[Mono-list] Frequent segfaults in c# running on mono

Matt Calder mvcalder at gmail.com
Fri Nov 27 12:10:50 UTC 2015 

River,

We see seg-faults in multi-threaded code on Ubuntu 14.04 and 14.10. These
can be reproduced relatively easily. We believe it is this bug:

https://bugzilla.xamarin.com/show_bug.cgi?id=29692

The discussions there will lead you to some discussions about kernel
versions. We still see this bug using 3.13 or 3.19 kernels.

Matt


On Fri, Nov 27, 2015 at 5:31 AM, River Satya <river.satya at gmail.com> wrote:

>
>  Stack trace flavour 3:
>
> --
> :Stacktrace:
> -
> -  at <unknown> <0xffffffff>
> -  at (wrapper managed-to-native) System.Delegate.CreateDelegate_internal
> (System.Type,object,System.Reflection.MethodInfo,bool) <IL 0x00010,
> 0xffffffff>
> -  at System.Delegate.CreateDelegate
> (System.Type,object,System.Reflection.MethodInfo,bool,bool) <0x0071a>
> -  at System.Delegate.CreateDelegate
> (System.Type,object,System.Reflection.MethodInfo) <0x00021>
> -  at System.Reflection.Emit.DynamicMethod.CreateDelegate
> (System.Type,object) <0x00043>
> -  at System.Linq.Expressions.Compiler.LambdaCompiler.CreateDelegate ()
> <IL 0x00022, 0x00094>
> -  at System.Linq.Expressions.Compiler.LambdaCompiler.Compile
> (System.Linq.Expressions.LambdaExpression,System.Runtime.CompilerServices.DebugInfoGenerator)
> <IL 0x0001e, 0x000d6>
> -  at System.Linq.Expressions.Expression`1.Compile () <IL 0x00002, 0x00016>
> -  at ServiceStack.OrmLite.SqlExpressionVisitor`1.VisitMemberAccess
> (System.Linq.Expressions.MemberExpression) <IL 0x000fe, 0x003a6>
> -  at ServiceStack.OrmLite.SqlExpressionVisitor`1.Visit
> (System.Linq.Expressions.Expression) <IL 0x000e3, 0x000e1>
> -  at ServiceStack.OrmLite.SqlExpressionVisitor`1.VisitBinary
> (System.Linq.Expressions.BinaryExpression) <IL 0x0016d, 0x00736>
> -  at ServiceStack.OrmLite.SqlExpressionVisitor`1.Visit
> (System.Linq.Expressions.Expression) <IL 0x000fd, 0x0016b>
> -  at ServiceStack.OrmLite.SqlExpressionVisitor`1.VisitLambda
> (System.Linq.Expressions.LambdaExpression) <IL 0x0005a, 0x00169>
> -  at ServiceStack.OrmLite.SqlExpressionVisitor`1.Visit
> (System.Linq.Expressions.Expression) <IL 0x000d6, 0x0009c>
> -  at
> ServiceStack.OrmLite.SqlExpressionVisitor`1.ProcessInternalExpression ()
> <IL 0x0001a, 0x00073>
> -  at ServiceStack.OrmLite.SqlExpressionVisitor`1.And
> (System.Linq.Expressions.Expression`1<System.Func`2<T, bool>>) <IL 0x00027,
> 0x000f4>
> -  at ServiceStack.OrmLite.SqlExpressionVisitor`1.Where
> (System.Linq.Expressions.Expression`1<System.Func`2<T, bool>>) <IL 0x00005,
> 0x00028>
> -  at ServiceStack.OrmLite.ReadExtensions.Select<T>
> (System.Data.IDbCommand,System.Linq.Expressions.Expression`1<System.Func`2<T,
> bool>>) <IL 0x0000d, 0x0006e>
> -  at
> ServiceStack.OrmLite.ReadConnectionExtensions/<>c__DisplayClassd`1.<Select>b__c
> (System.Data.IDbCommand) <IL 0x00007, 0x00046>
> --
> :Native stacktrace:
> -
> -       /usr/bin/mono() [0x4b23dc]
> -       /lib/x86_64-linux-gnu/libpthread.so.0(+0x10340) [0x7fc516460340]
> -       /lib/x86_64-linux-gnu/libc.so.6(gsignal+0x39) [0x7fc5160c1cc9]
> -       /lib/x86_64-linux-gnu/libc.so.6(abort+0x148) [0x7fc5160c50d8]
> -       /usr/bin/mono() [0x629999]
> -       /usr/bin/mono() [0x629ba7]
> -       /usr/bin/mono() [0x629cf6]
> -       /usr/bin/mono(mono_class_from_mono_type+0x37) [0x519b07]
> -       /usr/bin/mono() [0x42a69a]
> -       /usr/bin/mono() [0x42b301]
> -       /usr/bin/mono() [0x42c89f]
> -       /usr/bin/mono() [0x42d29b]
> -       /usr/bin/mono() [0x5398f8]
> -       [0x4115956d]
> -
> -Debug info from gdb:
> -
> -Could not attach to process.  If your uid matches the uid of the target
> -process, check the setting of /proc/sys/kernel/yama/ptrace_scope, or try
> --
>
>
> Stack trace flavour 4 (no non-native stacktrace):
> These vary a bit, but look pretty much like the below.
