1f57057 | Michal Simek | 02 March 2015, 13:39:43 UTC | usb: Fix compilation failure for CONFIG_USB_ZYNQ_PHY=y Fix "Merge tag 'v3.18' into zynqmp/master" (sha1: af1937c8095c189b9322d5a6d912b4bcafd68b19) because of mainline patch "usb: rename phy to usb_phy in HCD" (sha1: 3d46e73dfdb840f460e5b06416965d132570ec33) Signed-off-by: Michal Simek <michal.simek@xilinx.com> Tested-by: Nathan Rossi <nathan.rossi@xilinx.com> | 03 March 2015, 08:24:37 UTC |
1d2b4c4 | Punnaiah Choudary Kalluri | 20 February 2015, 16:24:45 UTC | nand: pl353: Configure the address cycles as per the device onfi parameter page Since the address cycles may vary for different parts, instead of hard coding the number address cycles required for read/write/erase operations, Calculate and configure them as per the address cycles value read from the device onfi parameter page. Signed-off-by: Punnaiah Choudary Kalluri <punnaia@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 26 February 2015, 10:39:52 UTC |
364259a | Hyun Kwon | 24 February 2015, 19:49:26 UTC | v4l: xilinx: vip: Document 'clk' of struct xvip_device Document description of 'clk' member of struct xvip_device. Signed-off-by: Hyun Kwon <hyun.kwon@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 25 February 2015, 07:51:06 UTC |
62c635d | Hyun Kwon | 24 February 2015, 19:49:25 UTC | v4l: xilinx: tpg: Add a goto label for error handling Calling media_entity_cleanup() when subdev == NULL results in kernel panic. Add a goto label for error handling with subdev == NULL. Signed-off-by: Hyun Kwon <hyun.kwon@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 25 February 2015, 07:50:42 UTC |
dd55646 | Michal Simek | 25 February 2015, 07:23:59 UTC | Merge branch 'zynqmp/master' into master Merge ZynqMP branch to main xilinx Linux repository and use only one repo from now. Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 25 February 2015, 07:23:59 UTC |
9cb4f12 | Anurag Kumar Vulisha | 24 February 2015, 14:54:42 UTC | mtd:spi-nor: Added support for issi flash parts Added support for issi flash parts is25lp128, is25lp064 & is25lp032 Signed-off-by: Anurag Kumar Vulisha <anuragku@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 25 February 2015, 07:23:27 UTC |
9320a8e | Anurag Kumar Vulisha | 24 February 2015, 14:54:41 UTC | mtd:spi-nor: Added Quad IO support Added Quad IO support Signed-off-by: Anurag Kumar Vulisha <anuragku@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 25 February 2015, 07:23:26 UTC |
1716024 | Punnaiah Choudary Kalluri | 20 February 2015, 17:31:38 UTC | net: emacps: Avoid Rx engine reset for 10 and 100mbps speeds The Rx engine lock down issue, occurs when there is a heavy rx traffic of smaller packets, is observed when the controller is operating at 1000 mbps speeds. But the driver workaround for this issue is added for all the speeds. The recent series of tests (in house and customer level) shows that the lock down issues was not observed for 10 and 100 speeds and the existing workaround causes unpredictable results. so, restring the workaround to 1000mbps speed only. Signed-off-by: Punnaiah Choudary Kalluri <punnaia@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 25 February 2015, 07:23:14 UTC |
0eb0e97 | Christian Kohn | 23 February 2015, 18:16:07 UTC | ARM: dts: zynq: remove duplicate leds entry Signed-off-by: Christian Kohn <christian.kohn@xilinx.com> Reviewed-by: Sören Brinkmann <soren.brinkmann@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 25 February 2015, 07:23:05 UTC |
c31fb6b | Subbaraya Sundeep Bhatta | 28 January 2015, 13:42:48 UTC | misc: xilinx_jesd204b: Add driver for Xilinx JESD204B IP. This patch adds driver support for JESD204B IP. Signed-off-by: Subbaraya Sundeep Bhatta <sbhatta@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 19 February 2015, 08:19:50 UTC |
0dd1a95 | Michal Simek | 22 January 2015, 12:46:33 UTC | xen: Add support for earlycon This is early version how to add earlycon support for xen. Just pass earlycon=xen on command line to enable it. Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 29 January 2015, 12:01:21 UTC |
b1d65e9 | Soren Brinkmann | 27 January 2015, 19:05:27 UTC | clk: zynq: Force CPU_2X clock to be ungated The CPU_2X clock does not have a classical in-kernel user, but is, amongst other things, required for OCM and debug access. Make sure this clock is not mistakenly disabled during boot up by enabling it in the platform's clock driver. Cc: stable@vger.kernel.org # 3.11+ Fixes: 0ee52b157b8e 'clk: zynq: Add clock controller driver' Signed-off-by: Soren Brinkmann <soren.brinkmann@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 29 January 2015, 06:04:56 UTC |
16b6c4c | Michal Simek | 20 January 2015, 12:44:16 UTC | i2c: cadence: Sync with mainline No functional changes. Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 22 January 2015, 09:15:23 UTC |
