Sign Up
Log In
Log In
or
Sign Up
Places
All Projects
Status Monitor
Collapse sidebar
home:marec2000:cryptocoins
bfgminer
bfgminer.changes
Overview
Repositories
Revisions
Requests
Users
Attributes
Meta
File bfgminer.changes of Package bfgminer
------------------------------------------------------------------- Sun Jan 7 11:27:04 UTC 2018 - marec@detebe.org - Update to version 5.5.0: * Bump version to 5.5.0 * Update copyrights * AUTHORS: Update * README.ASIC: Update Alcheminer * alchemist: Reset using DTR as well * alchemist: Experimental USB->UART Dongle Support on RPi * openwrt: Update for LEDE 17.01 * README: Expand on --set-device documentation * Update libblkmaker to 0.6.0 * openwrt: Update for LEDE 17.01 * ccan: Update to latest version (which fixes musl libc support) * Bugfix: probe_device_thread: Fix threadname buffer size * Bugfix: appdata_file_call: Fix buffer size and loop * Bugfix: SSM: Clear desired_default_share_pdiff when initialising desired_share_pdiff before authentication * cpu: Avoid relocations for x86_64 assembly that interfere with building BFGMiner as a PIE * Bugfix: SSM: We already hold the pool lock when we need to call stale_work, so avoid trying to lock it a second time inside the function * Bugfix: Hold the pool lock when updating lp_id * Bugfix: Hold the control lock when updating the generate-to script * Bugfix: icarus: Explicitly specify type wide enough for bitshift in icarus_proc_for_nonce * configure: BFG_FIND_INCLUDE_PATH: Handle more unusual syntax correctly * Bugfix: SSM: Handle network problems safely, and use thread-safe libevent * README.OpenWrt: Update for 15.05 ------------------------------------------------------------------- Thu Mar 31 15:18:36 UTC 2016 - marec@detebe.org - Update to version 5.4.2: * Bump version to 5.4.2 * Update libblkmaker to 0.5.3 * Update copyrights * Bugfix: SSM: Never send mining.notify for stale work2d * Bugfix: Sanitise lock situation for work2d * Bugfix: SSM: Never issue the same work2d twice in a row * SSM: Refactor work2d check and job pruning from _stratumsrv_update_notify into stratumsrv_update_notify_str * Updates for OpenWrt 15.05 * Bugfix: Correct work2d handling for GBT servers ------------------------------------------------------------------- Sun Nov 22 11:11:27 UTC 2015 - marec@detebe.org - Update to version 5.4.1: * Bump version to 5.4.1 * Bugfix: openwrt: Fix bitmain option * icarus: Use more portable PRIu64 rather than %llu (which fails on Windows) * bitmain: Use more portable PRIu64 rather than %llu (which fails on Windows) * Bugfix: antminer: Avoid strstr(NULL, ...) when device has no [available] product string * Update libblkmaker to 0.5.2 * openwrt: Add option to enable bitmain driver * openwrt: Update libusb device list ------------------------------------------------------------------- Wed Nov 18 20:24:39 UTC 2015 - marec@detebe.org - i2c-tools headers fixed ------------------------------------------------------------------- Fri Oct 23 15:43:13 UTC 2015 - marec@detebe.org - Update to version 5.4.0: * Bump version to 5.4.0 * AUTHORS: Move jstefanop to current maintainers * Disable alchemist by default, add Travis build, and document configure option in README * alchemist: New scrypt ASIC driver * hashfast: Support setting clock speed from TUI Manage * avalon: Support setting clock speed from TUI Manage * antminer: Support setting clock speed from TUI Manage (hidden for non-BM1382/4 devices) * jingtian: Allow changing clock speed from TUI * Debuglog for retrodiff being disabled * Disable retrodiff when it was not explicitly enabled, and the pool doesn't seem to like it * Bugfix: Only adjust work_difficulty on retrodiff submissions, so we can still detect them on rejection * (Re-)enable retrodiff by default for stratum pools, since some servers implement mining.set_difficulty wrong in this way * DevAPI: Remove unused temporary variable from driver iteration * Bugfix: DevAPI: Update includes from uthash to utlist and support older versions of uthash * Bugfix: RPC: Need to include utlist.h for config file list * Bugfix: antminer: Avoid NULL dereference for non-BM1382/4 devices * README.ASIC: Update documentation for BM1382/4 clock calculation support * Bugfix: Keep JSON from getaccountaddress around long enough for debug messages * Bugfix: Run gen-version.sh from source directory * Bugfix: Clean up JSON from getaccountaddress * Bugfix: Do not use hash tables for driver lists, since they require unique keys * README.ASIC: Expand Antminer S5 section to S1-S5 for completeness, since in theory they should work * bitmain: Make reg_data optional for S4 and S5, calculating it from clock * antminer: Support for setting BM1382/4 clock by frequency MHz * Bugfix: compac: Set dname so assignments match * Pool option #cksuggest to use CKPool-compatible mining.suggest_difficulty * RPC: Add "Rotate Period" to config * RPC: Extend setconfig to change strategy * Allow setting strategy by name * README.RPC: Add missing setconfig|stratum-port documentation * README.RPC: Add missing setconfig|http-port documentation * Document Bitmain Antminer S5 support * bitmain: Remove poll_prio_threshold and just ensure we poll at a regular interval to avoid nonce buffer overruns * bitmain: Remove dead per-device constants * bitmain: Only have one set of actual maximums * bitmain: Migrate BITMAIN_MAX_NONCE_NUM to runtime packet_max_nonce * bitmain: Migrate BITMAIN_MAX_WORK_QUEUE_NUM/bitmain_work_poll_prio to runtime poll_prio_threshold * bitmain: Migrate BITMAIN_MAX_WORK_NUM to runtime packet_max_work * bitmain: Dummy model parameter * bitmain: Remove baud/flush, as the fake VCOM kernel driver ignores (and complains about) both * bitmain: Poll after queuing work, once enough is queued * bitmain: Sleep a little between polling * bitmain: Flush queues for work restarts * bitmain: Queue multiple works at a time * bitmain: Remove more dead code * bitmain: Prune old work eventually * Bugfix: bitmain: Properly count work-search failures as HW errors * bitmain: Refactor to work with minerloop_queue (leaks work memory) * bitmain: Avoid busy-looping on network devices * bitmain: Support for talking to a device over the network * Bugfix: lowl-vcom: Check for tcgetattr/tcsetattr/tcflush failure * bitmain: Use the hottest reading for device temperature * Bugfix: bitmain: Actually use baud setting * bitmain: Remove more dead code * bitmain: Handle stale results normally * bitmain: Simplify num2bit * bitmain: Initial set-device based configuration * bitmain: Use lowl-vcom abstractions * Update copyrights * Bugfix: bitmain: Type-safe printf-format usage * Travis: Test bitmain-only build * Build bitmain driver * bitmain: Hardcode configuration for now * bitmain: Rework nonce_diff handling * bitmain: Update headers * bitmain: Deal with unused variable warnings sanely * bitmain: Explicitly de-const device_path for freeing * Bugfix: bitmain: Correct type of cgpu->temp reference * Bugfix: bitmain: Use uint8_t consistently for binary data * Bugfix: bitmain: Clear opt_bitmain_dev string properly * bitmain: Remove non-existent bitmain_detect parameter * bitmain: Cleanup unnecessary code * bitmain: Use uthash to find queued work * bitmain: Implement inc_dev_status within record_temp_fan * bitmain: Remove non-applicable limits * bitmain: Remove obsolete get_statline_before function * bitmain: Move g_miner_version to bitmain_info struct * bitmain: Remove