summaryrefslogtreecommitdiffstats
path: root/dma/ipe.c
Commit message (Collapse)AuthorAgeFilesLines
* Correct empty_detected location for gen3 IPEDMASuren A. Chilingaryan2015-11-191-2/+3
|
* Enforce 64-bit dma mask from IPEDMA if supportedSuren A. Chilingaryan2015-11-181-0/+12
|
* Support 64-bit addressing mode in IPEDMASuren A. Chilingaryan2015-11-181-37/+80
|
* Adjust IPEDMA register model to address new revisionSuren A. Chilingaryan2015-11-181-20/+48
|
* IPEDMA UpdateSuren A. Chilingaryan2015-11-171-18/+52
|
* Documentation updateSuren A. Chilingaryan2015-10-221-0/+1
|
* Fixes out-of-source builds and minor build issuesSuren A. Chilingaryan2015-09-011-0/+1
|
* In IPEDMA streaming mode put aside a single empty buffer to distinguish ↵Suren A. Chilingaryan2015-08-061-3/+14
| | | | between completely empty and full states of kernel ring buffer
* Detect if IPEDMA operates in streaming modeSuren A. Chilingaryan2015-08-061-49/+24
|
* Use 64-bit addressing in IPEDMA only for gen3 boards or if enforcedSuren A. Chilingaryan2015-07-241-7/+11
|
* Support streaming mode in IPEDMASuren A. Chilingaryan2015-06-121-1/+15
|
* Small clean up in IPEDMA driverSuren A. Chilingaryan2015-05-081-36/+2
|
* In IPEDMA use negotiated payload instead of maximum supported by deviceSuren A. Chilingaryan2015-05-071-1/+1
|
* Dynamicly set TLP(Payload) size in IPEDMASuren A. Chilingaryan2015-05-071-5/+19
|
* Add fields reporting consumed buffers and space to the dma_engine_status and ↵Suren A. Chilingaryan2015-05-061-92/+51
| | | | provide better ipedma benchmarking
* Support gen3 DMA engine and provide work-arround for hardware mishandling ↵Suren A. Chilingaryan2015-05-051-17/+57
| | | | last_descriptor_read register
* Provide an interface for logging debug messagesSuren A. Chilingaryan2015-04-301-9/+6
|
* Merge r241 fixing reporting of IPEDMA statusSuren A. Chilingaryan2015-04-201-30/+43
|
* Merge: disable complaining on empty_detected unless in debuggingSuren A. Chilingaryan2015-04-201-3/+6
|\
| * Only complain on empty_detected flag if debugging is oneSuren A. Chilingaryan2015-02-101-2/+2
| |
| * Better handle empty_detected flagSuren A. Chilingaryan2015-02-101-1/+4
| |
* | Big redign of model structuresSuren A. Chilingaryan2015-04-201-34/+49
|/
* Don't complain on empty_detected if flag was not set at all because no data ↵Suren A. Chilingaryan2015-02-061-1/+1
| | | | have arrived
* Use empty_detected flag to reduce timeoutSuren A. Chilingaryan2015-02-061-10/+22
|
* Increase DMA timeout to 100ms for ipedmaSuren A. Chilingaryan2015-02-061-1/+1
|
* Try to detect IPEDMA packets, but this probably will not work anyway as ↵Suren A. Chilingaryan2015-02-031-5/+23
| | | | often data is just streamed and pci is not make for infinite packets
* Support 64 bit mode, but disable due to problems in hardwareSuren A. Chilingaryan2015-02-031-2/+9
|
* Several fixes in IPEDMA driver and its benchmarking codeSuren A. Chilingaryan2015-02-031-6/+71
|
* Add delays to workaround problems in IPEDMA. Also provide other ↵Suren A. Chilingaryan2015-01-251-8/+21
| | | | modifications to reflect minor changes in hardware
* Initial implementation of IPEDMA, dummy driver for KAPTURE, start of API changesSuren A. Chilingaryan2014-12-191-0/+412