bf5e3c8 | Harini Katakam | 13 January 2015, 18:34:59 UTC | i2c: cadence: Check for errata condition involving master receive Cadence I2C controller has the following bugs: - completion indication is not given to the driver at the end of a read/receive transfer with HOLD bit set. - Invalid read transaction are generated on the bus when HW timeout condition occurs with HOLD bit set. As a result of the above, if a set of messages to be transferred with repeated start includes any message following a read message, completion is never indicated and timeout occurs. Hence a check is implemented to return -EOPNOTSUPP for such sequences. Signed-off-by: Harini Katakam <harinik@xilinx.com> Signed-off-by: Vishnu Motghare <vishnum@xilinx.com> [wsa: fixed some whitespaces] Signed-off-by: Wolfram Sang <wsa@the-dreams.de> | 22 January 2015, 09:15:22 UTC |
149bb3d | Michal Simek | 19 January 2015, 08:08:55 UTC | ARM: zynq: Fix tracing support for SMP v3.17 was adding tracepoint support which hasn't been properly resolved by merge "Merge tag 'v3.17' into master-next" (sha1: 9c270b33bef81466bf04284370df07cbd58ff111) Testing sequence: mount -t debugfs debufs /sys/kernel/debug/ cat /sys/kernel/debug/tracing/printk_formats Signed-off-by: Daniel Ladd <daniel.m.ladd@navy.mil> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 22 January 2015, 09:15:22 UTC |
59bf2a9 | Michal Simek | 16 January 2015, 15:13:11 UTC | ARM: zynq: Remove PCIe driver from mach-zynq Use mainline pcie driver instead (drivers/pci/host/pcie-xilinx.c). Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 22 January 2015, 09:15:19 UTC |
c22614b | Daniel M. Ladd | 21 January 2015, 23:05:48 UTC | ARM: zynq: pci: Modify xaxipcie.c read/write config to 32bit only Some PCIe cards only support 32bit access to configuration space and will return bogus values using 16 and 8 bit read/write accessors. Modified arch/arm/mach-zynq/xaxipcie.c so the read_config and write_config functions only use 32bits. Modified arch/arm/mach-zynq/xaxipcie-msi.c to correctly keep track of which MSI-X interrupts are allocated and deallocated. Signed-off-by: Daniel M. Ladd <daniel.m.ladd@navy.mil> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 22 January 2015, 09:12:14 UTC |
2f752af | Punnaiah Choudary Kalluri | 20 January 2015, 09:58:36 UTC | net: macb: Correct the MID field length value The latest spec "I-IPA01-0266-USR Rev 10" limit the MID field length to 12 bit value. For previous versions it is 16 bit value. This change will not break the backward compatibility as the latest ID value is 7 and with in the 12 bit value limit. Signed-off-by: Punnaiah Choudary Kalluri <punnaia@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 20 January 2015, 09:59:53 UTC |
837829e | Michal Simek | 16 January 2015, 15:09:44 UTC | ARM: zynq: Use architecture custom gpio.h Let's try to use what others are using. Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 19 January 2015, 12:04:03 UTC |
b271247 | Michal Simek | 16 January 2015, 14:37:20 UTC | remoteproc: mb: Simplify error path Checkpatch warning: WARNING: else is not generally useful after a break or return + return 0; + } else { Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 19 January 2015, 12:04:03 UTC |
5ee379e | Michal Simek | 16 January 2015, 14:36:37 UTC | remoteproc: mb: Remove whitespaces in front of tab Trivial change. Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 19 January 2015, 12:04:03 UTC |
391bfcb | Michal Simek | 16 January 2015, 14:32:46 UTC | i2c: cadence: Sync comments with mainline driver No functional changes. Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 19 January 2015, 12:04:03 UTC |
d28c534 | Michal Simek | 16 January 2015, 13:01:35 UTC | remoteproc: xilinx: Remove .owner field owner field is filled by module_platform_driver macro. Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 19 January 2015, 12:04:02 UTC |
d6aae49 | Michal Simek | 16 January 2015, 12:46:02 UTC | uio: xilinx_apm: Remove .owner field It is automatically filled by module_platform_driver macro. Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 19 January 2015, 12:04:02 UTC |
7ebd62d | Michal Simek | 16 January 2015, 12:43:31 UTC | Revert "microblaze: UIO setup compatible property" This reverts commit 3976b667483766aec3174fa01cdf2b668c4de19f. This patch removes "generic-uio" compatible string from the driver. This hack is in our tree for a long time and it is time to fix it. For ensuring the same behaviour please add uio_pdrv_genirq.of_id="generic-uio" to bootargs. Feel free to use different compatible string for your purpose. Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 19 January 2015, 12:04:02 UTC |
