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

[PATCH v2] HID: memory leak in dualshock4_get_calibration_data

by Eslam Khafagy

function dualshock4_get_calibration_data allocates memory to pointer buf. however the function may exit prematurely due to transfer_failure in this case it does not handle freeing memory. this patch handles memory deallocation at exit. Reported-by: syzbot+4f5f81e1456a1f645bf8(a)syzkaller.appspotmail.com Tested-by: syzbot+4f5f81e1456a1f645bf8(a)syzkaller.appspotmail.com Closes: https://lore.kernel.org/all/691560c4.a70a0220.3124cb.0019.GAE@google.com/T/ Fixes: 947992c7fa9e0 ("HID: playstation: DS4: Fix calibration workaround for clone devices") Cc: stable(a)vger.kernel.org Signed-off-by: Eslam Khafagy <eslam.medhat1993(a)gmail.com> --- v2: * Adding tag "Cc: stable(a)vger.kernel.org" v1: https://lore.kernel.org/all/20251115022323.1395726-1-eslam.medhat1993@gmail… --- drivers/hid/hid-playstation.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/drivers/hid/hid-playstation.c b/drivers/hid/hid-playstation.c index 63f6eb9030d1..fef81b7e27c1 100644 --- a/drivers/hid/hid-playstation.c +++ b/drivers/hid/hid-playstation.c @@ -1992,9 +1992,6 @@ static int dualshock4_get_calibration_data(struct dualshock4 *ds4) acc_z_plus = get_unaligned_le16(&buf[31]); acc_z_minus = get_unaligned_le16(&buf[33]); - /* Done parsing the buffer, so let's free it. */ - kfree(buf); - /* * Set gyroscope calibration and normalization parameters. * Data values will be normalized to 1/DS4_GYRO_RES_PER_DEG_S degree/s. @@ -2041,6 +2038,10 @@ static int dualshock4_get_calibration_data(struct dualshock4 *ds4) ds4->accel_calib_data[2].sens_denom = range_2g; transfer_failed: + /* First free buf if still allocated */ + if(buf) + kfree(buf); + /* * Sanity check gyro calibration data. This is needed to prevent crashes * during report handling of virtual, clone or broken devices not implementing -- 2.43.0

8 minutes

1
0
0 0

[PATCH] USB: ohci-nxp: Fix error handling in ohci-hcd-nxp driver

by Ma Ke

When obtaining the ISP1301 I2C client through the device tree, the driver does not release the device reference in the probe failure path or in the remove function. This could cause a reference count leak, which may prevent the device from being properly unbound or freed, leading to resource leakage. Add put_device() to release the reference in the probe failure path and in the remove function. Found by code review. Cc: stable(a)vger.kernel.org Fixes: 73108aa90cbf ("USB: ohci-nxp: Use isp1301 driver") Signed-off-by: Ma Ke <make24(a)iscas.ac.cn> --- drivers/usb/host/ohci-nxp.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/drivers/usb/host/ohci-nxp.c b/drivers/usb/host/ohci-nxp.c index 24d5a1dc5056..f79558ef0b45 100644 --- a/drivers/usb/host/ohci-nxp.c +++ b/drivers/usb/host/ohci-nxp.c @@ -223,6 +223,8 @@ static int ohci_hcd_nxp_probe(struct platform_device *pdev) fail_resource: usb_put_hcd(hcd); fail_disable: + if (isp1301_i2c_client) + put_device(&isp1301_i2c_client->dev); isp1301_i2c_client = NULL; return ret; } @@ -234,6 +236,8 @@ static void ohci_hcd_nxp_remove(struct platform_device *pdev) usb_remove_hcd(hcd); ohci_nxp_stop_hc(); usb_put_hcd(hcd); + if (isp1301_i2c_client) + put_device(&isp1301_i2c_client->dev); isp1301_i2c_client = NULL; } -- 2.17.1

