headers
Cliff Hones | 17 Aug 2012 03:56
Picon

False BLODA warning concerning cygusb?

I am seeing the following when invoking a utility which uses libusb to
drive a USB JTAG interface dongle:

  Potential BLODA detected!  Thread function called outside of Cygwin DLL:
    C:\cygwin1.7\bin\cygusb-1.0.dll

I've had CYGWIN=detect_bloda set for some time (months), and the
utility (developed locally) has never triggered this before, so
I think a recent Cygwin update has triggered this. (Note cygusb
has not changed since last year.)

I'll happily submit full cygcheck details if that would prove useful -
the relevant details (AFAICS) are:

  Windows XP Professional Ver 5.1 Build 2600 Service Pack 3
  PROCESSOR_IDENTIFIER = 'x86 Family 15 Model 15 Stepping 0, AuthenticAMD'

  73k 2011/08/16 C:\cygwin1.7\bin\cygusb-1.0.dll - os=4.0 img=1.0 sys=4.0
                  "cygusb-1.0.dll" v0.0 ts=2011/8/16 21:45
  2789k 2012/07/20 C:\cygwin1.7\bin\cygwin1.dll - os=4.0 img=1.0 sys=4.0
                  "cygwin1.dll" v0.0 ts=2012/7/20 21:55

    Cygwin DLL version info:
        DLL version: 1.7.16
        DLL epoch: 19
        DLL old termios: 5
        DLL malloc env: 28
        Cygwin conv: 181
        API major: 0
        API minor: 262
(Continue reading)

Permalink | Reply |
headers
Corinna Vinschen | 17 Aug 2012 11:39
Favicon

Re: False BLODA warning concerning cygusb?

On Aug 17 02:56, Cliff Hones wrote:
> I am seeing the following when invoking a utility which uses libusb to
> drive a USB JTAG interface dongle:
> 
>   Potential BLODA detected!  Thread function called outside of Cygwin DLL:
>     C:\cygwin1.7\bin\cygusb-1.0.dll
> 
> I've had CYGWIN=detect_bloda set for some time (months), and the
> utility (developed locally) has never triggered this before, so
> I think a recent Cygwin update has triggered this. (Note cygusb
> has not changed since last year.)
> 
> I'll happily submit full cygcheck details if that would prove useful -
> the relevant details (AFAICS) are:

I have no problems with occassional false postives.  The detect_bloda
option is meant for diagnostic purposes only.  It's not meant to used
all the time.  It slows down *every* thread start up!

Corinna

--

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Project Co-Leader          cygwin AT cygwin DOT com
Red Hat
Permalink | Reply |
headers
Cliff Hones | 17 Aug 2012 12:47
Picon

Re: False BLODA warning concerning cygusb?

On 17/08/2012 10:39, Corinna Vinschen wrote:

> I have no problems with occassional false postives.  The detect_bloda
> option is meant for diagnostic purposes only.  It's not meant to used
> all the time.  It slows down *every* thread start up!

Thanks.  I was only reporting it as I thought I remembered a request
for false positives to be reported.  I think that's why I turned the
detect_bloda option on in the first place.  I must admit I was also
slightly curious - what is an open-source library (libusb) built
under Cygwin doing that makes it look like BLODA? [Of course I
could spend time looking at the source to find out, but time is
in short supply, so I guess I'll just have to stay curious.]

-- Cliff
Permalink | Reply |
headers
Christopher Faylor | 17 Aug 2012 16:37
Favicon

Re: False BLODA warning concerning cygusb?

On Fri, Aug 17, 2012 at 11:47:10AM +0100, Cliff Hones wrote:
>On 17/08/2012 10:39, Corinna Vinschen wrote:
>
>> I have no problems with occassional false postives.  The detect_bloda
>> option is meant for diagnostic purposes only.  It's not meant to used
>> all the time.  It slows down *every* thread start up!
>
>Thanks.  I was only reporting it as I thought I remembered a request
>for false positives to be reported.  I think that's why I turned the
>detect_bloda option on in the first place.  I must admit I was also
>slightly curious - what is an open-source library (libusb) built
>under Cygwin doing that makes it look like BLODA? [Of course I
>could spend time looking at the source to find out, but time is
>in short supply, so I guess I'll just have to stay curious.]

Likely it's not starting threads using posix calls (pthread_create) but
using Windows ones (CreateThread).  That would imply that libusb assumes
that a Cygwin compilation means "build for Windows" rather than "built
for Linux".  Given that it is trying to access usb functionality, maybe
that makes sense and it might be a permanent false positive.

cgf
Permalink | Reply |

Gmane