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

[PATCH] drm/xe/bo: Don't include the CCS metadata in the dma-buf sg-table

by Thomas Hellström

Some Xe bos are allocated with extra backing-store for the CCS metadata. It's never been the intention to share the CCS metadata when exporting such bos as dma-buf. Don't include it in the dma-buf sg-table. Fixes: dd08ebf6c352 ("drm/xe: Introduce a new DRM driver for Intel GPUs") Cc: Rodrigo Vivi <rodrigo.vivi(a)intel.com> Cc: Matthew Brost <matthew.brost(a)intel.com> Cc: Maarten Lankhorst <maarten.lankhorst(a)linux.intel.com> Cc: <stable(a)vger.kernel.org> # v6.8+ Signed-off-by: Thomas Hellström <thomas.hellstrom(a)linux.intel.com> --- drivers/gpu/drm/xe/xe_dma_buf.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/xe/xe_dma_buf.c b/drivers/gpu/drm/xe/xe_dma_buf.c index 54e42960daad..7c74a31d4486 100644 --- a/drivers/gpu/drm/xe/xe_dma_buf.c +++ b/drivers/gpu/drm/xe/xe_dma_buf.c @@ -124,7 +124,7 @@ static struct sg_table *xe_dma_buf_map(struct dma_buf_attachment *attach, case XE_PL_TT: sgt = drm_prime_pages_to_sg(obj->dev, bo->ttm.ttm->pages, - bo->ttm.ttm->num_pages); + obj->size >> PAGE_SHIFT); if (IS_ERR(sgt)) return sgt; -- 2.51.1

2 weeks, 2 days

3
2
0 0

[PATCH v2] media: cx25821: Add missing unmap in snd_cx25821_hw_params()

by Haoxiang Li

In error path, add cx25821_alsa_dma_unmap() to release the resource acquired by cx25821_alsa_dma_map() Fixes: 8d8e6d6005de ("[media] cx28521: drop videobuf abuse in cx25821-alsa") Cc: stable(a)vger.kernel.org Signed-off-by: Haoxiang Li <lihaoxiang(a)isrc.iscas.ac.cn> --- Changes in v2: - Fix the "Fixes" tag error. --- drivers/media/pci/cx25821/cx25821-alsa.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/media/pci/cx25821/cx25821-alsa.c b/drivers/media/pci/cx25821/cx25821-alsa.c index a42f0c03a7ca..f463365163b7 100644 --- a/drivers/media/pci/cx25821/cx25821-alsa.c +++ b/drivers/media/pci/cx25821/cx25821-alsa.c @@ -535,6 +535,7 @@ static int snd_cx25821_hw_params(struct snd_pcm_substream *substream, chip->period_size, chip->num_periods, 1); if (ret < 0) { pr_info("DEBUG: ERROR after cx25821_risc_databuffer_audio()\n"); + cx25821_alsa_dma_unmap(chip); goto error; } -- 2.25.1

2 weeks, 2 days

1
0
0 0

[PATCH net 2/2] can: gs_usb: gs_can_open(): fix error handling

by Marc Kleine-Budde

Commit 2603be9e8167 ("can: gs_usb: gs_can_open(): improve error handling") added missing error handling to the gs_can_open() function. The driver uses 2 USB anchors to track the allocated URBs: the TX URBs in struct gs_can::tx_submitted for each netdev and the RX URBs in struct gs_usb::rx_submitted for the USB device. gs_can_open() allocates the RX URBs, while TX URBs are allocated during gs_can_start_xmit(). The cleanup in gs_can_open() kills all anchored dev->tx_submitted URBs (which is not necessary since the netdev is not yet registered), but misses the parent->rx_submitted URBs. Fix the problem by killing the rx_submitted instead of the tx_submitted. Fixes: 2603be9e8167 ("can: gs_usb: gs_can_open(): improve error handling") Cc: stable(a)vger.kernel.org Link: https://patch.msgid.link/20251210-gs_usb-fix-error-handling-v1-1-d6a5a03f10… Signed-off-by: Marc Kleine-Budde <mkl(a)pengutronix.de> --- drivers/net/can/usb/gs_usb.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/can/usb/gs_usb.c b/drivers/net/can/usb/gs_usb.c index e29e85b67fd4..a0233e550a5a 100644 --- a/drivers/net/can/usb/gs_usb.c +++ b/drivers/net/can/usb/gs_usb.c @@ -1074,7 +1074,7 @@ static int gs_can_open(struct net_device *netdev) usb_free_urb(urb); out_usb_kill_anchored_urbs: if (!parent->active_channels) { - usb_kill_anchored_urbs(&dev->tx_submitted); + usb_kill_anchored_urbs(&parent->rx_submitted); if (dev->feature & GS_CAN_FEATURE_HW_TIMESTAMP) gs_usb_timestamp_stop(parent); -- 2.51.0

2 weeks, 2 days

1
0
0 0

[PATCH can] can: gs_usb: gs_can_open(): fix error handling

by Marc Kleine-Budde

Commit 2603be9e8167 ("can: gs_usb: gs_can_open(): improve error handling") added missing error handling to the gs_can_open() function. The driver uses 2 USB anchors to track the allocated URBs: the TX URBs in struct gs_can::tx_submitted for each netdev and the RX URBs in struct gs_usb::rx_submitted for the USB device. gs_can_open() allocates the RX URBs, while TX URBs are allocated during gs_can_start_xmit(). The cleanup in gs_can_open() kills all anchored dev->tx_submitted URBs (which is not necessary since the netdev is not yet registered), but misses the parent->rx_submitted URBs. Fix the problem by killing the rx_submitted instead of the tx_submitted. Fixes: 2603be9e8167 ("can: gs_usb: gs_can_open(): improve error handling") Cc: stable(a)vger.kernel.org Signed-off-by: Marc Kleine-Budde <mkl(a)pengutronix.de> --- drivers/net/can/usb/gs_usb.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/can/usb/gs_usb.c b/drivers/net/can/usb/gs_usb.c index e29e85b67fd4..a0233e550a5a 100644 --- a/drivers/net/can/usb/gs_usb.c +++ b/drivers/net/can/usb/gs_usb.c @@ -1074,7 +1074,7 @@ static int gs_can_open(struct net_device *netdev) usb_free_urb(urb); out_usb_kill_anchored_urbs: if (!parent->active_channels) { - usb_kill_anchored_urbs(&dev->tx_submitted); + usb_kill_anchored_urbs(&parent->rx_submitted); if (dev->feature & GS_CAN_FEATURE_HW_TIMESTAMP) gs_usb_timestamp_stop(parent); --- base-commit: 186468c67fc687650b7fb713d8c627d5c8566886 change-id: 20251210-gs_usb-fix-error-handling-4f980294424c Best regards, -- Marc Kleine-Budde <mkl(a)pengutronix.de>

2 weeks, 2 days

1
0
0 0

[PATCH] net: nfc: nci: Fix parameter validation for packet data

by Michael Thalmeier

Since commit 8fcc7315a10a ("net: nfc: nci: Add parameter validation for packet data") communication with nci nfc chips is not working any more. The mentioned commit tries to fix access of uninitialized data, but failed to understand that in some cases the data packet is of variable length and can therefore not be compared to the maximum packet length given by the sizeof(struct). For these cases it is only possible to check for minimum packet length. Fixes: 8fcc7315a10a ("net: nfc: nci: Add parameter validation for packet data") Cc: stable(a)vger.kernel.org Signed-off-by: Michael Thalmeier <michael.thalmeier(a)hale.at> --- net/nfc/nci/ntf.c | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-) diff --git a/net/nfc/nci/ntf.c b/net/nfc/nci/ntf.c index 418b84e2b260..5161e94f067f 100644 --- a/net/nfc/nci/ntf.c +++ b/net/nfc/nci/ntf.c @@ -58,7 +58,8 @@ static int nci_core_conn_credits_ntf_packet(struct nci_dev *ndev, struct nci_conn_info *conn_info; int i; - if (skb->len < sizeof(struct nci_core_conn_credit_ntf)) + /* Minimal packet size for num_entries=1 is 1 x __u8 + 1 x conn_credit_entry */ + if (skb->len < (sizeof(__u8) + sizeof(struct conn_credit_entry))) return -EINVAL; ntf = (struct nci_core_conn_credit_ntf *)skb->data; @@ -364,7 +365,8 @@ static int nci_rf_discover_ntf_packet(struct nci_dev *ndev, const __u8 *data; bool add_target = true; - if (skb->len < sizeof(struct nci_rf_discover_ntf)) + /* Minimal packet size is 5 if rf_tech_specific_params_len=0 */ + if (skb->len < (5 * sizeof(__u8))) return -EINVAL; data = skb->data; @@ -596,7 +598,10 @@ static int nci_rf_intf_activated_ntf_packet(struct nci_dev *ndev, const __u8 *data; int err = NCI_STATUS_OK; - if (skb->len < sizeof(struct nci_rf_intf_activated_ntf)) + /* Minimal packet size is 11 if + * f_tech_specific_params_len=0 and activation_params_len=0 + */ + if (skb->len < (11 * sizeof(__u8))) return -EINVAL; data = skb->data; -- 2.52.0

2 weeks, 2 days

2
2
0 0

Re: Patch "dma-mapping: Allow use of DMA_BIT_MASK(64) in global scope" has been added to the 6.17-stable tree

by Nathan Chancellor

On Sun, Dec 07, 2025 at 10:07:49PM -0500, Sasha Levin wrote: > This is a note to let you know that I've just added the patch titled > > dma-mapping: Allow use of DMA_BIT_MASK(64) in global scope > > to the 6.17-stable tree which can be found at: > http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=sum… > > The filename of the patch is: > dma-mapping-allow-use-of-dma_bit_mask-64-in-global-s.patch > and it can be found in the queue-6.17 subdirectory. > > If you, or anyone else, feels it should not be added to the stable tree, > please let <stable(a)vger.kernel.org> know about it. > > > > commit 51a1912d46ceb70602de50c167e440966b9836f1 > Author: James Clark <james.clark(a)linaro.org> > Date: Thu Oct 30 14:05:27 2025 +0000 > > dma-mapping: Allow use of DMA_BIT_MASK(64) in global scope > > [ Upstream commit a50f7456f853ec3a6f07cbe1d16ad8a8b2501320 ] This change has a pending bug fix, consider waiting to apply this to the stable trees. https://lore.kernel.org/20251207184756.97904-1-johannes.goede@oss.qualcomm.… Cheers, Nathan

2 weeks, 2 days

2
1
0 0

[PATCH AUTOSEL 6.18-5.10] serial: sprd: Return -EPROBE_DEFER when uart clock is not ready

by Sasha Levin

From: Wenhua Lin <Wenhua.Lin(a)unisoc.com> [ Upstream commit 29e8a0c587e328ed458380a45d6028adf64d7487 ] In sprd_clk_init(), when devm_clk_get() returns -EPROBE_DEFER for either uart or source clock, we should propagate the error instead of just warning and continuing with NULL clocks. Currently the driver only emits a warning when clock acquisition fails and proceeds with NULL clock pointers. This can lead to issues later when the clocks are actually needed. More importantly, when the clock provider is not ready yet and returns -EPROBE_DEFER, we should return this error to allow deferred probing. This change adds explicit checks for -EPROBE_DEFER after both: 1. devm_clk_get(uport->dev, uart) 2. devm_clk_get(uport->dev, source) When -EPROBE_DEFER is encountered, the function now returns -EPROBE_DEFER to let the driver framework retry probing later when the clock dependencies are resolved. Signed-off-by: Wenhua Lin <Wenhua.Lin(a)unisoc.com> Link: https://patch.msgid.link/20251022030840.956589-1-Wenhua.Lin@unisoc.com Reviewed-by: Cixi Geng <cixi.geng(a)linux.dev> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- LLM Generated explanations, may be completely bogus: ## Commit Analysis: serial: sprd: Return -EPROBE_DEFER when uart clock is not ready ### 1. COMMIT MESSAGE ANALYSIS **Subject**: Fixes incorrect handling of `-EPROBE_DEFER` error from `devm_clk_get()` **Key observations**: - No `Cc: stable(a)vger.kernel.org` tag present - No `Fixes:` tag pointing to when the bug was introduced - Has `Reviewed-by:` from Cixi Geng - Accepted by Greg Kroah-Hartman (serial subsystem maintainer) ### 2. CODE CHANGE ANALYSIS The change adds two checks in `sprd_clk_init()`: ```c clk_uart = devm_clk_get(uport->dev, "uart"); if (IS_ERR(clk_uart)) { + if (PTR_ERR(clk_uart) == -EPROBE_DEFER) + return -EPROBE_DEFER; dev_warn(...); clk_uart = NULL; } clk_parent = devm_clk_get(uport->dev, "source"); if (IS_ERR(clk_parent)) { + if (PTR_ERR(clk_parent) == -EPROBE_DEFER) + return -EPROBE_DEFER; dev_warn(...); clk_parent = NULL; } ``` **Technical bug mechanism**: When clock providers aren't ready yet, `devm_clk_get()` returns `-EPROBE_DEFER`. The existing code ignores this error, sets the clock pointer to NULL, and continues. This bypasses the kernel's deferred probing mechanism which exists precisely to handle this dependency ordering scenario. **Existing pattern**: The function already has identical handling for the "enable" clock (visible in the context lines). This fix makes the handling consistent for all three clocks. ### 3. CLASSIFICATION **Type**: Bug fix (not a feature addition) This fixes incorrect error handling. The `-EPROBE_DEFER` mechanism is a fundamental kernel feature for handling driver load order dependencies. Not propagating this error is a bug. ### 4. SCOPE AND RISK ASSESSMENT - **Lines changed**: 6 lines added (two 3-line checks) - **Files touched**: 1 file (`drivers/tty/serial/sprd_serial.c`) - **Complexity**: Very low - follows identical pattern already in the same function - **Subsystem**: Hardware-specific serial driver for Spreadtrum/Unisoc UARTs **Risk**: Very low - Pattern is already established and proven in the same function - Only affects error handling during probe - No changes to normal operation when clocks are available - Worst case: probe failure happens earlier/more explicitly ### 5. USER IMPACT **Affected users**: Users of Spreadtrum/Unisoc UART hardware (embedded devices, some Android phones) **Severity**: Medium-High for affected users - without this fix, the serial port may not work correctly on systems where clock providers load after the serial driver. This is common in embedded systems with device tree-based configurations. **Real bug**: This is a practical issue in probe ordering scenarios. The driver would proceed with NULL clocks instead of waiting for dependencies to be ready. ### 6. STABILITY INDICATORS - Reviewed-by tag indicates code review - Maintainer accepted the change - Simple, straightforward change following existing code patterns - No complex logic introduced ### 7. DEPENDENCY CHECK - **Dependencies**: None - self-contained fix - **Code existence in stable**: The sprd_serial driver has existed since ~2015 (commit 3e1f2029a4b40), so it's present in all active stable trees ### SUMMARY **What it fixes**: Incorrect handling of `-EPROBE_DEFER` from `devm_clk_get()` for two clocks ("uart" and "source"), causing the driver to proceed with NULL clocks instead of deferring probe when clock providers aren't ready. **Stable kernel criteria**: - ✅ Obviously correct (follows identical pattern already in function) - ✅ Fixes a real bug (broken deferred probing) - ✅ Small and contained (6 lines, 1 file) - ✅ No new features - ✅ Low risk of regression **Concerns**: - No explicit `Cc: stable` tag (author/maintainer didn't flag for stable) - Relatively niche driver (Spreadtrum hardware) **Risk vs Benefit**: The fix is minimal risk (identical pattern to existing code) and addresses a real bug that could leave serial hardware non-functional on embedded systems. The benefit outweighs the minimal risk. The lack of `Cc: stable` tag is notable but not determinative - this is a straightforward bug fix that meets all stable criteria. The fix is small, obviously correct, and addresses a real probe ordering issue that embedded system users could encounter. **YES** drivers/tty/serial/sprd_serial.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/drivers/tty/serial/sprd_serial.c b/drivers/tty/serial/sprd_serial.c index 8c9366321f8e7..092755f356836 100644 --- a/drivers/tty/serial/sprd_serial.c +++ b/drivers/tty/serial/sprd_serial.c @@ -1133,6 +1133,9 @@ static int sprd_clk_init(struct uart_port *uport) clk_uart = devm_clk_get(uport->dev, "uart"); if (IS_ERR(clk_uart)) { + if (PTR_ERR(clk_uart) == -EPROBE_DEFER) + return -EPROBE_DEFER; + dev_warn(uport->dev, "uart%d can't get uart clock\n", uport->line); clk_uart = NULL; @@ -1140,6 +1143,9 @@ static int sprd_clk_init(struct uart_port *uport) clk_parent = devm_clk_get(uport->dev, "source"); if (IS_ERR(clk_parent)) { + if (PTR_ERR(clk_parent) == -EPROBE_DEFER) + return -EPROBE_DEFER; + dev_warn(uport->dev, "uart%d can't get source clock\n", uport->line); clk_parent = NULL; -- 2.51.0

2 weeks, 2 days

1
7
0 0

[PATCH v2] media: dw2102: validate I2C messages in su3000_i2c_transfer()

by Dharanitharan R

syzbot reports a general protection fault caused by su3000_i2c_transfer() dereferencing msg->buf without validating the message length or buffer pointer. Although i2c-dev blocks zero-length messages, malformed I²C messages can still reach the driver through the DVB USB subsystem. Add strict validation of each message to prevent NULL-pointer dereferences. Reported-by: syzbot+d99f3a288cc7d8ef60fb(a)syzkaller.appspotmail.com Fixes: 0e148a522b84 ("media: dw2102: Don't translate i2c read into write") Closes: https://syzkaller.appspot.com/bug?extid=d99f3a288cc7d8ef60fb Cc: stable(a)vger.kernel.org Signed-off-by: Dharanitharan R <dharanitharan725(a)gmail.com> --- drivers/media/usb/dvb-usb/dw2102.c | 30 ++++++++++++++++++++++++++++++ 1 file changed, 30 insertions(+) diff --git a/drivers/media/usb/dvb-usb/dw2102.c b/drivers/media/usb/dvb-usb/dw2102.c index 4fecf2f965e9..0dd210ea16f3 100644 --- a/drivers/media/usb/dvb-usb/dw2102.c +++ b/drivers/media/usb/dvb-usb/dw2102.c @@ -733,6 +733,36 @@ static int su3000_i2c_transfer(struct i2c_adapter *adap, struct i2c_msg msg[], return -EAGAIN; } + /* Validate incoming I²C messages */ + if (!msg || num <= 0) { + mutex_unlock(&d->data_mutex); + mutex_unlock(&d->i2c_mutex); + return -EINVAL; + } + + for (j = 0; j < num; j++) { + /* msg buffer must exist */ + if (!msg[j].buf) { + mutex_unlock(&d->data_mutex); + mutex_unlock(&d->i2c_mutex); + return -EINVAL; + } + + /* zero or negative length is invalid */ + if (msg[j].len <= 0) { + mutex_unlock(&d->data_mutex); + mutex_unlock(&d->i2c_mutex); + return -EINVAL; + } + + /* protect against unreasonable sizes */ + if (msg[j].len > 256) { + mutex_unlock(&d->data_mutex); + mutex_unlock(&d->i2c_mutex); + return -EOPNOTSUPP; + } + } + j = 0; while (j < num) { switch (msg[j].addr) { -- 2.43.0

2 weeks, 2 days

1
0
0 0

[PATCH v2] powerpc: Add reloc_offset() to font bitmap pointer used for bootx_printf()

by Finn Thain

Since Linux v6.7, booting using BootX on an Old World PowerMac produces an early crash. Stan Johnson writes, "the symptoms are that the screen goes blank and the backlight stays on, and the system freezes (Linux doesn't boot)." Further testing revealed that the failure can be avoided by disabling CONFIG_BOOTX_TEXT. Bisection revealed that the regression was caused by a change to the font bitmap pointer that's used when btext_init() begins painting characters on the display, early in the boot process. Christophe Leroy explains, "before kernel text is relocated to its final location ... data is addressed with an offset which is added to the Global Offset Table (GOT) entries at the start of bootx_init() by function reloc_got2(). But the pointers that are located inside a structure are not referenced in the GOT and are therefore not updated by reloc_got2(). It is therefore needed to apply the offset manually by using PTRRELOC() macro." Cc: Cedar Maxwell <cedarmaxwell(a)mac.com> Cc: Stan Johnson <userm57(a)yahoo.com> Cc: "Dr. David Alan Gilbert" <linux(a)treblig.org> Cc: Benjamin Herrenschmidt <benh(a)kernel.crashing.org> Cc: stable(a)vger.kernel.org Link: https://lists.debian.org/debian-powerpc/2025/10/msg00111.html Link: https://lore.kernel.org/linuxppc-dev/d81ddca8-c5ee-d583-d579-02b19ed95301@y… Reported-by: Cedar Maxwell <cedarmaxwell(a)mac.com> Closes: https://lists.debian.org/debian-powerpc/2025/09/msg00031.html Bisected-by: Stan Johnson <userm57(a)yahoo.com> Tested-by: Stan Johnson <userm57(a)yahoo.com> Fixes: 0ebc7feae79a ("powerpc: Use shared font data") Suggested-by: Christophe Leroy <christophe.leroy(a)csgroup.eu> Signed-off-by: Finn Thain <fthain(a)linux-m68k.org> --- Changed since v1: - Improved commit log entry to better explain the need for PTRRELOC(). --- arch/powerpc/kernel/btext.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/arch/powerpc/kernel/btext.c b/arch/powerpc/kernel/btext.c index 7f63f1cdc6c3..ca00c4824e31 100644 --- a/arch/powerpc/kernel/btext.c +++ b/arch/powerpc/kernel/btext.c @@ -20,6 +20,7 @@ #include <asm/io.h> #include <asm/processor.h> #include <asm/udbg.h> +#include <asm/setup.h> #define NO_SCROLL @@ -463,7 +464,7 @@ static noinline void draw_byte(unsigned char c, long locX, long locY) { unsigned char *base = calc_base(locX << 3, locY << 4); unsigned int font_index = c * 16; - const unsigned char *font = font_sun_8x16.data + font_index; + const unsigned char *font = PTRRELOC(font_sun_8x16.data) + font_index; int rb = dispDeviceRowBytes; rmci_maybe_on(); -- 2.49.1

2 weeks, 2 days

3
3
0 0

[PATCH AUTOSEL 6.18-6.17] functionfs: fix the open/removal races

by Sasha Levin

From: Al Viro <viro(a)zeniv.linux.org.uk> [ Upstream commit e5bf5ee266633cb18fff6f98f0b7d59a62819eee ] ffs_epfile_open() can race with removal, ending up with file->private_data pointing to freed object. There is a total count of opened files on functionfs (both ep0 and dynamic ones) and when it hits zero, dynamic files get removed. Unfortunately, that removal can happen while another thread is in ffs_epfile_open(), but has not incremented the count yet. In that case open will succeed, leaving us with UAF on any subsequent read() or write(). The root cause is that ffs->opened is misused; atomic_dec_and_test() vs. atomic_add_return() is not a good idea, when object remains visible all along. To untangle that * serialize openers on ffs->mutex (both for ep0 and for dynamic files) * have dynamic ones use atomic_inc_not_zero() and fail if we had zero ->opened; in that case the file we are opening is doomed. * have the inodes of dynamic files marked on removal (from the callback of simple_recursive_removal()) - clear ->i_private there. * have open of dynamic ones verify they hadn't been already removed, along with checking that state is FFS_ACTIVE. Reviewed-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> Signed-off-by: Al Viro <viro(a)zeniv.linux.org.uk> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- LLM Generated explanations, may be completely bogus: ## Summary ### What the Commit Fixes This commit fixes a **Use-After-Free (UAF)** vulnerability in USB gadget functionfs. The race condition occurs between `ffs_epfile_open()` and file removal: 1. Thread A is in `ffs_epfile_open()` but hasn't incremented `ffs->opened` yet 2. Thread B closes the last file handle, triggering removal (since `opened` is 0) 3. Thread A's open succeeds with `file->private_data` pointing to freed memory 4. Subsequent `read()`/`write()` operations cause UAF ### Fix Mechanism The fix implements proper synchronization: - Serializes openers using `ffs->mutex` - Uses `atomic_inc_not_zero()` to fail if counter already zero - Uses `smp_load_acquire()`/`smp_store_release()` for memory ordering - Clears `i_private` during removal via `simple_recursive_removal()` callback - Uses `file->private_data` instead of `inode->i_private` in release path ### Stable Kernel Criteria Assessment | Criterion | Assessment | |-----------|------------| | Obviously correct | ✅ Uses standard kernel primitives, proper locking patterns | | Fixes real bug | ✅ UAF vulnerability, security-relevant | | Important issue | ✅ Security bug, potential for crashes/exploitation | | Small and contained | ✅ Single file, +43/-10 lines, localized changes | | No new features | ✅ Pure bug fix, no new APIs | ### Risk vs Benefit **Benefits:** - Fixes serious UAF vulnerability - USB gadget functionfs used in Android, embedded systems - Reviewed by Greg Kroah-Hartman (USB maintainer, stable maintainer) - Written by Al Viro (highly respected kernel developer) **Risks:** - Moderate complexity (changes locking behavior) - Recent commit (November 2025), limited mainline soak time - No explicit `Cc: stable(a)vger.kernel.org` tag ### Dependencies - `ffs_mutex_lock()` - exists in functionfs since early versions - `simple_recursive_removal()` with callback - available since ~5.x kernels - Standard kernel APIs (`atomic_inc_not_zero`, memory barriers) - universally available ### Concerns 1. **No Fixes: tag** - Makes it harder to determine which stable trees need this fix 2. **No Cc: stable tag** - May indicate maintainers wanted soak time, or an oversight given Greg KH reviewed it 3. **Backport effort** - May need adjustment for older stable