direct USB code * bitmain: Update driver registration * bitmain: Update symbol names * bitmain: Remove dead code * bitmain: Replace custom hexdump with normal bin2hex * bitmain: Get network difficulty from first work task * bitmain: No-op htole8 * bitmain: Simulate cgminer work_block somewhat * bitmain: Update bin2hex usage for malloc-free version * bitmain: Copy low-level code from Bitmain cgminer usbutils * Copy working driver-bitmain.c from Bitmain's cgminer at commit 4ecf89341657ea7efecdf588586ca3f068ab17ab * Add api_add_percent to miner.h * configure: List compac driver in output * Travis: Workaround Travis bug * configure: Use AC_PROG_SED to find sed ------------------------------------------------------------------- Sat Sep 5 13:32:15 UTC 2015 - marec@detebe.org - Update to version 5.3.0: * Bump version to 5.3.0 * README.ASIC: Compac docs * antminer: Explicit support for GekkoScience's Compac BM1384 Bitcoin Miner * icarus: Use all null padding when probing work division (BM1384 reacts strangely - using part as start nonce?) * antminer: Match Product strings including "Antminer" * Fix to enable building into directory out of source tree * Bugfix: icarus: Never set timeout to 0, since it disables the timeout altogether ------------------------------------------------------------------- Fri Jun 5 09:26:59 UTC 2015 - marec@detebe.org - Update to version 5.2.0: * Bump version to 5.2.0 * Update bundled libbase58 to 0.1.4 * README.ASIC: Document usage with AntMiner U3 * icarus: Include ns/hash estimates for each nonce result, in debug logging * icarus: Cleanup dev vs proc repr in logging * Bugfix: update_block_display: Ensure we have console lock, and avoid rare relock when resizing windows inside curses_print_status * gitmodules: Change libblkmaker URI to use GitHub since Gitorious is defunct * Bugfix: DevAPI: Avoid infinite hang in set_device help * bitforce: Wait until pre-initialisation jobs all flush completely before starting to avoid sanity check issues * bitforce: Wait to clear job queue (at init) until we have actually opened the device * bitforce: Remove unnecessary delay * Bugfix: vcom: Avoid overflowing cc_t before division to deciseconds * Bugfix: bitforce: Ensure hashes_done is called for every completed job * configure: Improve path finding macro to work with more possible CPP output * configure: Macroify header path search * Reduce HTTP request failure log level to DEBUG since it often occurs probing stratum pools * Bugfix: DevAPI: Guarantee set_device functions never get passed a NULL pointer for newvalue * Always update block display after pool display in case of username wrapping around * antminer: Adapt default configuration to work with U3 as well as U1/U2 * antminer: Add U3 voltage setting (x??? format only for now) * antminer: getstatus needs to read a result for every chip, but we don't use the output anyway, so just skip it * icarus: Modify work division detection packet to be compatible with Antminer U3 * icarus: Replace decisecond-precision read_count with read_timeout_ms (millisecond precision) to handle faster devices like the Antminer U3 that complete works in under 1ds * lowl-vcom: vcom_set_timeout_ms function (increases precision to ms on Windows only) * util: timer_remaining_us function * icarus: Pass device representation to icarus_{gets,write} so logging can use it rather than fd numbers * Bugfix: icarus: Check for impossibly fast hw errors (and don't report Eh/s when they occur) * Bugfix: icarus: Avoid a race with watchdog applying settings by setting device_data before add_cgpu * RPC: Clean up api_add_* somewhat * Bugfix: util: Fix is_power_of_two * Port antminer driver to use set_device_funcs ------------------------------------------------------------------- Fri Feb 20 13:49:11 UTC 2015 - marec@detebe.org - Update to version 5.1.0: * Bump version to 5.1.0 * Bugfix: extract_domain: Correctly handle query/fragment in URI without path * New pool option "retrodiff" * If a share passes difficulty check for the updated pool target, record that target for correct accounting of diff_stale. * Fix broken driver options in configure script * kncasic: Do flush! * kncasic: avoid duplicate nonces to be reported * kncasic: Fix logging levels Also reduce severity of "Communication error" and "timeout" messages * kncasic: Neptune-compatible API interface for enabling/disabling dies * kncasic: Use one bfgminer processor per die (not per core) This reduces CPU load dramatically. The same was done for Titan. * Update libblkmaker to 0.5.1 * get_master_rolling_hashrate -> get_proc_rolling_hashrate The concept of calculate driver's own rolling hashrate only for the "master" device is flawed. Better to do it in more general way and implement driver's own rolling hashrate on per-processor basis. * Update webconfig code to latest BFGMiner * allow url based config files * bitforce: Enable changing voltage from the Manage TUI * avalonmm: Use proc_set_device_tui_wrapper instead of avalonmm_tui_wrapper * DevAPI: Add generic proc_set_device_tui_wrapper * bitforce: Add "voltage" setting to send V?X commands * kncasic: Fix "double free" error * kncasic: We need to free queued work to avoid memory leaks * Bugfix: minion: Don't try to add 0-chip devices * Titan: Lower log level when reconfiguring dies. This gets rid of annoying "KnC 2-0: No KnC chip found", when the die is actually OFF * Titan: does not need SHA256d algo * configure: Avoid BASHisms for better portability ------------------------------------------------------------------- Sat Nov 29 13:44:40 UTC 2014 - marec@detebe.org - Update to version 5.0.0: * Bump version to 5.0.0 * opencl: Fail gracefully if clGetProgramInfo/CL_PROGRAM_NUM_DEVICES returns zero * opencl: Remember results of global offset testing * opencl: Only save kernel binary when we built from source * opencl: Only try to patch BFI_INT when compiling a kernel from source * opencl: Major refactor, splitting up opencl_load_kernel into many new functions * Bugfix: bytes_free: Set buf to NULL so the bytes_t is reusable * Titan: lower log level for flush messages * travis: Initialise variables to satisfy compilers * Titan: adjust log levels in knc_titan_get_info * Bugfix: opencl: Correctly handle non-goffset kernels * Bugfix: minergate: Avoid malloc(0) if stats file has too few values * Bugfix: opencl: free memory if clState creation fails * Bugfix: modminer: Avoid leaking devname when detecting claimed devices * Bugfix: free memory allocated for config file loading/writing * Bugfix: nanofury: Avoid NULL pointer dereference on init failure * Titan: Do not do actual configuration job in API thread context. Instead, send commands to the worker thread to do the actual job. That prevents API stalls and helps with BFGMiner frequent restarts. * minion: Reinitialise chips if they don't seem to be progressing * minion: Read temperature sensor * minion: TUI clock speed control * minion: Include "Frequency" in RPC status * minion: Add "clock" setting * minion: Set configuration parameters we care about * minion: Reset chip at initialization * minion: Configure PLL to 900 MHz at startup * minion: Report hashes done as accurately as we can * minion: Core enable/disable control * minion: Implement queue flushing * minion: Implement basic mining * lowl-spi: linux_spi_txrx2 to include device protocol dumping * minion: Detection code for the Prospero X1 * Actually add