-ck (OP)
Legendary
Offline
Activity: 4242
Merit: 1644
Ruu \o/
|
|
October 29, 2015, 10:55:29 PM |
|
I run cgminer 4.9.2 on Raspberry Pi 2 and experience constant segfaults (once in every half hour) when Antminer U3 is attached.
I use only minimum set of command line options to start cgminer, just a server URL and login/password (everything else is by default). Source code was taken from official git and compiled with gcc 4.9.2. I used CFLAGS="-O2 -Wall --march=native" compiler option and "--enable-icarus" configuration option. Linux version is 4.1.7.
Other instances of cgminer (configured for Avalon Nano ASICs) work just fine at the same time. Also, I tested same version of cgminer under MacOS X with U3, and it seems to work stable, too. So it looks like some issue with specific setup of cgminer + Antminer U3 + Raspberry Pi 2.
Is it a known problem? Is there any patch around? Anybody interested in core dumps?
If it's a crash only with RPi then it's either a specific software bug that only shows up with the RPi build - or - it's an RPi hardware problem. It's very easy to get unreliability with the RPi simply with a dodgy flash card that's hard to figure out. Get a backtrace if you can by following the directions here: http://ck.kolivas.org/apps/cgminer/debug/README-debugIf it changes every time it crashes and seems to be random memory corruption it's almost certainly an RPi problem and not a software problem. Many people have run this version of cgminer on RPi without any issues so my bet is on that.
|
Developer/maintainer for cgminer, ckpool/ckproxy, and the -ck kernel 2% Fee Solo mining at solo.ckpool.org -ck
|
|
|
cameloid
Newbie
Offline
Activity: 32
Merit: 0
|
|
October 30, 2015, 02:10:47 AM |
|
Actually, crashes do not look random, they always happen at the same place. Here is a couple of examples. 1st: GNU gdb (Raspbian 7.7.1+dfsg-5) 7.7.1 Copyright (C) 2014 Free Software Foundation, Inc. License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html> This is free software: you are free to change and redistribute it. There is NO WARRANTY, to the extent permitted by law. Type "show copying" and "show warranty" for details. This GDB was configured as "arm-linux-gnueabihf". Type "show configuration" for configuration details. For bug reporting instructions, please see: <http://www.gnu.org/software/gdb/bugs/>. Find the GDB manual and other documentation resources online at: <http://www.gnu.org/software/gdb/documentation/>. For help, type "help". Type "apropos word" to search for commands related to "word"... Reading symbols from cgminer...done. [New LWP 12843] [New LWP 12852] [New LWP 12847] [New LWP 12850] [New LWP 12844] [New LWP 12849] [New LWP 12848] [New LWP 12846] [New LWP 12851] [New LWP 12838] [New LWP 12839] [New LWP 12841] [New LWP 12840] [Thread debugging using libthread_db enabled] Using host libthread_db library "/lib/arm-linux-gnueabihf/libthread_db.so.1". Core was generated by `cgminer -o stratum+tcp://eu.stratum.bitcoin.cz:3333 -u cameloid.worker2 -p x --'. Program terminated with signal SIGSEGV, Segmentation fault. #0 0x0004e110 in usb_perform_transfer (cgpu=0xfbec88, intinfo=<optimized out>, epinfo=0, data=0x0, length=-1166715880, transferred=0x1, timeout=0, mode=0, cmd=C_REJECTED, seq=0, cancellable=24, tt=true, usbdev=<optimized out>, usbdev=<optimized out>) at usbutils.c:3097 3097 else if ((endpoint & LIBUSB_ENDPOINT_DIR_MASK) == LIBUSB_ENDPOINT_IN && *transferred) (gdb) #0 0x0004e110 in usb_perform_transfer (cgpu=0xfbec88, intinfo=<optimized out>, epinfo=0, data=0x0, length=-1166715880, transferred=0x1, timeout=0, mode=0, cmd=C_REJECTED, seq=0, cancellable=24, tt=true, usbdev=<optimized out>, usbdev=<optimized out>) at usbutils.c:3097 callback_timeout = <optimized out> err_retries = <optimized out> dev_handle = 0xf882b0 usb_epinfo = <optimized out> ut = {cgsem = {__size = "\000\000\000\000\200\000\000\000\000\000\000\000\000\000\000", __align = 0}, transfer = 0x6fd0051c, cancellable = false, list = { next = 0x0, prev = 0x0}} endpoint = 129 '\201' interrupt = <optimized out> err = 0 errn = 110 tv_start = {tv_sec = 1446167091, tv_usec = 182376} tv_finish = {tv_sec = 1446167091, tv_usec = 217754} buf = "\000\000\000\000\205\361\340N\031\234\351\\", '\000' <repeats 499 times> #1 0x00000000 in ?? () No symbol table info available. (gdb) quit