6fea4f8 | Michal Simek | 15 January 2015, 14:41:38 UTC | Merge tag 'v3.18' into master Linux 3.18 - Sync zynq defconfig - Remove unneeded Kconfig entries for Zynq - Use savedefconfig for ppc defconfigs - Use kstrtoul instead of strict_strtoul in xilinx_devcfg - Sync i2c-cadence.c with mainline - Remove of_irq.h from xilinx_emaclite.c - Sync SPI Kconfig with mainline - Sync USB gadget Kconfig with mainline - Remove CONFIG_CONTRUCTORS_NAME because of mainline arm gcov - Fix xilinx_drm_dp.c "drm: Remove old defines for vswing and pre-emph values" (sha1: 26db63e3cb7af431707cd0467c7bcfefad25aa39) Signed-off-by: Michal Simek <michal.simek@xilinx.com> Conflicts: arch/arm/boot/dts/zynq-7000.dtsi arch/arm/boot/dts/zynq-zed.dts arch/arm/mach-zynq/Makefile arch/arm/mach-zynq/common.c arch/arm/mach-zynq/common.h arch/arm/mach-zynq/hotplug.c arch/arm/mach-zynq/platsmp.c arch/arm/mach-zynq/pm.c arch/arm/mach-zynq/slcr.c arch/powerpc/platforms/40x/virtex.c drivers/edac/Makefile drivers/mtd/spi-nor/spi-nor.c drivers/net/ethernet/xilinx/xilinx_axienet_main.c drivers/tty/serial/8250/8250_core.c drivers/usb/gadget/udc/Makefile drivers/watchdog/cadence_wdt.c kernel/gcov/Kconfig | 19 January 2015, 12:03:45 UTC |
57ba60a | Michal Simek | 16 January 2015, 09:56:06 UTC | Revert "drivers: misc: eeprom: at24: Add support for Microchip 24LC64 chip" This reverts commit ba2117abd421be9404176d867207390aba69a2b3. No reason to have one more name in MODULE_DEVICE_TABLE(). Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 16 January 2015, 09:56:14 UTC |
af1937c | Michal Simek | 16 January 2015, 08:53:48 UTC | Merge tag 'v3.18' into zynqmp/master Linux 3.18 No issues with merge. Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 16 January 2015, 08:53:48 UTC |
306c70c | Michal Simek | 16 January 2015, 08:48:20 UTC | tty/serial: cadence: Fix coding style This should be the part of: "tty/serial: cadence: Add generic earlycon support" (sha1: e0ddfb701d76db3b86c24d08c2b689f204ea5570) Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 16 January 2015, 08:50:06 UTC |
951e592 | Michal Simek | 15 January 2015, 13:40:15 UTC | ARM: zynq: Sync kernel-doc in platsmp with mainline No functional changes. Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 15 January 2015, 13:40:15 UTC |
f467c62 | Michal Simek | 29 August 2014, 14:08:58 UTC | ARM: zynq: Remove useless L2C AUX setting AUX setting has no effect that's why remove it. Warning log: L2C: platform provided aux values match the hardware, so have no effect. Please remove them. Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 15 January 2015, 13:34:54 UTC |
1b42567 | Michal Simek | 15 January 2015, 13:34:19 UTC | ARM: zynq: Move zynq_init_late to the same possition as in mainline No functional chagnes. Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 15 January 2015, 13:34:19 UTC |
73a44f8 | Michal Simek | 15 January 2015, 13:30:43 UTC | ARM: zynq: Clean comments in zynq_memory_init() Sync with mainline version. Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 15 January 2015, 13:30:43 UTC |
b3332cf | Michal Simek | 15 January 2015, 13:28:53 UTC | ARM: zynq: Sync zynq_pm_late_init() with mainline Function is returning int. Also add kernel-doc. Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 15 January 2015, 13:28:53 UTC |
4228152 | Soren Brinkmann | 20 August 2014, 15:56:59 UTC | ARM: zynq: DT: Sync ethernet phys with mainline. Only changes in name. Based on mainline patch: "ARM: zynq: DT: Add Ethernet phys" (sha1: f62f40475137c905d8e4bcbe9870b4d7e4fed91f) Also fix cc108, zc770-xm010, zc770-xm013 Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 15 January 2015, 13:03:44 UTC |
370e87e | Harini Katakam | 22 August 2014, 09:28:01 UTC | watchdog: Sync cadence driver with mainline version "devicetree: Add Cadence WDT devicetree bindings documentation" (sha1: 191891c0378f44aec8e06e889a08d0b76fe6c5cb) and "watchdog: Add Cadence WDT driver" (sha1: 58bf016426594e5370e7e7059698a278294db997) - DT binding is align with mainline driver - Use the same Kconfig entry as in mainline - Use the same position in Makefile - Remove .owner property from the driver Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 15 January 2015, 12:56:26 UTC |
c679433 | Michal Simek | 15 January 2015, 07:42:42 UTC | ARM: zynq: Fix ARCH_NR_GPIOS for Zynq "Merge tag 'v3.17' into master-next" (sha1: 9c270b33bef81466bf04284370df07cbd58ff111) accidentially remove ARCH_NR_GPIOS setup for Zynq. Reported-by: John Linn <linnj@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 15 January 2015, 12:56:25 UTC |
d369cd2 | Michal Simek | 05 January 2015, 12:43:32 UTC | ARM: zynq: Synchronize DDRC controller with mainline Use mainline compatible string. Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 15 January 2015, 12:56:20 UTC |
2046aa5 | Soren Brinkmann | 13 January 2015, 17:15:52 UTC | zynqmp: Rename pm.c -> sleep.c Free the name pm.c for the implementation for the PM firmware interface. Signed-off-by: Soren Brinkmann <soren.brinkmann@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 14 January 2015, 06:44:22 UTC |
