user/sven/linux.git/drivers/mtd, branch v5.12.8

mtd: rawnand: gpmi: Fix a double free in gpmi_nand_init

2021-05-14T08:52:26Z

[ Upstream commit 076de75de1e53160e9b099f75872c1f9adf41a0b ] If the callee gpmi_alloc_dma_buffer() failed to alloc memory for this->raw_buffer, gpmi_free_dma_buffer() will be called to free this->auxiliary_virt. But this->auxiliary_virt is still a non-NULL and valid ptr. Then gpmi_alloc_dma_buffer() returns err and gpmi_free_dma_buffer() is called again to free this->auxiliary_virt in err_out. This causes a double free. As gpmi_free_dma_buffer() has already called in gpmi_alloc_dma_buffer's error path, so it should return err directly instead of releasing the dma buffer again. Fixes: 4d02423e9afe6 ("mtd: nand: gpmi: Fix gpmi_nand_init() error path") Signed-off-by: Lv Yunlong Signed-off-by: Miquel Raynal Link: https://lore.kernel.org/linux-mtd/20210403060905.5251-1-lyl2019@mail.ustc.edu.cn Signed-off-by: Sasha Levin

mtd: require write permissions for locking and badblock ioctls

2021-05-14T08:52:23Z

[ Upstream commit 1e97743fd180981bef5f01402342bb54bf1c6366 ] MEMLOCK, MEMUNLOCK and OTPLOCK modify protection bits. Thus require write permission. Depending on the hardware MEMLOCK might even be write-once, e.g. for SPI-NOR flashes with their WP# tied to GND. OTPLOCK is always write-once. MEMSETBADBLOCK modifies the bad block table. Fixes: f7e6b19bc764 ("mtd: properly check all write ioctls for permissions") Signed-off-by: Michael Walle Reviewed-by: Greg Kroah-Hartman Acked-by: Rafał Miłecki Acked-by: Richard Weinberger Signed-off-by: Miquel Raynal Link: https://lore.kernel.org/linux-mtd/20210303155735.25887-1-michael@walle.cc Signed-off-by: Sasha Levin

mtd: maps: fix error return code of physmap_flash_remove()

2021-05-14T08:52:19Z

[ Upstream commit 620b90d30c08684dc6ebee07c72755d997f9d1f6 ] When platform_get_drvdata() returns NULL to info, no error return code of physmap_flash_remove() is assigned. To fix this bug, err is assigned with -EINVAL in this case Fixes: 73566edf9b91 ("[MTD] Convert physmap to platform driver") Reported-by: TOTE Robot Signed-off-by: Jia-Ju Bai Signed-off-by: Miquel Raynal Link: https://lore.kernel.org/linux-mtd/20210308034446.3052-1-baijiaju1990@gmail.com Signed-off-by: Sasha Levin

mtd: parsers: qcom: incompatible with spi-nor 4k sectors

2021-05-14T08:52:19Z

[ Upstream commit 8f62f59f83c3bc902af91c80732cfcd17e0d7069 ] Partition size and offset value are in block size units, which is the same as 'erasesize'. But when 4K sectors are enabled erasesize is set to 4K. Bail out in that case. Fixes: 803eb124e1a64 ("mtd: parsers: Add Qcom SMEM parser") Reviewed-by: Manivannan Sadhasivam Signed-off-by: Baruch Siach Signed-off-by: Miquel Raynal Link: https://lore.kernel.org/linux-mtd/0a2611f885b894274436ded3ca78bc0440fca74a.1614790096.git.baruch@tkos.co.il Signed-off-by: Sasha Levin

mtd: parsers: qcom: Fix error condition

2021-05-14T08:52:18Z

[ Upstream commit c95310e1b33eae9767af9698aa976d5301f37203 ] qcom_smem_get() does not return NULL, and even if it did, the NULL condition is usually not an error but a success condition and should not trigger an error trace. Let's replace IS_ERR_OR_NULL() by IS_ERR(). This fixes the following smatch warning: drivers/mtd/parsers/qcomsmempart.c:109 parse_qcomsmem_part() warn: passing zero to 'PTR_ERR' Reported-by: kernel test robot Reported-by: Dan Carpenter Fixes: 803eb124e1a6 ("mtd: parsers: Add Qcom SMEM parser") Signed-off-by: Miquel Raynal Link: https://lore.kernel.org/linux-mtd/20210303084634.12796-1-miquel.raynal@bootlin.com Signed-off-by: Sasha Levin

