kernel/drivers/mtd, branch linux-6.1.y

mtd: rawnand: pl353: Fix software ECC support

2026-03-04T12:20:52Z

[ Upstream commit 89b831ebdaca0df4ca3b226f7e7a1d1db1629060 ] We need to set also write_page_raw in ecc structure to allow choosing SW ECC instead of HW one, otherwise write operation fail. Fixes: 08d8c62164a322 ("mtd: rawnand: pl353: Add support for the ARM PL353 SMC NAND controller") Signed-off-by: Andrea Scian Cc: stable@kernel.org Signed-off-by: Miquel Raynal Signed-off-by: Sasha Levin

mtd: parsers: ofpart: fix OF node refcount leak in parse_fixed_partitions()

2026-03-04T12:19:52Z

[ Upstream commit 7cce81df7d26d44123bd7620715c8349d96793d7 ] of_get_child_by_name() returns a node pointer with refcount incremented, which must be released with of_node_put() when done. However, in parse_fixed_partitions(), when dedicated is true (i.e., a "partitions" subnode was found), the ofpart_node obtained from of_get_child_by_name() is never released on any code path. Add of_node_put(ofpart_node) calls on all exit paths when dedicated is true to fix the reference count leak. This bug was detected by our static analysis tool. Fixes: 562b4e91d3b2 ("mtd: parsers: ofpart: fix parsing subpartitions") Signed-off-by: Weigang He Signed-off-by: Miquel Raynal Signed-off-by: Sasha Levin

mtd: rawnand: cadence: Fix return type of CDMA send-and-wait helper

2026-03-04T12:19:50Z

[ Upstream commit 6d8226cbbf124bb5613b532216b74c886a4361b7 ] cadence_nand_cdma_send_and_wait() propagates negative errno values from cadence_nand_cdma_send(), returns -ETIMEDOUT on failure and -EIO when the CDMA engine reports a command failure. However, it is declared as u32, causing error codes to wrap. Change the return type to int to correctly propagate errors. Fixes: ec4ba01e894d ("mtd: rawnand: Add new Cadence NAND driver to MTD subsystem") Signed-off-by: Alok Tiwari Signed-off-by: Miquel Raynal Signed-off-by: Sasha Levin

mtd: lpddr_cmds: fix signed shifts in lpddr_cmds

2026-01-11T14:18:33Z

[ Upstream commit c909fec69f84b39e63876c69b9df2c178c6b76ba ] There are several places where a value of type 'int' is shifted by lpddr->chipshift. lpddr->chipshift is derived from QINFO geometry and might reach 31 when QINFO reports a 2 GiB size - the maximum supported by LPDDR(1) compliant chips. This may cause unexpected sign-extensions when casting the integer value to the type of 'unsigned long'. Use '1UL << lpddr->chipshift' and cast 'j' to unsigned long before shifting so the computation is performed at the destination width. Found by Linux Verification Center (linuxtesting.org) with SVACE. Fixes: c68264711ca6 ("[MTD] LPDDR Command set driver") Signed-off-by: Ivan Stepchenko Signed-off-by: Miquel Raynal Signed-off-by: Sasha Levin

mtd: rawnand: renesas: Handle devm_pm_runtime_enable() errors

2026-01-11T14:18:33Z

[ Upstream commit a3623e1ae1ed6be4d49b2ccb9996a9d2b65c1828 ] devm_pm_runtime_enable() can fail due to memory allocation failures. The current code ignores its return value and proceeds with pm_runtime_resume_and_get(), which may operate on incorrectly initialized runtime PM state. Check the return value of devm_pm_runtime_enable() and return the error code if it fails. Fixes: 6a2277a0ebe7 ("mtd: rawnand: renesas: Use runtime PM instead of the raw clock API") Signed-off-by: Haotian Zhang Reviewed-by: Geert Uytterhoeven Signed-off-by: Miquel Raynal Signed-off-by: Sasha Levin

mtdchar: fix integer overflow in read/write ioctls

2025-12-06T21:12:37Z

commit e4185bed738da755b191aa3f2e16e8b48450e1b8 upstream. The "req.start" and "req.len" variables are u64 values that come from the user at the start of the function. We mask away the high 32 bits of "req.len" so that's capped at U32_MAX but the "req.start" variable can go up to U64_MAX which means that the addition can still integer overflow. Use check_add_overflow() to fix this bug. Fixes: 095bb6e44eb1 ("mtdchar: add MEMREAD ioctl") Fixes: 6420ac0af95d ("mtdchar: prevent unbounded allocation in MEMWRITE ioctl") Cc: stable@vger.kernel.org Signed-off-by: Dan Carpenter Signed-off-by: Miquel Raynal Signed-off-by: Greg Kroah-Hartman

mtd: rawnand: cadence: fix DMA device NULL pointer dereference

2025-12-06T21:12:37Z

commit 5c56bf214af85ca042bf97f8584aab2151035840 upstream. The DMA device pointer `dma_dev` was being dereferenced before ensuring that `cdns_ctrl->dmac` is properly initialized. Move the assignment of `dma_dev` after successfully acquiring the DMA channel to ensure the pointer is valid before use. Fixes: d76d22b5096c ("mtd: rawnand: cadence: use dma_map_resource for sdma address") Cc: stable@vger.kernel.org Signed-off-by: Niravkumar L Rabara Signed-off-by: Miquel Raynal Signed-off-by: Greg Kroah-Hartman

mtd: onenand: Pass correct pointer to IRQ handler

2025-12-06T21:12:34Z

[ Upstream commit 97315e7c901a1de60e8ca9b11e0e96d0f9253e18 ] This was supposed to pass "onenand" instead of "&onenand" with the ampersand. Passing a random stack address which will be gone when the function ends makes no sense. However the good thing is that the pointer is never used, so this doesn't cause a problem at run time. Fixes: e23abf4b7743 ("mtd: OneNAND: S5PC110: Implement DMA interrupt method") Signed-off-by: Dan Carpenter Signed-off-by: Miquel Raynal Signed-off-by: Sasha Levin

mtd: rawnand: fsmc: Default to autodetect buswidth

2025-10-19T14:23:16Z

commit b8df622cf7f6808c85764e681847150ed6d85f3d upstream. If you don't specify buswidth 2 (16 bits) in the device tree, FSMC doesn't even probe anymore: fsmc-nand 10100000.flash: FSMC device partno 090, manufacturer 80, revision 00, config 00 nand: device found, Manufacturer ID: 0x20, Chip ID: 0xb1 nand: ST Micro 10100000.flash nand: bus width 8 instead of 16 bits nand: No NAND device found fsmc-nand 10100000.flash: probe with driver fsmc-nand failed with error -22 With this patch to use autodetection unless buswidth is specified, the device is properly detected again: fsmc-nand 10100000.flash: FSMC device partno 090, manufacturer 80, revision 00, config 00 nand: device found, Manufacturer ID: 0x20, Chip ID: 0xb1 nand: ST Micro NAND 128MiB 1,8V 16-bit nand: 128 MiB, SLC, erase size: 128 KiB, page size: 2048, OOB size: 64 fsmc-nand 10100000.flash: Using 1-bit HW ECC scheme Scanning device for bad blocks I don't know where or how this happened, I think some change in the nand core. Cc: stable@vger.kernel.org Signed-off-by: Linus Walleij Signed-off-by: Miquel Raynal Signed-off-by: Greg Kroah-Hartman

mtd: rawnand: atmel: Fix error handling path in atmel_nand_controller_add_nands

2025-10-15T09:56:38Z

[ Upstream commit 8ed4728eb9f10b57c3eb02e0f6933a89ffcb8a91 ] In case of a jump to the err label due to atmel_nand_create() or atmel_nand_controller_add_nand() failure, the reference to nand_np need to be released Use for_each_child_of_node_scoped() to fix the issue. Fixes: f88fc122cc34 ("mtd: nand: Cleanup/rework the atmel_nand driver") Signed-off-by: Erick Karanja Signed-off-by: Miquel Raynal Signed-off-by: Sasha Levin