da2d296 | Nicolae Rosia | 05 January 2015, 16:12:13 UTC | ARM: zynq: xilinx_axidma: fix race condition in dma_intr_handler On SMP platform when in IRQ, another CPU may be executing different parts of the driver, which may read/write registers. Fix this by acquiring the lock before read/write-ing the DMA registers. Signed-off-by: Nicolae Rosia <nicolae.rosia@certsign.ro> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 08 January 2015, 07:53:17 UTC |
4de67c2 | Nicolae Rosia | 05 January 2015, 16:12:10 UTC | ARM: zynq: xilinx_axidma: move locking out of xilinx_dma_start_transfer Signed-off-by: Nicolae Rosia <nicolae.rosia@certsign.ro> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 08 January 2015, 07:53:17 UTC |
32db567 | Nicolae Rosia | 05 January 2015, 16:12:06 UTC | ARM: zynq: xilinx_axidma: move locking out of xilinx_dma_update_completed_cookie Signed-off-by: Nicolae Rosia <nicolae.rosia@certsign.ro> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 08 January 2015, 07:53:16 UTC |
8271885 | Nicolae Rosia | 05 January 2015, 16:11:27 UTC | ARM: zynq: xilinx_axidma: refactor if/else in xilinx_dma_update_completed_cookie Refactor the if/else so it improves clarity Signed-off-by: Nicolae Rosia <nicolae.rosia@certsign.ro> Acked-by: Kedareswara rao Appana <appanad@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 08 January 2015, 07:53:15 UTC |
8fe69de | Nicolae Rosia | 05 January 2015, 16:11:23 UTC | ARM: zynq: xilinx_axidma: get rid of redundant variables in xilinx_dma_update_completed_cookie cookie and done variables only confuses things. Signed-off-by: Nicolae Rosia <nicolae.rosia@certsign.ro> Acked-by: Kedareswara rao Appana <appanad@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 08 January 2015, 07:53:14 UTC |
1b28f17 | Nicolae Rosia | 05 January 2015, 16:11:02 UTC | ARM: zynq: xilinx_axidma: fix race condition in xilinx_dma_tx_submit The error flag should be protected against race conditions as other member variables are. Signed-off-by: Nicolae Rosia <nicolae.rosia@certsign.ro> Acked-by: Kedareswara rao Appana <appanad@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 08 January 2015, 07:53:14 UTC |
b00d8a3 | Nicolae Rosia | 05 January 2015, 16:10:23 UTC | ARM: zynq: xilinx_axidma: make xilinx_dma_chan err variable bool err from struct xilinx_dma_chan is being used as a boolean variable. Make it bool. Signed-off-by: Nicolae Rosia <nicolae.rosia@certsign.ro> Acked-by: Kedareswara rao Appana <appanad@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 08 January 2015, 07:53:13 UTC |
d1d52cf | Nicolae Rosia | 05 January 2015, 16:10:17 UTC | ARM: zynq: xilinx_axidma: remove redundant register read from dma_intr_handler Since we enable all interrupts, there's no reason to mask any of them. Signed-off-by: Nicolae Rosia <nicolae.rosia@certsign.ro> Acked-by: Kedareswara rao Appana <appanad@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 08 January 2015, 07:53:12 UTC |
41b0300 | Nicolae Rosia | 05 January 2015, 16:10:10 UTC | ARM: zynq: xilinx_axidma: get rid of start_transfer function pointer of struct xilinx_dma_chan start_transfer always get assigned the same function, there's no reason to keep it. Signed-off-by: Nicolae Rosia <nicolae.rosia@certsign.ro> Acked-by: Kedareswara rao Appana <appanad@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 08 January 2015, 07:53:12 UTC |
1428cc3 | Nicolae Rosia | 05 January 2015, 16:10:04 UTC | ARM: zynq: xilinx_axidma: refactor xilinx_dma_chan_probe logic into if/else branches Signed-off-by: Nicolae Rosia <nicolae.rosia@certsign.ro> Acked-by: Kedareswara rao Appana <appanad@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 08 January 2015, 07:53:11 UTC |
2d806da | Nicolae Rosia | 05 January 2015, 16:09:54 UTC | ARM: zynq: xilinx_axidma: fix xilinx_dma_chan_probe handling of invalid compatible string Signed-off-by: Nicolae Rosia <nicolae.rosia@certsign.ro> Acked-by: Kedareswara rao Appana <appanad@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 08 January 2015, 07:53:10 UTC |
4a00b2c | Nicolae Rosia | 05 January 2015, 16:09:49 UTC | ARM: zynq: xilinx_axidma: fix race condition in xilinx_dma_start_transfer The error flag should be protected against race conditions as other member variables are. Signed-off-by: Nicolae Rosia <nicolae.rosia@certsign.ro> Acked-by: Kedareswara rao Appana <appanad@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 08 January 2015, 07:53:10 UTC |
28049a5 | Nicolae Rosia | 05 January 2015, 16:09:45 UTC | ARM: zynq: xilinx_axidma: refactor code in xilinx_dma_start_transfer Improve clarity by refactoring the code into an if/else clause. Also remove redundant comment. Signed-off-by: Nicolae Rosia <nicolae.rosia@certsign.ro> Acked-by: Kedareswara rao Appana <appanad@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 08 January 2015, 07:53:09 UTC |
b2fe572 | Nicolae Rosia | 05 January 2015, 16:07:13 UTC | ARM: zynq: xilinx_axidma: remove superfluous check in simple mode The channel's state is checked twice in simple mode. Fix this. Signed-off-by: Nicolae Rosia <nicolae.rosia@certsign.ro> Acked-by: Kedareswara rao Appana <appanad@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 08 January 2015, 07:53:08 UTC |