trees depending on functionfs evolution ### Conclusion This is a legitimate UAF security fix that affects real-world users (Android, embedded USB gadgets). Despite moderate complexity, the fix: - Addresses a serious vulnerability class (UAF) - Uses correct synchronization patterns - Has been reviewed by the appropriate maintainer who also maintains stable trees - Is self-contained with no feature additions The lack of explicit stable tags appears to be an oversight given the security nature of the bug and Greg KH's review. UAF vulnerabilities typically warrant expedited backporting. **YES** drivers/usb/gadget/function/f_fs.c | 53 ++++++++++++++++++++++++------ 1 file changed, 43 insertions(+), 10 deletions(-) diff --git a/drivers/usb/gadget/function/f_fs.c b/drivers/usb/gadget/function/f_fs.c index 47cfbe41fdff8..69f6e3c0f7e00 100644 --- a/drivers/usb/gadget/function/f_fs.c +++ b/drivers/usb/gadget/function/f_fs.c @@ -640,13 +640,22 @@ static ssize_t ffs_ep0_read(struct file *file, char __user *buf, static int ffs_ep0_open(struct inode *inode, struct file *file) { - struct ffs_data *ffs = inode->i_private; + struct ffs_data *ffs = inode->i_sb->s_fs_info; + int ret; - if (ffs->state == FFS_CLOSING) - return -EBUSY; + /* Acquire mutex */ + ret = ffs_mutex_lock(&ffs->mutex, file->f_flags & O_NONBLOCK); + if (ret < 0) + return ret; - file->private_data = ffs; ffs_data_opened(ffs); + if (ffs->state == FFS_CLOSING) { + ffs_data_closed(ffs); + mutex_unlock(&ffs->mutex); + return -EBUSY; + } + mutex_unlock(&ffs->mutex); + file->private_data = ffs; return stream_open(inode, file); } @@ -1193,14 +1202,33 @@ static ssize_t ffs_epfile_io(struct file *file, struct ffs_io_data *io_data) static int ffs_epfile_open(struct inode *inode, struct file *file) { - struct ffs_epfile *epfile = inode->i_private; + struct ffs_data *ffs = inode->i_sb->s_fs_info; + struct ffs_epfile *epfile; + int ret; - if (WARN_ON(epfile->ffs->state != FFS_ACTIVE)) + /* Acquire mutex */ + ret = ffs_mutex_lock(&ffs->mutex, file->f_flags & O_NONBLOCK); + if (ret < 0) + return ret; + + if (!atomic_inc_not_zero(&ffs->opened)) { + mutex_unlock(&ffs->mutex); + return -ENODEV; + } + /* + * we want the state to be FFS_ACTIVE; FFS_ACTIVE alone is + * not enough, though - we might have been through FFS_CLOSING + * and back to FFS_ACTIVE, with our file already removed. + */ + epfile = smp_load_acquire(&inode->i_private); + if (unlikely(ffs->state != FFS_ACTIVE || !epfile)) { + mutex_unlock(&ffs->mutex); + ffs_data_closed(ffs); return -ENODEV; + } + mutex_unlock(&ffs->mutex); file->private_data = epfile; - ffs_data_opened(epfile->ffs); - return stream_open(inode, file); } @@ -1332,7 +1360,7 @@ static void ffs_dmabuf_put(struct dma_buf_attachment *attach) static int ffs_epfile_release(struct inode *inode, struct file *file) { - struct ffs_epfile *epfile = inode->i_private; + struct ffs_epfile *epfile = file->private_data; struct ffs_dmabuf_priv *priv, *tmp; struct ffs_data *ffs = epfile->ffs; @@ -2352,6 +2380,11 @@ static int ffs_epfiles_create(struct ffs_data *ffs) return 0; } +static void clear_one(struct dentry *dentry) +{ + smp_store_release(&dentry->d_inode->i_private, NULL); +} + static void ffs_epfiles_destroy(struct ffs_epfile *epfiles, unsigned count) { struct ffs_epfile *epfile = epfiles; @@ -2359,7 +2392,7 @@ static void ffs_epfiles_destroy(struct ffs_epfile *epfiles, unsigned count) for (; count; --count, ++epfile) { BUG_ON(mutex_is_locked(&epfile->mutex)); if (epfile->dentry) { - simple_recursive_removal(epfile->dentry, NULL); + simple_recursive_removal(epfile->dentry, clear_one); epfile->dentry = NULL; } } -- 2.51.0

2 weeks, 2 days

1
19
0 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

Linux-stable-mirror