32 minutes

2
1
0 0

[PATCH] USB: Fix error handling in gadget driver

by Ma Ke

lpc32xx_udc_probe() acquires an i2c_client reference through isp1301_get_client() but fails to release it in both error handling paths and the normal removal path. This could result in a reference count leak for the I2C device, preventing proper cleanup and potentially leading to resource exhaustion. Add put_device() to release the reference in the probe failure path and in the remove function. Calling path: isp1301_get_client() -> of_find_i2c_device_by_node() -> i2c_find_device_by_fwnode(). As comments of i2c_find_device_by_fwnode() says, 'The user must call put_device(&client->dev) once done with the i2c client.' Found by code review. Cc: stable(a)vger.kernel.org Fixes: 24a28e428351 ("USB: gadget driver for LPC32xx") Signed-off-by: Ma Ke <make24(a)iscas.ac.cn> --- drivers/usb/gadget/udc/lpc32xx_udc.c | 35 +++++++++++++++++++++++----- 1 file changed, 29 insertions(+), 6 deletions(-) diff --git a/drivers/usb/gadget/udc/lpc32xx_udc.c b/drivers/usb/gadget/udc/lpc32xx_udc.c index 1a7d3c4f652f..b6fddfff712d 100644 --- a/drivers/usb/gadget/udc/lpc32xx_udc.c +++ b/drivers/usb/gadget/udc/lpc32xx_udc.c @@ -2986,6 +2986,7 @@ static int lpc32xx_udc_probe(struct platform_device *pdev) int retval, i; dma_addr_t dma_handle; struct device_node *isp1301_node; + bool isp1301_acquired = false; udc = devm_kmemdup(dev, &controller_template, sizeof(*udc), GFP_KERNEL); if (!udc) @@ -3013,6 +3014,7 @@ static int lpc32xx_udc_probe(struct platform_device *pdev) if (!udc->isp1301_i2c_client) { return -EPROBE_DEFER; } + isp1301_acquired = true; dev_info(udc->dev, "ISP1301 I2C device at address 0x%x\n", udc->isp1301_i2c_client->addr); @@ -3020,7 +3022,7 @@ static int lpc32xx_udc_probe(struct platform_device *pdev) pdev->dev.dma_mask = &lpc32xx_usbd_dmamask; retval = dma_set_coherent_mask(&pdev->dev, DMA_BIT_MASK(32)); if (retval) - return retval; + goto i2c_fail; udc->board = &lpc32xx_usbddata; @@ -3038,28 +3040,32 @@ static int lpc32xx_udc_probe(struct platform_device *pdev) /* Get IRQs */ for (i = 0; i < 4; i++) { udc->udp_irq[i] = platform_get_irq(pdev, i); - if (udc->udp_irq[i] < 0) - return udc->udp_irq[i]; + if (udc->udp_irq[i] < 0) { + retval = udc->udp_irq[i]; + goto i2c_fail; + } } udc->udp_baseaddr = devm_platform_ioremap_resource(pdev, 0); if (IS_ERR(udc->udp_baseaddr)) { dev_err(udc->dev, "IO map failure\n"); - return PTR_ERR(udc->udp_baseaddr); + retval = PTR_ERR(udc->udp_baseaddr); + goto i2c_fail; } /* Get USB device clock */ udc->usb_slv_clk = devm_clk_get(&pdev->dev, NULL); if (IS_ERR(udc->usb_slv_clk)) { dev_err(udc->dev, "failed to acquire USB device clock\n"); - return PTR_ERR(udc->usb_slv_clk); + retval = PTR_ERR(udc->usb_slv_clk); + goto i2c_fail; } /* Enable USB device clock */ retval = clk_prepare_enable(udc->usb_slv_clk); if (retval < 0) { dev_err(udc->dev, "failed to start USB device clock\n"); - return retval; + goto i2c_fail; } /* Setup deferred workqueue data */ @@ -3161,6 +3167,8 @@ static int lpc32xx_udc_probe(struct platform_device *pdev) dma_free_coherent(&pdev->dev, UDCA_BUFF_SIZE, udc->udca_v_base, udc->udca_p_base); i2c_fail: + if (isp1301_acquired && udc->isp1301_i2c_client) + put_device(&udc->isp1301_i2c_client->dev); clk_disable_unprepare(udc->usb_slv_clk); dev_err(udc->dev, "%s probe failed, %d\n", driver_name, retval); @@ -3170,6 +3178,18 @@ static int lpc32xx_udc_probe(struct platform_device *pdev) static void lpc32xx_udc_remove(struct platform_device *pdev) { struct lpc32xx_udc *udc = platform_get_drvdata(pdev); + struct device *dev = &pdev->dev; + struct device_node *isp1301_node; + bool isp1301_acquired = false; + + /* Check if we acquired isp1301 via device tree */ + if (dev->of_node) { + isp1301_node = of_parse_phandle(dev->of_node, "transceiver", 0); + if (isp1301_node) { + isp1301_acquired = true; + of_node_put(isp1301_node); + } + } usb_del_gadget_udc(&udc->gadget); if (udc->driver) { @@ -3189,6 +3209,9 @@ static void lpc32xx_udc_remove(struct platform_device *pdev) dma_free_coherent(&pdev->dev, UDCA_BUFF_SIZE, udc->udca_v_base, udc->udca_p_base); + if (isp1301_acquired && udc->isp1301_i2c_client) + put_device(&udc->isp1301_i2c_client->dev); + clk_disable_unprepare(udc->usb_slv_clk); } -- 2.17.1

45 minutes

1
0
0 0

[PATCH v2 0/3] SDM630/660: Add missing MDSS reset

by Alexey Minnekhanov

Since kernel 6.17 display stack needs to reset the hardware properly to ensure that we don't run into issues with the hardware configured by the bootloader. MDSS reset is necessary to have working display when the bootloader has already initialized it for the boot splash screen. Signed-off-by: Alexey Minnekhanov <alexeymin(a)postmarketos.org> --- Changes in v2: - Added "Fixes" tag, pointing to commit which is the root cause for exposing this "bug", which is only present since 6.17 - Extended commit messages - Prepared series using correct user.email git setting - Link to v1: https://lore.kernel.org/r/20251031-sdm660-mdss-reset-v1-0-14cb4e6836f2@post… --- Alexey Minnekhanov (3): dt-bindings: clock: mmcc-sdm660: Add missing MDSS reset clk: qcom: mmcc-sdm660: Add missing MDSS reset arm64: dts: qcom: sdm630: Add missing MDSS reset arch/arm64/boot/dts/qcom/sdm630.dtsi | 1 + drivers/clk/qcom/mmcc-sdm660.c | 1 + include/dt-bindings/clock/qcom,mmcc-sdm660.h | 1 + 3 files changed, 3 insertions(+) --- base-commit: 1cc41c88ef00de0f3216c5f4b9cfab47de1c49d3 change-id: 20251031-sdm660-mdss-reset-015a46a238b5 Best regards, -- Alexey Minnekhanov <alexeymin(a)postmarketos.org>