76408eb | Nicolae Rosia | 05 January 2015, 16:07:12 UTC | ARM: zynq: xilinx_axidma: remove superfluous dma_halt In simple mode, the DMA Engine is halted twice. Fix this. Signed-off-by: Nicolae Rosia <nicolae.rosia@certsign.ro> Acked-by: Kedareswara rao Appana <appanad@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 08 January 2015, 07:53:08 UTC |
00ff5b2 | Nicolae Rosia | 05 January 2015, 16:07:11 UTC | ARM: zynq: xilinx_axidma: set default value for DMA Channel Coalesce Signed-off-by: Nicolae Rosia <nicolae.rosia@certsign.ro> Reported-by: Kedareswara rao Appana <appanad@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 08 January 2015, 07:53:07 UTC |
5b4bae9 | Nicolae Rosia | 05 January 2015, 16:07:10 UTC | ARM: zynq: xilinx_axidma: move channel configuration logic into xilinx_dma_chan_config xilinx_dma_chan_config configures a channel's IRQThreshold and IRQDelay. Signed-off-by: Nicolae Rosia <nicolae.rosia@certsign.ro> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 08 January 2015, 07:53:06 UTC |
1095858 | Nicolae Rosia | 05 January 2015, 16:07:09 UTC | ARM: zynq: xilinx_axidma: use list_last_entry instead of container_of Signed-off-by: Nicolae Rosia <nicolae.rosia@certsign.ro> Acked-by: Kedareswara rao Appana <appanad@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 08 January 2015, 07:53:06 UTC |
8e7ddd6 | Nicolae Rosia | 05 January 2015, 16:07:08 UTC | ARM: zynq: xilinx_axidma: Fix NULL dereference bug in xilinx_dma_free_channels If only one channel is defined, xilinx_dma_free_channels will dereference a NULL pointer when iterating over the channels. Signed-off-by: Nicolae Rosia <nicolae.rosia@certsign.ro> Acked-by: Kedareswara rao Appana <appanad@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 08 January 2015, 07:53:05 UTC |
0a8718a | Nicolae Rosia | 05 January 2015, 16:02:37 UTC | ARM: zynq: xilinx_axidma: remove superfluous caused by xilinx_dma_alloc_descriptor failure xilinx_dma_alloc_descriptor prints a message on failure, but so does xilinx_dma_prep_slave_sg which calls it. Remove the debug message from xilinx_dma_alloc_descriptor. Signed-off-by: Nicolae Rosia <nicolae.rosia@certsign.ro> Acked-by: Kedareswara rao Appana <appanad@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 08 January 2015, 07:53:05 UTC |
6ac0b33 | Nicolae Rosia | 05 January 2015, 16:02:36 UTC | ARM: zynq: xilinx_axidma: check if there are channels defined Since it makes no sense to register a DMA Engine without any channels, add a check for this. Signed-off-by: Nicolae Rosia <nicolae.rosia@certsign.ro> Acked-by: Kedareswara rao Appana <appanad@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 08 January 2015, 07:53:04 UTC |
b9ed63f | Nicolae Rosia | 05 January 2015, 16:02:35 UTC | ARM: zynq: xilinx_axidma: get rid of useless and confusing variables update_cookie and to_transfer are being initialized with the same value and assigned with a different value under the same condition. Get rid of them and move their action when testing the condition Signed-off-by: Nicolae Rosia <nicolae.rosia@certsign.ro> Acked-by: Kedareswara rao Appana <appanad@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 08 January 2015, 07:53:03 UTC |
72c975d | Nicolae Rosia | 05 January 2015, 16:02:34 UTC | ARM: zynq: xilinx_axidma: Update Device Tree Documentation xlnx,sg-include-stscntrl-strm property is not used. Signed-off-by: Nicolae Rosia <nicolae.rosia@certsign.ro> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 08 January 2015, 07:53:03 UTC |
c239960 | Nicolae Rosia | 05 January 2015, 16:02:33 UTC | ARM: zynq: xilinx_axidma: get rid of useless 'feature' variable The structures xilinx_dma_chan and xilinx_dma_device have a variable which encode 2 capabilities: - device has Scatter Gatter - Control Stream and Status Stream. For Scatter Gatter, there already was a variable for this, called has_sg and Control Stream and Status Stream is not being used. With this ocassion, we also get rid of the defines which encode capabilities. Signed-off-by: Nicolae Rosia <nicolae.rosia@certsign.ro> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 08 January 2015, 07:53:02 UTC |
2e0eef9 | Nicolae Rosia | 02 December 2014, 19:33:34 UTC | arm: zynq: xilinx_axidma: fix race condition in xilinx_chan_desc_cleanup xilinx_chan_desc_cleanup uses a spinlock to prevent multiple threads from accessing the same descriptor. The problem arrises when there are at least two descriptors in active_list which require callback and are both done because before callback, the spinlock is released. Suppose the following: - Thread 1 acquires the lock, enters the loop with desc pointing to A and _desc pointing to B. Deletes A from the list, releases the lock, and makes the callback. - Thread 2 acquires the lock, enters the loop with desc pointing to B and _desc pointing to next element. Deletes B from the list, releases the lock. - Thread 1 reacquires the lock, does the next iteration of the loop with desc pointing to B (which was deleted by Thread 2) - race condition. The proposed solution is to iterate while there are elements in the list, get the head descriptor and remove it from list before callback. Signed-off-by: Nicolae Rosia <nicolae.rosia@certsign.ro> Reported-by: Joseph Zavodny <joseph.zavodny@jhuapl.edu> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 08 January 2015, 07:39:45 UTC |
83e50a2 | Nicolae Rosia | 02 December 2014, 19:33:33 UTC | arm: zynq: xilinx_axidma: replace pr_debug with dev_err Signed-off-by: Nicolae Rosia <nicolae.rosia@certsign.ro> Acked-by: Soren Brinkmann <soren.brinkmann@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 08 January 2015, 07:39:45 UTC |
59d81e7 | Nicolae Rosia | 02 December 2014, 19:33:32 UTC | arm: zynq: xilinx_axidma: remove superfluous pr_info message This message was present only on Simple DMA transfer path. Signed-off-by: Nicolae Rosia <nicolae.rosia@certsign.ro> Acked-by: Soren Brinkmann <soren.brinkmann@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 08 January 2015, 07:39:45 UTC |
f6fde7c | Nicolae Rosia | 02 December 2014, 19:33:31 UTC | arm: zynq: xilinx_axidma: replace while loop with do while Signed-off-by: Nicolae Rosia <nicolae.rosia@certsign.ro> Acked-by: Soren Brinkmann <soren.brinkmann@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 08 January 2015, 07:39:45 UTC |
fd30175 | Nicolae Rosia | 02 December 2014, 19:33:30 UTC | arm: zynq: xilinx_axidma: dma_start and dma_stop wait for the wrong condition According to the LogiCORE IP AXI DMA v7.1, the halted bit in the DMA Status Register deasserts to 0 when the DMA engine begins operations and the Halted bit is set after some time the RunStop bit is cleared. Signed-off-by: Nicolae Rosia <nicolae.rosia@certsign.ro> Reported-by: Joseph Zavodny <joseph.zavodny@jhuapl.edu> Acked-by: Soren Brinkmann <soren.brinkmann@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 08 January 2015, 07:39:45 UTC |
334cf53 | Soren Brinkmann | 06 January 2015, 16:15:09 UTC | zynqmp: Add suspend support Minimal implementation of suspend ops. Not really functional, but basic suspend paths should be exercisible. Signed-off-by: Soren Brinkmann <soren.brinkmann@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 07 January 2015, 07:21:44 UTC |
de13611 | Appana Durga Kedareswara Rao | 23 December 2014, 12:37:55 UTC | net: xilinx: Remove unnecessary temac_property in the driver This property is no longer used in the code yet the code looks for it in the device tree. It does not cause an error if it's not in the tree. Signed-off-by: Kedareswara rao Appana <appanad@xilinx.com> Signed-off-by: David S. Miller <davem@davemloft.net> | 05 January 2015, 16:02:55 UTC |
eb3b729 | Michal Simek | 05 December 2014, 08:25:08 UTC | edac: zynq: Do not show ecc not enable as error Use dev_info instead. Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 05 January 2015, 12:14:36 UTC |
4fd7cd2 | Kedareswara rao Appana | 05 January 2015, 04:31:19 UTC | dma: Remove errors reported by coccicheck coccicheck errors reported by Michal: ./drivers/dma/xilinx/axidmatest.c:545:16-22: ERROR: reference preceded by free on line 540 ./drivers/dma/xilinx/vdmatest.c:535:16-22: ERROR: reference preceded by free on line 532 Don't add the thread to the list of threads for thread failed case. Reported-by: Michal Simek <michal.simek@xilinx.com> Signed-off-by: Kedareswara rao Appana <appanad@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 05 January 2015, 12:11:25 UTC |
7763fdb | Matthias Auchmann | 11 December 2014, 08:51:13 UTC | net: can: xilinx_can: Fixed ksoftirq CPU monopolization When using both the RXOK and the RXNEMP interrupt, when there were more than one receive messages in the FIFO, ksoftirqd started to go crazy and monopolize one CPU. The reason being is that RXOK just fires once even if there are multiple frames sitting in the RX FIFO, so the code didn't reflect that properly. Changed that to only use the RXNEMP interrupt and thereby got rid of the bug. Signed-off-by: Matthias Auchmann <m.auchmann@artech.at> Acked-by: Kedareswara rao Appana <appanad@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 05 January 2015, 12:11:24 UTC |
7fb5fe8 | Soren Brinkmann | 18 December 2014, 01:22:03 UTC | arm64: Add Kconfig entry for Xilinx Zynq MP SoCs Signed-off-by: Soren Brinkmann <soren.brinkmann@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 18 December 2014, 05:34:44 UTC |
467c055 | Mike Looijmans | 15 December 2014, 10:07:31 UTC | xilinx_emacps.c: Fix three random MAC address assignment issues When a random address has been assigned to the controller, the entry addr_assign_type in sysfs still claims the address has been assigned by other means. Fix this by replacing "random_ether_addr" by "eth_hw_addr_random", which also sets the NET_ADDR_RANDOM value. Use the more efficient "ether_addr_copy" instead of memcpy(), because the addr variable is u16 aligned. Fix the incorrect message "invalid address, use assigned" which is the very opposite of what is going to happen. Log "invalid address, use random" instead. Signed-off-by: Mike Looijmans <mike.looijmans@topic.nl> Acked-by: Punnaiah Choudary Kalluri <punnaia@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com> | 16 December 2014, 09:33:33 UTC |
82550e3 | Davor Joja | 10 December 2014, 09:23:59 UTC | gpu: drm: xylon: Updated logiCVC devicetree bool properties Removed "is-" string from the beginning of logiCVC devicetree bool properties. Signed-off-by: Davor Joja <davorjoja@logicbricks.com> Reviewed-by: Hyun Kwon <hyun.kwon@xilinx.com> Tested-by: Radhey Shyam Pandey <radheys@xilinx.com> Tested-by: Christian Kohn <christian.kohn@xilinx.com> | 15 December 2014, 06:19:15 UTC |
89cccd3 | Davor Joja | 10 December 2014, 08:37:13 UTC | gpu: drm: xylon: Removed CRTC control, position_x,y properties Removed CRTC control, position_x,y properties. Universal plane functionality exports CRTC primary plane in the same way as overlay plane to userspace application, and this gives ability to control primary plane enabling and position through standard DRM API. Signed-off-by: Davor Joja <davorjoja@logicbricks.com> Reviewed-by: Hyun Kwon <hyun.kwon@xilinx.com> Tested-by: Radhey Shyam Pandey <radheys@xilinx.com> Tested-by: Christian Kohn <christian.kohn@xilinx.com> | 15 December 2014, 06:19:15 UTC |
c0e6d29 | Davor Joja | 10 December 2014, 08:14:27 UTC | gpu: drm: xylon: Removed unneeded destroy functionality Removed unneeded plane and plane manager destroy function calls from crtc and plane driver, and functions itself. Removed destroy functions exports from all header files. Signed-off-by: Davor Joja <davorjoja@logicbricks.com> Reviewed-by: Hyun Kwon <hyun.kwon@xilinx.com> Tested-by: Radhey Shyam Pandey <radheys@xilinx.com> Tested-by: Christian Kohn <christian.kohn@xilinx.com> | 15 December 2014, 06:19:15 UTC |
cbdc2b1 | Davor Joja | 18 November 2014, 19:03:55 UTC | gpu: drm: xylon: Registered plane update Set crtc base by calling registered plane driver function. This removes direct plane driver function calls and removes functionality redundancy. Signed-off-by: Davor Joja <davorjoja@logicbricks.com> Reviewed-by: Hyun Kwon <hyun.kwon@xilinx.com> Tested-by: Radhey Shyam Pandey <radheys@xilinx.com> Tested-by: Christian Kohn <christian.kohn@xilinx.com> | 15 December 2014, 06:19:14 UTC |
6ea2c94 | Davor Joja | 18 November 2014, 18:58:46 UTC | gpu: drm: xylon: Cleanup plane driver Remove unneeded plane driver parameters and replace their usage with parameters provided in local driver or drm subsystem structures. Signed-off-by: Davor Joja <davorjoja@logicbricks.com> Reviewed-by: Hyun Kwon <hyun.kwon@xilinx.com> Tested-by: Radhey Shyam Pandey <radheys@xilinx.com> Tested-by: Christian Kohn <christian.kohn@xilinx.com> | 15 December 2014, 06:19:14 UTC |
0c49ff4 | Davor Joja | 18 November 2014, 18:41:51 UTC | gpu: drm: xylon: Added universal plane functionality support Added universal plane functionality support. Crtc private plane replaced with primary plane. Signed-off-by: Davor Joja <davorjoja@logicbricks.com> Reviewed-by: Hyun Kwon <hyun.kwon@xilinx.com> Tested-by: Radhey Shyam Pandey <radheys@xilinx.com> Tested-by: Christian Kohn <christian.kohn@xilinx.com> | 15 December 2014, 06:19:14 UTC |
d2206c2 | Davor Joja | 18 November 2014, 17:12:42 UTC | gpu: drm: xylon: Fixed memory registers access Fixed wrong functionality in access functions for memory registers available when hw registers are not readable. Signed-off-by: Davor Joja <davorjoja@logicbricks.com> Tested-by: Radhey Shyam Pandey <radheys@xilinx.com> Tested-by: Christian Kohn <christian.kohn@xilinx.com> | 15 December 2014, 06:19:14 UTC |
3425217 | Hyun Kwon | 11 December 2014, 22:32:34 UTC | drm: xilinx: encoder: Check device->driver pointer for i2c encoder slave When probe is deferred in i2c device driver, device->driver is set to NULL. So check if the device->driver pointer is NULL before accessing the i2c encoder slave driver. Reported-by: Rob Armstrong JR <ra@xilinx.com> Signed-off-by: Hyun Kwon <hyun.kwon@xilinx.com> Tested-by: Rob Armstrong JR <ra@xilinx.com> | 15 December 2014, 06:19:14 UTC |
980b2ad | Christian Kohn | 05 November 2014, 21:05:12 UTC | media: adv7604: select default input | 15 December 2014, 06:19:13 UTC |
8cde006 | Christian Kohn | 03 December 2014, 19:33:45 UTC | v4l: xilinx: hls: Rename compatible string to xlnx,v-hls This matches the compatible string values used by the DT generator. Signed-off-by: Christian Kohn <christian.kohn@xilinx.com> | 15 December 2014, 06:19:13 UTC |
f6e8a7e | Laurent Pinchart | 28 November 2014, 07:58:13 UTC | v4l: xilinx: remapper: Handle video core clock The video core clock is mandatory, manage it explicitly by requesting and enabling it at probe time and disabling it a removal time. Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> | 15 December 2014, 06:19:13 UTC |