2nd: GNU gdb (Raspbian 7.7.1+dfsg-5) 7.7.1 Copyright (C) 2014 Free Software Foundation, Inc. License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html> This is free software: you are free to change and redistribute it. There is NO WARRANTY, to the extent permitted by law. Type "show copying" and "show warranty" for details. This GDB was configured as "arm-linux-gnueabihf". Type "show configuration" for configuration details. For bug reporting instructions, please see: <http://www.gnu.org/software/gdb/bugs/>. Find the GDB manual and other documentation resources online at: <http://www.gnu.org/software/gdb/documentation/>. For help, type "help". Type "apropos word" to search for commands related to "word"... Reading symbols from cgminer...done. [New LWP 13102] [New LWP 13106] [New LWP 13099] [New LWP 13107] [New LWP 13101] [New LWP 13100] [New LWP 13103] [New LWP 13105] [New LWP 13111] [New LWP 13110] [New LWP 13109] [New LWP 13108] [New LWP 13098] [Thread debugging using libthread_db enabled] Using host libthread_db library "/lib/arm-linux-gnueabihf/libthread_db.so.1". Core was generated by `cgminer -o stratum+tcp://eu.stratum.bitcoin.cz:3333 -u cameloid.worker2 -p x --'. Program terminated with signal SIGSEGV, Segmentation fault. #0 0x0004e110 in usb_perform_transfer (cgpu=0x1622c88, intinfo=<optimized out>, epinfo=0, data=0x0, length=0, transferred=0x0, timeout=0, mode=0, cmd=C_REJECTED, seq=0, cancellable=false, tt=false, usbdev=<optimized out>, usbdev=<optimized out>) at usbutils.c:3097 3097 else if ((endpoint & LIBUSB_ENDPOINT_DIR_MASK) == LIBUSB_ENDPOINT_IN && *transferred) (gdb) #0 0x0004e110 in usb_perform_transfer (cgpu=0x1622c88, intinfo=<optimized out>, epinfo=0, data=0x0, length=0, transferred=0x0, timeout=0, mode=0, cmd=C_REJECTED, seq=0, cancellable=false, tt=false, usbdev=<optimized out>, usbdev=<optimized out>) at usbutils.c:3097 callback_timeout = <optimized out> err_retries = <optimized out> dev_handle = 0x15ec2b0 usb_epinfo = <optimized out> ut = {cgsem = {__size = "\000\000\000\000\200\000\000\000\000\000\000\000\000\000\000", __align = 0}, transfer = 0x6fd0051c, cancellable = false, list = { next = 0x0, prev = 0x0}} endpoint = 129 '\201' interrupt = <optimized out> err = 0 errn = 110 tv_start = {tv_sec = 1446169096, tv_usec = 721899} tv_finish = {tv_sec = 1446169096, tv_usec = 774762} buf = "\000\354A\270\231\227\212~1\221\261\222\370?\002\000\b\000\000\000\b\000\000\000\004\000\000\000\004\000\000\000\006\000\000\000\064\000\000\000\064\000\001\000\064\000\001\000 \001\000\000 \001\000\000\005\000\000\000\004\000\000\000\003\000\000\000T\001\000\000T\001\001\000T\001\001\000\031\000\000\000\031\000\000\000\004\000\000\000\001\000\000\000\001\000\000\000\000\000\000\000\000\000\001\000\000\000\001\000\004@\001\000\004@\001\000\005\000\000\000\000\000\001\000\001\000\000\000\374N\001\000\374N\003\000\374N\003\000\324\003\000\000\300\n\000\000\006\000\000\000\000\000\001\000\002\000\000\000\bO\001\000\bO\003\000\bO\003\000\370\000\000\000\370\000\000\000\006\000\000\000\004\000\000\000\004\000\000\000p\001\000\000"... #1 0x00000000 in ?? () No symbol table info available. (gdb) quit
|
|
|
|
kano
Legendary
Offline
Activity: 4592
Merit: 1851
Linux since 1997 RedHat 4
|
|
October 31, 2015, 03:02:37 AM |
|
You'd need to actually compile it -g -O0 to get some useful info out of gdb The gdb dump shows no info about why the stack was corrupted. As a wild guess try changing line 3000 to buf[1024] instead of buf[512] This, however, can also simply move the problem if something else is causing it. (and compiling the code differently can also move a problem)
|
|
|
|
mobiletalk
Member
Offline
Activity: 98
Merit: 10
|
|
October 31, 2015, 06:04:15 PM |
|
You'd need to actually compile it -g -O0 to get some useful info out of gdb The gdb dump shows no info about why the stack was corrupted. As a wild guess try changing line 3000 to buf[1024] instead of buf[512] This, however, can also simply move the problem if something else is causing it. (and compiling the code differently can also move a problem)
Do you have skype or icq. I need help some problem with it.