1 hour, 22 minutes

1
3
0 0

[PATCH v2 1/2] crypto: scatterwalk - Fix memcpy_sglist() to always succeed

by Eric Biggers

The original implementation of memcpy_sglist() was broken because it didn't handle scatterlists that describe exactly the same memory, which is a case that many callers rely on. The current implementation is broken too because it calls the skcipher_walk functions which can fail. It ignores any errors from those functions. Fix it by replacing it with a new implementation written from scratch. It always succeeds. It's also a bit faster, since it avoids the overhead of skcipher_walk. skcipher_walk includes a lot of functionality (such as alignmask handling) that's irrelevant here. Reported-by: Colin Ian King <coking(a)nvidia.com> Closes: https://lore.kernel.org/r/20251114122620.111623-1-coking@nvidia.com Fixes: 131bdceca1f0 ("crypto: scatterwalk - Add memcpy_sglist") Fixes: 0f8d42bf128d ("crypto: scatterwalk - Move skcipher walk and use it for memcpy_sglist") Cc: stable(a)vger.kernel.org Signed-off-by: Eric Biggers <ebiggers(a)kernel.org> --- crypto/scatterwalk.c | 97 +++++++++++++++++++++++++++++++----- include/crypto/scatterwalk.h | 52 +++++++++++-------- 2 files changed, 115 insertions(+), 34 deletions(-) diff --git a/crypto/scatterwalk.c b/crypto/scatterwalk.c index 1d010e2a1b1a..b95e5974e327 100644 --- a/crypto/scatterwalk.c +++ b/crypto/scatterwalk.c @@ -99,30 +99,101 @@ void memcpy_to_sglist(struct scatterlist *sg, unsigned int start, scatterwalk_start_at_pos(&walk, sg, start); memcpy_to_scatterwalk(&walk, buf, nbytes); } EXPORT_SYMBOL_GPL(memcpy_to_sglist); +/** + * memcpy_sglist() - Copy data from one scatterlist to another + * @dst: The destination scatterlist. Can be NULL if @nbytes == 0. + * @src: The source scatterlist. Can be NULL if @nbytes == 0. + * @nbytes: Number of bytes to copy + * + * The scatterlists can describe exactly the same memory, in which case this + * function is a no-op. No other overlaps are supported. + * + * Context: Any context + */ void memcpy_sglist(struct scatterlist *dst, struct scatterlist *src, unsigned int nbytes) { - struct skcipher_walk walk = {}; + unsigned int src_offset, dst_offset; - if (unlikely(nbytes == 0)) /* in case sg == NULL */ + if (unlikely(nbytes == 0)) /* in case src and/or dst is NULL */ return; - walk.total = nbytes; - - scatterwalk_start(&walk.in, src); - scatterwalk_start(&walk.out, dst); + src_offset = src->offset; + dst_offset = dst->offset; + for (;;) { + /* Compute the length to copy this step. */ + unsigned int len = min3(src->offset + src->length - src_offset, + dst->offset + dst->length - dst_offset, + nbytes); + struct page *src_page = sg_page(src); + struct page *dst_page = sg_page(dst); + const void *src_virt; + void *dst_virt; + + if (IS_ENABLED(CONFIG_HIGHMEM)) { + /* HIGHMEM: we may have to actually map the pages. */ + const unsigned int src_oip = offset_in_page(src_offset); + const unsigned int dst_oip = offset_in_page(dst_offset); + const unsigned int limit = PAGE_SIZE; + + /* Further limit len to not cross a page boundary. */ + len = min3(len, limit - src_oip, limit - dst_oip); + + /* Compute the source and destination pages. */ + src_page += src_offset / PAGE_SIZE; + dst_page += dst_offset / PAGE_SIZE; + + if (src_page != dst_page) { + /* Copy between different pages. */ + memcpy_page(dst_page, dst_oip, + src_page, src_oip, len); + flush_dcache_page(dst_page); + } else if (src_oip != dst_oip) { + /* Copy between different parts of same page. */ + dst_virt = kmap_local_page(dst_page); + memcpy(dst_virt + dst_oip, dst_virt + src_oip, + len); + kunmap_local(dst_virt); + flush_dcache_page(dst_page); + } /* Else, it's the same memory. No action needed. */ + } else { + /* + * !HIGHMEM: no mapping needed. Just work in the linear + * buffer of each sg entry. Note that we can cross page + * boundaries, as they are not significant in this case. + */ + src_virt = page_address(src_page) + src_offset; + dst_virt = page_address(dst_page) + dst_offset; + if (src_virt != dst_virt) { + memcpy(dst_virt, src_virt, len); + if (ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE) + __scatterwalk_flush_dcache_pages( + dst_page, dst_offset, len); + } /* Else, it's the same memory. No action needed. */ + } + nbytes -= len; + if (nbytes == 0) /* No more to copy? */ + break; - skcipher_walk_first(&walk, true); - do { - if (walk.src.virt.addr != walk.dst.virt.addr) - memcpy(walk.dst.virt.addr, walk.src.virt.addr, - walk.nbytes); - skcipher_walk_done(&walk, 0); - } while (walk.nbytes); + /* + * There's more to copy. Advance the offsets by the length + * copied this step, and advance the sg entries as needed. + */ + src_offset += len; + if (src_offset >= src->offset + src->length) { + src = sg_next(src); + src_offset = src->offset; + } + dst_offset += len; + if (dst_offset >= dst->offset + dst->length) { + dst = sg_next(dst); + dst_offset = dst->offset; + } + } } EXPORT_SYMBOL_GPL(memcpy_sglist); struct scatterlist *scatterwalk_ffwd(struct scatterlist dst[2], struct scatterlist *src, diff --git a/include/crypto/scatterwalk.h b/include/crypto/scatterwalk.h index 83d14376ff2b..f485454e3955 100644 --- a/include/crypto/scatterwalk.h +++ b/include/crypto/scatterwalk.h @@ -225,10 +225,38 @@ static inline void scatterwalk_done_src(struct scatter_walk *walk, { scatterwalk_unmap(walk); scatterwalk_advance(walk, nbytes); } +/* + * Flush the dcache of any pages that overlap the region + * [offset, offset + nbytes) relative to base_page. + * + * This should be called only when ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE, to ensure + * that all relevant code (including the call to sg_page() in the caller, if + * applicable) gets fully optimized out when !ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE. + */ +static inline void __scatterwalk_flush_dcache_pages(struct page *base_page, + unsigned int offset, + unsigned int nbytes) +{ + unsigned int num_pages; + + base_page += offset / PAGE_SIZE; + offset %= PAGE_SIZE; + + /* + * This is an overflow-safe version of + * num_pages = DIV_ROUND_UP(offset + nbytes, PAGE_SIZE). + */ + num_pages = nbytes / PAGE_SIZE; + num_pages += DIV_ROUND_UP(offset + (nbytes % PAGE_SIZE), PAGE_SIZE); + + for (unsigned int i = 0; i < num_pages; i++) + flush_dcache_page(base_page + i); +} + /** * scatterwalk_done_dst() - Finish one step of a walk of destination scatterlist * @walk: the scatter_walk * @nbytes: the number of bytes processed this step, less than or equal to the * number of bytes that scatterwalk_next() returned. @@ -238,31 +266,13 @@ static inline void scatterwalk_done_src(struct scatter_walk *walk, */ static inline void scatterwalk_done_dst(struct scatter_walk *walk, unsigned int nbytes) { scatterwalk_unmap(walk); - /* - * Explicitly check ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE instead of just - * relying on flush_dcache_page() being a no-op when not implemented, - * since otherwise the BUG_ON in sg_page() does not get optimized out. - * This also avoids having to consider whether the loop would get - * reliably optimized out or not. - */ - if (ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE) { - struct page *base_page; - unsigned int offset; - int start, end, i; - - base_page = sg_page(walk->sg); - offset = walk->offset; - start = offset >> PAGE_SHIFT; - end = start + (nbytes >> PAGE_SHIFT); - end += (offset_in_page(offset) + offset_in_page(nbytes) + - PAGE_SIZE - 1) >> PAGE_SHIFT; - for (i = start; i < end; i++) - flush_dcache_page(base_page + i); - } + if (ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE) + __scatterwalk_flush_dcache_pages(sg_page(walk->sg), + walk->offset, nbytes); scatterwalk_advance(walk, nbytes); } void scatterwalk_skip(struct scatter_walk *walk, unsigned int nbytes); -- 2.51.2

3 hours, 27 minutes

1
0
0 0

[PATCH 1/2] crypto: scatterwalk - Fix memcpy_sglist() to always succeed

by Eric Biggers

The original implementation of memcpy_sglist() was broken because it didn't handle overlapping scatterlists. The current implementation is broken too because it calls the skcipher_walk functions which can fail. It ignores any errors from those functions. Fix it by replacing it with a new implementation written from scratch. It always succeeds. It's also a bit faster, since it avoids the overhead of skcipher_walk. skcipher_walk includes a lot of functionality (such as alignmask handling) that's irrelevant here. Reported-by: Colin Ian King <coking(a)nvidia.com> Closes: https://lore.kernel.org/r/20251114122620.111623-1-coking@nvidia.com Fixes: 131bdceca1f0 ("crypto: scatterwalk - Add memcpy_sglist") Fixes: 0f8d42bf128d ("crypto: scatterwalk - Move skcipher walk and use it for memcpy_sglist") Cc: stable(a)vger.kernel.org Signed-off-by: Eric Biggers <ebiggers(a)kernel.org> --- crypto/scatterwalk.c | 103 ++++++++++++++++++++++++++++++----- include/crypto/scatterwalk.h | 52 +++++++++++------- 2 files changed, 121 insertions(+), 34 deletions(-) diff --git a/crypto/scatterwalk.c b/crypto/scatterwalk.c index 1d010e2a1b1a..af6c17bfcadb 100644 --- a/crypto/scatterwalk.c +++ b/crypto/scatterwalk.c @@ -99,30 +99,107 @@ void memcpy_to_sglist(struct scatterlist *sg, unsigned int start, scatterwalk_start_at_pos(&walk, sg, start); memcpy_to_scatterwalk(&walk, buf, nbytes); } EXPORT_SYMBOL_GPL(memcpy_to_sglist); +/** + * memcpy_sglist() - Copy data from one scatterlist to another + * @dst: The destination scatterlist. Can be NULL if @nbytes == 0. + * @src: The source scatterlist. Can be NULL if @nbytes == 0. + * @nbytes: Number of bytes to copy + * + * The scatterlists can overlap. Hence this really acts like memmove(), not + * memcpy(). + * + * Context: Any context + */ void memcpy_sglist(struct scatterlist *dst, struct scatterlist *src, unsigned int nbytes) { - struct skcipher_walk walk = {}; + unsigned int src_offset, dst_offset; - if (unlikely(nbytes == 0)) /* in case sg == NULL */ + if (unlikely(nbytes == 0)) /* in case src and/or dst is NULL */ return; - walk.total = nbytes; - - scatterwalk_start(&walk.in, src); - scatterwalk_start(&walk.out, dst); + src_offset = src->offset; + dst_offset = dst->offset; + for (;;) { + /* Compute the length to copy this step. */ + unsigned int len = min3(src->offset + src->length - src_offset, + dst->offset + dst->length - dst_offset, + nbytes); + struct page *src_page = sg_page(src); + struct page *dst_page = sg_page(dst); + const void *src_virt; + void *dst_virt; + + if (IS_ENABLED(CONFIG_HIGHMEM)) { + /* HIGHMEM: we may have to actually map the pages. */ + const unsigned int src_oip = offset_in_page(src_offset); + const unsigned int dst_oip = offset_in_page(dst_offset); + const unsigned int limit = PAGE_SIZE; + + /* Further limit len to not cross a page boundary. */ + len = min3(len, limit - src_oip, limit - dst_oip); + + /* Compute the source and destination pages. */ + src_page += src_offset / PAGE_SIZE; + dst_page += dst_offset / PAGE_SIZE; + + if (src_page != dst_page) { + /* + * Copy between different pages. + * No need for memmove(), as the pages differ. + */ + src_virt = kmap_local_page(src_page); + dst_virt = kmap_local_page(dst_page); + memcpy(dst_virt + dst_oip, src_virt + src_oip, + len); + flush_dcache_page(dst_page); + kunmap_local(dst_virt); + kunmap_local(src_virt); + } else if (src_oip != dst_oip) { + /* Copy between different parts of same page */ + dst_virt = kmap_local_page(dst_page); + memmove(dst_virt + dst_oip, dst_virt + src_oip, + len); + flush_dcache_page(dst_page); + kunmap_local(dst_virt); + } /* Exact overlap. No action needed. */ + } else { + /* + * !HIGHMEM: no mapping needed. Just work in the linear + * buffer of