mtd: don't lock when recursively deleting partitions

2021-05-14T08:52:18Z

[ Upstream commit cb4543054c5c4fd33df960b41d7b483ebca8e786 ] When recursively deleting partitions, don't acquire the masters partition lock twice. Otherwise the process ends up in a deadlocked state. Fixes: 46b5889cc2c5 ("mtd: implement proper partition handling") Signed-off-by: David Bauer Signed-off-by: Miquel Raynal Link: https://lore.kernel.org/linux-mtd/20210217195320.893253-1-mail@david-bauer.net Signed-off-by: Sasha Levin

mtd: rawnand: qcom: Return actual error code instead of -ENODEV

2021-05-14T08:52:18Z

[ Upstream commit 55fbb9ba4f06cb6aff32daca1e1910173c13ec51 ] In qcom_probe_nand_devices() function, the error code returned by qcom_nand_host_init_and_register() is converted to -ENODEV in the case of failure. This poses issue if -EPROBE_DEFER is returned when the dependency is not available for a component like parser. So let's restructure the error handling logic a bit and return the actual error code in case of qcom_nand_host_init_and_register() failure. Fixes: c76b78d8ec05 ("mtd: nand: Qualcomm NAND controller driver") Signed-off-by: Manivannan Sadhasivam Signed-off-by: Miquel Raynal Signed-off-by: Sasha Levin

mtd: Handle possible -EPROBE_DEFER from parse_mtd_partitions()

2021-05-14T08:52:18Z

[ Upstream commit 08608adb520e51403be7592c2214846fa440a23a ] There are chances that the parse_mtd_partitions() function will return -EPROBE_DEFER in mtd_device_parse_register(). This might happen when the dependency is not available for the parser. For instance, on SDX55 the MTD_QCOMSMEM_PARTS parser depends on the QCOM_SMEM driver to parse the partitions defined in the shared memory region. With the current flow, the error returned from parse_mtd_partitions() will be discarded in favor of trying to add the fallback partition. This will prevent the driver to end up in probe deferred pool and the partitions won't be parsed even after the QCOM_SMEM driver is available. Fix this issue by bailing out of mtd_device_parse_register() when -EPROBE_DEFER error is returned from parse_mtd_partitions() function and propagate the error code to the driver core for probing later. Fixes: 5ac67ce36cfe ("mtd: move code adding (registering) partitions to the parse_mtd_partitions()") Signed-off-by: Manivannan Sadhasivam Signed-off-by: Miquel Raynal Signed-off-by: Sasha Levin

mtd: rawnand: brcmnand: fix OOB R/W with Hamming ECC

2021-05-14T08:52:18Z

[ Upstream commit f5200c14242fb8fa4a9b93f7fd4064d237e58785 ] Hamming ECC doesn't cover the OOB data, so reading or writing OOB shall always be done without ECC enabled. This is a problem when adding JFFS2 cleanmarkers to erased blocks. If JFFS2 clenmarkers are added to the OOB with ECC enabled, OOB bytes will be changed from ff ff ff to 00 00 00, reporting incorrect ECC errors. Fixes: 27c5b17cd1b1 ("mtd: nand: add NAND driver "library" for Broadcom STB NAND controller") Signed-off-by: Álvaro Fernández Rojas Acked-by: Brian Norris Signed-off-by: Miquel Raynal Link: https://lore.kernel.org/linux-mtd/20210224080210.23686-1-noltari@gmail.com Signed-off-by: Sasha Levin

mtd: rawnand: fsmc: Fix error code in fsmc_nand_probe()

2021-05-14T08:52:18Z

[ Upstream commit e7a97528e3c787802d8c643d6ab2f428511bb047 ] If dma_request_channel() fails then the probe fails and it should return a negative error code, but currently it returns success. fixes: 4774fb0a48aa ("mtd: nand/fsmc: Add DMA support") Signed-off-by: Dan Carpenter Signed-off-by: Miquel Raynal Link: https://lore.kernel.org/linux-mtd/YCqaOZ83OvPOzLwh@mwanda Signed-off-by: Sasha Levin