More dumpeths

I’ve attached a core-dump “bt” trace that was produced exiting GRC–I
wasn’t even actively running a flow-graph. Interestingly enough, PMT
seems to have been involved at the very top of the stack–just like
the dumps I’ve been getting exiting from running flow-graphs.

On Sat, Apr 14, 2012 at 10:17, Marcus D. Leech [email protected]
wrote:

I’ve attached a core-dump “bt” trace that was produced exiting GRC–I wasn’t
even actively running a flow-graph. Interestingly enough, PMT
seems to have been involved at the very top of the stack–just like the
dumps I’ve been getting exiting from running flow-graphs.

Since you can reliably generate these failures, could you use git
bisect to narrow it down?

Also, can you document gcc version, boost version, OS, etc.?

Johnathan

On Sat, Apr 14, 2012 at 10:40, Johnathan C.
[email protected] wrote:

Also, can you document gcc version, boost version, OS, etc.?

I see this from your other email, thanks.

Johnathan

On 04/14/2012 01:41 PM, Johnathan C. wrote:

On Sat, Apr 14, 2012 at 10:40, Johnathan C.
[email protected] wrote:

Also, can you document gcc version, boost version, OS, etc.?
I see this from your other email, thanks.

Johnathan

Somewhere between commit 2f73fe22c5e1591242a9556221ec8487170a7b00 and
27dd737afdab4123e40a1b886635b387b3c063c6

Those commits were only a couple of commits apart, but I didn’t bisect
it further.


Marcus L.
Principal Investigator
Shirleys Bay Radio Astronomy Consortium

On Sat, Apr 14, 2012 at 12:11, Johnathan C.
[email protected] wrote:

On Sat, Apr 14, 2012 at 11:57, Marcus D. Leech [email protected] wrote:

Somewhere between commit 2f73fe22c5e1591242a9556221ec8487170a7b00 and
27dd737afdab4123e40a1b886635b387b3c063c6

Both of these commits fail, correct? What was the last good one you tried?

In particular, I’d like to confirm that f88b8cf5 (3.5 API) or c787fb66
(3.6 API) are working correctly. It’s that last time we changed the
PMT library.

Johnathan

On 04/14/2012 03:18 PM, Johnathan C. wrote:

Johnathan

Both of those commits fail to build for me in gr-trellis


Marcus L.
Principal Investigator
Shirleys Bay Radio Astronomy Consortium

On Sat, Apr 14, 2012 at 11:57, Marcus D. Leech [email protected]
wrote:

Somewhere between commit 2f73fe22c5e1591242a9556221ec8487170a7b00 and
27dd737afdab4123e40a1b886635b387b3c063c6

Both of these commits fail, correct? What was the last good one you
tried?

Johnathan

On 04/14/2012 04:03 PM, Johnathan C. wrote:

of the working commit 27dd737a.
This commit fails.

Please try commit c787fb66, which is the merge on the former ‘next’
branch of a PMT library modification (this is one of the ones that
failed in trellis for you, hopefully you can make it work.)

This commit fails. But I had to roll-back to an older version of UHD to
get it to build. Still failed though.

If this last fails, please try branch ‘test/undo-pmt-deleter’, which
I’ve pushed to the main gnuradio repository.

That works. No segfaults on exit.

I appreciate all the testing, since I can’t duplicate the problem, and
obviously this is important to fix as soon as possible.

Johnathan


Marcus L.
Principal Investigator
Shirleys Bay Radio Astronomy Consortium

On Sat, Apr 14, 2012 at 13:30, Marcus D. Leech [email protected]
wrote:

That works. No segfaults on exit.

Excellent, thanks. As this is in the 3.5.3 release, I’m going to
rebase this ‘fix’ off the last 3.5.3 maint, but hold off merging it.
I’m hoping Josh can supply an actual fix soon so we don’t have to take
that functionality out altogether.

