driver-hashfast.c


Log

Author Commit Date CI Message
Con Kolivas edcec1f2 2014-06-21T11:30:10 Revert "Hashfast voltage support" This reverts commit ef97e80a14240c95a7d2966abe462b12f7d4bf03.
Con Kolivas a7b5762e 2014-06-21T11:30:00 Revert "Set hashfast voltage settings only when really needed" This reverts commit 7cd8076875e8fef978e32b077aaf6bbf06316944.
jaketri 7cd80768 2014-06-17T20:49:53 Set hashfast voltage settings only when really needed
jaketri ef97e80a 2014-06-16T13:39:44 Hashfast voltage support
Con Kolivas 1759ffc2 2014-06-16T10:27:03 Revert "Merge pull request #609 from jaketri/master" This reverts commit a8857bb27d6f14fb178bf71b956af864fd23af11, reversing changes made to 2f007cb685ab8c5728ece2acb269d6eec3e888e5.
Con Kolivas 0f05774e 2014-06-16T10:27:03 Revert "Merge pull request #609 from jaketri/master" This reverts commit a8857bb27d6f14fb178bf71b956af864fd23af11, reversing changes made to 2f007cb685ab8c5728ece2acb269d6eec3e888e5.
jaketri e57d7825 2014-06-10T15:35:02 Hashfast hack to add per die voltage and clock offset (part2) - enable code only for devices with 4 dies
jaketri 4aed9315 2014-06-10T01:20:56 Hashfast hack to add per die voltage and clock offset
ckolivas e1d98902 2014-06-05T09:46:03 Avoid dereference on getting API stats on partially initialised HFA instances
ckolivas 78f246fd 2014-05-07T12:14:24 Rationalise diffs stored in the work struct and document them to avoid further confusion
Con Kolivas 3a36d246 2014-05-04T19:50:05 Make hashfast reset counter rise on old instances when inheriting the value on new ones
Con Kolivas e5741146 2014-04-06T14:40:49 Avoid derefrence when calling statline before on hfa device during init sequence
Con Kolivas 88b7c0d0 2014-04-06T14:25:42 Calloc the info structures even on failed hfa reset to prevent later possible dereference
Con Kolivas cdd1787f 2014-04-06T14:15:01 Load all hfa devices based on identification alone and defer init sequence till mining thread init sequence to allow all devices to be recognised rapidly but each device initialisation not delay others
Con Kolivas 0d484fc9 2014-03-28T20:40:54 Use hashfast unique id instead of number since the unique id is displayed
Con Kolivas f314f272 2014-03-28T17:19:56 Use the hashfast opname as its unique identifier
Con Kolivas 67085a81 2014-03-18T22:39:52 Software ntime roll for all hashfast devices.
Kano 3c6cf56d 2014-03-13T21:48:48 hashfast make api stats field names unique
ckolivas c92fe1cf 2014-03-12T08:33:55 Check device data exists on a hfa instance before trying to reinit it.
Con Kolivas f164adaf 2014-03-11T11:09:50 Print off what quadrant regulator failed if known in hfa driver.
Con Kolivas 5dbb5e7f 2014-03-11T00:32:41 Whitelist all firmwares <= 0.5 on hfa for software rolling of ntime.
Con Kolivas ff6a4aba 2014-03-08T12:15:26 Add an hfa-options command line that allows the clockspeed to be chosen per device by name comma separated, with a function that can be expanded with more options in the future.
Con Kolivas 3c0b2812 2014-03-08T10:55:42 Off by one drv_rolllimit check against jobs
Con Kolivas ea86a1af 2014-03-08T10:53:55 Free the work that may be lost, leaking memory, in a failed hfa_send_frame
Con Kolivas be95fa90 2014-03-08T10:52:11 Roll the ntime for work within the hfa driver for firmware we know doesn't do it internally as an optimisation.
Con Kolivas cdf2e2f9 2014-03-06T21:29:57 Disable the MAX_CLOCK_DIFF check for newer hashfast firmwares since it's not required.
Con Kolivas d8c09364 2014-03-06T19:08:24 Store the hardware and firmware revision in the info struct for easy use in the hfa driver.
Con Kolivas f078bea0 2014-03-06T18:58:10 Only decrease the hfa clock rate if the device has been running for less than an hour before dying.
Con Kolivas 8d8095fa 2014-03-06T13:05:02 Change lack of op name response message in hfa driver
Con Kolivas 092277a7 2014-03-06T12:56:29 Check for lost devices at every write/read in hfa_detect_common
ckolivas ed0f9cd8 2014-03-04T12:27:28 Cope with older hfa firmware not even responding to op_name.
Con Kolivas 3ff8c500 2014-02-28T11:31:12 Starvation is not the problem with a bad sequence tail on hfa so revert the attempt to detect it.
