[ANNOUNCEMENT] clamav 0.99.1-1

[ANNOUNCEMENT] clamav 0.99.1-1

[Yaakov Selkowitz]

The following packages have been uploaded to the Cygwin distribution:

* clamav-0.99.1-1
* clamav-doc-0.99.1-1
* libclamav7-0.99.1-1
* libclamav-devel-0.99.1-1
* clamav-db-main-57-1
* clamav-db-daily-21477.253-1

Clam AntiVirus is a GPL anti-virus toolkit for UNIX, featuring command-line 
scanner, fast, multi-threaded daemon, database updater with support for 
digital signatures, virus scanner C library, detection of over 20000 
viruses, worms and trojans, built-in support for RAR (2.0), Zip, Gzip, 
Bzip2, built-in support for Mbox, Maildir and raw mail files

This is an update to the latest upstream release.

--
Yaakov

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: [ANNOUNCEMENT] clamav 0.99.1-1

[Mike Bonnet]

On 3/30/16 10:08 AM, Yaakov Selkowitz wrote:
> The following packages have been uploaded to the Cygwin distribution:
>
> * clamav-0.99.1-1
> * clamav-doc-0.99.1-1
> * libclamav7-0.99.1-1
> * libclamav-devel-0.99.1-1
> * clamav-db-main-57-1
> * clamav-db-daily-21477.253-1
>
> Clam AntiVirus is a GPL anti-virus toolkit for UNIX, featuring command-line
> scanner, fast, multi-threaded daemon, database updater with support for
> digital signatures, virus scanner C library, detection of over 20000
> viruses, worms and trojans, built-in support for RAR (2.0), Zip, Gzip,
> Bzip2, built-in support for Mbox, Maildir and raw mail files
>
> This is an update to the latest upstream release.

I'm seeing segfaults with this version when recursively scanning large 
directories. Running Cygwin x86_64 on Windows Server 2012.

In the past this happened when some 32-bit files were in the source tree 
when the 64-bit version was built. It's possible this happened again. 
Any chance we could get a new 64-bit build?


--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: [ANNOUNCEMENT] clamav 0.99.1-1

[Yaakov Selkowitz]

On 2016-05-11 09:35, Mike Bonnet wrote:
> I'm seeing segfaults with this version when recursively scanning large
> directories. Running Cygwin x86_64 on Windows Server 2012.

But not with 0.99?  Have you been able to reproduce this on any other 
systems?

> In the past

When?

> this happened when some 32-bit files were in the source tree
> when the 64-bit version was built. It's possible this happened again.

Extremely unlikely.  cygport makes 32-bit and 64-bit builds in their own 
directories.

> Any chance we could get a new 64-bit build?

We'd need to find the real cause of this before it would be of any help. 
  I'd start with updating to 0.99.2, however we borrow Fedora's -norar 
sources but they haven't bumped yet (#1333949).

-- 
Yaakov

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: [ANNOUNCEMENT] clamav 0.99.1-1

[Mike Bonnet]

On 5/11/16 8:18 AM, Yaakov Selkowitz wrote:
> On 2016-05-11 09:35, Mike Bonnet wrote:
>> I'm seeing segfaults with this version when recursively scanning large
>> directories. Running Cygwin x86_64 on Windows Server 2012.
>
> But not with 0.99?  Have you been able to reproduce this on any other
> systems?

Only saw it with 0.99.1. Yes, saw it on several systems, all running 0.99.1.

>> In the past
>
> When?

This was several years ago.

>> this happened when some 32-bit files were in the source tree
>> when the 64-bit version was built. It's possible this happened again.
>
> Extremely unlikely.  cygport makes 32-bit and 64-bit builds in their own
> directories.
>
>> Any chance we could get a new 64-bit build?
>
> We'd need to find the real cause of this before it would be of any help.
>  I'd start with updating to 0.99.2, however we borrow Fedora's -norar
> sources but they haven't bumped yet (#1333949).

I'll retry with 0.99.2 when it's available.


