- 06 Aug, 2021 11 commits
-
-
Dave Jiang authored
The list lock is never acquired in interrupt context. Therefore there is no need to disable interrupts. Remove interrupt flags for lock operations. Reviewed-by:
Dan Williams <dan.j.williams@intel.com> Signed-off-by:
Dave Jiang <dave.jiang@intel.com> Link: https://lore.kernel.org/r/162826417450.3454650.3733188117742416238.stgit@djiang5-desk3.ch.intel.comSigned-off-by:
Vinod Koul <vkoul@kernel.org>
-
Andy Shevchenko authored
While IRQ test agaist the returned variable in practice is a good enough there is still a room for theoretical mistake in case the vIRQ of the device contains the same error code that acpi_register_gsi() may return. Due to this, check for error code separately from matching the vIRQs. Besides that, append documentation to tell why acpi_gsi_to_irq() can't be used and we call acpi_register_gsi() instead. Signed-off-by:
Andy Shevchenko <andriy.shevchenko@linux.intel.com> Link: https://lore.kernel.org/r/20210802175532.54311-1-andriy.shevchenko@linux.intel.comSigned-off-by:
-
Andy Shevchenko authored
Currently the CRST parsing relies on the fact that on most of x86 devices the IRQ mapping is 1:1 with Linux vIRQ. However, it may be not true for some. Fix this by converting GSI to Linux vIRQ before checking it. Fixes: ee8209fd ("dma: acpi-dma: parse CSRT to extract additional resources") Signed-off-by:
-
Dave Jiang authored
The interrupt thread handler currently loops forever to process outstanding completions. This causes either an "irq X: nobody cared" kernel splat or the NMI watchdog kicks in due to running too long in the function. The irq thread handler is expected to run again after exiting if there are interrupts fired while the thread handler is running. So the handler code can process all the completed I/O in a single pass and exit without losing the follow on completed I/O. Reviewed-by:
-
Dave Jiang authored
The block on fault flag is not cleared when we disable or reset wq. This causes it to remain set if the user does not clear it on the next configuration load. Add clear of flag in dxd_wq_disable_cleanup() routine. Fixes: da32b28c ("dmaengine: idxd: cleanup workqueue config after disabling") Signed-off-by:
-
Dave Jiang authored
The device general capability has a bit that indicate whether 'block on fault' is supported. Add check to wq sysfs knob to check if cap exists before allowing user to toggle. Signed-off-by:
-
Marek Vasut authored
Various DMA users call the dmaengine_slave_config() and expect it to succeed, but that can only succeed if .device_config is implemented. Add empty device_config function rather than patching all the places which use dmaengine_slave_config(). Signed-off-by:
Marek Vasut <marex@denx.de> Cc: Akinobu Mita <akinobu.mita@gmail.com> Cc: Kedareswara rao Appana <appana.durga.rao@xilinx.com> Cc: Michal Simek <monstr@monstr.eu> Cc: Vinod Koul <vkoul@kernel.org> Link: https://lore.kernel.org/r/20210804195140.61396-1-marex@denx.deSigned-off-by:
-
Andy Shevchenko authored
Since we converted internal data types to match DT, there is no need to have an intermediate conversion layer, hence drop a few conditionals and for loops for good. Signed-off-by:
Serge Semin <fancer.lancer@gmail.com> Tested-by:
-
Andy Shevchenko authored
u32 is a type that is used for properties retrieval from DT. With the type change it allows to clean up properties reading routine. While at it, order the fields in way how they are parsed. Signed-off-by:
-
Andy Shevchenko authored
Users are a bit frightened of the harmless message that tells that DT is missed on ACPI-based platforms. Remove it for good, it will simplify the future conversion to fwnode and device property APIs. Fixes: a9ddb575 ("dmaengine: dw_dmac: Enhance device tree support") Depends-on: f5e84eae ("dmaengine: dw: platform: Split OF helpers to separate module") BugLink: https://bugzilla.kernel.org/show_bug.cgi?id=199379Signed-off-by:
-
Nathan Chancellor authored
status is no longer used within this block: drivers/dma/idxd/irq.c:255:6: warning: unused variable 'status' [-Wunused-variable] u8 status = desc->completion->status & DSA_COMP_STATUS_MASK; ^ 1 warning generated. Fixes: b60bb6e2 ("dmaengine: idxd: fix abort status check") Signed-off-by:Nathan Chancellor <nathan@kernel.org> Acked-by:
-
- 02 Aug, 2021 9 commits
-
-
Christophe JAILLET authored
'device_driver_attach()' dereferences its first argument (i.e. 'alt_drv') so it must not be NULL. Simplify the error handling logic about NULL 'alt_drv' in order to be more robust and future-proof. Fixes: 568b2126 ("dmaengine: idxd: fix uninit var for alt_drv") Fixes: 6e7f3ee9 ("dmaengine: idxd: move dsa_drv support to compatible mode") Signed-off-by:
Christophe JAILLET <christophe.jaillet@wanadoo.fr> Acked-by:
-
Vinod Koul authored
-
Juergen Borleis authored
Commit dea7a9fb ("dmaengine: imx-dma: remove dma_slave_config direction usage") changes the method from a "configuration when called" to an "configuration when used". Due to this, only the cyclic DMA type gets configured correctly, while the generic DMA type is left non-configured. Without this additional call, the struct imxdma_channel::word_size member is stuck at DMA_SLAVE_BUSWIDTH_UNDEFINED and imxdma_prep_slave_sg() always returns NULL. Signed-off-by:
Juergen Borleis <jbe@pengutronix.de> Fixes: dea7a9fb ("dmaengine: imx-dma: remove dma_slave_config direction usage") Link: https://lore.kernel.org/r/20210729071821.9857-1-jbe@pengutronix.deSigned-off-by:
-
Juergen Borleis authored
Commit dea7a9fb dmaengine: imx-dma: remove dma_slave_config direction usage changes the method from a "configuration when called" to an "configuration when used". Due to this, only the cyclic DMA type gets configured correctly, while the generic DMA type is left non-configured. Without this additional call, the struct imxdma_channel::word_size member is stuck at DMA_SLAVE_BUSWIDTH_UNDEFINED and imxdma_prep_slave_sg() always returns NULL. Signed-off-by:
-
Alexander Sverdlin authored
Use clk_prepare_enable()/clk_disable_unprepare() in preparation for switch to Common Clock Framework, otherwise the following is visible: WARNING: CPU: 0 PID: 1 at drivers/clk/clk.c:1011 clk_core_enable+0x9c/0xbc Enabling unprepared m2p0 ... Hardware name: Cirrus Logic EDB9302 Evaluation Board ... clk_core_enable clk_core_enable_lock ep93xx_dma_alloc_chan_resources dma_chan_get find_candidate __dma_request_channel snd_dmaengine_pcm_request_channel dmaengine_pcm_new snd_soc_pcm_component_new soc_new_pcm snd_soc_bind_card edb93xx_probe ... ep93xx-i2s ep93xx-i2s: Missing dma channel for stream: 0 ep93xx-i2s ep93xx-i2s: ASoC: error at snd_soc_pcm_component_new on ep93xx-i2s: -22 edb93xx-audio edb93xx-audio: ASoC: can't create pcm CS4271 HiFi :-22 edb93xx-audio edb93xx-audio: snd_soc_register_card() failed: -22 edb93xx-audio: probe of edb93xx-audio failed with error -22 Signed-off-by:
Alexander Sverdlin <alexander.sverdlin@gmail.com> Signed-off-by:
Nikita Shubin <nikita.shubin@maquefel.me> Reviewed-by:
Linus Walleij <linus.walleij@linaro.org> Link: https://lore.kernel.org/r/20210726140001.24820-6-nikita.shubin@maquefel.meSigned-off-by:
-
Pandith N authored
Burst length, DMA HW capability set in dt-binding is now used in driver. Signed-off-by:
Pandith N <pandith.n@intel.com> Tested-by:
Pan Kris <kris.pan@intel.com> Link: https://lore.kernel.org/r/20210802055454.15192-4-pandith.n@intel.comSigned-off-by:
-
Pandith N authored
Added support for multiple DMA_MEM_TO_DEV, DMA_DEV_TO_MEM transfers in parallel. This is required for peripherals using DMA for transmit and receive operations at the same time. APB slot number needs to be programmed in channel hardware handshaking interface Signed-off-by:
-
Pandith N authored
Removed free slot check algorithm in dw_axi_dma_set_hw_channel. For 8 DMA channels, use respective handshake slot in DMA_HS_SEL APB register. For every channel, an dedicated slot is provided in hardware handshake register AXIDMA_CTRL_DMA_HS_SEL_n. Peripheral source number is programmed in respective channel slots. Signed-off-by:
-
Jordy Zomer authored
The usb_dmac_get_current_residue function used to take a signed integer as a pos parameter. The only callers of this function passes an unsigned integer to it. Therefore to make it obviously safe, let's just make this an unsgined integer as this is used in pointer arithmetics. Signed-off-by:
Jordy Zomer <jordy@pwning.systems> Reviewed-by:
Greg Kroah-Hartman <gregkh@linuxfoundation.org> Link: https://lore.kernel.org/r/20210731091939.510816-1-jordy@pwning.systemsSigned-off-by:
-
- 29 Jul, 2021 1 commit
-
-
Dave Jiang authored
Coverity static analysis of linux-next found issue. The check (status == IDXD_COMP_DESC_ABORT) is always false since status was previously masked with 0x7f and IDXD_COMP_DESC_ABORT is 0xff. Fixes: 6b4b87f2 ("dmaengine: idxd: fix submission race window") Reported-by:
Colin Ian King <colin.king@canonical.com> Signed-off-by:
-
- 28 Jul, 2021 19 commits
-
-
Dave Jiang authored
Enabling device and wq returns standard errno and that does not provide enough details to indicate what exactly failed. The hardware command status is only 8bits. Expand the command status to 32bits and use the upper 16 bits to define software errors to provide more details on the exact failure. Bit 31 will be used to indicate the error is software set as the driver is using some of the spec defined hardware error as well. Cc: Ramesh Thomas <ramesh.thomas@intel.com> Signed-off-by:
-
Dave Jiang authored
The device submission portal is on a 4k page and any of those 64bit aligned address on the page can be used for descriptor submission. By rotating the offset through the 4k range and prevent successive writes to the same MMIO address, performance improvement is observed through testing. Signed-off-by:
-
Dave Jiang authored
The sbitmap wait and allocate routine checks the index that is returned from sbitmap_queue_get(). It should be idxd >= 0 as 0 is also a valid index. This fixes issue where submission path hangs when WQ size is 1. Fixes: 0705107f ("dmaengine: idxd: move submission to sbitmap_queue") Signed-off-by:
-
Dave Jiang authored
0-day detected uninitialized alt_drv variable in the bind_store() function. The branch can be taken when device is not idxd device or wq 'struct device'. Init alt_drv to NULL. Fixes: 6e7f3ee9 ("dmaengine: idxd: move dsa_drv support to compatible mode") Reported-by:
kernel test robot <lkp@intel.com> Signed-off-by:
-
Dave Jiang authored
Set GRPCFG traffic class to value of 1 for best performance on current generation of accelerators. Also add override option to allow experimentation. Sysfs knobs are disabled for DSA/IAX gen1 devices. Signed-off-by:
-
Clément Léger authored
When using SCMI clocks, the clocks are probed later than subsys initcall level. This driver uses platform_driver_probe which is not compatible with deferred probing and won't be probed again later if probe function fails due to clocks not being available at that time. This patch replaces the use of platform_driver_probe with platform_driver_register which will allow probing the driver later again when clocks will be available. Signed-off-by:
Clément Léger <clement.leger@bootlin.com> Link: https://lore.kernel.org/r/20210728094607.50589-1-clement.leger@bootlin.comSigned-off-by:
-
Peter Ujfalusi authored
If the router_xlate can not find the controller in the available DMA devices then it should return with -EPORBE_DEFER in a same way as the of_dma_request_slave_channel() does. The issue can be reproduced if the event router is registered before the DMA controller itself and a driver would request for a channel before the controller is registered. In of_dma_request_slave_channel(): 1. of_dma_find_controller() would find the dma_router 2. ofdma->of_dma_xlate() would fail and returned NULL 3. -ENODEV is returned as error code with this patch we would return in this case the correct -EPROBE_DEFER and the client can try to request the channel later. Signed-off-by:
Peter Ujfalusi <peter.ujfalusi@gmail.com> Link: https://lore.kernel.org/r/20210717190021.21897-1-peter.ujfalusi@gmail.comSigned-off-by:
-
Zhang Qilong authored
use pm_runtime_resume_and_get() to replace pm_runtime_get_sync and pm_runtime_put_noidle. this change is just to simplify the code, there is no actual functional change. Signed-off-by:
Zhang Qilong <zhangqilong3@huawei.com> Link: https://lore.kernel.org/r/20210607064640.121394-4-zhangqilong3@huawei.comSigned-off-by:
-
Salah Triki authored