version.c file * Only rebuild a minimal version.c file when git commit changes * Bugfix: Use HASH_ADD_KEYPTR for const char *, not HASH_ADD_STR * Bugfix: Re-lock stgd_lock when we don't have a malgo-specific pool to use * Titan: bugfix: dies are not configured if die#0 is OFF * Titan: unused function argument * Titan: unused variables * opencl: Remove support for "OCL1" kernel define (indicated OpenCL 1.1+) * opencl: Autodetect whether global offset actually works * Fixing column spacing of non-curses device output * opencl: Add no-goffset support to scrypt kernels * opencl: Determine kernel support for goffset by code analysis * opencl: Add "goffset" setting to override detection of support for global work offsets * Bugfix: varint_decode: Correctly decode multibyte numbers * util: seek_data_cb: Check validity of offset * Allow libcurl to rewind the upload buffer * Set the priority of miner threads to idle on Windows * Use _SC_NPROCESSORS_CONF instead of _SC_NPROCESSORS_ONLN * Do not set thread affinity on single-processor systems * Fix processor count detection with HW_NCPU * SCHED_IDLE and SCHED_BATCH can only be used with sched_priority 0 * fix sched_setaffinity parameter : sizeof(cpu_set_t) * Adding additional boolean keywords enable and disable * scrypt: Disable OpenCL by default * Bugfix: opencl: Correctly accept negative intensities and set dynamic intensity without changing current value * Since longpoll connections are never reused, explicitly forbid reuse so libcurl cleans them up immediately * Bugfix: Correct already-in-use getcbaddr message * Demote some Bitfury debug log messages to only with device protocol dumping * Update knc-asic: Titan FPGA image with spi->i2c commands removed and fixed some lock-up scenarios * Titan code rework: Use one bfgminer processor per die (instead of core) * AUTHORS: Move Pelle and Vitalii to "Current Maintainers" since they actively maintain the Titan driver * AUTHORS: Add Ronny Van Keer (Keccak C implementation) * Bugfix: configure: kncasic does not actually need i2c-tools * Update bundled libbase58 to 0.1.3 * Bump version to 4.99.1 in preparation for 5.0 * Bump bundled libbase58 to 0.1.2 * Update stratum mining.capabilities method to use a single Object for all parameters * Keccak: Simplify keccak_hash_data * keccak: Adapt opencl intensity interpretation to have approximately the same desktop interactivity affect as SHA256d * Keccak: Include support in various builds * Travis: Update with Keccak * opencl/keccak: Add non-goffset support * opencl/keccak: Unroll all the hash rounds properly * opencl: Add Keccak support via "fullheader" kernel interface * keccak: Adapt for BFGMiner * Keccak: Import algorithm code as-is * opencl: Add a simple "fullheader" kernel interface * Bugfix: Correct USE_SHA256D macro name to ensure SHA256d is always preferred as default algorithm * cpu: Generic scanhash that can support any PoW algorithm * RPC: Include Hash Method in minecoin for new algorithms * opencl: Teach findnonce to behave based on kernel interface rather than mining algorithm * opencl: Get min_nonce_diff from struct mining_algorithm * Move malgo-specific code to dedicated source files * opencl: Simplify kernel-specific data handling * Update knc-asic: Fix issue with detect_die crashing due to errors in response * Init last_nonce on die reconfiguration to avoid lots of "unknown work" messages * Bugfix: Extra garbage was sent on SPI with new commands * README.ASIC: Minor cleanups to Zeusminer section * README.ASIC: Explain Zeusminer options * Update knc-asic: Fix issue with detect_die crashing due to errors in response * Init last_nonce on die reconfiguration to avoid lots of "unknown work" messages * opencl: Remove redundant checks for USE_OPENCL * cpu: Remove redundant checks for USE_CPUMINING * opencl: Only attempt to BFI_INT patch SHA256d kernels * Significantly rewrite configure script to handle driver dependencies cleaner * Bugfix: cointerra: Can handle nonces down to pdiff 1 * Make SHA256d mining optional * Makefile: Specify .cl files explicitly, and only include scrypt ones when configured with scrypt support * kncasic: Initialise first_cgpu to silence false warning * Bugfix: Clarify goffset decision code, and refuse to compile kernels in situations where they won't work * Bugfix: opencl: Output buffer must be readable by kernel since it is used with a count/position iterator * Bugfix: opencl: Skip NULL kernel interface entry * DevAPI: hashes_done: Simplify and improve precision of max_nonce calculation * opencl: Avoid duplicating kernel_interfaces mapping in select_kernel * opencl: Avoid duplicating kernel_interfaces mapping in opencl_scanhash_get_kernel * cpu: Drivers should not be doing fulltest on their own, so remove it * cpu: Pass full work struct to scanhash functions * Bugfix: DevAPI: hashes_done: Explicitly cast to uint64_t for big calculations (LLVM was allowing overflow) * AUTHORS: Add Pelle Nilsson * kncasic: Split up each ASIC channel to its own device * Bugfix: Extra garbage was sent on SPI with new commands * Titan: supply its own rolling hashrate implementation * A driver can supply its own rolling hashrate function (in case generic rolling hashrate implementation does not work) * Remove temporary log messages for manual report checks * Wrong variable used in new per-die iteration * Do manual core checks in parallel with normal ones * Fix: Only one die was checked after flush in last commit * Manually check all cores for reports after flush * Move thread_reportin() call to poll function * titan: reportin every process_report to avoid falsely tripping the watchdog * Revert "Increase watchdog sick time to 300 seconds" * titan: Remove unused variables * Bump knc-asic submodule * Increase FPGA to ASIC SPI frequency to 6 MHz * Use new get_work_status prototype and print FPGA CRC error counters to log * Titan flush optimization fix: Don't send get_info commands when FPGA is speaking to ASIC * Increase watchdog sick time to 300 seconds * All dies use same nonce working range * Bugfix: First core in each die got too big nonce area to work on * Titan flush optimization bugfix: FPGA status was checked once per poll instead of once per ASIC * Titan flush optimization: Bugfix: Flush time measurement only worked with one ASIC * Titan flush optimization: Point knc-asic submodule to include new FPGA image * Titan flush optimization: Remove temporary debug messages * Titan flush optimization: Keep works in local queue until slot number is reused * Titan flush optimization: Point knc-asic submodule to needed revision * Titan flush optimization: Measure new flush time * Titan flush optimization: Send works to all dies after flush * Titan flush optimization: First rough implementation * Remove temporary log messages for manual report checks * KnC: Use consistent naming for KNC_MAX_DIES_PER_CORE * Refactoring: coreid is not used, remove it * Refactoring: put all device architecture defines in one place (knc-asic.h) * Add lock to protect concurrent accesses to knc_state * Wrong variable used in new per-die iteration * Do manual core checks in parallel with normal ones * Fix: Only one die was checked after flush in last commit * Manually check all cores for reports after flush * Move thread_reportin() call to poll function * configure: Bump to 4.99.0 in preparation for 5.0 * titan: reportin every process_report to avoid falsely tripping the watchdog * Revert "Increase watchdog sick time to 300 seconds" * titan: Remove unused variables * Bump knc-asic submodule * Increase FPGA to ASIC SPI frequency to 6 MHz * Bugfix: Prefer not using work created just to ensure a specific algorithm is queued, so strategies work as much as possible * Stratum: Enable mining.set_goal to change parameters on the current goal * Travis: Build MinGW64 with ncurses * Use new get_work_status prototype and print FPGA CRC error counters to log * cpu: Fail gracefully if unsupported mining algo gets into scanhash * opencl: Fail gracefully if unsupported mining algo gets into opencl_scanhash_get_kernel * Bugfix: Declare manual_enable_pool in miner.h, for api.c use * Titan flush optimization fix: Don't send get_info commands when FPGA is speaking to ASIC * SSM: Send goal malgo to goal-enabled clients * Allow pools with #change_goal_malgo attribute to change the mining algorithm used by their assigned goal * Gracefully fail when no pool can be found to generate specific-algo work * Remove opt_scrypt * dualminer: Replace opt_scrypt with a per-device "scrypt" option * dualminer: Make dual_mode a per-device option as it should be * opencl: Nearly complete migration to per-work mining algorithms * pool_actively_desired: If we are the highest priority, workable pool for a given algorithm, we are needed * Ignore opt_queue for unused mining algorithms * Set name and aliases on mining algorithms * When hotplug is enabled and a mining algorithm is configured for the first time, schedule a rescan of hardware to pick up anything now applicable that may not have been before * Increase watchdog sick time to 300 seconds * Bugfix: bitfury: Use long long and llabs to avoid under/over flows misbehaving * Try to keep enough work queued for each mining algorithm in use * Move select_loadbalance and select_failover logic into their own functions * Keep track of how much work is staged per-algorithm * Keep track of how many goals reference each mining algorithm * Abstract goal_set_malgo function * Build a mining_algorithms list * get_work: Restore previous getwork rollntime behaviour * get_work: Only return work items compatible with processor (degrades getwork rollntime support) * opencl: Support for per-work mining algorithms * All dies use same nonce working range * Bugfix: First core in each die got too big nonce area to work on * Build without POW_SCRYPT at all * opencl: Defer loading kernel until it is needed * opencl: Split initCl into opencl_create_clState and opencl_load_kernel * SSM: Implement mining.capabilities including proxying mining.set_goal * Replace mining.goal.subscribe with mining.capabilities * Stratum: Avoid resetting the goal if the old and new name match * Allow specifying goal options as eg, --pool-goal name:malgo=scrypt * Bugfix: Highlight all active pools, in pool listing * Set quota to 1 when enabling a failover-only pool, and disable failover-only when setting quota * Bugfix: pool_actively_desired: Detect failover-only conditions for loadbalance and balance strategies * Bugfix: Only display pools as Failover state when it is effectively so * Make sure KEY_EVENT from wincon is ignored (it conflicts with ncurses) * Titan flush optimization bugfix: FPGA status was checked once per poll instead of once per ASIC * cpu: Support for per-work mining algorithms * Replace cgpu.min_nonce_diff with drv.drv_min_nonce_diff(cgpu, malgo) * proxy: Avoid assumptions about mining algorithm at initialisation * Core only: Partially move most of opt_scrypt to mining_goal * opencl: Reintroduce independent intensity setting internally * Move MAX_*_INTENSITY stuff to driver-opencl.h * proxy: Minimise minimum difficulty for proxy clients * Titan flush optimization: Bugfix: Flush time measurement only worked with one ASIC * README.ASIC: Add details for BFx2 USB * Save pool goals in written config files * Enable --generate-to option (was --coinbase-addr) to work with non-default goals * Titan flush optimization: Point knc-asic submodule to include new FPGA image * Titan flush optimization: Remove temporary debug messages * Titan flush optimization: Keep works in local queue until slot number is reused * Titan flush optimization: Point knc-asic submodule to needed revision * README: Update documentation for multiple blockchain support * RPC: Add "Difficulty Accepted" to "coin" command * RPC: Add "Mining Goal" to "pools" command * RPC: Accept an additional argument for "addpool" to indicate mining goal by name * Stratum: Support for mining.set_goal("goal name") - currently just resetting the user-configured goal * Teach longpoll logic about multiple mining goals * RPC: Include non-default goals in reply to "coin" command * Ignore generation address/script on pools using non-default mining goals * Only include block display lines for active goals * Move income to block display line, and show extra block display lines for each additional mining goal * New --pool-goal option to set a distinct named goal per-pool * Parameterise most references to global mining_goal_info * Move block_time to be per block_info * Show "?" for income if block height is unknown * Move block height tracking onto block_info * Calculate current_fullhash only when needed (for RPC 'coins') * Replace current_block_id with blkchain->currentblk[->block_id] * Move current_hash to goal->current_goal_detail * Clean up struct block_info names * Use full prevblock hash as block key (also gets rid of dead code) * Move global variables related to the blockchain or mining goals on to global structs * Bugfix: Need unicode_micro in non-Unicode curses builds * Remove unused variables from curses-less builds * Titan flush optimization: Measure new flush time * Titan flush optimization: Send works to all dies after flush * Titan flush optimization: First rough implementation * Bugfix: Cannot duplicate userpass option or we end up with it processed double * Titan: Less clobber on the screen: group some "failure" messages * Titan: Submit stale shares while doing flushes ------------------------------------------------------------------- Sun Nov 9 19:04:11 UTC 2014 - marec@detebe.org - systemd service added ------------------------------------------------------------------- Tue Oct 21 07:17:20 UTC 2014 - marec@detebe.org - Update to version 4.10.0: * Bump version to 4.10.0 * i2c-tools are not required by Titan * kncasic: Fix auto * minergate: Fix hashmeter * minergate: Support minergate-side ntime rolling for SP30 only * Build system fixes for kncasic driver * configure: Deduplicate Linux i2c-dev.h checks for KnCMiner drivers * Travis: Update for kncasic driver * kncasic: Only show the relevant die for RPC stats * kncasic: Use proc_repr for logging * kncasic: Split up logical processors * kncasic: Silence warning about unhandled enumeration values * Bugfix: kncasic: Proper format specifications * kncasic: Remove unused code * kncasic: Minimal changes to get it compiling * kncasic: Import gen 2 driver from cgminer commit bc153552be8591250cb3214bf5202501d4a39922 * i2c-tools are not required by Titan * Stop using uthash as a submodule. (for compatibility witrh upstream. After all, we have uthash installed on RPi) * Titan: fix compiler warning "maybe-uninitialized" * Titan: fix compiler warning "maybe-uninitialized" * Titan: Increase die inactivity timeout to 20 secs * Titan: Slightly improve some debug messages * Titan: Use multi-part batched SPI transfers for flushes (saves 2 secs on each flush) * Titan: Fix buffer overflow * Titan: Increase queue prefill value up to 20 * Titan: Flag for fast broadcast flushes. Not enabled: DC/DCs trip off easily! * Titan: Monitor die health, reconfigure it if no shares in 10 seconds * Titan: Make log level of some messages lower, to not clobber the screen * Titan: Work assignment and flushing is per-die, not per-ASIC * Titan: Set flush flag after re-configuring the die. This implies one-by-one bring-up of cores, not by broadcast. (setting first work by broadcast to previously idle die kills the DC/DCs) * Titan: Refactoring: intermediate variables for first_proc and repr * Titan: Flush cores one-by-one right before reconfiguring them * Titan: Define for the broadcast core address * Titan: Re-configuring dies through API command "procset" * Typo in comment * Titan: Use correct version of knc_titan_setup_core * Titan: configure_one_die func for configuring single die * Titan: fill all non-found ASIC structs with the same (invalid) data * Titan: Core init parameters (nonce range) independent of number of found dies. That allows later (re-)initialization of newly found dies in uniform manner. * Increase die inactivity timeout to 20 secs * Slightly improve some debug messages * Use multi-part batched SPI transfers for flushes (saves 2 secs on each flush) * Fix buffer overflow * titan: Increase queue prefill value up to 20 * Whitespaces * Do not return error on RECONFIGURE command or the whole bfgminer will be restarted * Flag for fast broadcast flushes. Not enabled: DC/DCs trip off easily! * Monitor die health, reconfigure it if no shares in 10 seconds * Make log level of some messages lower, to not clobber the screen * Work assignment and flushing is per-die, not per-ASIC * Set flush flag after re-configuring the die. This implies one-by-one bring-up of cores, not by broadcast. (setting first work by broadcast to previously idle die kills the DC/DCs) * Refactoring: intermediate variables for first_proc and repr * Flush cores one-by-one right before reconfiguring them * Define for the broadcast core address * Don't forget to setup minerloop locking ("control_request") mechanism * First clumsy attempt to reconfigure die * Use correct version of knc_titan_setup_core * titan: configure_one_die func for configuring single die * titan: fill all non-found ASIC structs with the same (invalid) data * titan: stubs for re-configuring dies through API command "procset" * Core init parameters (nonce range) independent of number of found dies. That allows later (re-)initialization of newly found dies in uniform manner. ------------------------------------------------------------------- Sun Oct 5 09:17:44 UTC 2014 - marec@detebe.org - Update to version 4.9.0: * Bump version to 4.9.0 * NEWS: Missing Windows library updates * minergate: Autodetect SP30 on /tmp/connection_pipe_sp30 * minergate: Make stats file configurable * minergate: SP30 only wants max 10 queue requests at a time * minergate: Use work_completed flag for SP30 * minergate: Only SP10 has a second winner_nonce * minergate: Simplify multi-winner_nonce handling * Bugfix: minergate: Correct endian for 2nd winner_nonce * minergate: Vary max jobs queued * minergate: Vary number of requests/responses per packet * minergate: Support --set MGT:protover=N * Add uthash as a submodule * libbase58: Use git URI for submodule to avoid failure on systems without HTTPS support * Travis: Update for titan driver * configure: Accept --enable-titan=CONTROLLER to select controller * make-release: Remove unnecessary knc-asic/{*.rbf,*system,waas} from release source * extra_work_queue so devices can influence their effect on the central work queue somewhat (titan needs less than 1-per-proc) * Avoid adding include paths for titan driver * Bugfix: titan: Add missing printf formatting for core busy status * avalon: Drop custom hexdump logging * Build titan driver independently from knc (Jupiter) driver * titan: Do not fill up next slot immediately after urgent setwork * titan: Pre-fill work queue so that all ASICs have fresh jobs after a flush * Build instructions for KnC Titan * Doesn't compile without explicitly included inttypes.h on some machines * knc-asic: Updated to e5c986d3c44fde8c5b069508ef6979f2f2be92d6 * Fix Makefile.am to build bfgminer for titan * titan: Subdivide full nonce range only between cores in one ASIC (because works are now distributed per-ASIC too) * titan: DC/DCs does not like broadcast flushes (urgent setwork). Do not do it! * titan: Preparation to setting threads-per-core externally, by user * titan: Re-flush cores in case of slot number collision * titan: Per-ASIC flush, per-ASIC work management * titan: Start cores after flush individually, not by broadcast. * titan: Default frequency is 275 MHz * titan: Difficulty is offset by one in ASIC cores. * titan: Fix first_proc pointer * titan: Use 2 threads per core * titan: Use setup_core from knc-asic library * titan: Poll all enabled ASICs amd dies, not only one * titan: Properly set work_accepted flag * titan: Hint detection function about expected device type * titan: Fix setup_core command * titan: Use knc-asic library for transport layer * Add knc-asic as submodule * titan: Change spi device to spidev1.0 * titan: Add define to .h file * titan: Increase workqueue size up to number of slots per core * titan: Send data to hashmeter * titan: Disregard stale reports after flush * titan: Check for next asic/die switch when processing info results * Bugfix: titan: Fix segfault * titan: Set actual hardware nonce_diff for works in prepare_work * titan: Do clean flush ("purge") on init * titan: Store last_nonce right * titan: First attempt to process nonce responses * titan: Change 'scanhash' minerloop to 'queue' * titan: Init all cores for their own nonce ranges * titan: For RPi we use spidev0.1 * titan: Setup_core command implemented * titan: New commands set_work & get_report * titan: Move asic-specific functionality to the separate file (titan-asic.c) * titan: First ugly detect of Titan chip over SPI * knc-titan: Begin work on Titan (scrypt miner) driver * Travis: Cross-compile a Win64 build * RPC: Initialise json_config to silence false warning * Make sure MOUSE_MOVED from wincon is ignored (it conflicts with curses) * Travis: Perform full builds with libbase58's base58 tool (which is used for tests) * Travis: Test many configuration variations * Travis: Build with libsensors and VFIO * Travis: Upgrading GCC triggers locale rebuild, so just do the one in use * Travis: No need to upgrade GCC for LLVM build * Travis build configuration * Run BFGMiner's unit tests for 'make check', and have --unittest exit with failure if any problems occur * libbase58: Update to pick up on LLVM fixes * Bugfix: configure: Affect gridseed driver with --disable-other-drivers * Bugfix: configure: minergate driver needs lowlevel for claiming sockets * Bugfix: configure: --disable-other-drivers should not affect non-driver options * Bugfix: configure: --with[out]-vfio needs $withval, not $enableval * Bugfix: rockminer: Correct types for short read error message * Bugfix: cointerra: Check lowlevel device is USB before trying to probe it (as USB) * Bugfix: icarus: fix the STATS RPC API call crashes with a multi-proc device * bitforce: Reinstate device work inprogress count sanity check for 28nm devices * littlefury: Read uC temperature sensor * littlefury: Keep track of enabled chips and power state explicitly in case of trouble * Bugfix: async minerloop fix for devices disabled at start * twinfury: Implement device protocol dump more low-level ------------------------------------------------------------------- Wed Sep 10 19:08:26 UTC 2014 - marec@detebe.org - Update to version 4.8.0: * Bump version to 4.8.0 * Improve precision of total_secs used in (at least) RPC summary Elapsed * Bump embedded libblkmaker to 0.5.0 * Bump embedded libbase58 to 0.1.1 * Bugfix: Reorder LDADD and such for priority * Bugfix: bitforce: Initialise variable to NULL * Bugfix: SSM: Use client_next member consistently when working with stratumsrv_connlist * Remove now-unused bfg_cond_timedwait which cannot be made portable * Spawn a new thread for cmd-idle rather than relying on problematic pthread timedwait * Bugfix: Need signed types for ntime min/max offsets * cointerra: Ensure devlog messages cannot overflow * Bugfix: cointerra: Defer setting USB timeout until after initialisation * cointerra: Set configuration and claim interface * Bugfix: cointerra: Check ep is open before trying to talk to it (crash at init failure) * cointerra: Support for --set cta:load=N * cointerra: Store load setting on struct cointerra_info * cointerra: Operate within a single thread * Bugfix: rockminer: Fix processor disabling * Silently ignore shares rejected if they were above target and only got submitted "just in case" * Abstract put_in_parens function * Abstract extract_reject_reason function * Remove dead code * cointerra: Update to minerloop_queue * cointerra: Split work packet into cointerra_queue_append function * cointerra: Prepare for splitting work packet into cointerra_queue_append function * cointerra: Use more fresh code for work packet * cointerra: Use fresh code for work packet * Bugfix: cointerra: Use bfg_cond_timedwait to avoid spinning * Bugfix: bfg_cond_timedwait providing semantics expected for --cmd-idle implementation * tq_pop: Remove abstime argument since nothing used it and it wouldn't work anyway (uses CLOCK_REALTIME while we use CLOCK_MONOTONIC[_RAW] when possible) * cointerra: Claim and release lowlevel device * Per-processor TUI: Align columns for more-than-one proc letter * cointerra: Propagate per-core temperatures to each processor * cointerra: Reduce redundant stats information * cointerra: Correctly divide up individual processors * cointerra: Update to latest BFGMiner * work_ntime_range helper function * work_{get,set}_ntime inline functions * Store a reference timeval with ntime_roll_limits * util: min macro * m4/bundled_lib: Workaround bug in autoconf <2.64 * README: --coinbase-check-* options * Bugfix: Accept actual percentages for --coinbase-check-percent * Optimise coinbase check logic by using actual script bytes everywhere possible * If full version is too long, try truncating it at '-' * Bugfix: Check last solo generation tx against new template rather than most recent * Pool option #skipcbcheck to disable new coinbase checks * pool_check_coinbase: Avoid redisabling an already misbehaving pool * Bugfix: Keep connection active for rejecting and misbehaving pools so we can detect when they recover * Share pool coinbase check reaction code * Initial version of coinbase checking function for GBT and stratum * Bugfix: Recheck current_pool after calling pool_died * Bugfix: Stable pool recovery: Only care if the pool is enabled * Import libbase58 for base58 encoding/decoding * rockminer: Limit even unsafe frequencies to 640 MHz, since above that overflows frequency bits and triggers fan control * README.ASIC: Add a section for Gridseed * Cleanup libblkmaker bundling code to mostly live in autoconf macros * Bugfix: Fix CPU miner benchmarking within benchmark-intense mode * benchmark: Detect duplicate shares within 5 minutes * benchmark-intense: benchmark_update_interval constant in code * benchmark-intense: Detect stale results * benchmark-intense: Update work every second * benchmark-intense: Simulate 250 KB generation transaction * benchmark-intense: Generate individual work items from 2D work (tests host CPU rate of work production) * Introduce --benchmark-intense option * Bugfix: benchmark: Free json_null() after use * Bugfix: minergate: Claim socket before we initialise a cgpu for it * Bugfix: avalonmm: Claim device before we initialise a cgpu for it * Update NEWS * Always check if we should switch pools when enabling one, and always enable pools we want to switch to * Bugfix: switch_pools: Broadcast lp_cond outside of control_lock to avoid deadlocking * Protect enabled_pools by a mutex in disable_pool function * Combine reject_pool into disable_pool function, and don't allow it to override a manual disable * Call disable_pool() at the begin of remove_pool() and combine them when appears together in the code * Bugfix: Always call enable_pool and disable_pool to ensure consistent handling of the situations * Remove dead CPU mining code to silence warnings ------------------------------------------------------------------- Sun Aug 17 14:04:37 UTC 2014 - marec@detebe.org - Update to version 4.7.0: * openwrt/multibuild: Avoid copying bitforce-firmware-flash, which is no longer included * Bump version to 4.7.0 * AUTHORS: Simpler design, alphabetise lists * Remove obsolete SuSE packaging * AUTHORS: Add Vitalii Demianets * Bugfix: test_hash: Fix Htarg calculation for exact binary divisions of diff 1 * avalonmm: Try to autodetect a sane default voltage and clock from MM version * avalonmm: Only poll one module at a time, so they don't try to talk over each other * avalonmm: Set baud rate for detection * Stratum: Ignore response from mining.extranonce.subscribe * Stratum: Implement mining.set_extranonce, and advertise support for it to pools with a #xnsub flag * Stratum: Keep nonce1 and n2size in a "staging" state until mining.notify is received * Fix target calculation from diff < 1.0 in test_hash() * Stratum: Avoid putting new targets on pool stratum_work until mining.notify is received for it * cointerra: Cleanup debugging * minergate: Flush during init * minergate: Flush ready-to-queue list * minergate: Avoid preparing new jobs to queue while we are pending a flush, since the flush would occur after queuing the new jobs * minergate: Check for reused job ids * minergate: Avoid counting flushed jobs toward hashes done * minergate: Report all temperatures individually to RPC * minergate: Load temperatures from stats file * minergate: Implement hashmeter * minergate: Scale nonce difficulty up to pdiff 32 * minergate: Implement mining * minergate: Support for detecting a SP10 * Bugfix: configure: avalonmm needs work2d * Bugfix: bitforce: Remove sanity check triggered by race conditions on 65nm devices * Bugfix: bitforce: Avoid double-free of voltage data ------------------------------------------------------------------- Sat Aug 2 11:25:01 UTC 2014 - marec@detebe.org - Update to version 4.6.0: * Bump version to 4.6.0 * avalonmm: Even if no fans report speed, display set % * Bugfix: avalonmm: Fix fan speed setting * Bugfix: avalonmm: Actually read the result needed to get the correct module id * README.ASIC: Document AvalonMM driver usage with Avalon 2/3 rigs * Bugfix: Makefile.am: Remove reference to non-existent driver-avalonmm.h * Use BUILT_SOURCES to ensure version.h is always built first * avalonmm: Safely handle an improper job id that is within the last 2 sent * avalonmm: Include asserted fan speed in RPC * avalonmm: Include asserted fan speed in ManageTUI * avalonmm: Silence warning about detect ack at runtime * avalonmm: Make fan speed an option (both RPC and TUI) * avalonmm: Allow changing clock speed and voltage from Manage TUI * Bugfix: avalonmm: Show proper units for fans & voltage * avalonmm: Support for disabling the entire chain * configure option --with-udevrules-group to allow customising the group name used * Implement broad_udevrules for avalonmm * avalonmm: Show extranonce1, module id, temperatures, fans, clock, and voltage in Manage TUI * avalonmm: Include Module Id and ExtraNonce1 in RPC devdetails * avalonmm: Add Temperature0/1 and Fan Percent 0/1 to RPC * avalonmm: Add Frequency and Voltage to RPC * avalonmm: Add voltage setting (defaults to 0.6625 V) * avalonmm: Add clock setting and try to autodetect it if not provided * avalonmm: Implement hashmeter * avalonmm: Adjust device target up to pdiff 32 when possible * avalonmm: Update job when current pool changes * Bugfix: avalonmm: MM flips the xnonce2, so we need to do the same * avalonmm: Implement mining logic * lowl-spi: Move bit order reverse to bitflip8 function in util * avalonmm: Treat multiple chained modules as slaves rather than processors * avalonmm: Probing for devices using Avalon Miner Manager (Avalon2/3 rigs) * littlefury: Move crc16 logic to util * Bugfix: zero_stats: Only call cgpu function if it exists * Remove FPGA-only bitforce-firmware-flash tool (now located at https://github.com/luke-jr/bitforce-fpga-firmware-flash ) ------------------------------------------------------------------- Mon Jul 28 14:34:06 UTC 2014 - marec@detebe.org - Update to version 4.5.0: * Bump version to 4.5.0 * Bugfix: gridseed: Fix crash in TUI with newly introduced feature * Bugfix: zeusminer: Fix crash in TUI with newly introduced feature * gridseed: Add the ability to set clock speed via the TUI * zeusminer: Add the ability to set clock speed via the TUI * Bugfix: gridseed: Only set PLL frequency once (not once per processor) * Bugfix: zeusminer: Retry detection on failure - ZM fails detection 1 / ~30 times * Bugfix: zeusminer: Enabling the ZeusMiner driver should flag needing lowl-vcom * Bugfix: gridseed: Implement a proper scanhash routine for GridSeeds * aan: Silence false warnings * Include broad udev rules when configured with --enable-broad-udevrules option * gridseed: Document GridSeed specifics with comments and URLs * Bugfix: util: Lower the threshold for considering pool Scrypt diff broken * Copyright cleanups and updates * zeusminer: Return the Chip # via the RPC API call procdetails * zeusminer: Display the Chip # in the UI when viewing per-proc details * Bugfix: miner: Check work->blk.nonce to see if work should be abandoned * rockminer: Allow setting clock 300-2560 as long as prefixed by "unsafe:" * bitforce: Support for Monarch devices on Windows using new "mswin" lowlevel interface to scan for KMDF driver * gridseed: Refactor code to clean up and organize * Bugfix: gc3355: Do not send work_id (or anything) in the last 4 bytes of work * util: Work around broken Scrypt pools automatically using realistic pdiff assumptions * scanhash: Document scanhash related methods with comments * Bugfix: gridseed: Report communication errors using dev_error() * gc3355: Return bytes read to distinguish 0 bytes from error * Bugfix: gridseed: Add delay between Scrypt reset and sending work * Bugfix: zeusminer: using icarus timing can lead to false positive Idle detection * gridseed: Bugfix: fix the constant used for estimated hash rates * klondike: zero_stats support * bitforce: zero_stats support (Temperature, Voltage, and Avg Wait) * avalon: zero_stats support (temp_max, no_matching_work, match_work_count%d) * Add driver interface for zero_stats function * Bugfix: bitforce: Short-circuit bitforce_read(0) to avoid lowlif issues * Bugfix: bitforce: Ensure data for the current command is always read (and only once), with lowl-pci * gridseed: Bugfix: Revert earlier changes to scanhash based on feedback * Bugfix: setup-vfio: Add device ids to vfio-pci even if they didn't have a driver previously * Bugfix: lowl-pci: Use volatile qualifier for PCI mmaps * hashfast: Support for changing clock at runtime * hashfast: Store a copy of what we believe to be the chip configuration data * hashfast: Store firmware revision * README.ASIC: KnCMiner: Talk about additional i2c-tools dependency * README.ASIC: KnCMiner: Remove mention of now-obsolete Bertmod * README.ASIC: KnCMiner: Grammatical corrections * Bugfix: RPC: Avoid processor-specific details in devdetails * Bugfix: bitforce: Use lowlevel interface device was probed with, for actual mining * setup-vfio: New shell script to automatically configure VFIO * gridseed: Bugfix: log invalid detect responses as debug info, not an error * gridseed: Bugfix: Fix hash-rate calculation when no nonces are being found * README: Add Controla and Minera controller software links & info * zeusminer: Support setting clock via RPC API procset call * dualminer: Bugfix: assign the proper value for work_division * gridseed: Refactor to clarify GridSeed terminology * gridseed: Add support for displaying per-chip statistics * gridseed: Convert to use newer set_device_funcs interface * Bugfix: gridseed: Don't leak cgpu_info when devices are already claimed * icarus: Processor detail support * DevAPI: Allow using device_proc_by_id with const struct cgpu_info * icarus: Ensure fpga_count is finalised before add_cgpu is called * icarus: Move work_division autodetection code into new icarus_probe_work_division function * icarus: Remove ability to change fpga_count at runtime * Bugfix: gen-version.sh: Replace bashism ("here string") * gridseed: Support devices with more than 255 physical processors * Use `git describe` to generate version.h to avoid user confusion over builds from git * zeusminer: Support devices with more than 255 physical processors * README.ASIC: Monarch: Need to do new_id for every affected device, and only after unbinding them * README.ASIC: Monarch: Use shell variables for commands * util: pk_uNle macro for changing a number inside a packed structure * hashfast: Include chip/core address in RPC procdetails * gridseed: Reduce traffic sending work to GridSeed devices * gridseed: Bugfix: previous refactor (unpublished) changed logic * gc3355: Refactor to remove SHA2 code specific to the 5-Chip GridSeed Orb * lowl-vcom: Bugfix: do not attempt to open the USB devices when enumerating ------------------------------------------------------------------- Mon Jul 7 17:28:52 UTC 2014 - marec@detebe.org - Update to version 4.4.0: * Bump version to 4.4.0 * lowl-vcom: Added support for auto scanning (-S auto) on Mac OS X Uses I/O Kit to enumerate USB devices https://developer.apple.com/library/mac/documentation/DeviceDrivers/Conceptual/IOKitFundamentals/Introduction/Introduction.html http://en.wikipedia.org/wiki/I/O_Kit * README: Update for commandline options * README: Update configure options * add_local_gbt: Avoid adding servers already configured * Bugfix: bfg_gpio_setpin_output: Clear alt-function bits via INP_GPIO macro * jingtian: Explicitly configure SPI device while opening * jingtian: Toggle ASIC reset GPIO at startup * aan: Set defaults as soon as a proc is initialised * aan: Allow specifying clock as xHEXX for a raw PLL register config * aan: Include current frequency in RPC status * aan: Implement --set jtn:clock=MHz * aan: Logic to calculate PLL configurations for a given frequency * jingtian: Use SPI enable GPIO to disconnect SPI during chipselect changes * aan: Set PLL to 850 MHz * aan: Simplify register buffer * jingtian: Decode extra temperature bytes in read_reg * aan: Add a read_reg hook * aan: Enable configuring nonce diff with (eg) --set jtn:diff=32 * aan: Properly handle nonce_diff * aan: Implement mining * DevAPI: Some designs set the main thr tv_poll from secondary thrs, so check it after the loop * aan: aan_spi_parse_rx implies spi_clear_buf * jingtian: Do detection asynchronously across all possible chipselects so they complete in parallel * aan: Refactor aan_spi_cmd a bit * rockminer: implement --device-protocol-dump for debugging * rockminer: Bugfix: must specify a baud rate (maximum of 115200) to get a read response Potentially a Silicon Labs driver on Mac OS X - not an issue on Linux Tried (and failed) baud rates: 128000, 153600, 230400, 256000 * Bugfix: Use atexit() to ensure a final \n is always printed at exit to work cleanly with new logging design * SGW: Support for proxy-share difficulty preferences * SSM: Propagate proxy-share difficulty changes to established connections * SSM: Track stratum connections for each proxy user * SSM: Track authorised users for each connection * SSM: Initialise proxyshare difficulty from --set pxy:diff=N * proxy: Accept --set pxy:diff=N to set preferred proxyshare difficulty * proxy: Provide a place to store desired proxyshare difficulty on a per-username basis, and copy it to SSM connections when authorising them * SSM: Track proxy share difficulties * Expose target_diff function and add pdiff_to_bdiff macro * util: double_find_precision function to identify ideal precision for a fp number * work2d: Expose WORK2D_MAX_DIVISIONS in header * Bugfix: add_pool: If no current pool set, initialise it (otherwise pool testing may start a longpoll thread which tries to access currentpool uninitialised) * devpath_to_devid: *nix: Reject anything that doesn't begin with a / * Bugfix: probe for ZeusMiner before probing for DualMiner ------------------------------------------------------------------- Tue Jul 1 20:55:37 UTC 2014 - marec@detebe.org - Update to version 4.3.0: * Update NEWS * Bugfix: zeusminer: Include config.h * Bump version to 4.3.0 * openwrt: Disable libusb via --without-libusb * lowl-pci: Silence warning on big endian * bitfury: Portability fixes, including big endian compatibility * Bugfix: bifury: strtoll already implies native endian, no need to flip it again * bitforce: Additional sanity checks, including ensuring our final queued count matches that seen by the device * configure: Clean up zeusminer logic * icarus: Use estimated hashes for hw errors (more accurate hashrate for devices where hw error values vary significantly from hashes performed) * openwrt: Provide scrypt option * Benchmark: Debuglog headers generated * Bugfix: cpu: Fix scrypt CPU miner * Bugfix: scrypt: Fix scrypt hash postprocessing for big endian * scrypt: Add unit test for scrypt_1024_1_1_256_sp * Bugfix: Display proper statline R and HW error percents for values < 1 * scrypt: Add unit tests for PBKDF2_SHA256_80_128 and salsa20_8 * Bugfix: scrypt: Correct endian handling in PBKDF2_SHA256_80_128 * Bugfix: Copy ALL the data with swap32to<native> * zeusminer: Bugfix: calculate work_division based on chips and cores * bitforce: Go back to ignoring race-triggered sanity check warnings for ZqX * bitforce: Refactor FLB and queue flush sanity checks * ccan: Update to latest version (which includes floatval option support) * bitforce: Use binary FLB protocol for Monarch * bitforce: Add bitforce_read function to deal with binary responses * Bugfix: Adapt quit function to new logging design * Workaround GCC 4.6 initialisation bugs with BYTES_INIT * Bugfix: Use int rather than ssize_t to type-match %d expectation * Minor refactor of text-only mode: avoid wasting bottom line of console, and save status lines as they are replaced by log items * Replace swabn with bswap_32mult (matching swap32yes parameters and performance) * Bugfix: zeusminer: Correctly encode job header regardless of native endian * zeusminer: support added for ZeusMiner scrypt ASICs * icarus: refactoring: commenting ICARUS_INFO members * icarus: Accept any power of two as a valid work_division * icarus: option added for Icarus-based drivers to ignore the golden nonce during probe * util: is_power_of_two function * bitforce: Support for queue id based jobs * Make work->device_id unsigned * bitforce: Support device explicit "Scan Interval" to override hardcoded sleep times * bitforce: Allow queued_max to exceed BITFORCE_MAX_QUEUED_MAX iff the firmware provides an explicit Queue Depth * bitforce: Fix bitforce_chips_to_plan_for to work beyond 32 chips * jingtian: Implement device protocol dump * jingtian: Detection-only code for new driver * lowl-spi: GPIO access functions * lowl-spi: Move knc_spi_txrx to linux_spi_txrx ------------------------------------------------------------------- Wed Jun 11 18:54:19 UTC 2014 - marec@detebe.org - Update to version 4.2.0: * Bump version to 4.2.0 * SGW: Workaround BE Cube bugs with new libmicrohttpd by forcing HTTP 1.0 response * bfg_socket wrapper to ensure sockets are close-on-exec * set_cloexec_socket: Explicitly cast SOCKET to HANDLE * rockminer: Increase hw error count for problems * rockminer: Autodetect with "R-BOX miner" or "RX-BOX miner" product strings * rockminer: Support for seeing and setting clock frequency in Manage TUI ------------------------------------------------------------------- Fri Jun 6 07:08:27 UTC 2014 - marec@detebe.org - Update to version 4.1.0: * Bugfix: Ensure variables are declared even without ADL support * Bump version to 4.1.0 * RPC: Include a list of config files loaded in "config" reply * Bugfix: Save a linked list of config files loaded so output makes sense (previously only the most recent config file was named, and errors were reported inconsistently) * README.RPC: Document Coinbase-Sig in config reply * Bugfix: Safely handle pool status line when no pools are alive * bitforce: Refactor bitforce_vcom_gets slightly to be more sane ------------------------------------------------------------------- Tue May 27 20:14:06 UTC 2014 - marec@detebe.org - Update to version 4.0.0: * Update openwrt Makefile for xz source * Bump version to 4.0.0 * Update libblkmaker to 0.4.0 * Workaround Homebrew-induced segfault during device scan * Allow hotplug with scrypt * Bugfix: test_domain_funcs: printf precisions must be int, not size_t * make-release: Copy submodules from local system * make-release: Automatically determine DLL dependencies to include ------------------------------------------------------------------- Fri Mar 21 15:27:03 UTC 2014 - marec@detebe.org - bumped to version 3.99.0 ------------------------------------------------------------------- Thu Mar 20 22:45:34 UTC 2014 - marec@detebe.org - initial packge build
Locations
Projects
Search
Status Monitor
Help
OpenBuildService.org
Documentation
API Documentation
Code of Conduct
Contact
Support
@OBShq
Terms
openSUSE Build Service is sponsored by
The Open Build Service is an
openSUSE project
.
Sign Up
Log In
Places
Places
All Projects
Status Monitor