Re: Ryzen specific issues

From: Mike Tancsa <mike_at_sentex.net>
Date: Wed, 31 Jan 2018 08:49:29 -0500
On 1/31/2018 8:37 AM, Mike Tancsa wrote:
> Hi folks,
> 	I have been running into some deadlock issues on Ryzen boards.  The
> most easy way to trigger the problem is to compile net/samba47 from the
> ports.
> 
> https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=225584
> 
> Varying the load sometimes allows the compile to finish, but most of the
> time the compile deadlocks.  I have seen this both on RELENG_11 as well
> as HEAD from r328611
> 
> Any ideas how to possibly work around this ?

sysctl kern.eventtimer.periodic=1

doesnt seem to help on HEAD

First build attempt and it hangs in a similar / same way

G -D_SAMBA_BUILD_=4 -DHAVE_CONFIG_H=1 -D_GNU_SOURCE=1
-D_XOPEN_SOURCE_EXTENDED=1
default/source4/heimdal/lib/hx509/asn1_ocsp_asn1.c -c -o
default/source4/heimdal/lib/hx509/asn1_ocsp_asn1_92.o
[1008/3804] Compiling source4/heimdal/lib/hcrypto/libtommath/bncore.c
runner cc -pipe -g -ggdb -gdwarf-2 -gstrict-dwarf -DLIBICONV_PLUG
-fno-color-diagnostics -D_FUNCTION_DEF -g -fstack-protector
-DLDAP_DEPRECATED -fno-strict-aliasing -fno-omit-frame-pointer -fPIC
-D_REENTRANT -D_POSIX_PTHREAD_SEMANTICS -DSTATIC_tommath_MODULES=NULL
-DSTATIC_tommath_MODULES_PROTO=extern void
__tommath_dummy_module_proto(void) -MD -Idefault/source4/heimdal_build
-I../source4/heimdal_build
-Idefault/source4/heimdal/lib/hcrypto/libtommath
-I../source4/heimdal/lib/hcrypto/libtommath -Idefault/include/public
-I../include/public -Idefault/source4 -I../source4 -Idefault/lib
-I../lib -Idefault/source4/lib -I../source4/lib
-Idefault/source4/include -I../source4/include -Idefault/include
-I../include -Idefault/lib/replace -I../lib/replace -Idefault -I..
-I/usr/local/include -I/usr/local/include -DLIBICONV_PLUG
-D_SAMBA_BUILD_=4 -DHAVE_CONFIG_H=1 -D_GNU_SOURCE=1
-D_XOPEN_SOURCE_EXTENDED=1
../source4/heimdal/lib/hcrypto/libtommath/bncore.c -c -o
default/source4/heimdal/lib/hcrypto/libtommath/bncore_83.o

ctrl+t

load: 3.14  cmd: sh 94611 [wait] 490.27r 0.00u 0.00s 0% 2240k
make: Working in: /usr/ports/net/samba47
make[1]: Working in: /usr/ports/net/samba47





load: 3.14  cmd: sh 94611 [wait] 490.27r 0.00u 0.00s 0% 2240k
make: Working in: /usr/ports/net/samba47
make[1]: Working in: /usr/ports/net/samba47


root_at_amdtestr12:/tmp # procstat -kk 14062
  PID    TID COMM                TDNAME              KSTACK