Johnathan

On Sat, Apr 14, 2012 at 12:25, Marcus D. Leech [email protected]
wrote:

Both of those commits fail to build for me in gr-trellis

Not sure why that is, might be a result of repeated git checkouts on
different tags, etc., without cleaning up in between. It compiles ok
here.

Two more tests will help me narrow this down:

Please try commit 29c887da, which is the merge on former ‘next’ branch
of the working commit 27dd737a.

Please try commit c787fb66, which is the merge on the former ‘next’
branch of a PMT library modification (this is one of the ones that
failed in trellis for you, hopefully you can make it work.)

If this last fails, please try branch ‘test/undo-pmt-deleter’, which
I’ve pushed to the main gnuradio repository.

I appreciate all the testing, since I can’t duplicate the problem, and
obviously this is important to fix as soon as possible.

Johnathan

Hi folks,
Yesterday after a long wait I received my DVB-T unit in the mail (it
arrived on my birthday). After some initial work, I got it to work with
GNU Radio and GRC and today I made my first working FM receiver using
the same blocks as the video from Ettus R. (building GR FM Radio
in 10 minutes).

The design is using the Linux driver and the GR blocks made by Balint
Seeber (thanks a million mate).

I have attached the GRC with this e-mail, hope you find it useful. I am
now going to work on the RTL2832 block to fine tune it (I have the
complete datasheet) and add more feature to it.

The problem I have is that the signals are not very strong, no matter
how much I change the gain, the signal level is pretty low and the
received audio is pretty noisy. I have managed to sample at 2Msps
without any problems but when I sample more than that on my laptop, I
get a lot of overflow.

I keep you posted on my progress, this is really exciting :slight_smile:

Best regards,
Farhad

On Sun, Apr 15, 2012 at 10:50 AM, Farhad A. [email protected]
wrote:

Hi folks,
Yesterday after a long wait I received my DVB-T unit in the mail (it
arrived
on my birthday). After some initial work, I got it to work with GNU Radio
and GRC and today I made my first working FM receiver using the same
blocks
as the video from Ettus R. (building GR FM Radio in 10 minutes).

The design is using the Linux driver and the GR blocks made by Balint
Seeber
(thanks a million mate).

I have attached the GRC with this e-mail, hope you find it useful. I am
now
going to work on the RTL2832 block to fine tune it (I have the complete
datasheet) and add more feature to it.
The problem I have is that the signals are not very strong, no matter how
much I change the gain, the signal level is pretty low and the received
audio is pretty noisy. I have managed to sample at 2Msps without any
problems but when I sample more than that on my laptop, I get a lot of
overflow.

Hi Farhad,

As you may have seen in my on air tests I just
postedhttp://www.oz9aec.net/index.php/gnu-radio/gnu-radio-blog/465-eztv666-dvb-t-dongle-on-the-air,
I have no problem with low signal levels, however, there is an active
hardware AGC which may be causing you troubles. If you have a strong
signal
in the analog passband of the tuner it will drive the gain down and the
weaker signal you are trying to receive will become noisy. You can see
that
in all 3 videos which show an EzCAP666 with E4000 tuner in action.

I don’t know if this AGC is in the RTL chip or the E4000 tuner. I
suspect
it is in the tuner.

Alex

On 04/14/2012 01:38 PM, Johnathan C. wrote:

If this last fails, please try branch ‘test/undo-pmt-deleter’, which
I’ve pushed to the main gnuradio repository.

That works. No segfaults on exit.

Excellent, thanks. As this is in the 3.5.3 release, I’m going to
rebase this ‘fix’ off the last 3.5.3 maint, but hold off merging it.
I’m hoping Josh can supply an actual fix soon so we don’t have to take
that functionality out altogether.

Maybe its deleting itself: diff --git a/gruel/src/lib/pmt/pmt.cc b/gruel/src/lib/pmt/pmt.ccindex a7e0fd1. - Pastebin.com

-Josh

Hi Alex,
Thanks for the videos, I had seen some of them before, I am not sure we
are using the same modules, the one i am using a much smaller one like
the image I have attached.

