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

[PATCH 1/2] platform/x86: ISST: Add missing write block check

by Srinivas Pandruvada

If writes are blocked, then return error during SST-CP enable command. Add missing write block check in this code path. Fixes: 8bed9ff7dbcc ("platform/x86: ISST: Process read/write blocked feature status") Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada(a)linux.intel.com> Cc: <stable(a)vger.kernel.org> --- drivers/platform/x86/intel/speed_select_if/isst_tpmi_core.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/platform/x86/intel/speed_select_if/isst_tpmi_core.c b/drivers/platform/x86/intel/speed_select_if/isst_tpmi_core.c index 34bff2f65a83..f587709ddd47 100644 --- a/drivers/platform/x86/intel/speed_select_if/isst_tpmi_core.c +++ b/drivers/platform/x86/intel/speed_select_if/isst_tpmi_core.c @@ -612,6 +612,9 @@ static long isst_if_core_power_state(void __user *argp) return -EINVAL; if (core_power.get_set) { + if (power_domain_info->write_blocked) + return -EPERM; + _write_cp_info("cp_enable", core_power.enable, SST_CP_CONTROL_OFFSET, SST_CP_ENABLE_START, SST_CP_ENABLE_WIDTH, SST_MUL_FACTOR_NONE) _write_cp_info("cp_prio_type", core_power.priority_type, SST_CP_CONTROL_OFFSET, -- 2.52.0

1 day

1
0
0 0

[PATCH net] gve: defer interrupt enabling until NAPI registration

by Harshitha Ramamurthy

From: Ankit Garg <nktgrg(a)google.com> Currently, interrupts are automatically enabled immediately upon request. This allows interrupt to fire before the associated NAPI context is fully initialized and cause failures like below: [ 0.946369] Call Trace: [ 0.946369] <IRQ> [ 0.946369] __napi_poll+0x2a/0x1e0 [ 0.946369] net_rx_action+0x2f9/0x3f0 [ 0.946369] handle_softirqs+0xd6/0x2c0 [ 0.946369] ? handle_edge_irq+0xc1/0x1b0 [ 0.946369] __irq_exit_rcu+0xc3/0xe0 [ 0.946369] common_interrupt+0x81/0xa0 [ 0.946369] </IRQ> [ 0.946369] <TASK> [ 0.946369] asm_common_interrupt+0x22/0x40 [ 0.946369] RIP: 0010:pv_native_safe_halt+0xb/0x10 Use the `IRQF_NO_AUTOEN` flag when requesting interrupts to prevent auto enablement and explicitly enable the interrupt in NAPI initialization path (and disable it during NAPI teardown). This ensures that interrupt lifecycle is strictly coupled with readiness of NAPI context. Cc: stable(a)vger.kernel.org Fixes: 1dfc2e46117e ("gve: Refactor napi add and remove functions") Signed-off-by: Ankit Garg <nktgrg(a)google.com> Reviewed-by: Jordan Rhee <jordanrhee(a)google.com> Reviewed-by: Joshua Washington <joshwash(a)google.com> Signed-off-by: Harshitha Ramamurthy <hramamurthy(a)google.com> --- drivers/net/ethernet/google/gve/gve_main.c | 2 +- drivers/net/ethernet/google/gve/gve_utils.c | 2 ++ 2 files changed, 3 insertions(+), 1 deletion(-) diff --git a/drivers/net/ethernet/google/gve/gve_main.c b/drivers/net/ethernet/google/gve/gve_main.c index 5a74760..96adbbe 100644 --- a/drivers/net/ethernet/google/gve/gve_main.c +++ b/drivers/net/ethernet/google/gve/gve_main.c @@ -558,7 +558,7 @@ static int gve_alloc_notify_blocks(struct gve_priv *priv) block->priv = priv; err = request_irq(priv->msix_vectors[msix_idx].vector, gve_is_gqi(priv) ? gve_intr : gve_intr_dqo, - 0, block->name, block); + IRQF_NO_AUTOEN, block->name, block); if (err) { dev_err(&priv->pdev->dev, "Failed to receive msix vector %d\n", i); diff --git a/drivers/net/ethernet/google/gve/gve_utils.c b/drivers/net/ethernet/google/gve/gve_utils.c index ace9b86..b53b7fc 100644 --- a/drivers/net/ethernet/google/gve/gve_utils.c +++ b/drivers/net/ethernet/google/gve/gve_utils.c @@ -112,11 +112,13 @@ void gve_add_napi(struct gve_priv *priv, int ntfy_idx, netif_napi_add_locked(priv->dev, &block->napi, gve_poll); netif_napi_set_irq_locked(&block->napi, block->irq); + enable_irq(block->irq); } void gve_remove_napi(struct gve_priv *priv, int ntfy_idx) { struct gve_notify_block *block = &priv->ntfy_blocks[ntfy_idx]; + disable_irq(block->irq); netif_napi_del_locked(&block->napi); } -- 2.52.0.322.g1dd061c0dc-goog