Con Kolivas f1c508c5 2014-02-28T11:20:42 Check for starved condition in hfa sequence progression and allow the tail to catch up.
Con Kolivas 64b03af8 2014-02-28T11:09:57 Check for work sequence starvation as a false detection of a bad sequence tail on hfa.
Con Kolivas d963689e 2014-02-28T10:39:27 Add more information to a hfa bad sequence tail event.
Con Kolivas 27430408 2014-02-28T00:27:54 Bringing each hfa device online takes a lot of work generation so only ever do one at a time.
Con Kolivas 7ce7de29 2014-02-27T22:14:25 Keep the old hfa device data intact with a clean thread shutdown to allow it to be re-hotplugged with the old information.
Con Kolivas 8d18327e 2014-02-27T22:02:55 Cope with the API calling hfa on partially initialised devices having no info.
Con Kolivas 35fb1727 2014-02-27T21:38:17 Cold plug only one hashfast device to get started, and then hotplug many to minimise startup delays and possible communication delays causing failed first starts.
Con Kolivas 7755b610 2014-02-27T21:24:25 Send a shutdown and do a usb_nodev if hfa_reset fails.
ckolivas 266b5791 2014-02-27T16:57:52 Make all device_data dereferences in the hfa driver safe by not accessing it in statline before when it's non-existent.
ckolivas 84364a18 2014-02-27T16:18:56 Add an option to disable dynamic core shedding on hashfast devices.
ckolivas 5ed0e195 2014-02-27T15:13:50 Do not remove the info struct on a failure to hfa prepare.
ckolivas 8137e813 2014-02-27T14:46:06 Detect an hfa device purely on the basis of getting a valid header response to an OP_NAME query, leaving init to hfa_prepare which will allow multiple devices to start without holding each other up at startup.
ckolivas b6457fc4 2014-02-27T14:12:23 Store the presence and validity of opname in the hfa info.
Con Kolivas 92bf971f 2014-02-26T23:31:49 Use usb_detect_one in the hfa driver.
Con Kolivas d114c852 2014-02-26T23:18:56 Revert "Only start one hfa device at startup, leaving the rest to hotplug." This reverts commit 13c199c5ab39b4a03f087d5a4e412d61bd97c6c0.
Con Kolivas 13c199c5 2014-02-26T23:09:48 Only start one hfa device at startup, leaving the rest to hotplug.
Con Kolivas 6321995c 2014-02-26T22:46:20 Issue a usb_nodev on a bad work sequence tail in hfa
ckolivas 1bd011c2 2014-02-26T09:26:12 Read in hfa stream until we get a HF_PREAMBLE
Con Kolivas 2786526d 2014-02-25T21:36:31 Add shed count to hfa API stats output.
Con Kolivas 032b01eb 2014-02-25T21:11:00 Display the base clockrate for hfa devices with a different name to per die clockrates to be able to easily distinguish them.
Con Kolivas 433479da 2014-02-25T20:17:21 Use op_name if possible first with hfa devices to detect old instances and be able to choose the starting clockspeed before sending an init sequence, reverting to setting op name and serial number as fallbacks.
Con Kolivas e8ae1d58 2014-02-25T19:27:49 Make hfa resets properly inherit across a shutdown.
Con Kolivas 2e31c97f 2014-02-25T17:51:29 Don't break out of hfa_old_device early if there's no serial number.
ckolivas a1917795 2014-02-25T15:03:04 Fix harmless warning.
ckolivas 70210434 2014-02-25T12:59:58 Allow the drop in MHz per hfa failure to be specified on the command line.
ckolivas be526ef3 2014-02-25T10:03:20 Enable the hfa shed supported feature by default.
Con Kolivas 41226f18 2014-02-24T19:51:48 Add a --hfa-name command that allows one to specify the unique opname for a hashfast device.
Con Kolivas 2f454d86 2014-02-24T11:49:17 Set the clock rate with a work restart instead of an init when changing to old clocks for hfa
Con Kolivas ade4b0a6 2014-02-24T10:36:44 Set opname on hfa devices without a serial number to a hex value based on time to not overflow the field.
Con Kolivas 687c1705 2014-02-24T10:33:53 Add op name to hfa API stats output if it exists.
Con Kolivas a621e2b8 2014-02-24T09:54:23 Set the actual op_name in hfa devices if cgminer is choosing it itself due to it being invalid.
Con Kolivas 67041fbc 2014-02-24T09:00:53 Re-init an hfa device to its old data before setting up info structures as their sizes may change.
Con Kolivas c0e65369 2014-02-24T08:11:36 Remove the usb device whenever we do a running shutdown on hfa and do a shutdown as the imitated reinit to allow it to hotplug again.
Con Kolivas 6d33bf7e 2014-02-24T08:03:25 Reset opt hfa dfu boot after it's used.