I am ‘playing’ with the different parameters today and hopefully I can
get some good results from it. My signals are really weak (about -90db
with strong FM channel).

BR,
Farhad

On Sun, Apr 15, 2012 at 00:08, Josh B. [email protected] wrote:

Maybe its deleting itself: diff --git a/gruel/src/lib/pmt/pmt.cc b/gruel/src/lib/pmt/pmt.ccindex a7e0fd1. - Pastebin.com

Marcus, this is now on branch ‘test/fix-pmt-deleter’ for testing.

Johnathan

On 04/15/2012 12:08 PM, Johnathan C. wrote:

Marcus, this is now on branch ‘test/fix-pmt-deleter’ for testing.

Johnathan
That seems to work.


Marcus L.
Principal Investigator
Shirleys Bay Radio Astronomy Consortium

HI,
I am sorry but the datasheet I have is specifically made for my company
after we signed the NDA with them and I can’t share it with anyone
outside the company.

You can ask the company for the info and I am sure you will be able get
one from them.

Best regards,
Farhad


From: “Zhigang ZHOU (SIMIT, CAS)” [email protected]
To: Farhad A. [email protected]; “[email protected]
[email protected]
Sent: Sunday, April 15, 2012 12:46 PM
Subject: Re: [Discuss-gnuradio] First try with RTL2832

Could you email us the RTL2832 data sheet for reference.
I think many people in this list is seeking this data sheet, thanks a
lot!


Zhigang ZHOU (SIMIT, CAS)

From: Farhad
Abdolian
Date: 2012-04-15 16:50
To: [email protected]
Subject: [Discuss-gnuradio] First try with
RTL2832
Hi folks,
Yesterday after a long wait I received my DVB-T unit in the mail (it arrived on
my birthday). After some initial work,
I got it to work with GNU Radio and GRC and today I made my first
working FM
receiver using the same blocks as the video from Ettus R.
(building GR FM
Radio in 10 minutes).

The design is using the Linux driver and the GR blocks made by Balint S.
(thanks a million mate).

I have attached the GRC with this e-mail, hope you find it useful. I
am now going to work on the RTL2832 block to fine tune it (I have the
complete
datasheet) and add more feature to it.

The problem I have is that the signals are not very strong, no matter how
much I change the gain, the signal level is pretty low and the received
audio is
pretty noisy. I have managed to sample at 2Msps without any problems but
when I
sample more than that on my laptop, I get a lot of overflow.

I keep you posted on my progress, this is really exciting
:slight_smile:

To: Johnathan C. [email protected]
Cc: “[email protected][email protected]
Sent: Sunday, April 15, 2012 9:08 AM
Subject: Re: [Discuss-gnuradio] More dumpeths

On 04/14/2012
01:38 PM, Johnathan C. wrote:

On Sat, Apr 14, 2012 at 13:30,
Marcus D. Leech [email protected] wrote:

Please try commit c787fb66, which is the merge on the former
‘next’
branch of a PMT library modification (this is one of
the ones that
failed in trellis for you, hopefully you can
make it work.)

This commit fails. But I had
to roll-back to an older version of UHD to get
it to build.
Still failed though.

If this last
fails, please try branch ‘test/undo-pmt-deleter’, which
I’ve
pushed to the main gnuradio repository.

That
works. No segfaults on exit.

Excellent, thanks. As this
is in the 3.5.3 release, I’m going to
rebase this ‘fix’ off the last
3.5.3 maint, but hold off merging it.
I’m hoping Josh can supply an
actual fix soon so we don’t have to take
that functionality out
altogether.

Maybe its deleting itself:
diff --git a/gruel/src/lib/pmt/pmt.cc b/gruel/src/lib/pmt/pmt.ccindex a7e0fd1. - Pastebin.com

-Josh


Discuss-gnuradio
mailing list