1 day, 1 hour

2
1
0 0

[PATCH] gfs2: Fix use-after-free in gfs2_fill_super

by Ryota Sakamoto

The issue occurs when gfs2_freeze_lock_shared() fails in gfs2_fill_super(). If !sb_rdonly(sb), threads for the quotad and logd were started, however, in the error path for gfs2_freeze_lock_shared(), the threads are not stopped by gfs2_destroy_threads() before jumping to fail_per_node. This patch introduces fail_threads to handle stopping the threads if the threads were started. Reported-by: syzbot+4cb0d0336db6bc6930e9(a)syzkaller.appspotmail.com Closes: https://syzkaller.appspot.com/bug?extid=4cb0d0336db6bc6930e9 Fixes: a28dc123fa66 ("gfs2: init system threads before freeze lock") Cc: stable(a)vger.kernel.org Signed-off-by: Ryota Sakamoto <sakamo.ryota(a)gmail.com> --- fs/gfs2/ops_fstype.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/fs/gfs2/ops_fstype.c b/fs/gfs2/ops_fstype.c index e7a88b717991ae3647c1da039636daef7005a7f0..4b5ac1a7050f1fd34e10be4100a2bc381f49c83d 100644 --- a/fs/gfs2/ops_fstype.c +++ b/fs/gfs2/ops_fstype.c @@ -1269,21 +1269,23 @@ static int gfs2_fill_super(struct super_block *sb, struct fs_context *fc) error = gfs2_freeze_lock_shared(sdp); if (error) - goto fail_per_node; + goto fail_threads; if (!sb_rdonly(sb)) error = gfs2_make_fs_rw(sdp); if (error) { gfs2_freeze_unlock(sdp); - gfs2_destroy_threads(sdp); fs_err(sdp, "can't make FS RW: %d\n", error); - goto fail_per_node; + goto fail_threads; } gfs2_glock_dq_uninit(&mount_gh); gfs2_online_uevent(sdp); return 0; +fail_threads: + if (!sb_rdonly(sb)) + gfs2_destroy_threads(sdp); fail_per_node: init_per_node(sdp, UNDO); fail_inodes: --- base-commit: 7839932417dd53bb09eb5a585a7a92781dfd7cb2 change-id: 20251230-fix-use-after-free-gfs2-66cfbe23baa8 Best regards, -- Ryota Sakamoto <sakamo.ryota(a)gmail.com>

1 day, 1 hour

2
2
0 0

[PATCH] rust_binder: correctly handle FDA objects of length zero

by Alice Ryhl

Fix a bug where an empty FDA (fd array) object with 0 fds would cause an out-of-bounds error. The previous implementation used `skip == 0` to mean "this is a pointer fixup", but 0 is also the correct skip length for an empty FDA. If the FDA is at the end of the buffer, then this results in an attempt to write 8-bytes out of bounds. This is caught and results in an EINVAL error being returned to userspace. The pattern of using `skip == 0` as a special value originates from the C-implementation of Binder. As part of fixing this bug, this pattern is replaced with a Rust enum. I considered the alternate option of not pushing a fixup when the length is zero, but I think it's cleaner to just get rid of the zero-is-special stuff. The root cause of this bug was diagnosed by Gemini CLI on first try. I used the following prompt: > There appears to be a bug in @drivers/android/binder/thread.rs where > the Fixups oob bug is triggered with 316 304 316 324. This implies > that we somehow ended up with a fixup where buffer A has a pointer to > buffer B, but the pointer is located at an index in buffer A that is > out of bounds. Please investigate the code to find the bug. You may > compare with @drivers/android/binder.c that implements this correctly. Cc: stable(a)vger.kernel.org Reported-by: DeepChirp <DeepChirp(a)outlook.com> Closes: https://github.com/waydroid/waydroid/issues/2157 Fixes: eafedbc7c050 ("rust_binder: add Rust Binder driver") Tested-by: DeepChirp <DeepChirp(a)outlook.com> Signed-off-by: Alice Ryhl <aliceryhl(a)google.com> --- drivers/android/binder/thread.rs | 59 +++++++++++++++++++++++----------------- 1 file changed, 34 insertions(+), 25 deletions(-) diff --git a/drivers/android/binder/thread.rs b/drivers/android/binder/thread.rs index 1a8e6fdc0dc42369ee078e720aa02b2554fb7332..dcd47e10aeb8c748d04320fbbe15ad35201684b9 100644 --- a/drivers/android/binder/thread.rs +++ b/drivers/android/binder/thread.rs @@ -69,17 +69,24 @@ struct ScatterGatherEntry { } /// This entry specifies that a fixup should happen at `target_offset` of the -/// buffer. If `skip` is nonzero, then the fixup is a `binder_fd_array_object` -/// and is applied later. Otherwise if `skip` is zero, then the size of the -/// fixup is `sizeof::<u64>()` and `pointer_value` is written to the buffer. -struct PointerFixupEntry { - /// The number of bytes to skip, or zero for a `binder_buffer_object` fixup. - skip: usize, - /// The translated pointer to write when `skip` is zero. - pointer_value: u64, - /// The offset at which the value should be written. The offset is relative - /// to the original buffer. - target_offset: usize, +/// buffer. +enum PointerFixupEntry { + /// A fixup for a `binder_buffer_object`. + Fixup { + /// The translated pointer to write. + pointer_value: u64, + /// The offset at which the value should be written. The offset is relative + /// to the original buffer. + target_offset: usize, + }, + /// A skip for a `binder_fd_array_object`. + Skip { + /// The number of bytes to skip. + skip: usize, + /// The offset at which the skip should happen. The offset is relative + /// to the original buffer. + target_offset: usize, + }, } /// Return type of `apply_and_validate_fixup_in_parent`. @@ -762,8 +769,7 @@ fn translate_object( parent_entry.fixup_min_offset = info.new_min_offset; parent_entry.pointer_fixups.push( - PointerFixupEntry { - skip: 0, + PointerFixupEntry::Fixup { pointer_value: buffer_ptr_in_user_space, target_offset: info.target_offset, }, @@ -807,9 +813,8 @@ fn translate_object( parent_entry .pointer_fixups .push( - PointerFixupEntry { + PointerFixupEntry::Skip { skip: fds_len, - pointer_value: 0, target_offset: info.target_offset, }, GFP_KERNEL, @@ -871,17 +876,21 @@ fn apply_sg(&self, alloc: &mut Allocation, sg_state: &mut ScatterGatherState) -> let mut reader = UserSlice::new(UserPtr::from_addr(sg_entry.sender_uaddr), sg_entry.length).reader(); for fixup in &mut sg_entry.pointer_fixups { - let fixup_len = if fixup.skip == 0 { - size_of::<u64>() - } else { - fixup.skip + let (fixup_len, fixup_offset) = match fixup { + PointerFixupEntry::Fixup { target_offset, .. } => { + (size_of::<u64>(), *target_offset) + } + PointerFixupEntry::Skip { + skip, + target_offset, + } => (*skip, *target_offset), }; - let target_offset_end = fixup.target_offset.checked_add(fixup_len).ok_or(EINVAL)?; - if fixup.target_offset < end_of_previous_fixup || offset_end < target_offset_end { + let target_offset_end = fixup_offset.checked_add(fixup_len).ok_or(EINVAL)?; + if fixup_offset < end_of_previous_fixup || offset_end < target_offset_end { pr_warn!( "Fixups oob {} {} {} {}", - fixup.target_offset, + fixup_offset, end_of_previous_fixup, offset_end, target_offset_end @@ -890,13 +899,13 @@ fn apply_sg(&self, alloc: &mut Allocation, sg_state: &mut ScatterGatherState) -> } let copy_off = end_of_previous_fixup; - let copy_len = fixup.target_offset - end_of_previous_fixup; + let copy_len = fixup_offset - end_of_previous_fixup; if let Err(err) = alloc.copy_into(&mut reader, copy_off, copy_len) { pr_warn!("Failed copying into alloc: {:?}", err); return Err(err.into()); } - if fixup.skip == 0 { - let res = alloc.write::<u64>(fixup.target_offset, &fixup.pointer_value); + if let PointerFixupEntry::Fixup { pointer_value, .. } = fixup { + let res = alloc.write::<u64>(fixup_offset, pointer_value); if let Err(err) = res { pr_warn!("Failed copying ptr into alloc: {:?}", err); return Err(err.into()); --- base-commit: 8f0b4cce4481fb22653697cced8d0d04027cb1e8 change-id: 20251229-fda-zero-4e46e56be58d Best regards, -- Alice Ryhl <aliceryhl(a)google.com>

1 day, 2 hours

2
2
0 0

[PATCH] dmaengine: fix a reference leak in __dma_async_device_channel_register()

by Haoxiang Li

After device_register() is called, put_device() is required to drop the device reference. In this case, put_device() -> chan_dev_release() will finally release chan->dev. Thus there is no need to call kfree again to release the chan->dev. Found by code review. Fixes: d2fb0a043838 ("dmaengine: break out channel registration") Cc: stable(a)vger.kernel.org Signed-off-by: Haoxiang Li <lihaoxiang(a)isrc.iscas.ac.cn> --- drivers/dma/dmaengine.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/drivers/dma/dmaengine.c b/drivers/dma/dmaengine.c index ca13cd39330b..9d7cea1d6e91 100644 --- a/drivers/dma/dmaengine.c +++ b/drivers/dma/dmaengine.c @@ -1071,6 +1071,7 @@ static int __dma_async_device_channel_register(struct dma_device *device, const char *name) { int rc; + bool dev_registered = false; chan->local = alloc_percpu(typeof(*chan->local)); if (!chan->local) @@ -1102,6 +1103,7 @@ static int __dma_async_device_channel_register(struct dma_device *device, else dev_set_name(&chan->dev->device, "%s", name); rc = device_register(&chan->dev->device); + dev_registered = true; if (rc) goto err_out_ida; chan->client_count = 0; @@ -1112,7 +1114,10 @@ static int __dma_async_device_channel_register(struct dma_device *device, err_out_ida: ida_free(&device->chan_ida, chan->chan_id); err_free_dev: - kfree(chan->dev); + if (dev_registered) + put_device(&chan->dev->device); + else + kfree(chan->dev); err_free_local: free_percpu(chan->local); chan->local = NULL; -- 2.25.1

1 day, 2 hours

2
1
0 0

[report] Performance regressions introduced via "cpuidle: menu: Remove iowait influence" on 6.12.y

by ALOK TIWARI

Hi, I’m reporting a performance regression of up to 6% sequential I/O vdbench regression observed on 6.12.y kernel. While running performance benchmarks on v6.12.60 kernel the sequential I/O vdbench metrics are showing a 5-6% performance regression when compared to v6.12.48 Bisect root cause commit ======================== - commit b39b62075ab4 ("cpuidle: menu: Remove iowait influence") Things work fine again when the previously removed performance-multiplier code is added back. Test details ============ The system is connected to a number of disks in disk array using multipathing and directio configuration in the vdbench profile. wd=wd1,sd=sd*,rdpct=0,seekpct=sequential,xfersize=128k rd=128k64T,wd=wd1,iorate=max,elapsed=600,interval=1,warmup=300,threads=64 Thanks, Alok

1 day, 2 hours

3
3
0 0

FAILED: patch "[PATCH] fuse: fix readahead reclaim deadlock" failed to apply to 5.15-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 5.15-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. To reproduce the conflict and resubmit, you may use the following commands: git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-5.15.y git checkout FETCH_HEAD git cherry-pick -x bd5603eaae0aabf527bfb3ce1bb07e979ce5bd50 # <resolve conflicts, build, test, etc.> git commit -s git send-email --to '<stable(a)vger.kernel.org>' --in-reply-to '2025122915-discover-sediment-3371@gregkh' --subject-prefix 'PATCH 5.15.y' HEAD^.. Possible dependencies: thanks, greg k-h ------------------ original commit in Linus's tree ------------------ From bd5603eaae0aabf527bfb3ce1bb07e979ce5bd50 Mon Sep 17 00:00:00 2001 From: Joanne Koong <joannelkoong(a)gmail.com> Date: Fri, 10 Oct 2025 15:07:38 -0700 Subject: [PATCH] fuse: fix readahead reclaim deadlock Commit e26ee4efbc79 ("fuse: allocate ff->release_args only if release is needed") skips allocating ff->release_args if the server does not implement open. However in doing so, fuse_prepare_release() now skips grabbing the reference on the inode, which makes it possible for an inode to be evicted from the dcache while there are inflight readahead requests. This causes a deadlock if the server triggers reclaim while servicing the readahead request and reclaim attempts to evict the inode of the file being read ahead. Since the folio is locked during readahead, when reclaim evicts the fuse inode and fuse_evict_inode() attempts to remove all folios associated with the inode from the page cache (truncate_inode_pages_range()), reclaim will block forever waiting for the lock since readahead cannot relinquish the lock because it is itself blocked in reclaim: >>> stack_trace(1504735) folio_wait_bit_common (mm/filemap.c:1308:4) folio_lock (./include/linux/pagemap.h:1052:3) truncate_inode_pages_range (mm/truncate.c:336:10) fuse_evict_inode (fs/fuse/inode.c:161:2) evict (fs/inode.c:704:3) dentry_unlink_inode (fs/dcache.c:412:3) __dentry_kill (fs/dcache.c:615:3) shrink_kill (fs/dcache.c:1060:12) shrink_dentry_list (fs/dcache.c:1087:3) prune_dcache_sb (fs/dcache.c:1168:2) super_cache_scan (fs/super.c:221:10) do_shrink_slab (mm/shrinker.c:435:9) shrink_slab (mm/shrinker.c:626:10) shrink_node (mm/vmscan.c:5951:2) shrink_zones (mm/vmscan.c:6195:3) do_try_to_free_pages (mm/vmscan.c:6257:3) do_swap_page (mm/memory.c:4136:11) handle_pte_fault (mm/memory.c:5562:10) handle_mm_fault (mm/memory.c:5870:9) do_user_addr_fault (arch/x86/mm/fault.c:1338:10) handle_page_fault (arch/x86/mm/fault.c:1481:3) exc_page_fault (arch/x86/mm/fault.c:1539:2) asm_exc_page_fault+0x22/0x27 Fix this deadlock by allocating ff->release_args and grabbing the reference on the inode when preparing the file for release even if the server does not implement open. The inode reference will be dropped when the last reference on the fuse file is dropped (see fuse_file_put() -> fuse_release_end()). Fixes: e26ee4efbc79 ("fuse: allocate ff->release_args only if release is needed") Cc: stable(a)vger.kernel.org Signed-off-by: Joanne Koong <joannelkoong(a)gmail.com> Reported-by: Omar Sandoval <osandov(a)fb.com> Signed-off-by: Miklos Szeredi <mszeredi(a)redhat.com> diff --git a/fs/fuse/file.c b/fs/fuse/file.c index f1ef77a0be05..4d96e684d736 100644 --- a/fs/fuse/file.c +++ b/fs/fuse/file.c @@ -110,7 +110,9 @@ static void fuse_file_put(struct fuse_file *ff, bool sync) fuse_file_io_release(ff, ra->inode); if (!args) { - /* Do nothing when server does not implement 'open' */ + /* Do nothing when server does not implement 'opendir' */ + } else if (args->opcode == FUSE_RELEASE && ff->fm->fc->no_open) { + fuse_release_end(ff->fm, args, 0); } else if (sync) { fuse_simple_request(ff->fm, args); fuse_release_end(ff->fm, args, 0); @@ -131,8 +133,17 @@ struct fuse_file *fuse_file_open(struct fuse_mount *fm, u64 nodeid, struct fuse_file *ff; int opcode = isdir ? FUSE_OPENDIR : FUSE_OPEN; bool open = isdir ? !fc->no_opendir : !fc->no_open; + bool release = !isdir || open; - ff = fuse_file_alloc(fm, open); + /* + * ff->args->release_args still needs to be allocated (so we can hold an + * inode reference while there are pending inflight file operations when + * ->release() is called, see fuse_prepare_release()) even if + * fc->no_open is set else it becomes possible for reclaim to deadlock + * if while servicing the readahead request the server triggers reclaim + * and reclaim evicts the inode of the file being read ahead. + */ + ff = fuse_file_alloc(fm, release); if (!ff) return ERR_PTR(-ENOMEM); @@ -152,13 +163,14 @@ struct fuse_file *fuse_file_open(struct fuse_mount *fm, u64 nodeid, fuse_file_free(ff); return ERR_PTR(err); } else { - /* No release needed */ - kfree(ff->args); - ff->args = NULL; - if (isdir) + if (isdir) { + /* No release needed */ + kfree(ff->args); + ff->args = NULL; fc->no_opendir = 1; - else + } else { fc->no_open = 1; + } } }

