Bug in LR-WPAN model: noise PSD not updated when phyCurrentChannel is set

28 views
Skip to first unread message

Guilherme Lello

unread,
Nov 15, 2018, 9:22:02 AM11/15/18
to ns-3-users
I'm currently working on simulating 802.15.4 performance under 802.11 interference with ns-3. I have found a bug in the lr-wpan model and I was actually able to fix it.

According to the ns-3 guidelines, bugs and bug fixes should be reported in https://www.nsnam.org/bugzilla/. I have sent a Bugzilla account creation request to webm...@nsnam.org, but I have received no reply within 2 days. Is this delay common?

I would be glad if I have a new account granted so I can start contributing to this awesome project by filing the bug report and the patch. Can any admin help me with the Bugzilla account creation?

Thanks in advance.

Best regards,

Guilherme Lello
Researcher @ LEMT/UFRJ

Tom Henderson

unread,
Nov 15, 2018, 10:25:44 AM11/15/18
to ns-3-...@googlegroups.com
On 11/15/18 6:22 AM, Guilherme Lello wrote:
I'm currently working on simulating 802.15.4 performance under 802.11 interference with ns-3. I have found a bug in the lr-wpan model and I was actually able to fix it.

According to the ns-3 guidelines, bugs and bug fixes should be reported in https://www.nsnam.org/bugzilla/. I have sent a Bugzilla account creation request to webm...@nsnam.org, but I have received no reply within 2 days. Is this delay common?

I would be glad if I have a new account granted so I can start contributing to this awesome project by filing the bug report and the patch. Can any admin help me with the Bugzilla account creation?

Hi Guilherme, I've created an account for you (sorry for the delay).  Thanks in advance for filing a patch to fix an issue.

- Tom

Guilherme Lello

unread,
Nov 15, 2018, 12:08:43 PM11/15/18
to ns-3-users
Hi Tom, thanks for the fast reply.

I have a patch for ns-3.29, but I'm running the fix on ns-3-dev before submitting to Bugzilla just to make sure it has not already been fixed.

I noticed the dev build is currently broken. Here is the compilation error caused by the wifi module:

[1387/1837] Compiling src/wifi/model/he-configuration.cc
../src/wifi/model/he-configuration.cc:51:60: error: no member named 'm_bssColor' in 'ns3::HeConfiguration'
                   MakeUintegerAccessor (&HeConfiguration::m_bssColor),
                                          ~~~~~~~~~~~~~~~~~^
1 error generated.

In order to fix the build, I removed all references to HeConfiguration::m_bssColor from files src/wifi/model/he-configuration.cc / .h, but I'm not sure this should be the right approach.

Regards,

Guilherme Lello
Researcher @ LEMT/UFRJ

Tom Henderson

unread,
Nov 15, 2018, 1:25:22 PM11/15/18
to Guilherme Lello, ns-3-...@googlegroups.com
On 11/15/18 9:08 AM, Guilherme Lello wrote:
> Hi Tom, thanks for the fast reply.
>
> I have a patch for ns-3.29, but I'm running the fix on ns-3-dev before
> submitting to Bugzilla just to make sure it has not already been fixed.
>
> I noticed the dev build is currently broken. Here is the compilation
> error caused by the wifi module:
>
> [1387/1837] Compiling src/wifi/model/he-configuration.cc
> ../src/wifi/model/he-configuration.cc:51:60: error: no member named
> 'm_bssColor' in 'ns3::HeConfiguration'
>                    MakeUintegerAccessor (&HeConfiguration::m_bssColor),
>                                           ~~~~~~~~~~~~~~~~~^
> 1 error generated.

This shouldn't be the case; are you sure that you have a clean build?
We are not seeing this error and I see this variable declared in the
header. This variable was introduced 11 days ago.

- Tom

Guilherme Lello

unread,
Nov 15, 2018, 1:33:31 PM11/15/18
to ns-3-users
Yes Tom, I have a clean build.

Maybe this error is related to my compiler version.

I'm running Apple LLVM version 8.0.0 (clang-800.0.42.1) on OSX 10.12.6.

I have already filed the issue on Bugzilla. The patch attached to the issue fixes both lr-wpan noise PSD when switching channels (the real bug) and the build for my compiler.

Thanks,

Guilherme Lello
Researcher @ LEMT/UFRJ

Guilherme Lello

unread,
Nov 15, 2018, 2:05:55 PM11/15/18
to ns-3-users
Tom, I have double checked the output error. There is indeed a build error for my compiler, but I have pasted the wrong build error message earlier.

The correct build error message follows below:

In file included from ../src/wifi/model/he-configuration.cc:23:
../src/wifi/model/he-configuration.h:53:11: error: private field 'm_bssColor' is not used [-Werror,-Wunused-private-field]
  uint8_t m_bssColor;   //!< BSS color

I'm sorry for that.

Regards,

Guilherme Lello
Researcher @ LEMT/UFRJ

Tom Henderson

unread,
Nov 15, 2018, 2:49:58 PM11/15/18
to ns-3-...@googlegroups.com
On 11/15/18 11:05 AM, Guilherme Lello wrote:
Tom, I have double checked the output error. There is indeed a build error for my compiler, but I have pasted the wrong build error message earlier.

The correct build error message follows below:

In file included from ../src/wifi/model/he-configuration.cc:23:
../src/wifi/model/he-configuration.h:53:11: error: private field 'm_bssColor' is not used [-Werror,-Wunused-private-field]
  uint8_t m_bssColor;   //!< BSS color

OK, we were not testing that version of clang (8.0.0), but I have an older Mac and I was able to reproduce the above.

I will work with Sebastien on a workaround for it to suppress the warning.

- Tom

Reply all
Reply to author
Forward
0 new messages