summary refs log tree commit diff stats
path: root/results/classifier/gemma3:27b/runtime/1763536
diff options
context:
space:
mode:
authorChristian Krinitsin <mail@krinitsin.com>2025-07-06 16:43:19 +0000
committerChristian Krinitsin <mail@krinitsin.com>2025-07-06 16:43:19 +0000
commit238ec2b7cc1557d6f34c33cc482e4d0cd3e266dd (patch)
treecd8a1b75ba7b3543eb7fe6857f408e7be4d9fd0b /results/classifier/gemma3:27b/runtime/1763536
parent96049c939b1916d80532630d63c14e04d5244f1d (diff)
downloademulator-bug-study-238ec2b7cc1557d6f34c33cc482e4d0cd3e266dd.tar.gz
emulator-bug-study-238ec2b7cc1557d6f34c33cc482e4d0cd3e266dd.zip
add results
Diffstat (limited to 'results/classifier/gemma3:27b/runtime/1763536')
-rw-r--r--results/classifier/gemma3:27b/runtime/176353686
1 files changed, 86 insertions, 0 deletions
diff --git a/results/classifier/gemma3:27b/runtime/1763536 b/results/classifier/gemma3:27b/runtime/1763536
new file mode 100644
index 00000000..d18dcbcc
--- /dev/null
+++ b/results/classifier/gemma3:27b/runtime/1763536
@@ -0,0 +1,86 @@
+
+
+
+go build fails under qemu-ppc64le-static (qemu-user)
+
+I am using qemu-user (built static) in a docker container environment.  When running multi-threaded go commands in the container (go build for example) the process may hang, report segfaults or other errors.  I built qemu-ppc64le from the upstream git (master).
+
+I see the problem running on a multi core system with Intel i7 processors.
+# cat /proc/cpuinfo | grep "model name"
+model name	: Intel(R) Core(TM) i7-2760QM CPU @ 2.40GHz
+model name	: Intel(R) Core(TM) i7-2760QM CPU @ 2.40GHz
+model name	: Intel(R) Core(TM) i7-2760QM CPU @ 2.40GHz
+model name	: Intel(R) Core(TM) i7-2760QM CPU @ 2.40GHz
+model name	: Intel(R) Core(TM) i7-2760QM CPU @ 2.40GHz
+model name	: Intel(R) Core(TM) i7-2760QM CPU @ 2.40GHz
+model name	: Intel(R) Core(TM) i7-2760QM CPU @ 2.40GHz
+model name	: Intel(R) Core(TM) i7-2760QM CPU @ 2.40GHz
+
+Steps to reproduce:
+1) Build qemu-ppc64le as static and copy into docker build directory named it qemu-ppc64le-static.
+
+2) Add hello.go to docker build dir.
+
+package main
+import "fmt"
+func main() {
+	fmt.Println("hello world")
+}
+
+3) Create the Dockerfile from below:
+
+FROM ppc64le/golang:1.10.1-alpine3.
+COPY qemu-ppc64le-static /usr/bin/
+COPY hello.go /go
+
+4) Build container
+$ docker build -t qemutest -f Dockerfile ./go 
+
+5) Run test
+$ docker run -it qemutest
+
+/go # /usr/bin/qemu-ppc64le-static --version
+qemu-ppc64le version 2.11.93 (v2.12.0-rc3-dirty)
+Copyright (c) 2003-2017 Fabrice Bellard and the QEMU Project developers
+
+/go # go version
+go version go1.10.1 linux/ppc64le
+
+/go # go build hello.go
+fatal error: fatal error: stopm holding locksunexpected signal during runtime execution
+
+panic during panic
+[signal SIGSEGV: segmentation violation code=0x1 addr=0x0 pc=0x1003528c]
+
+runtime stack:
+runtime: unexpected return pc for syscall.Syscall6 called from 0xc42007f500
+stack: frame={sp:0xc4203be840, fp:0xc4203be860} stack=[0x4000b7ecf0,0x4000b928f0)
+
+syscall.Syscall6(0x100744e8, 0x3d, 0xc42050c140, 0x20, 0x18, 0x10422b80, 0xc4203be968[signal , 0x10012d88SIGSEGV: segmentation violation, 0xc420594000 code=, 0x00x1 addr=0x0 pc=0x1003528c)
+]
+
+runtime stack:
+	/usr/local/go/src/syscall/asm_linux_ppc64x.s:61runtime.throw(0x10472d19, 0x13)
+ +	/usr/local/go/src/runtime/panic.go:0x6c616 +0x68
+
+
+runtime.stopm()
+	/usr/local/go/src/runtime/proc.go:1939goroutine  +10x158
+ [runtime.exitsyscall0semacquire(0xc42007f500)
+	/usr/local/go/src/runtime/proc.go:3129 +]:
+0x130
+runtime.mcall(0xc42007f500)
+	/usr/local/go/src/runtime/asm_ppc64x.s:183 +0x58sync.runtime_Semacquire
+(0xc4201fab1c)
+	/usr/local/go/src/runtime/sema.go:56 +0x38
+
+----
+Note the results may differ between attempts,  hangs and other faults sometimes happen.
+----
+If I run "go: single threaded I don't see the problem, for example:
+
+/go # GOMAXPROCS=1 go build -p 1 hello.go 
+/go # ./hello
+hello world
+
+I see the same issue with arm64.  I don't think this is a go issue, but don't have a real evidence to prove that.  This problem looks similar to other problem I have seen reported against qemu running multi-threaded applications.
\ No newline at end of file