async_tx/dmaengine fixes and new drivers for 2.6.29-rc3
Dan Williams
Acked-by Sascha Hauer Acked-by Sascha Hauer Acked-by Timur Tabi Acked-by Li Yang Acked-by Timur Tabi Acked-by Li Yang Acked-by Haavard Skinnemoen Acked-by Maciej Sosnowski Tested-by Ira SnyderHi Linus, please pull from:
git://git.kernel.org/pub/scm/linux/kernel/git/djbw/async_tx.git fixes
...to receive:
Atsushi Nemoto (1):
dmatest: Use custom map/unmap for destination buffer
Dan Williams (3):
fsldma: use a valid 'device' for dma_pool_create
dmaengine: kill some dubious WARN_ONCEs
dmaengine: dma_issue_pending_all == nop when CONFIG_DMA_ENGINE=n
Guennadi Liakhovetski (3):
dmaengine: add async_tx_clear_ack() macro i.MX31: Image Processing Unit DMA and IRQ drivers i.MX31: framebuffer driver
Peter Korsgaard (2):
fsldma: check for NO_IRQ in fsl_dma_chan_remove()
fsldma: print correct IRQ on mpc83xx
Yuri Tikhonov (1):
dmaengine: fix dependency chaining
1/ Fix async_tx dependency chaining. This currently impacts an out of tree driver, but that driver is approaching mainline inclusion. 2/ Some fsldma fixups including one regression fix
3/ A fix for dmatest to catch errant dma-memcpy operations 4/ Core dmaengine fixups to address Bug #12511 on Rafael's list 5/ New drivers from Guennadi. The dma driver is a low regression risk because all of its channels are tagged DMA_PRIVATE which means they will not be picked up by NET_DMA or MD RAID. The framebuffer driver is included as a consumer of the new dma channels. It is a bit odd to see drivers/video/ in a dmaengine pull request but I included it because:
a) drivers/video/ does not appear to have an active maintainer
b) merging the dma driver without an in-tree user is pointless
c) i.MX31 maintainer, Sascha, has acked both drivers
arch/arm/plat-mxc/include/mach/ipu.h | 181 ++++ arch/arm/plat-mxc/include/mach/irqs.h | 10 +- arch/arm/plat-mxc/include/mach/mx3fb.h | 38 + drivers/dma/Kconfig | 19 + drivers/dma/Makefile | 1 + drivers/dma/dmaengine.c | 8 +- drivers/dma/dmatest.c | 35 +- drivers/dma/fsldma.c | 8 +- drivers/dma/ipu/Makefile | 1 + drivers/dma/ipu/ipu_idmac.c | 1740 ++++++++++++++++++++++++++++++++ drivers/dma/ipu/ipu_intern.h | 176 ++++ drivers/dma/ipu/ipu_irq.c | 413 ++++++++ drivers/video/Kconfig | 12 + drivers/video/Makefile | 1 + drivers/video/mx3fb.c | 1555 ++++++++++++++++++++++++++++ include/linux/dmaengine.h | 11 +- 16 files changed, 4194 insertions(+), 15 deletions(-) create mode 100644 arch/arm/plat-mxc/include/mach/ipu.h create mode 100644 arch/arm/plat-mxc/include/mach/mx3fb.h create mode 100644 drivers/dma/ipu/Makefile create mode 100644 drivers/dma/ipu/ipu_idmac.c create mode 100644 drivers/dma/ipu/ipu_intern.h create mode 100644 drivers/dma/ipu/ipu_irq.c create mode 100644 drivers/video/mx3fb.c
commit 86528da229a448577a8401a17c295883640d336c
Author: Guennadi Liakhovetski <[email protected]> Date: Wed Jan 21 10:32:34 2009 -0700
i.MX31: framebuffer driver
This is a framebuffer driver for i.MX31 SoCs. It only supports synchronous displays, vertical panning supported, no overlay support.
commit 5296b56d1b2000b60fb966be161c1f8fb629786b
Author: Guennadi Liakhovetski <[email protected]> Date: Mon Jan 19 15:36:21 2009 -0700
i.MX31: Image Processing Unit DMA and IRQ drivers
i.MX3x SoCs contain an Image Processing Unit, consisting of a Control Module (CM), Display Interface (DI), Synchronous Display Controller (SDC), Asynchronous Display Controller (ADC), Image Converter (IC), Post-Filter (PF), Camera Sensor Interface (CSI), and an Image DMA Controller (IDMAC). CM contains, among other blocks, an Interrupt Generator (IG) and a Clock and Reset Control Unit (CRCU). This driver serves IDMAC and IG. They are supported over dmaengine and irq-chip APIs respectively.