each sg entry. + */ + src_virt = page_address(src_page) + src_offset; + dst_virt = page_address(dst_page) + dst_offset; + if (src_virt != dst_virt) { + memmove(dst_virt, src_virt, len); + if (ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE) + __scatterwalk_flush_dcache_pages( + dst_page, dst_offset, len); + } + } + nbytes -= len; + if (nbytes == 0) /* No more to copy? */ + break; - skcipher_walk_first(&walk, true); - do { - if (walk.src.virt.addr != walk.dst.virt.addr) - memcpy(walk.dst.virt.addr, walk.src.virt.addr, - walk.nbytes); - skcipher_walk_done(&walk, 0); - } while (walk.nbytes); + /* + * There's more to copy. Advance the offsets by the length + * copied this step, and advance the sg entries as needed. + */ + src_offset += len; + if (src_offset >= src->offset + src->length) { + src = sg_next(src); + src_offset = src->offset; + } + dst_offset += len; + if (dst_offset >= dst->offset + dst->length) { + dst = sg_next(dst); + dst_offset = dst->offset; + } + } } EXPORT_SYMBOL_GPL(memcpy_sglist); struct scatterlist *scatterwalk_ffwd(struct scatterlist dst[2], struct scatterlist *src, diff --git a/include/crypto/scatterwalk.h b/include/crypto/scatterwalk.h index 83d14376ff2b..f485454e3955 100644 --- a/include/crypto/scatterwalk.h +++ b/include/crypto/scatterwalk.h @@ -225,10 +225,38 @@ static inline void scatterwalk_done_src(struct scatter_walk *walk, { scatterwalk_unmap(walk); scatterwalk_advance(walk, nbytes); } +/* + * Flush the dcache of any pages that overlap the region + * [offset, offset + nbytes) relative to base_page. + * + * This should be called only when ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE, to ensure + * that all relevant code (including the call to sg_page() in the caller, if + * applicable) gets fully optimized out when !ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE. + */ +static inline void __scatterwalk_flush_dcache_pages(struct page *base_page, + unsigned int offset, + unsigned int nbytes) +{ + unsigned int num_pages; + + base_page += offset / PAGE_SIZE; + offset %= PAGE_SIZE; + + /* + * This is an overflow-safe version of + * num_pages = DIV_ROUND_UP(offset + nbytes, PAGE_SIZE). + */ + num_pages = nbytes / PAGE_SIZE; + num_pages += DIV_ROUND_UP(offset + (nbytes % PAGE_SIZE), PAGE_SIZE); + + for (unsigned int i = 0; i < num_pages; i++) + flush_dcache_page(base_page + i); +} + /** * scatterwalk_done_dst() - Finish one step of a walk of destination scatterlist * @walk: the scatter_walk * @nbytes: the number of bytes processed this step, less than or equal to the * number of bytes that scatterwalk_next() returned. @@ -238,31 +266,13 @@ static inline void scatterwalk_done_src(struct scatter_walk *walk, */ static inline void scatterwalk_done_dst(struct scatter_walk *walk, unsigned int nbytes) { scatterwalk_unmap(walk); - /* - * Explicitly check ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE instead of just - * relying on flush_dcache_page() being a no-op when not implemented, - * since otherwise the BUG_ON in sg_page() does not get optimized out. - * This also avoids having to consider whether the loop would get - * reliably optimized out or not. - */ - if (ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE) { - struct page *base_page; - unsigned int offset; - int start, end, i; - - base_page = sg_page(walk->sg); - offset = walk->offset; - start = offset >> PAGE_SHIFT; - end = start + (nbytes >> PAGE_SHIFT); - end += (offset_in_page(offset) + offset_in_page(nbytes) + - PAGE_SIZE - 1) >> PAGE_SHIFT; - for (i = start; i < end; i++) - flush_dcache_page(base_page + i); - } + if (ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE) + __scatterwalk_flush_dcache_pages(sg_page(walk->sg), + walk->offset, nbytes); scatterwalk_advance(walk, nbytes); } void scatterwalk_skip(struct scatter_walk *walk, unsigned int nbytes); -- 2.51.2