1 day, 3 hours

1
0
0 0

FAILED: patch "[PATCH] fuse: fix readahead reclaim deadlock" failed to apply to 6.1-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 6.1-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. To reproduce the conflict and resubmit, you may use the following commands: git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-6.1.y git checkout FETCH_HEAD git cherry-pick -x bd5603eaae0aabf527bfb3ce1bb07e979ce5bd50 # <resolve conflicts, build, test, etc.> git commit -s git send-email --to '<stable(a)vger.kernel.org>' --in-reply-to '2025122914-hundredth-munchkin-531c@gregkh' --subject-prefix 'PATCH 6.1.y' HEAD^.. Possible dependencies: thanks, greg k-h ------------------ original commit in Linus's tree ------------------ From bd5603eaae0aabf527bfb3ce1bb07e979ce5bd50 Mon Sep 17 00:00:00 2001 From: Joanne Koong <joannelkoong(a)gmail.com> Date: Fri, 10 Oct 2025 15:07:38 -0700 Subject: [PATCH] fuse: fix readahead reclaim deadlock Commit e26ee4efbc79 ("fuse: allocate ff->release_args only if release is needed") skips allocating ff->release_args if the server does not implement open. However in doing so, fuse_prepare_release() now skips grabbing the reference on the inode, which makes it possible for an inode to be evicted from the dcache while there are inflight readahead requests. This causes a deadlock if the server triggers reclaim while servicing the readahead request and reclaim attempts to evict the inode of the file being read ahead. Since the folio is locked during readahead, when reclaim evicts the fuse inode and fuse_evict_inode() attempts to remove all folios associated with the inode from the page cache (truncate_inode_pages_range()), reclaim will block forever waiting for the lock since readahead cannot relinquish the lock because it is itself blocked in reclaim: >>> stack_trace(1504735) folio_wait_bit_common (mm/filemap.c:1308:4) folio_lock (./include/linux/pagemap.h:1052:3) truncate_inode_pages_range (mm/truncate.c:336:10) fuse_evict_inode (fs/fuse/inode.c:161:2) evict (fs/inode.c:704:3) dentry_unlink_inode (fs/dcache.c:412:3) __dentry_kill (fs/dcache.c:615:3) shrink_kill (fs/dcache.c:1060:12) shrink_dentry_list (fs/dcache.c:1087:3) prune_dcache_sb (fs/dcache.c:1168:2) super_cache_scan (fs/super.c:221:10) do_shrink_slab (mm/shrinker.c:435:9) shrink_slab (mm/shrinker.c:626:10) shrink_node (mm/vmscan.c:5951:2) shrink_zones (mm/vmscan.c:6195:3) do_try_to_free_pages (mm/vmscan.c:6257:3) do_swap_page (mm/memory.c:4136:11) handle_pte_fault (mm/memory.c:5562:10) handle_mm_fault (mm/memory.c:5870:9) do_user_addr_fault (arch/x86/mm/fault.c:1338:10) handle_page_fault (arch/x86/mm/fault.c:1481:3) exc_page_fault (arch/x86/mm/fault.c:1539:2) asm_exc_page_fault+0x22/0x27 Fix this deadlock by allocating ff->release_args and grabbing the reference on the inode when preparing the file for release even if the server does not implement open. The inode reference will be dropped when the last reference on the fuse file is dropped (see fuse_file_put() -> fuse_release_end()). Fixes: e26ee4efbc79 ("fuse: allocate ff->release_args only if release is needed") Cc: stable(a)vger.kernel.org Signed-off-by: Joanne Koong <joannelkoong(a)gmail.com> Reported-by: Omar Sandoval <osandov(a)fb.com> Signed-off-by: Miklos Szeredi <mszeredi(a)redhat.com> diff --git a/fs/fuse/file.c b/fs/fuse/file.c index f1ef77a0be05..4d96e684d736 100644 --- a/fs/fuse/file.c +++ b/fs/fuse/file.c @@ -110,7 +110,9 @@ static void fuse_file_put(struct fuse_file *ff, bool sync) fuse_file_io_release(ff, ra->inode); if (!args) { - /* Do nothing when server does not implement 'open' */ + /* Do nothing when server does not implement 'opendir' */ + } else if (args->opcode == FUSE_RELEASE && ff->fm->fc->no_open) { + fuse_release_end(ff->fm, args, 0); } else if (sync) { fuse_simple_request(ff->fm, args); fuse_release_end(ff->fm, args, 0); @@ -131,8 +133,17 @@ struct fuse_file *fuse_file_open(struct fuse_mount *fm, u64 nodeid, struct fuse_file *ff; int opcode = isdir ? FUSE_OPENDIR : FUSE_OPEN; bool open = isdir ? !fc->no_opendir : !fc->no_open; + bool release = !isdir || open; - ff = fuse_file_alloc(fm, open); + /* + * ff->args->release_args still needs to be allocated (so we can hold an + * inode reference while there are pending inflight file operations when + * ->release() is called, see fuse_prepare_release()) even if + * fc->no_open is set else it becomes possible for reclaim to deadlock + * if while servicing the readahead request the server triggers reclaim + * and reclaim evicts the inode of the file being read ahead. + */ + ff = fuse_file_alloc(fm, release); if (!ff) return ERR_PTR(-ENOMEM); @@ -152,13 +163,14 @@ struct fuse_file *fuse_file_open(struct fuse_mount *fm, u64 nodeid, fuse_file_free(ff); return ERR_PTR(err); } else { - /* No release needed */ - kfree(ff->args); - ff->args = NULL; - if (isdir) + if (isdir) { + /* No release needed */ + kfree(ff->args); + ff->args = NULL; fc->no_opendir = 1; - else + } else { fc->no_open = 1; + } } }

1 day, 3 hours

1
0
0 0

FAILED: patch "[PATCH] fuse: fix readahead reclaim deadlock" failed to apply to 6.6-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 6.6-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. To reproduce the conflict and resubmit, you may use the following commands: git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-6.6.y git checkout FETCH_HEAD git cherry-pick -x bd5603eaae0aabf527bfb3ce1bb07e979ce5bd50 # <resolve conflicts, build, test, etc.> git commit -s git send-email --to '<stable(a)vger.kernel.org>' --in-reply-to '2025122913-nucleus-deny-f4d4@gregkh' --subject-prefix 'PATCH 6.6.y' HEAD^.. Possible dependencies: thanks, greg k-h ------------------ original commit in Linus's tree ------------------ From bd5603eaae0aabf527bfb3ce1bb07e979ce5bd50 Mon Sep 17 00:00:00 2001 From: Joanne Koong <joannelkoong(a)gmail.com> Date: Fri, 10 Oct 2025 15:07:38 -0700 Subject: [PATCH] fuse: fix readahead reclaim deadlock Commit e26ee4efbc79 ("fuse: allocate ff->release_args only if release is needed") skips allocating ff->release_args if the server does not implement open. However in doing so, fuse_prepare_release() now skips grabbing the reference on the inode, which makes it possible for an inode to be evicted from the dcache while there are inflight readahead requests. This causes a deadlock if the server triggers reclaim while servicing the readahead request and reclaim attempts to evict the inode of the file being read ahead. Since the folio is locked during readahead, when reclaim evicts the fuse inode and fuse_evict_inode() attempts to remove all folios associated with the inode from the page cache (truncate_inode_pages_range()), reclaim will block forever waiting for the lock since readahead cannot relinquish the lock because it is itself blocked in reclaim: >>> stack_trace(1504735) folio_wait_bit_common (mm/filemap.c:1308:4) folio_lock (./include/linux/pagemap.h:1052:3) truncate_inode_pages_range (mm/truncate.c:336:10) fuse_evict_inode (fs/fuse/inode.c:161:2) evict (fs/inode.c:704:3) dentry_unlink_inode (fs/dcache.c:412:3) __dentry_kill (fs/dcache.c:615:3) shrink_kill (fs/dcache.c:1060:12) shrink_dentry_list (fs/dcache.c:1087:3) prune_dcache_sb (fs/dcache.c:1168:2) super_cache_scan (fs/super.c:221:10) do_shrink_slab (mm/shrinker.c:435:9) shrink_slab (mm/shrinker.c:626:10) shrink_node (mm/vmscan.c:5951:2) shrink_zones (mm/vmscan.c:6195:3) do_try_to_free_pages (mm/vmscan.c:6257:3) do_swap_page (mm/memory.c:4136:11) handle_pte_fault (mm/memory.c:5562:10) handle_mm_fault (mm/memory.c:5870:9) do_user_addr_fault (arch/x86/mm/fault.c:1338:10) handle_page_fault (arch/x86/mm/fault.c:1481:3) exc_page_fault (arch/x86/mm/fault.c:1539:2) asm_exc_page_fault+0x22/0x27 Fix this deadlock by allocating ff->release_args and grabbing the reference on the inode when preparing the file for release even if the server does not implement open. The inode reference will be dropped when the last reference on the fuse file is dropped (see fuse_file_put() -> fuse_release_end()). Fixes: e26ee4efbc79 ("fuse: allocate ff->release_args only if release is needed") Cc: stable(a)vger.kernel.org Signed-off-by: Joanne Koong <joannelkoong(a)gmail.com> Reported-by: Omar Sandoval <osandov(a)fb.com> Signed-off-by: Miklos Szeredi <mszeredi(a)redhat.com> diff --git a/fs/fuse/file.c b/fs/fuse/file.c index f1ef77a0be05..4d96e684d736 100644 --- a/fs/fuse/file.c +++ b/fs/fuse/file.c @@ -110,7 +110,9 @@ static void fuse_file_put(struct fuse_file *ff, bool sync) fuse_file_io_release(ff, ra->inode); if (!args) { - /* Do nothing when server does not implement 'open' */ + /* Do nothing when server does not implement 'opendir' */ + } else if (args->opcode == FUSE_RELEASE && ff->fm->fc->no_open) { + fuse_release_end(ff->fm, args, 0); } else if (sync) { fuse_simple_request(ff->fm, args); fuse_release_end(ff->fm, args, 0); @@ -131,8 +133,17 @@ struct fuse_file *fuse_file_open(struct fuse_mount *fm, u64 nodeid, struct fuse_file *ff; int opcode = isdir ? FUSE_OPENDIR : FUSE_OPEN; bool open = isdir ? !fc->no_opendir : !fc->no_open; + bool release = !isdir || open; - ff = fuse_file_alloc(fm, open); + /* + * ff->args->release_args still needs to be allocated (so we can hold an + * inode reference while there are pending inflight file operations when + * ->release() is called, see fuse_prepare_release()) even if + * fc->no_open is set else it becomes possible for reclaim to deadlock + * if while servicing the readahead request the server triggers reclaim + * and reclaim evicts the inode of the file being read ahead. + */ + ff = fuse_file_alloc(fm, release); if (!ff) return ERR_PTR(-ENOMEM); @@ -152,13 +163,14 @@ struct fuse_file *fuse_file_open(struct fuse_mount *fm, u64 nodeid, fuse_file_free(ff); return ERR_PTR(err); } else { - /* No release needed */ - kfree(ff->args); - ff->args = NULL; - if (isdir) + if (isdir) { + /* No release needed */ + kfree(ff->args); + ff->args = NULL; fc->no_opendir = 1; - else + } else { fc->no_open = 1; + } } }