IDMAC is a specialised DMA controller, its DMA channels cannot be used for general-purpose operations, even though it might be possible to configure a memory-to-memory channel for memcpy operation. This driver will not work with generic dmaengine clients, clients, wishing to use it must use respective wrapper structures, they also must specify which channels they require, as channels are hard-wired to specific IPU functions.
commit ef560682a97491f62ef538931a4861b57d66c52c
Author: Guennadi Liakhovetski <[email protected]> Date: Mon Jan 19 15:36:21 2009 -0700
dmaengine: add async_tx_clear_ack() macro
To complete the DMA_CTRL_ACK handling API add a async_tx_clear_ack() macro.
commit c50331e8be32eaba5e1949f98c70d50b891262db
Author: Dan Williams <[email protected]> Date: Mon Jan 19 15:33:14 2009 -0700
dmaengine: dma_issue_pending_all == nop when CONFIG_DMA_ENGINE=n
The device list will always be empty in this configuration, so no need to walk the list.
commit 83436a0560e9ef8af2f0796264dde4bed1415359
Author: Dan Williams <[email protected]> Date: Mon Jan 19 14:39:10 2009 -0700
dmaengine: kill some dubious WARN_ONCEs
dma_find_channel and dma_issue_pending_all are good places to warn about improper api usage. However, warning correctly means synchronizing with dma_list_mutex, i.e. too much overhead for these fast-path calls.
commit 169d5f663759ec494aa74a552ce99486235e6e50
Author: Peter Korsgaard <[email protected]> Date: Wed Jan 14 22:33:31 2009 -0700
fsldma: print correct IRQ on mpc83xx
The mpc83xx variant uses a shared IRQ for all channels, so the individual channel nodes don't have an interrupt property. Fix the code to print the controller IRQ instead if there isn't any for the channel.
commit 6782dfe44acedf1e583d84e9e0d4f966d8e9befa
Author: Peter Korsgaard <[email protected]> Date: Wed Jan 14 22:32:58 2009 -0700
fsldma: check for NO_IRQ in fsl_dma_chan_remove()
There's no per-channel IRQ on mpc83xx, so only call free_irq if we have one.
commit d86be86e9aab221089d72399072511f13fe2a771
Author: Atsushi Nemoto <[email protected]> Date: Tue Jan 13 09:22:20 2009 -0700
dmatest: Use custom map/unmap for destination buffer
The dmatest driver should use DMA_BIDIRECTIONAL on the destination buffer to ensure that the poison values are written to RAM and not just written to cache and discarded.
commit 6527de6d6d25ebfae7c7572cb7a4ed768e2e20a5
Author: Dan Williams <[email protected]> Date: Mon Jan 12 15:18:34 2009 -0700
fsldma: use a valid 'device' for dma_pool_create
The dmaengine sysfs implementation was fixed to support proper lifetime rules which means that the current:
new_fsl_chan->dev = &new_fsl_chan->common.dev->device;
...retrieves a NULL pointer because new_fsl_chan->common.dev has not been allocated at this point. So, set new_fsl_chan->dev to a valid device.
commit dd59b8537f6cb53ab863fafad86a5828f1e889a2
Author: Yuri Tikhonov <[email protected]> Date: Mon Jan 12 15:17:20 2009 -0700
dmaengine: fix dependency chaining
In dmaengine we track the dependencies between the descriptors using the 'next' pointers of the structure. These pointers are set to NULL as soon as the corresponding descriptor has been submitted to the channel (in dma_run_dependencies()).
But, the first 'next' in chain is still remaining set, regardless the fact, that tx->next has been already submitted. This may lead to multiple submissions of the same descriptor. This patch fixes this.
Actually, some previous implementation of the xxx_run_dependencies() function already had this fix in place. The fdb..0eaf3 commit, beside the correct things, broke this.