14062 100933 python2.7           -                   mi_switch+0x18c
sleepq_switch+0x10d sleepq_catch_signals+0x300 sleepq_wait_sig+0xf
_sleep+0x359 umtxq_sleep+0x163 do_sem2_wait+0x426
__umtx_op_sem2_wait+0x4b amd64_syscall+0x79b fast_syscall_common+0xfc
14062 101206 python2.7           -                   mi_switch+0x18c
sleepq_switch+0x10d sleepq_catch_signals+0x300 sleepq_wait_sig+0xf
_sleep+0x359 umtxq_sleep+0x163 do_lock_umutex+0x6e3
__umtx_op_wait_umutex+0x48 amd64_syscall+0x79b fast_syscall_common+0xfc
14062 101207 python2.7           -                   mi_switch+0x18c
sleepq_switch+0x10d sleepq_catch_signals+0x300 sleepq_wait_sig+0xf
_sleep+0x359 umtxq_sleep+0x163 do_sem2_wait+0x426
__umtx_op_sem2_wait+0x4b amd64_syscall+0x79b fast_syscall_common+0xfc
14062 101208 python2.7           -                   mi_switch+0x18c
sleepq_switch+0x10d sleepq_catch_signals+0x300 sleepq_wait_sig+0xf
_sleep+0x359 umtxq_sleep+0x163 do_sem2_wait+0x426
__umtx_op_sem2_wait+0x4b amd64_syscall+0x79b fast_syscall_common+0xfc
14062 101209 python2.7           -                   mi_switch+0x18c
sleepq_switch+0x10d sleepq_catch_signals+0x300 sleepq_wait_sig+0xf
_sleep+0x359 umtxq_sleep+0x163 do_sem2_wait+0x426
__umtx_op_sem2_wait+0x4b amd64_syscall+0x79b fast_syscall_common+0xfc
14062 101210 python2.7           -                   mi_switch+0x18c
sleepq_switch+0x10d sleepq_catch_signals+0x300 sleepq_wait_sig+0xf
_sleep+0x359 umtxq_sleep+0x163 do_sem2_wait+0x426
__umtx_op_sem2_wait+0x4b amd64_syscall+0x79b fast_syscall_common+0xfc
14062 101211 python2.7           -                   mi_switch+0x18c
sleepq_switch+0x10d sleepq_catch_signals+0x300 sleepq_wait_sig+0xf
_sleep+0x359 umtxq_sleep+0x163 do_sem2_wait+0x426
__umtx_op_sem2_wait+0x4b amd64_syscall+0x79b fast_syscall_common+0xfc
14062 101212 python2.7           -                   mi_switch+0x18c
sleepq_switch+0x10d sleepq_catch_signals+0x300 sleepq_wait_sig+0xf
_sleep+0x359 umtxq_sleep+0x163 do_sem2_wait+0x426
__umtx_op_sem2_wait+0x4b amd64_syscall+0x79b fast_syscall_common+0xfc
14062 101213 python2.7           -                   mi_switch+0x18c
sleepq_switch+0x10d sleepq_catch_signals+0x300 sleepq_wait_sig+0xf
_sleep+0x359 umtxq_sleep+0x163 do_sem2_wait+0x426
__umtx_op_sem2_wait+0x4b amd64_syscall+0x79b fast_syscall_common+0xfc
14062 101214 python2.7           -                   mi_switch+0x18c
sleepq_switch+0x10d sleepq_catch_signals+0x300 sleepq_wait_sig+0xf
_sleep+0x359 umtxq_sleep+0x163 do_sem2_wait+0x426
__umtx_op_sem2_wait+0x4b amd64_syscall+0x79b fast_syscall_common+0xfc
14062 101215 python2.7           -                   mi_switch+0x18c
sleepq_switch+0x10d sleepq_catch_signals+0x300 sleepq_wait_sig+0xf
_sleep+0x359 umtxq_sleep+0x163 do_sem2_wait+0x426
__umtx_op_sem2_wait+0x4b amd64_syscall+0x79b fast_syscall_common+0xfc
14062 101216 python2.7           -                   mi_switch+0x18c
sleepq_switch+0x10d sleepq_catch_signals+0x300 sleepq_wait_sig+0xf
_sleep+0x359 umtxq_sleep+0x163 do_sem2_wait+0x426
__umtx_op_sem2_wait+0x4b amd64_syscall+0x79b fast_syscall_common+0xfc
14062 101217 python2.7           -                   mi_switch+0x18c
sleepq_switch+0x10d sleepq_catch_signals+0x300 sleepq_wait_sig+0xf
_sleep+0x359 umtxq_sleep+0x163 do_sem2_wait+0x426
__umtx_op_sem2_wait+0x4b amd64_syscall+0x79b fast_syscall_common+0xfc
root_at_amdtestr12:/tmp # procstat -t 14062
  PID    TID COMM                TDNAME              CPU  PRI STATE
WCHAN
14062 100933 python2.7           -                    -1  152 sleep
usem
14062 101206 python2.7           -                    -1  126 sleep
umtxn
14062 101207 python2.7           -                    -1  124 sleep
usem
14062 101208 python2.7           -                    -1  125 sleep
usem
14062 101209 python2.7           -                    -1  125 sleep
usem
14062 101210 python2.7           -                    -1  124 sleep
usem
14062 101211 python2.7           -                    -1  126 sleep
usem
14062 101212 python2.7           -                    -1  123 sleep
usem
14062 101213 python2.7           -                    -1  125 sleep
usem
14062 101214 python2.7           -                    -1  125 sleep
usem
14062 101215 python2.7           -                    -1  125 sleep
usem
14062 101216 python2.7           -                    -1  125 sleep
usem
14062 101217 python2.7           -                    -1  126 sleep
usem


-- 
-------------------
Mike Tancsa, tel +1 519 651 3400
Sentex Communications, mike_at_sentex.net
Providing Internet services since 1994 www.sentex.net
Cambridge, Ontario Canada
Received on Wed Jan 31 2018 - 12:49:32 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:41:14 UTC