3 hours, 48 minutes

1
1
0 0

[merged mm-hotfixes-stable] mm-swap-fix-potential-uaf-issue-for-vma-readahead.patch removed from -mm tree

by Andrew Morton

The quilt patch titled Subject: mm, swap: fix potential UAF issue for VMA readahead has been removed from the -mm tree. Its filename was mm-swap-fix-potential-uaf-issue-for-vma-readahead.patch This patch was dropped because it was merged into the mm-hotfixes-stable branch of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm ------------------------------------------------------ From: Kairui Song <kasong(a)tencent.com> Subject: mm, swap: fix potential UAF issue for VMA readahead Date: Tue, 11 Nov 2025 21:36:08 +0800 Since commit 78524b05f1a3 ("mm, swap: avoid redundant swap device pinning"), the common helper for allocating and preparing a folio in the swap cache layer no longer tries to get a swap device reference internally, because all callers of __read_swap_cache_async are already holding a swap entry reference. The repeated swap device pinning isn't needed on the same swap device. Caller of VMA readahead is also holding a reference to the target entry's swap device, but VMA readahead walks the page table, so it might encounter swap entries from other devices, and call __read_swap_cache_async on another device without holding a reference to it. So it is possible to cause a UAF when swapoff of device A raced with swapin on device B, and VMA readahead tries to read swap entries from device A. It's not easy to trigger, but in theory, it could cause real issues. Make VMA readahead try to get the device reference first if the swap device is a different one from the target entry. Link: https://lkml.kernel.org/r/20251111-swap-fix-vma-uaf-v1-1-41c660e58562@tence… Fixes: 78524b05f1a3 ("mm, swap: avoid redundant swap device pinning") Suggested-by: Huang Ying <ying.huang(a)linux.alibaba.com> Signed-off-by: Kairui Song <kasong(a)tencent.com> Acked-by: Chris Li <chrisl(a)kernel.org> Cc: Baoquan He <bhe(a)redhat.com> Cc: Barry Song <baohua(a)kernel.org> Cc: Kemeng Shi <shikemeng(a)huaweicloud.com> Cc: Nhat Pham <nphamcs(a)gmail.com> Cc: <stable(a)vger.kernel.org> Signed-off-by: Andrew Morton <akpm(a)linux-foundation.org> --- mm/swap_state.c | 13 +++++++++++++ 1 file changed, 13 insertions(+) --- a/mm/swap_state.c~mm-swap-fix-potential-uaf-issue-for-vma-readahead +++ a/mm/swap_state.c @@ -748,6 +748,8 @@ static struct folio *swap_vma_readahead( blk_start_plug(&plug); for (addr = start; addr < end; ilx++, addr += PAGE_SIZE) { + struct swap_info_struct *si = NULL; + if (!pte++) { pte = pte_offset_map(vmf->pmd, addr); if (!pte) @@ -761,8 +763,19 @@ static struct folio *swap_vma_readahead( continue; pte_unmap(pte); pte = NULL; + /* + * Readahead entry may come from a device that we are not + * holding a reference to, try to grab a reference, or skip. + */ + if (swp_type(entry) != swp_type(targ_entry)) { + si = get_swap_device(entry); + if (!si) + continue; + } folio = __read_swap_cache_async(entry, gfp_mask, mpol, ilx, &page_allocated, false); + if (si) + put_swap_device(si); if (!folio) continue; if (page_allocated) { _ Patches currently in -mm which might be from kasong(a)tencent.com are mm-swap-do-not-perform-synchronous-discard-during-allocation.patch mm-swap-rename-helper-for-setup-bad-slots.patch mm-swap-cleanup-swap-entry-allocation-parameter.patch mm-migrate-swap-drop-usage-of-folio_index.patch mm-swap-remove-redundant-argument-for-isolating-a-cluster.patch

7 hours, 42 minutes

1
0
0 0

[merged mm-hotfixes-stable] selftests-user_events-fix-type-cast-for-write_index-packed-member-in-perf_test.patch removed from -mm tree