Fix the checkpatch.pl warning: "Prefer kstrto<type> to single variable sscanf". Signed-off-by:
Salah Triki <salah.triki@gmail.com> Link: https://lore.kernel.org/r/20210710165432.GA690401@pcSigned-off-by:
-
Amelie Delaunay authored
STM32 USART/UART is not managing correctly the default DMA REQ/ACK protocol leading to possibly lock the DMA stream. Default protocol consists in maintaining ACK signal up to the removal of REQuest and the transfer completion. In case of alternative REQ/ACK protocol, ACK de-assertion does not wait the removal of the REQuest, but only the transfer completion. This patch retrieves the need of the alternative protocol through the device tree, and sets the protocol accordingly. It also unwrap STM32_DMA_DIRECT_MODE_GET macro definition for consistency with new STM32_DMA_ALT_ACK_MODE_GET macro definition. Signed-off-by:
Amelie Delaunay <amelie.delaunay@foss.st.com> Link: https://lore.kernel.org/r/20210624093959.142265-3-amelie.delaunay@foss.st.comSigned-off-by:
-
Amelie Delaunay authored
Default REQ/ACK protocol consists in maintaining ACK signal up to the removal of REQuest and the transfer completion. In case of alternative REQ/ACK protocol, ACK de-assertion does not wait the removal of the REQuest, but only the transfer completion. Due to a possible DMA stream lock when transferring data to/from STM32 USART/UART, this new bindings allow to select this alternative protocol in device tree, especially for STM32 USART/UART nodes. Signed-off-by:
Rob Herring <robh@kernel.org> Link: https://lore.kernel.org/r/20210624093959.142265-2-amelie.delaunay@foss.st.comSigned-off-by:
-
Olivier Dautricourt authored
The response slave port can be disabled in some configuration [1] and csr + MSGDMA_CSR_RESP_FILL_LEVEL will be 0 even if transfer has suceeded. We have to only rely on the interrupts in that scenario. This was tested on cyclone V with the controller resp port disabled. [1] https://www.intel.com/content/www/us/en/programmable/documentation/sfo1400787952932.html 30.3.1.2 30.3.1.3 30.5.5 Fixes: https://forum.rocketboards.org/t/ip-msgdma-linux-driver/1919Signed-off-by:
Olivier Dautricourt <olivier.dautricourt@orolia.com> Reviewed-by:
Stefan Roese <sr@denx.de> Link: https://lore.kernel.org/r/8220756f2191ca08cb21702252d1f2d4f753a7f5.1623898678.git.olivier.dautricourt@orolia.comSigned-off-by:
-
Olivier Dautricourt authored
Response port is not required in some configuration of the IP core. Signed-off-by:
-
Christophe JAILLET authored
When using 'pcim_enable_device()', 'pci_alloc_irq_vectors()' is auto-magically a managed function. It is useless (but harmless) to record an action to explicitly call 'pci_free_irq_vectors()'. So keep things simple, comment why and how these resources are freed, axe some useless code and save some memory. Signed-off-by:
-
Baokun Li authored
Using list_move_tail() instead of list_del() + list_add_tail(). Reported-by:
Hulk Robot <hulkci@huawei.com> Signed-off-by:
Baokun Li <libaokun1@huawei.com> Link: https://lore.kernel.org/r/20210609071349.1336853-1-libaokun1@huawei.comSigned-off-by:
-
Baokun Li authored
Using list_move_tail() instead of list_del() + list_add_tail(). Reported-by:
-
Baokun Li authored
Using list_move_tail() instead of list_del() + list_add_tail(). Reported-by:
Radhey Shyam Pandey <radhey.shyam.pandey@xilinx.com> Link: https://lore.kernel.org/r/20210608030905.2818831-1-libaokun1@huawei.comSigned-off-by:
-
Zhang Qilong authored
pm_runtime_get_sync will increment pm usage counter even it failed. Forgetting to putting operation will result in reference leak here. We fix it by replacing it with pm_runtime_resume_and_get to keep usage counter balanced. Fixes: 4f3ceca2 ("dmaengine: stm32-dmamux: Add PM Runtime support") Signed-off-by:
-
Zhang Qilong authored
pm_runtime_get_sync will increment pm usage counter even it failed. Forgetting to putting operation will result in reference leak here. We fix it by replacing it with pm_runtime_resume_and_get to keep usage counter balanced. Fixes: 48bc73ba ("dmaengine: stm32-dma: Add PM Runtime support") Fixes: 05f8740a ("dmaengine: stm32-dma: add suspend/resume power management support") Signed-off-by:
-