c29cd9b | Laurent Pinchart | 27 November 2014, 02:03:15 UTC | v4l: xilinx: switch: Use xvip_init_resources() Replace the manual ioremap by a call to xvip_init_resources(), which also handles the functional clock. Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> | 15 December 2014, 06:19:13 UTC |
a5dabe7 | Laurent Pinchart | 27 November 2014, 02:03:15 UTC | v4l: xilinx: scaler: Use xvip_init_resources() Replace the manual ioremap by a call to xvip_init_resources(), which also handles the functional clock. Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> | 15 December 2014, 06:19:12 UTC |
1b82e54 | Laurent Pinchart | 27 November 2014, 02:03:15 UTC | v4l: xilinx: rgb2yuv: Use xvip_init_resources() Replace the manual ioremap by a call to xvip_init_resources(), which also handles the functional clock. Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> | 15 December 2014, 06:19:12 UTC |
5815500 | Laurent Pinchart | 27 November 2014, 02:03:15 UTC | v4l: xilinx: hls: Use xvip_init_resources() Replace the manual ioremap by a call to xvip_init_resources(), which also handles the functional clock. Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> | 15 December 2014, 06:19:12 UTC |
dd21099 | Laurent Pinchart | 27 November 2014, 02:03:15 UTC | v4l: xilinx: resample: Use xvip_init_resources() Replace the manual ioremap by a call to xvip_init_resources(), which also handles the functional clock. Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> | 15 December 2014, 06:19:12 UTC |
a23a699 | Laurent Pinchart | 27 November 2014, 02:03:15 UTC | v4l: xilinx: cfa: Use xvip_init_resources() Replace the manual ioremap by a call to xvip_init_resources(), which also handles the functional clock. Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> | 15 December 2014, 06:19:11 UTC |
a10bd6d | Laurent Pinchart | 27 November 2014, 02:03:15 UTC | v4l: xilinx: vtc: Use xvip_init_resources() Replace the manual ioremap by a call to xvip_init_resources(), which also handles the functional clock. Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> | 15 December 2014, 06:19:11 UTC |
bcb4312 | Laurent Pinchart | 27 November 2014, 02:03:15 UTC | v4l: xilinx: tpg: Use xvip_init_resources() Replace the manual ioremap by a call to xvip_init_resources(), which also handles the functional clock. Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> | 15 December 2014, 06:19:11 UTC |
8c7a713 | Laurent Pinchart | 27 November 2014, 02:02:18 UTC | v4l: xilinx: vip: Add resource init and cleanup helpers All video entities have at least one iomem resource and a functional clock. Add a function to remap the memory and retrieve and enable the clock. Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> | 15 December 2014, 06:19:11 UTC |
6678d9c | Laurent Pinchart | 28 November 2014, 08:35:57 UTC | v4l: xilinx: dma: Return buffers to vb2 when media pipeline start fails The start_streaming operation must return all queued buffers to vb2 in case of failure. This is already implemented except in case of media pipeline start failures. Fix it. Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> | 15 December 2014, 06:19:11 UTC |
348dd49 | Radhey Shyam Pandey | 18 November 2014, 14:13:14 UTC | v4l: xilinx: dma: Cleanup unused vdma configuration. Remove DMA_SLAVE_CONFIG call as setting vdma channel params 'reset, park and ext_fsync' is no longer required. However at a later point for configuring any vdma channel param (i.e non-cyclic mode) we should use xilinx_vdma_channel_set_config(). Signed-off-by: Radhey Shyam Pandey <radheys@xilinx.com> Reviewed-by: Hyun Kwon <hyun.kwon@xilinx.com> Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> | 15 December 2014, 06:19:10 UTC |
a467382 | Laurent Pinchart | 05 November 2014, 15:31:47 UTC | v4l: xilinx: dma: Don't hardcode colorspace Even though most applications don't care about the colorspace, its value matters and should be reported correctly by drivers. As userspace is responsible for propagating formats, and thus colorspaces, through the pipeline, just accept any colorspace value requested by userspace when trying or setting the format and validates at streamon time that the colorspace matches the one set up on the connected subdev. Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> | 15 December 2014, 06:19:10 UTC |
3421c74 | Laurent Pinchart | 05 November 2014, 15:14:25 UTC | v4l: xilinx: dma: Remove explicit trailing \0 after snprintf snprintf adds a trailing zero even when it truncates the printed chain. Don't add a trailing \0 manually. Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> | 15 December 2014, 06:19:10 UTC |
80ea963 | Laurent Pinchart | 02 November 2014, 14:17:51 UTC | v4l: xilinx: tpg: Fix typo Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> | 15 December 2014, 06:19:10 UTC |