by Andrew Morton

The quilt patch titled Subject: selftests/user_events: fix type cast for write_index packed member in perf_test has been removed from the -mm tree. Its filename was selftests-user_events-fix-type-cast-for-write_index-packed-member-in-perf_test.patch This patch was dropped because it was merged into the mm-hotfixes-stable branch of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm ------------------------------------------------------ From: Ankit Khushwaha <ankitkhushwaha.linux(a)gmail.com> Subject: selftests/user_events: fix type cast for write_index packed member in perf_test Date: Thu, 6 Nov 2025 15:25:32 +0530 Accessing 'reg.write_index' directly triggers a -Waddress-of-packed-member warning due to potential unaligned pointer access: perf_test.c:239:38: warning: taking address of packed member 'write_index' of class or structure 'user_reg' may result in an unaligned pointer value [-Waddress-of-packed-member] 239 | ASSERT_NE(-1, write(self->data_fd, &reg.write_index, | ^~~~~~~~~~~~~~~ Since write(2) works with any alignment. Casting '&reg.write_index' explicitly to 'void *' to suppress this warning. Link: https://lkml.kernel.org/r/20251106095532.15185-1-ankitkhushwaha.linux@gmail… Fixes: 42187bdc3ca4 ("selftests/user_events: Add perf self-test for empty arguments events") Signed-off-by: Ankit Khushwaha <ankitkhushwaha.linux(a)gmail.com> Cc: Beau Belgrave <beaub(a)linux.microsoft.com> Cc: "Masami Hiramatsu (Google)" <mhiramat(a)kernel.org> Cc: Steven Rostedt <rostedt(a)goodmis.org> Cc: sunliming <sunliming(a)kylinos.cn> Cc: Wei Yang <richard.weiyang(a)gmail.com> Cc: Shuah Khan <shuah(a)kernel.org> Cc: <stable(a)vger.kernel.org> Signed-off-by: Andrew Morton <akpm(a)linux-foundation.org> --- tools/testing/selftests/user_events/perf_test.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) --- a/tools/testing/selftests/user_events/perf_test.c~selftests-user_events-fix-type-cast-for-write_index-packed-member-in-perf_test +++ a/tools/testing/selftests/user_events/perf_test.c @@ -236,7 +236,7 @@ TEST_F(user, perf_empty_events) { ASSERT_EQ(1 << reg.enable_bit, self->check); /* Ensure write shows up at correct offset */ - ASSERT_NE(-1, write(self->data_fd, &reg.write_index, + ASSERT_NE(-1, write(self->data_fd, (void *)&reg.write_index, sizeof(reg.write_index))); val = (void *)(((char *)perf_page) + perf_page->data_offset); ASSERT_EQ(PERF_RECORD_SAMPLE, *val); _ Patches currently in -mm which might be from ankitkhushwaha.linux(a)gmail.com are selftest-mm-fix-pointer-comparison-in-mremap_test.patch

7 hours, 42 minutes

1
0
0 0

[merged mm-hotfixes-stable] lib-test_kho-check-if-kho-is-enabled.patch removed from -mm tree

by Andrew Morton

The quilt patch titled Subject: lib/test_kho: check if KHO is enabled has been removed from the -mm tree. Its filename was lib-test_kho-check-if-kho-is-enabled.patch This patch was dropped because it was merged into the mm-hotfixes-stable branch of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm ------------------------------------------------------ From: Pasha Tatashin <pasha.tatashin(a)soleen.com> Subject: lib/test_kho: check if KHO is enabled Date: Thu, 6 Nov 2025 17:06:35 -0500 We must check whether KHO is enabled prior to issuing KHO commands, otherwise KHO internal data structures are not initialized. Link: https://lkml.kernel.org/r/20251106220635.2608494-1-pasha.tatashin@soleen.com Fixes: b753522bed0b ("kho: add test for kexec handover") Signed-off-by: Pasha Tatashin <pasha.tatashin(a)soleen.com> Reported-by: kernel test robot <oliver.sang(a)intel.com> Closes: https://lore.kernel.org/oe-lkp/202511061629.e242724-lkp@intel.com Reviewed-by: Pratyush Yadav <pratyush(a)kernel.org> Reviewed-by: Mike Rapoport (Microsoft) <rppt(a)kernel.org> Cc: Alexander Graf <graf(a)amazon.com> Cc: <stable(a)vger.kernel.org> Signed-off-by: Andrew Morton <akpm(a)linux-foundation.org> --- lib/test_kho.c | 3 +++ 1 file changed, 3 insertions(+) --- a/lib/test_kho.c~lib-test_kho-check-if-kho-is-enabled +++ a/lib/test_kho.c @@ -301,6 +301,9 @@ static int __init kho_test_init(void) phys_addr_t fdt_phys; int err; + if (!kho_is_enabled()) + return 0; + err = kho_retrieve_subtree(KHO_TEST_FDT, &fdt_phys); if (!err) return kho_test_restore(fdt_phys); _ Patches currently in -mm which might be from pasha.tatashin(a)soleen.com are kho-make-debugfs-interface-optional.patch kho-add-interfaces-to-unpreserve-folios-page-ranges-and-vmalloc.patch memblock-unpreserve-memory-in-case-of-error.patch test_kho-unpreserve-memory-in-case-of-error.patch kho-dont-unpreserve-memory-during-abort.patch liveupdate-kho-move-to-kernel-liveupdate.patch liveupdate-kho-move-to-kernel-liveupdate-fix.patch maintainers-update-kho-maintainers.patch kho-fix-misleading-log-message-in-kho_populate.patch kho-convert-__kho_abort-to-return-void.patch kho-introduce-high-level-memory-allocation-api.patch kho-preserve-fdt-folio-only-once-during-initialization.patch kho-verify-deserialization-status-and-fix-fdt-alignment-access.patch kho-always-expose-output-fdt-in-debugfs.patch kho-simplify-serialization-and-remove-__kho_abort.patch kho-remove-global-preserved_mem_map-and-store-state-in-fdt.patch kho-remove-abort-functionality-and-support-state-refresh.patch kho-update-fdt-dynamically-for-subtree-addition-removal.patch kho-allow-kexec-load-before-kho-finalization.patch kho-allow-memory-preservation-state-updates-after-finalization.patch kho-add-kconfig-option-to-enable-kho-by-default.patch liveupdate-luo_core-luo_ioctl-live-update-orchestrator.patch liveupdate-luo_core-integrate-with-kho.patch liveupdate-luo_core-integrate-with-kho-fix.patch reboot-call-liveupdate_reboot-before-kexec.patch liveupdate-luo_session-add-sessions-support.patch liveupdate-luo_session-add-sessions-support-fix.patch liveupdate-luo_ioctl-add-user-interface.patch liveupdate-luo_file-implement-file-systems-callbacks.patch liveupdate-luo_session-add-ioctls-for-file-preservation-and-state-management.patch liveupdate-luo_flb-introduce-file-lifecycle-bound-global-state.patch docs-add-luo-documentation.patch maintainers-add-liveupdate-entry.patch mm-memfd_luo-allow-preserving-memfd-fix.patch selftests-liveupdate-add-userspace-api-selftests.patch selftests-liveupdate-add-kexec-based-selftest-for-session-lifecycle.patch selftests-liveupdate-add-kexec-test-for-multiple-and-empty-sessions.patch tests-liveupdate-add-in-kernel-liveupdate-test.patch

7 hours, 42 minutes

1
0
0 0

[merged mm-hotfixes-stable] mm-huge_memory-fix-folio-split-check-for-anon-folios-in-swapcache.patch removed from -mm tree

by Andrew Morton

The quilt patch titled Subject: mm/huge_memory: fix folio split check for anon folios in swapcache has been removed from the -mm tree. Its filename was mm-huge_memory-fix-folio-split-check-for-anon-folios-in-swapcache.patch This patch was dropped because it was merged into the mm-hotfixes-stable branch of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm ------------------------------------------------------ From: Zi Yan <ziy(a)nvidia.com> Subject: mm/huge_memory: fix folio split check for anon folios in swapcache Date: Wed, 5 Nov 2025 11:29:10 -0500 Both uniform and non uniform split check missed the check to prevent splitting anon folios in swapcache to non-zero order. Splitting anon folios in swapcache to non-zero order can cause data corruption since swapcache only support PMD order and order-0 entries. This can happen when one use split_huge_pages under debugfs to split anon folios in swapcache. In-tree callers do not perform such an illegal operation. Only debugfs interface could trigger it. I will put adding a test case on my TODO list. Fix the check. Link: https://lkml.kernel.org/r/20251105162910.752266-1-ziy@nvidia.com Fixes: 58729c04cf10 ("mm/huge_memory: add buddy allocator like (non-uniform) folio_split()") Signed-off-by: Zi Yan <ziy(a)nvidia.com> Reported-by: "David Hildenbrand (Red Hat)" <david(a)kernel.org> Closes: https://lore.kernel.org/all/dc0ecc2c-4089-484f-917f-920fdca4c898@kernel.org/ Acked-by: David Hildenbrand (Red Hat) <david(a)kernel.org> Cc: Baolin Wang <baolin.wang(a)linux.alibaba.com> Cc: Barry Song <baohua(a)kernel.org> Cc: Dev Jain <dev.jain(a)arm.com> Cc: Lance Yang <lance.yang(a)linux.dev> Cc: Liam Howlett <liam.howlett(a)oracle.com> Cc: Lorenzo Stoakes <lorenzo.stoakes(a)oracle.com> Cc: Nico Pache <npache(a)redhat.com> Cc: Ryan Roberts <ryan.roberts(a)arm.com> Cc: Wei Yang <richard.weiyang(a)gmail.com> Cc: <stable(a)vger.kernel.org> Signed-off-by: Andrew Morton <akpm(a)linux-foundation.org> --- mm/huge_memory.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) --- a/mm/huge_memory.c~mm-huge_memory-fix-folio-split-check-for-anon-folios-in-swapcache +++ a/mm/huge_memory.c @@ -3522,7 +3522,8 @@ bool non_uniform_split_supported(struct /* order-1 is not supported for anonymous THP. */ VM_WARN_ONCE(warns && new_order == 1, "Cannot split to order-1 folio"); - return new_order != 1; + if (new_order == 1) + return false; } else if (IS_ENABLED(CONFIG_READ_ONLY_THP_FOR_FS) && !mapping_large_folio_support(folio->mapping)) { /* @@ -3553,7 +3554,8 @@ bool uniform_split_supported(struct foli if (folio_test_anon(folio)) { VM_WARN_ONCE(warns && new_order == 1, "Cannot split to order-1 folio"); - return new_order != 1; + if (new_order == 1) + return false; } else if (new_order) { if (IS_ENABLED(CONFIG_READ_ONLY_THP_FOR_FS) && !mapping_large_folio_support(folio->mapping)) { _ Patches currently in -mm which might be from ziy(a)nvidia.com are mm-huge_memory-add-split_huge_page_to_order.patch mm-memory-failure-improve-large-block-size-folio-handling.patch mm-huge_memory-fix-kernel-doc-comments-for-folio_split-and-related.patch mm-huge_memory-fix-kernel-doc-comments-for-folio_split-and-related-fix.patch mm-huge_memory-fix-kernel-doc-comments-for-folio_split-and-related-fix-2.patch migrate-optimise-alloc_migration_target-fix.patch

7 hours, 42 minutes

1
0
0 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

Linux-stable-mirror