|
|
|
|
kano
Legendary
Offline
Activity: 4592
Merit: 1851
Linux since 1997 RedHat 4
|
|
October 31, 2015, 10:49:03 PM |
|
You'd need to actually compile it -g -O0 to get some useful info out of gdb The gdb dump shows no info about why the stack was corrupted. As a wild guess try changing line 3000 to buf[1024] instead of buf[512] This, however, can also simply move the problem if something else is causing it. (and compiling the code differently can also move a problem)
Do you have skype or icq. I need help some problem with it. Nope. Read the README https://github.com/ckolivas/cgminer/blob/master/READMEI also don't do any linux teaching.
|
|
|
|
p3yot33at3r
|
|
November 01, 2015, 12:10:22 PM |
|
Well supposedly I'm getting an S7 - so it would be soon after that if it happens.
Hi kano, A little birdy told me that Yoshi was going to contact you about the S7 actually - maybe drop him a line to remind him? He does forget things sometimes.... Hi kano, Did you receive your S7 yet - or are you still waiting like the rest of us?
|
|
|
|
kano
Legendary
Offline
Activity: 4592
Merit: 1851
Linux since 1997 RedHat 4
|
|
November 02, 2015, 12:12:47 AM |
|
S7? What's that?
|
|
|
|
Jay_Pal
Legendary
Offline
Activity: 1493
Merit: 1003
|
|
November 02, 2015, 01:08:29 PM |
|
|
|
|
|
mm5aes
|
|
November 02, 2015, 01:15:37 PM |
|
No it's a new entity, a Nativity Calendar Miner, a new heat sink falls off every day apparently .....
|
|
|
|
Jay_Pal
Legendary
Offline
Activity: 1493
Merit: 1003
|
|
November 02, 2015, 05:19:23 PM |
|
No it's a new entity, a Nativity Calendar Miner, a new heat sink falls off every day apparently ..... Uh!!! I'd love to buy such calendar!!! But they are still too expensive for me...
|
|
|
|
fullintegrity
Member
Offline
Activity: 110
Merit: 10
|
|
November 06, 2015, 06:10:22 AM |
|
where do i find how to create the file to run the miners? a cgminer .bat file runs inside the extracted cgminer program files? on c drive? i dont get how to initiate the miner? using x3 thunder and ispace.co.uk pool, but dont know the exact .bat file. Anyone have sample, i could plug my pool credentials in or even better, a simple set of instructions? thanks a mil
|
1Lfx2Dv69BUgs5v18LtcLqFYKuiJvhxPYh
|
|
|
kano
Legendary
Offline
Activity: 4592
Merit: 1851
Linux since 1997 RedHat 4
|
|
November 09, 2015, 04:35:04 AM |
|
where do i find how to create the file to run the miners? a cgminer .bat file runs inside the extracted cgminer program files? on c drive? i dont get how to initiate the miner? using x3 thunder and ispace.co.uk pool, but dont know the exact .bat file. Anyone have sample, i could plug my pool credentials in or even better, a simple set of instructions? thanks a mil
https://github.com/ckolivas/cgminer/blob/master/README#L55
|
|
|
|
helipotte
|
|
November 09, 2015, 05:25:21 AM |
|
Anyone had any problems with poor performance of Monarchs using firmware 1.4.5? Mine all hash at about 25% of their expected rate. This problem
does not happen with the units I have using firmware 1.4.2 or 1.4.3. I have tried CGminer 4.9.2 in Windows 7 X64 and with Minera on a beaglebone.
Same issue on both platforms.
|
|
|
|
kano
Legendary
Offline
Activity: 4592
Merit: 1851
Linux since 1997 RedHat 4
|
|
November 09, 2015, 05:53:18 AM |
|
My monarch died long ago ... that someone else (not BFL) supplied me with.
BFL never contacted us about firmware changes or anything like that (or even about the Monarch until LONG after they should have) so your on your own with support of newer firmware in them since I've no idea at all what might have changed and no way to test it.
|
|
|
|
helipotte
|
|
November 10, 2015, 02:07:44 AM |
|
My monarch died long ago ... that someone else (not BFL) supplied me with.