>
> --
> :Stacktrace:
> -
> -
> :Native stacktrace:
> -
> -       /usr/bin/mono() [0x4b23dc]
> -       /lib/x86_64-linux-gnu/libpthread.so.0(+0x10340) [0x7fb78e49c340]
> -       /lib/x86_64-linux-gnu/libc.so.6(gsignal+0x39) [0x7fb78e0fdcc9]
> -       /lib/x86_64-linux-gnu/libc.so.6(abort+0x148) [0x7fb78e1010d8]
> -       /usr/bin/mono() [0x629999]
> -       /usr/bin/mono() [0x629ba7]
> -       /usr/bin/mono() [0x629cf6]
> -       /usr/bin/mono() [0x6194dc]
> -       /usr/bin/mono() [0x422086]
> -       /usr/bin/mono() [0x5a7012]
> -       /usr/bin/mono() [0x5b0720]
> -       /usr/bin/mono() [0x5a1d99]
> -       /usr/bin/mono() [0x5a1dd0]
> -       /usr/bin/mono() [0x5a226d]
> -       /usr/bin/mono() [0x5875f8]
> -       /usr/bin/mono() [0x623b66]
> -       /lib/x86_64-linux-gnu/libpthread.so.0(+0x8182) [0x7fb78e494182]
> -       /lib/x86_64-linux-gnu/libc.so.6(clone+0x6d) [0x7fb78e1c147d]
> -
>
> On 27 November 2015 at 20:31, River Satya <river.satya at gmail.com> wrote:
>
>> Hi mono list,
>>
>> We see a lot of segfaults (~10 / day) in our program running on Ubuntu
>> Wheezy. I was hoping that upgrading to mono 4.2 would resolve it, but we're
>> still seeing them.
>>
>> The stack traces take a few different forms, so it's possible that
>> they're actually separate issues.
>>
>> We mostly see #1 and #4. #2 and #3 have only been seen in isolation, but
>> have more interesting stack traces.
>>
>> These look to me like mono bugs. Are any of these known issues? Is there
>> anything I can do to either work around them and/or assist in debugging
>> them? They're currently causing significant problems in a mission critical
>> piece of software for my client.
>>
>> (email split into pieces as the full version was rejected by the maillist
>> server).
>>
>> Thanks,
>>
>> River
>>
>>>
>>> Stack trace 1 (empty trace):
>>> :Stacktrace:
>>> -
>>> -
>>> :Native stacktrace:
>>> -
>>> Stack trace 2:
>>> --
>>> :Stacktrace:
>>> -
>>> -  at <unknown> <0xffffffff>
>>> -  at (wrapper managed-to-native)
>>> object.__icall_wrapper_mono_gc_alloc_vector (intptr,intptr,intptr) <IL
>>> 0x0000f, 0xffffffff>
>>> -  at (wrapper alloc) object.AllocVector (intptr,intptr) <IL 0x00088,
>>> 0xffffffff>
>>> -  at System.Xml.XmlUtf8RawTextWriter..ctor
>>> (System.IO.Stream,System.Xml.XmlWriterSettings) <IL 0x00039, 0x000cf>
>>> -  at System.Xml.XmlWriterSettings.CreateWriter (System.IO.Stream) <IL
>>> 0x00067, 0x00117>
>>> -  at System.Xml.XmlWriter.Create
>>> (System.IO.Stream,System.Xml.XmlWriterSettings) <IL 0x0000f, 0x00066>
>>> -  at
>>>  <snip> rest of stack trace is in our code</snip>
>>> --
>>> :Native stacktrace:
>>> -
>>> -       /usr/bin/mono() [0x49cf0c]
>>> -       /lib/x86_64-linux-gnu/libpthread.so.0(+0x10340) [0x7fb9e29e9340]
>>> -       /lib/x86_64-linux-gnu/libc.so.6(gsignal+0x39) [0x7fb9e264acc9]
>>> -       /lib/x86_64-linux-gnu/libc.so.6(abort+0x148) [0x7fb9e264e0d8]
>>> -       /usr/bin/mono() [0x62a329]
>>> -       /usr/bin/mono() [0x62a537]
>>> -       /usr/bin/mono() [0x62a5e2]
>>> -       /usr/bin/mono() [0x5ea7a6]
>>> -       /usr/bin/mono() [0x5ecfe0]
>>> -       /usr/bin/mono() [0x5f5af2]
>>> -       /usr/bin/mono() [0x5f67aa]
>>> -       /usr/bin/mono() [0x5ecb03]
>>> -       /usr/bin/mono() [0x5dbf76]
>>> -       /usr/bin/mono() [0x5e380d]
>>> -       /usr/bin/mono() [0x5f99f8]
>>> -       /usr/bin/mono() [0x5e5f63]
>>> -       /usr/bin/mono() [0x5e805f]
>>> -       /usr/bin/mono() [0x5db55d]
>>> -       /usr/bin/mono() [0x5ca11e]
>>> --
>>
>>
>>
>>
>
> _______________________________________________
> Mono-list maillist  -  Mono-list at lists.ximian.com
> http://lists.ximian.com/mailman/listinfo/mono-list
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ximian.com/pipermail/mono-list/attachments/20151127/3cdfdb05/attachment.html>

More information about the Mono-list mailing list