--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: [ANNOUNCEMENT] clamav 0.99.1-1

[Warren Young]

On May 11, 2016, at 10:10 AM, Mike Bonnet wrote:
> 
> On 5/11/16 8:18 AM, Yaakov Selkowitz wrote:
>> On 2016-05-11 09:35, Mike Bonnet wrote:
>> 
>>> Any chance we could get a new 64-bit build?
>> 
>> We'd need to find the real cause of this before it would be of any help.
>> I'd start with updating to 0.99.2, however we borrow Fedora's -norar
>> sources but they haven't bumped yet (#1333949).
> 
> I'll retry with 0.99.2 when it's available.

Or you could rebuild 0.99.1 from the sources and see if that fixes the symptom.  If it does, you have a reason to reject Yaakov’s answer.  If not, Yaakov was right to disbelieve your hypothesis.
--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: [ANNOUNCEMENT] clamav 0.99.1-1

[Mike Bonnet]

On 5/11/16 9:14 AM, Warren Young wrote:
> On May 11, 2016, at 10:10 AM, Mike Bonnet wrote:
>>
>> On 5/11/16 8:18 AM, Yaakov Selkowitz wrote:
>>> On 2016-05-11 09:35, Mike Bonnet wrote:
>>>
>>>> Any chance we could get a new 64-bit build?
>>>
>>> We'd need to find the real cause of this before it would be of any help.
>>> I'd start with updating to 0.99.2, however we borrow Fedora's -norar
>>> sources but they haven't bumped yet (#1333949).
>>
>> I'll retry with 0.99.2 when it's available.
>
> Or you could rebuild 0.99.1 from the sources and see if that fixes the symptom.  If it does, you have a reason to reject YaakovÂ’s answer.  If not, Yaakov was right to disbelieve your hypothesis.

The build issue was a red herring, sorry about that. It turns out a 
specific file, xdate.exe, is causing clamscan to segfault. This is an 
old file, from 2012, used for date formatting, but it has just started 
causing problems. I've reproduced the segfault in 0.99.1-1 and 0.98.7-2 
running on Cygwin64 on Windows Server 2012, but didn't go any further 
back than that. I've verified that there are no viruses in the file 
(according to virustotal.com):

https://virustotal.com/en/file/937800ed6b0408b9dba4e4dc507cd0b1962d3adb6285947db520009d19cbe24e/analysis/1463097443/

The file is here (it's listed as "free for any use"):

http://people.redhat.com/mikeb/clamscan-segfault/xdate.exe

The output from running "clamscan --debug xdate.exe" is here:

http://people.redhat.com/mikeb/clamscan-segfault/clamscan-debug-xdate.log

and the stackdump is here:

http://people.redhat.com/mikeb/clamscan-segfault/clamscan.exe.stackdump

Sounds like this should be reported to Clamav upstream. Is anyone 
involved with Clamav already, that could report this to the right 
people, or do I need to start signing up for mailing lists? :)


--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: [ANNOUNCEMENT] clamav 0.99.1-1

[Yaakov Selkowitz]

On 2016-05-13 13:51, Mike Bonnet wrote:
> The build issue was a red herring, sorry about that. It turns out a
> specific file, xdate.exe, is causing clamscan to segfault. This is an
> old file, from 2012, used for date formatting, but it has just started
> causing problems. I've reproduced the segfault in 0.99.1-1 and 0.98.7-2
> running on Cygwin64 on Windows Server 2012, but didn't go any further
> back than that.

Mike,

Thanks for narrowing this down to a test case.

I was able to reproduce this too on both Windows 7 (sporadically) and 
Server 2012 (consistently), but only after updating the "daily" 
signatures from 21513 to 21539.  Upgrading to 0.99.2 locally did not 
help, but disabling bytecode (e.g. --bytecode=no in CLI) seems to be a 
workaround.

Therefore, I suspect its either a bug in LLVM JIT (certainly possible) 
or in register handling in Cygwin.  Either way, it's going to be some 
"fun" to track this down further.

> Sounds like this should be reported to Clamav upstream. Is anyone
> involved with Clamav already, that could report this to the right
> people, or do I need to start signing up for mailing lists? :)

There is a bugzilla, but I suspect this will have to be fixed on our end 
somewhere.

-- 
Yaakov

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple