK_CANARY + stptr.d t8, t7, 0 #endif move tp, a2 cpu_restore_nonscratch a1 -- 2.47.3

2 weeks, 1 day

3
2
0 0

[PATCH 1/2] Revert "block: Move checking GENHD_FL_NO_PART to bdev_add_partition()"

by Gulam Mohamed

This reverts commit 7777f47f2ea64efd1016262e7b59fab34adfb869. The commit 1a721de8489f ("block: don't add or resize partition on the disk with GENHD_FL_NO_PART") and the commit 7777f47f2ea6 ("block: Move checking GENHD_FL_NO_PART to bdev_add_partition()") used the flag GENHD_FL_NO_PART to prevent the add or resize of partitions in 5.15 stable kernels.But in these 5.15 kernels, this is giving an issue with the following error where the loop driver wants to create a partition when the partscan is disabled on the loop device: dd if=/dev/zero of=loopDisk.dsk bs=1M count=1 seek=10240; losetup -f loopDisk.dsk;parted -s /dev/loop0 -- mklabel gpt mkpart primary 2048s 4096s 1+0 records in 1+0 records out 1048576 bytes (1.0 MB, 1.0 MiB) copied, 0.0016293 s, 644 MB/s "" Error: Partition(s) 1 on /dev/loop0 have been written, but we have been unable to inform the kernel of the change, probably because it/they are in use. As a result, the old partition(s) will remain in use. You should reboot now before making further changes. "" If the partition scan is not enabled on the loop device, this flag GENHD_FL_NO_PART is getting set and when partition creation is tried, it returns an error EINVAL thereby preventing the creation of partitions. So, there is no such distinction between disabling of partition scan and partition creation. Later in 6.xxx kernels, the commit b9684a71fca7 ("block, loop: support partitions without scanning") a new flag GD_SUPPRESS_PART_SCAN was introduced that just disables the partition scan and uses GENHD_FL_NO_PART only to prevent creating partition scan. So, the partition creationg can proceed with even if partition scan is disabled. As the commit b9684a71fca7 ("block, loop: support partitions without scanning") is not available in 5.15 stable kernel, and since there is no distinction between disabling of "partition scan" and "partition creation", we need to revert the commits 1a721de8489f and 7777f47f2ea6 from 5.15 stable kernel to allow partition creation when partscan is disabled. Cc: stable(a)vger.kernel.org Signed-off-by: Gulam Mohamed <gulam.mohamed(a)oracle.com> --- block/ioctl.c | 2 ++ block/partitions/core.c | 5 ----- 2 files changed, 2 insertions(+), 5 deletions(-) diff --git a/block/ioctl.c b/block/ioctl.c index a260e39e56a4..d25b84441237 100644 --- a/block/ioctl.c +++ b/block/ioctl.c @@ -20,6 +20,8 @@ static int blkpg_do_ioctl(struct block_device *bdev, struct blkpg_partition p; sector_t start, length; + if (disk->flags & GENHD_FL_NO_PART) + return -EINVAL; if (!capable(CAP_SYS_ADMIN)) return -EACCES; if (copy_from_user(&p, upart, sizeof(struct blkpg_partition))) diff --git a/block/partitions/core.c b/block/partitions/core.c index 0d1fe2b42b85..7b5750db7eaf 100644 --- a/block/partitions/core.c +++ b/block/partitions/core.c @@ -463,11 +463,6 @@ int bdev_add_partition(struct gendisk *disk, int partno, sector_t start, goto out; } - if (disk->flags & GENHD_FL_NO_PART) { - ret = -EINVAL; - goto out; - } - if (partition_overlaps(disk, start, length, -1)) { ret = -EBUSY; goto out; -- 2.47.3

2 weeks, 1 day

2
6
0 0

[PATCH v4] mailbox: Prevent out-of-bounds access in fw_mbox_index_xlate()

by Joonwon Kang

Although it is guided that `#mbox-cells` must be at least 1, there are many instances of `#mbox-cells = <0>;` in the device tree. If that is the case and the corresponding mailbox controller does not provide `fw_xlate` and of_xlate` function pointers, `fw_mbox_index_xlate()` will be used by default and out-of-bounds accesses could occur due to lack of bounds check in that function. Cc: stable(a)vger.kernel.org Signed-off-by: Joonwon Kang <joonwonkang(a)google.com> --- V3 -> V4: Prevented access to sp->args[0] if sp->nargs < 1 and rebased on the linux-next tree. For CVE review, below is a problematic control flow when `#mbox-cells = <0>;`: ``` static struct mbox_chan * fw_mbox_index_xlate(struct mbox_controller *mbox, const struct fwnode_reference_args *sp) { int ind = sp->args[0]; // (4) if (ind >= mbox->num_chans) // (5) return ERR_PTR(-EINVAL); return &mbox->chans[ind]; // (6) } struct mbox_chan *mbox_request_channel(struct mbox_client *cl, int index) { ... struct fwnode_reference_args fwspec; // (1) ... ret = fwnode_property_get_reference_args(fwnode, "mboxes", // (2) "#mbox-cells", 0, index, &fwspec); ... scoped_guard(mutex, &con_mutex) { ... list_for_each_entry(mbox, &mbox_cons, node) { if (device_match_fwnode(mbox->dev, fwspec.fwnode)) { if (mbox->fw_xlate) { chan = mbox->fw_xlate(mbox, &fwspec); // (3) if (!IS_ERR(chan)) break; } ... } } ... ret = __mbox_bind_client(chan, cl); // (7) ... } ... } static int __mbox_bind_client(struct mbox_chan *chan, struct mbox_client *cl) { if (chan->cl || ...) { // (8) } ``` (1) `fwspec.args[]` is filled with arbitrary leftover values in the stack. Let's say that `fwspec.args[0] == 0xffffffff`. (2) Since `#mbox-cells = <0>;`, `fwspec.nargs` is assigned 0 and `fwspec.args[]` are untouched. (3) Since the controller has not provided `fw_xlate` and `of_xlate`, `fw_mbox_index_xlate()` is used instead. (4) `idx` is assigned -1 due to the value of `fwspec.args[0]`. (5) Since `mbox->num_chans >= 0` and `idx == -1`, this condition does not filter out this case. (6) Out-of-bounds address is returned. Depending on what was left in `fwspec.args[0]`, it could be an arbitrary(but confined to a specific range) address. (7) A function is called with the out-of-bounds address in `chan`. (8) The out-of-bounds address is accessed. drivers/mailbox/mailbox.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) diff --git a/drivers/mailbox/mailbox.c b/drivers/mailbox/mailbox.c index 2acc6ec229a4..617ba505691d 100644 --- a/drivers/mailbox/mailbox.c +++ b/drivers/mailbox/mailbox.c @@ -489,12 +489,10 @@ EXPORT_SYMBOL_GPL(mbox_free_channel); static struct mbox_chan *fw_mbox_index_xlate(struct mbox_controller *mbox, const struct fwnode_reference_args *sp) { - int ind = sp->args[0]; - - if (ind >= mbox->num_chans) + if (sp->nargs < 1 || sp->args[0] >= mbox->num_chans) return ERR_PTR(-EINVAL); - return &mbox->chans[ind]; + return &mbox->chans[sp->args[0]]; } /** -- 2.52.0.487.g5c8c507ade-goog

2 weeks, 1 day

1
0
0 0

[PATCH 5.4-6.6 v2] uio_hv_generic: Enable user space to manage interrupt_mask for subchannels

by Naman Jain

From: Long Li <longli(a)microsoft.com> Enable the user space to manage interrupt_mask for subchannels through irqcontrol interface for uio device. Also remove the memory barrier when monitor bit is enabled as it is not necessary. This is a backport of the upstream commit d062463edf17 ("uio_hv_generic: Set event for all channels on the device") with some modifications to resolve merge conflicts and take care of missing support for slow devices on older kernels. Original change was not a fix, but it needs to be backported to fix a NULL pointer crash resulting from missing interrupt mask setting. Commit b15b7d2a1b09 ("uio_hv_generic: Let userspace take care of interrupt mask") removed the default setting of interrupt_mask for channels (including subchannels) in the uio_hv_generic driver, as it relies on the user space to take care of managing it. This approach works fine when user space can control this setting using the irqcontrol interface provided for uio devices. Support for setting the interrupt mask through this interface for subchannels came only after commit d062463edf17 ("uio_hv_generic: Set event for all channels on the device"). On older kernels, this change is not present. With uio_hv_generic no longer setting the interrupt_mask, and userspace not having the capability to set it, it remains unset, and interrupts can come for the subchannels, which can result in a crash in hv_uio_channel_cb. Backport the change to older kernels, where this change was not present, to allow userspace to set the interrupt mask properly for subchannels. Additionally, this patch also adds certain checks for primary vs subchannels in the hv_uio_channel_cb, which can gracefully handle these two cases and prevent the NULL pointer crashes. Signed-off-by: Long Li <longli(a)microsoft.com> Fixes: b15b7d2a1b09 ("uio_hv_generic: Let userspace take care of interrupt mask") Closes: https://bugs.debian.org/1120602 Cc: stable(a)vger.kernel.org # all LTS kernels from 5.4 till 6.6 Signed-off-by: Naman Jain <namjain(a)linux.microsoft.com> --- Changes since v1: https://lore.kernel.org/all/20251115085937.2237-1-namjain@linux.microsoft.c… * Changed commit being fixed, to reflect upstream commit * Added stable kernel versions info in Stable tag and email subject. This needs to be ported to all the stable kernels, where the Fixes patch went in - 5.4.x, 5.10.x, 5.15.x, 6.1.x, 6.6.x. Previous notes: Remove reviewed-by tags since the original code has changed quite a bit while backporting. Backported change for 6.12 kernel is sent separately. --- drivers/uio/uio_hv_generic.c | 21 +++++++++++++++++---- 1 file changed, 17 insertions(+), 4 deletions(-) diff --git a/drivers/uio/uio_hv_generic.c b/drivers/uio/uio_hv_generic.c index 137109f5f69b..95c1f08028a3 100644 --- a/drivers/uio/uio_hv_generic.c +++ b/drivers/uio/uio_hv_generic.c @@ -80,9 +80,15 @@ hv_uio_irqcontrol(struct uio_info *info, s32 irq_state) { struct hv_uio_private_data *pdata = info->priv; struct hv_device *dev = pdata->device; + struct vmbus_channel *primary, *sc; - dev->channel->inbound.ring_buffer->interrupt_mask = !irq_state; - virt_mb(); + primary = dev->channel; + primary->inbound.ring_buffer->interrupt_mask = !irq_state; + + mutex_lock(&vmbus_connection.channel_mutex); + list_for_each_entry(sc, &primary->sc_list, sc_list) + sc->inbound.ring_buffer->interrupt_mask = !irq_state; + mutex_unlock(&vmbus_connection.channel_mutex); return 0; } @@ -93,11 +99,18 @@ hv_uio_irqcontrol(struct uio_info *info, s32 irq_state) static void hv_uio_channel_cb(void *context) { struct vmbus_channel *chan = context; - struct hv_device *hv_dev = chan->device_obj; - struct hv_uio_private_data *pdata = hv_get_drvdata(hv_dev); + struct hv_device *hv_dev; + struct hv_uio_private_data *pdata; virt_mb(); + /* + * The callback may come from a subchannel, in which case look + * for the hv device in the primary channel + */ + hv_dev = chan->primary_channel ? + chan->primary_channel->device_obj : chan->device_obj; + pdata = hv_get_drvdata(hv_dev); uio_event_notify(&pdata->info); } -- 2.43.0

2 weeks, 1 day

1
0
0 0

[PATCH 6.6 and older] uio_hv_generic: Enable user space to manage interrupt_mask for subchannels

by Naman Jain

From: Long Li <longli(a)microsoft.com> Enable the user space to manage interrupt_mask for subchannels through irqcontrol interface for uio device. Also remove the memory barrier when monitor bit is enabled as it is not necessary. This is a backport of the upstream commit d062463edf17 ("uio_hv_generic: Set event for all channels on the device") with some modifications to resolve merge conflicts and take care of missing support for slow devices on older kernels. Original change was not a fix, but it needs to be backported to fix a NULL pointer crash resulting from missing interrupt mask setting. Commit 37bd91f22794 ("uio_hv_generic: Let userspace take care of interrupt mask") removed the default setting of interrupt_mask for channels (including subchannels) in the uio_hv_generic driver, as it relies on the user space to take care of managing it. This approach works fine when user space can control this setting using the irqcontrol interface provided for uio devices. Support for setting the interrupt mask through this interface for subchannels came only after commit d062463edf17 ("uio_hv_generic: Set event for all channels on the device"). On older kernels, this change is not present. With uio_hv_generic no longer setting the interrupt_mask, and userspace not having the capability to set it, it remains unset, and interrupts can come for the subchannels, which can result in a crash in hv_uio_channel_cb. Backport the change to older kernels, where this change was not present, to allow userspace to set the interrupt mask properly for subchannels. Additionally, this patch also adds certain checks for primary vs subchannels in the hv_uio_channel_cb, which can gracefully handle these two cases and prevent the NULL pointer crashes. Signed-off-by: Long Li <longli(a)microsoft.com> Fixes: 37bd91f22794 ("uio_hv_generic: Let userspace take care of interrupt mask") Closes: https://bugs.debian.org/1120602 Cc: <stable(a)vger.kernel.org> # 6.6.x and older Signed-off-by: Naman Jain <namjain(a)linux.microsoft.com> --- Remove reviewed-by tags since the original code has changed quite a bit while backporting. Backported change for 6.12 kernel is sent separately. --- drivers/uio/uio_hv_generic.c | 21 +++++++++++++++++---- 1 file changed, 17 insertions(+), 4 deletions(-) diff --git a/drivers/uio/uio_hv_generic.c b/drivers/uio/uio_hv_generic.c index 2724656bf634..69e5016ebd46 100644 --- a/drivers/uio/uio_hv_generic.c +++ b/drivers/uio/uio_hv_generic.c @@ -80,9 +80,15 @@ hv_uio_irqcontrol(struct uio_info *info, s32 irq_state) { struct hv_uio_private_data *pdata = info->priv; struct hv_device *dev = pdata->device; + struct vmbus_channel *primary, *sc; - dev->channel->inbound.ring_buffer->interrupt_mask = !irq_state; - virt_mb(); + primary = dev->channel; + primary->inbound.ring_buffer->interrupt_mask = !irq_state; + + mutex_lock(&vmbus_connection.channel_mutex); + list_for_each_entry(sc, &primary->sc_list, sc_list) + sc->inbound.ring_buffer->interrupt_mask = !irq_state; + mutex_unlock(&vmbus_connection.channel_mutex); return 0; } @@ -93,11 +99,18 @@ hv_uio_irqcontrol(struct uio_info *info, s32 irq_state) static void hv_uio_channel_cb(void *context) { struct vmbus_channel *chan = context; - struct hv_device *hv_dev = chan->device_obj; - struct hv_uio_private_data *pdata = hv_get_drvdata(hv_dev); + struct hv_device *hv_dev; + struct hv_uio_private_data *pdata; virt_mb(); + /* + * The callback may come from a subchannel, in which case look + * for the hv device in the primary channel + */ + hv_dev = chan->primary_channel ? + chan->primary_channel->device_obj : chan->device_obj; + pdata = hv_get_drvdata(hv_dev); uio_event_notify(&pdata->info); } -- 2.34.1

2 weeks, 1 day

3
3
0 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

Linux-stable-mirror