BFL never contacted us about firmware changes or anything like that (or even about the Monarch until LONG after they should have) so your on your own with support of newer firmware in them since I've no idea at all what might have changed and no way to test it.
I am just guessing here but it seem like CGminer is treating this unit (Monarch W/1.4.5 FW) as a "single". My units all hash at about 60-70Gh. Maybe incorrect identification in some way.
|
|
|
|
kano
Legendary
Offline
Activity: 4592
Merit: 1851
Linux since 1997 RedHat 4
|
|
November 10, 2015, 02:27:55 AM |
|
My monarch died long ago ... that someone else (not BFL) supplied me with.
BFL never contacted us about firmware changes or anything like that (or even about the Monarch until LONG after they should have) so your on your own with support of newer firmware in them since I've no idea at all what might have changed and no way to test it.
I am just guessing here but it seem like CGminer is treating this unit (Monarch W/1.4.5 FW) as a "single". My units all hash at about 60-70Gh. Maybe incorrect identification in some way. Well if you can try an older version of cgminer and if an older version works and a newer one doesn't then that would relate to identification issues. However, that won't be the cause of slowing it down, that would more be a case of it not working. Looking at https://github.com/ckolivas/cgminer/blob/master/NEWSI'd guess maybe try a real old one about the time it was working for me ... I was running an old one back on 19-Jan (last day my moth worked for me) but I was using 4.6.0 (or try 4.6.1) If either of them work OK, then the problem is something that someone changed in cgminer since then.
|
|
|
|
helipotte
|
|
November 10, 2015, 02:49:11 AM |
|
Older versions Identify it as a "BAS' aka single. I think back around March, or page 798 of this thread: https://bitcointalk.org/index.php?topic=28402.msg10844936#msg10844936-ck updated the driver to properly tell the difference between a monarch and the older 65nm stuff. CGminer works great with my older monarchs but these units all have later firmware (the latest i think) and they behave poorly when using cgminer. They mine at full speed with that "other" mining software but it has very high CPU usage with many monarchs connected to one host.
|
|
|
|
-ck (OP)
Legendary
Offline
Activity: 4242
Merit: 1644
Ruu \o/
|
|
November 10, 2015, 03:08:59 AM |
|
Older versions Identify it as a "BAS' aka single. I think back around March, or page 798 of this thread: https://bitcointalk.org/index.php?topic=28402.msg10844936#msg10844936-ck updated the driver to properly tell the difference between a monarch and the older 65nm stuff. CGminer works great with my older monarchs but these units all have later firmware (the latest i think) and they behave poorly when using cgminer. They mine at full speed with that "other" mining software but it has very high CPU usage with many monarchs connected to one host. This was reported before (I believe by Os2sam) where they updated the firmware and never told us anything, breaking cgminer. We've lost all contact with BFL (which was very tenuous at the best of times anyway) and neither Kano nor I have any working BFL hardware any more to update the software, so there really isn't much choice I'm afraid.
|
Developer/maintainer for cgminer, ckpool/ckproxy, and the -ck kernel 2% Fee Solo mining at solo.ckpool.org -ck
|
|
|
Bink
Newbie
Offline
Activity: 31
Merit: 0
|
|
November 11, 2015, 08:35:02 AM |
|
Hearing that some buyers of used S3s on fLeaBay are getting much less hash on their pools, than their rigs show. Seems another go-round of asshats going out of their way to trash legit miner SW devs to line their own pockets.
Let's face it, someone running 6 S3s would have MUCH more than 1.2 TH/s on a pool with a legit cgm on their rigs...
|
|
|
|
kano
Legendary
Offline
Activity: 4592
Merit: 1851
Linux since 1997 RedHat 4
|
|
November 11, 2015, 11:04:25 AM Last edit: November 11, 2015, 11:22:47 AM by kano |
|
Hearing that some buyers of used S3s on fLeaBay are getting much less hash on their pools, than their rigs show. Seems another go-round of asshats going out of their way to trash legit miner SW devs to line their own pockets.
Let's face it, someone running 6 S3s would have MUCH more than 1.2 TH/s on a pool with a legit cgm on their rigs...
Yeah there's one, very much appreciated by me, person on my pool running an S3 24/7 mining to my account. The best thing to do with the S3 is wipe it back to factory settings (with a firmware from bitmain) then add my update: https://github.com/kanoi/cgminer-binaries/tree/master/AntS3His averages 0.44TH/s Here's an accepted share rate graph of his worker over the last 12 days: Pretty boring graph - but also pretty consistent.
|
|
|
|
|