Con Kolivas eb5ad60e 2014-02-24T08:02:25 Comment out windows only transfer on hfa startup.
Con Kolivas 2ab31317 2014-02-24T08:01:32 Clean up structures unused in case of all failures in hfa detect common
Con Kolivas fa7f84e6 2014-02-24T02:24:33 Clear all structures should we fail to hfa reset on adjusting clock on a hotplug.
Con Kolivas f832ce5b 2014-02-24T02:10:01 Set master and copy cgpu hash clock rate for hfa when dropping it on a restart.
Con Kolivas 8b0fb208 2014-02-24T02:00:44 Set the master hfa clock speed to lower when shutting down a copy.
Con Kolivas 9eaf90b2 2014-02-24T01:43:05 Do a clear readbuf on any hfa reset in case the device has not yet cleanly shut down.
Con Kolivas 33db33dd 2014-02-24T00:28:16 Increase hfa fanspeed slightly more when it's rising in the optimal range than falling.
Con Kolivas e0b857ce 2014-02-24T00:17:16 Always decrease hfa clock speed on a running shutdown and don't try sending an init frame since it will be dropped regardless.
Con Kolivas 66b18833 2014-02-23T23:58:02 Match hfa devices to old ones based on OP_NAME values before serial numbers if possible.
Con Kolivas 21fa38d5 2014-02-23T23:52:54 Read off the OP_NAME if it exists and is supported on hfa devices, setting it to the device serial number or a timestamp if it is invalid.
Con Kolivas 50fa5885 2014-02-23T22:56:14 Updated hf protocol
Con Kolivas d2f253cd 2014-02-23T22:42:15 Check for an amount along with no error in hfa clear readbuf
Con Kolivas f44831db 2014-02-23T22:39:17 Hfa clear readbuf can return a nonsense amount when there's a return error so ignore the amount.
Con Kolivas f1c6058c 2014-02-23T22:21:01 Running resets always cause a shutdown on hfa meaning the device will disappear with modern firmware so always kill off the threads to allow re-hotplugging.
Con Kolivas 7f74e21d 2014-02-23T22:09:28 Try a restart instead of a shutdown prior to trying to re-initialise a non-working hfa device.
Con Kolivas f3080422 2014-02-23T22:02:47 Reset the hfa hash clock rate to the old one if we find an old instance, only setting the device id in hfa_prepare
Con Kolivas 680a885a 2014-02-23T21:48:05 Keep the device_id on the original zombie thread for HFA in case of further resets.
Con Kolivas e022dc9f 2014-02-23T21:07:30 Local cinfo extra
Con Kolivas c22e080b 2014-02-23T20:56:40 Break out of hfa inherit if there is no device data.
Con Kolivas 94dcfce0 2014-02-23T20:52:42 Inherit the hfa zombie instance after the device id has been allocated.
Con Kolivas 19d0efd7 2014-02-23T20:31:04 The list_for_each_cgpu macro will dereference when there are no mining threads yet.
Con Kolivas db8c31c8 2014-02-23T20:21:36 Make hfa hotplug inherit some parameters from a previous instance if the serial number exists and is matching, avoiding dropping the clock on all devices.
Con Kolivas c76647f2 2014-02-23T16:45:04 Use the share_work_tdiff function in the driver watchdogs.
Con Kolivas 55cd36c1 2014-02-23T14:06:44 Limit the decrease of hfa clock rate on reset to the default clockrate.
Con Kolivas 0d32dc64 2014-02-23T14:04:34 Base the hfa failure time on the current expected hashrate instead of a static 15 seconds.
Con Kolivas 456e3dd0 2014-02-23T13:52:02 We shouldn't be trying to read from the hfa_send_shutdown function itself.
Con Kolivas cf27324d 2014-02-22T10:51:29 Take into account shed cores on hfa devices when determining how many jobs to send.
Con Kolivas c9197a75 2014-02-21T02:44:09 Give a nodev error if we have already set nodev in hfa clear readbuf to avoid further usb comms attempts.
Con Kolivas dab4a154 2014-02-20T20:30:36 Update the per die hash clock data on a running reset on hfa devices.
Con Kolivas 9aa6c6a2 2014-02-20T11:54:48 Set the per die clock on hfa to the known starting base clock instead of our requested clock rate.
Con Kolivas 3bbb8976 2014-02-20T11:35:02 Hfa device failure can be detected within 15 seconds so we should try restarting it sooner to avoid tripping the device's own watchdog.
Con Kolivas cca9b500 2014-02-20T11:29:43 Check return result of hfa clear readbuf to minimise error messages on device failure.
Con Kolivas f8571b60 2014-02-20T11:15:32 Send a work restart with every shutdown message to hfa devices to clear any work that might be stale on the next restart.
Con Kolivas 0f4ad7f4 2014-02-20T11:07:06 Store the hfa hash_clock rate and display it in the statline.