1 day, 3 hours

1
0
0 0

FAILED: patch "[PATCH] iommu/mediatek: fix use-after-free on probe deferral" failed to apply to 6.1-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 6.1-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. To reproduce the conflict and resubmit, you may use the following commands: git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-6.1.y git checkout FETCH_HEAD git cherry-pick -x de83d4617f9fe059623e97acf7e1e10d209625b5 # <resolve conflicts, build, test, etc.> git commit -s git send-email --to '<stable(a)vger.kernel.org>' --in-reply-to '2025122941-reluctant-exhale-a49f@gregkh' --subject-prefix 'PATCH 6.1.y' HEAD^.. Possible dependencies: thanks, greg k-h ------------------ original commit in Linus's tree ------------------ From de83d4617f9fe059623e97acf7e1e10d209625b5 Mon Sep 17 00:00:00 2001 From: Johan Hovold <johan(a)kernel.org> Date: Mon, 20 Oct 2025 06:53:10 +0200 Subject: [PATCH] iommu/mediatek: fix use-after-free on probe deferral The driver is dropping the references taken to the larb devices during probe after successful lookup as well as on errors. This can potentially lead to a use-after-free in case a larb device has not yet been bound to its driver so that the iommu driver probe defers. Fix this by keeping the references as expected while the iommu driver is bound. Fixes: 26593928564c ("iommu/mediatek: Add error path for loop of mm_dts_parse") Cc: stable(a)vger.kernel.org Cc: Yong Wu <yong.wu(a)mediatek.com> Acked-by: Robin Murphy <robin.murphy(a)arm.com> Signed-off-by: Johan Hovold <johan(a)kernel.org> Reviewed-by: Yong Wu <yong.wu(a)mediatek.com> Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno(a)collabora.com> Signed-off-by: Joerg Roedel <joerg.roedel(a)amd.com> diff --git a/drivers/iommu/mtk_iommu.c b/drivers/iommu/mtk_iommu.c index 82a55fe19a62..54d8936d9d11 100644 --- a/drivers/iommu/mtk_iommu.c +++ b/drivers/iommu/mtk_iommu.c @@ -1213,16 +1213,19 @@ static int mtk_iommu_mm_dts_parse(struct device *dev, struct component_match **m } component_match_add(dev, match, component_compare_dev, &plarbdev->dev); - platform_device_put(plarbdev); } - if (!frst_avail_smicomm_node) - return -EINVAL; + if (!frst_avail_smicomm_node) { + ret = -EINVAL; + goto err_larbdev_put; + } pcommdev = of_find_device_by_node(frst_avail_smicomm_node); of_node_put(frst_avail_smicomm_node); - if (!pcommdev) - return -ENODEV; + if (!pcommdev) { + ret = -ENODEV; + goto err_larbdev_put; + } data->smicomm_dev = &pcommdev->dev; link = device_link_add(data->smicomm_dev, dev, @@ -1230,7 +1233,8 @@ static int mtk_iommu_mm_dts_parse(struct device *dev, struct component_match **m platform_device_put(pcommdev); if (!link) { dev_err(dev, "Unable to link %s.\n", dev_name(data->smicomm_dev)); - return -EINVAL; + ret = -EINVAL; + goto err_larbdev_put; } return 0; @@ -1402,8 +1406,12 @@ static int mtk_iommu_probe(struct platform_device *pdev) iommu_device_sysfs_remove(&data->iommu); out_list_del: list_del(&data->list); - if (MTK_IOMMU_IS_TYPE(data->plat_data, MTK_IOMMU_TYPE_MM)) + if (MTK_IOMMU_IS_TYPE(data->plat_data, MTK_IOMMU_TYPE_MM)) { device_link_remove(data->smicomm_dev, dev); + + for (i = 0; i < MTK_LARB_NR_MAX; i++) + put_device(data->larb_imu[i].dev); + } out_runtime_disable: pm_runtime_disable(dev); return ret; @@ -1423,6 +1431,9 @@ static void mtk_iommu_remove(struct platform_device *pdev) if (MTK_IOMMU_IS_TYPE(data->plat_data, MTK_IOMMU_TYPE_MM)) { device_link_remove(data->smicomm_dev, &pdev->dev); component_master_del(&pdev->dev, &mtk_iommu_com_ops); + + for (i = 0; i < MTK_LARB_NR_MAX; i++) + put_device(data->larb_imu[i].dev); } pm_runtime_disable(&pdev->dev); for (i = 0; i < data->plat_data->banks_num; i++) {

1 day, 3 hours

1
0
0 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

Linux-stable-mirror