Please note: This project is no longer active. The website is kept online for historic purposes only.
If you´re looking for a Linux driver for your Atheros WLAN device, you should continue here .

Ticket #850 (new enhancement)

Opened 16 years ago

Last modified 15 years ago

ATH_SUPERG_COMP not enabled by default

Reported by: *s*t*a*y*d* Assigned to:
Priority: minor Milestone:
Component: madwifi: driver Version: v0.9.3
Keywords: compression ahdemo not work Cc:
Patch is attached: 0 Pending:

Description (Last modified by mentor)

WITHOUT IFDOWN AND IFUP CARD

# iwpriv ath1 get_compression
ath1      get_compression:0
# iwpriv ath1 compression 1
# iwpriv ath1 get_compression
ath1      get_compression:0

WITH IFDOWN AND IFUP CARD

# ifdown ath1
# iwpriv ath1 compression 1
# ifup ath1
# iwpriv ath1 get_compression
ath1      get_compression:0

It is the same.

CARD type: WISTRON CM9 in minipci/pci reduction PCI 2.2

# uname -a
Linux 2.6.11 #1 Sun Oct 23 00:37:10 CEST 2005 i686 GNU/Linux
# iwconfig ath1
ath1      IEEE 802.11a  ESSID:"PtP"
          Mode:Ad-Hoc  Frequency:5.64 GHz  Cell: 00:00:00:00:00:00
          Bit Rate=24 Mb/s   Tx-Power:16 dBm   Sensitivity=0/3
          Retry:off   RTS thr:off   Fragment thr:off
          Encryption key:off
          Power Management:off
          Link Quality=48/94  Signal level=-47 dBm  Noise level=-95 dBm
          Rx invalid nwid:0  Rx invalid crypt:0  Rx invalid frag:0
          Tx excessive retries:0  Invalid misc:0   Missed beacon:0

madwifi-ng 0.9.1 release []

Change History

09/01/06 19:16:11 changed by *s*t*a*y*d*

Mode:Ad-Hoc Frequency:5.64 GHz Cell: 00:00:00:00:00:00 ---

Mode: Ad-Hoc ? but it sould be ahdemo or Ad-hoc demo or something else to differ from normal Ad-Hoc.

Is it possible to change it ? So the driver will show correct ahdemo or demo Ad-Hoc or Lucent ad-hoc or something diffrent from 802.11 adhoc mode.

09/03/06 07:48:48 changed by mrenzmann

  • priority changed from major to minor.
  • description changed.

Iirc the situation (compression can not be turned on) is similar for other modes of operation, can you check that please?

09/05/06 04:20:39 changed by *s*t*a*y*d*

HOSTAP# iwconfig ath1 ath1 IEEE 802.11a ESSID:"PtP"

Mode:Master Frequency:5.58 GHz Access Point: 00:0B:6B:2A:EA:12 Bit Rate=24 Mb/s Tx-Power:16 dBm Sensitivity=0/3 Retry:off RTS thr:off Fragment thr:off Encryption key:6865-736C-6F6D-726F-7A65 Security mode:restricted Power Management:off Link Quality=41/94 Signal level=-54 dBm Noise level=-95 dBm Rx invalid nwid:7 Rx invalid crypt:0 Rx invalid frag:0 Tx excessive retries:0 Invalid misc:0 Missed beacon:0

STA# iwconfig ath1 ath1 IEEE 802.11a ESSID:"PtP"

Mode:Managed Frequency:5.58 GHz Access Point: 00:0B:6B:2A:EA:12 Bit Rate=24 Mb/s Tx-Power:14 dBm Sensitivity=0/3 Retry:off RTS thr:off Fragment thr:off Encryption key:6865-736C-6F6D-726F-7A65 Security mode:restricted Power Management:off Link Quality=46/94 Signal level=-49 dBm Noise level=-95 dBm Rx invalid nwid:0 Rx invalid crypt:0 Rx invalid frag:0 Tx excessive retries:0 Invalid misc:0 Missed beacon:0

HOSTAP# iwpriv ath1 get_compression ath1 get_compression:0 HOSTAP# iwpriv ath1 compression 1 HOSTAP# iwpriv ath1 get_compression ath1 get_compression:0

STA# iwpriv ath1 get_compression ath1 get_compression:0 STA# iwpriv ath1 compression 1 STA# iwpriv ath1 get_compression ath1 get_compression:0

IT IS THE SAME EVEN WITH IFDOWN set compression and IFUP >> get 0 both sides HOSTAP and STA.

AND LAST USEFUL mode 802.11 ad-hoc has the same issue !!!

COMPRESSION DOES NOT WORK AT ALL !!! 0.9.0 announcement about working compression IS NOT TRUE AT ALL !!!

NOT WORKING IN: ahdemo adhoc demo mode adhoc 802.11 ad-hoc btw ad-hoc does not work at all, but it is known nowadays. ap 802.11 AP/HOSTAP sta 802.11 client/station monitor I did not tried, but I think it is not useful for this mode. :)

SET IT TO MAJOR BACK OR TAKE BACK THE ANNOUNCEMENT ! Do not waste time of people !

09/05/06 04:26:43 changed by *s*t*a*y*d*

I am sorry it is a little bit rude. But I do not like to waste my time to try to do work that is offically working and at the end to know, that perhaps nobody tried if it works or not.

09/05/06 07:22:27 changed by mrenzmann

Come back down to earth, bro. Remember the README:

WARNING: THIS IS A BETA DISTRIBUTION. THIS SOFTWARE HAS KNOWN PROBLEMS
WARNING: AND LIMITATIONS THAT WILL BE CORRECTED BEFORE A PRODUCTION
WARNING: RELEASE. DON'T BLAME US IF THE SOFTWARE EATS YOUR SYSTEM,
WARNING: DESTROYS YOUR DISK OR MAKES YOUR CORN-FLAKES SOGGY.
WARNING: USE AT YOUR OWN RISK!

While you are right that the compression feature has been announced in the releases as working, it seems to be broken. The reason has to be examined. It could be a bug in the driver, a problem in the HAL or an issue related to not having described the circumstances in detail when compression is working and when not - or something completely different.

And yes, there is room for improvement regarding testing the driver automatically, to ensure its quality and to make sure that such at such issues are detected early. However, this requires hardware we don't have yet (I'm working on that particular part) and an automated testing environment that needs to be implemented once the hardware is available. For now, we're limited to our own, manual tests and feedback from users - and since you're the first to report problems with compression, it seems that the feature either isn't used very often (which would lower the priority of this ticket compared to other open issues) or it is just working for others.

Rome wasn't build in one day, and we are all working in our spare time on this project. You're welcome to contribute your time and energy to help improve MadWifi. If you don't want to contribute and don't like the fact that this (free) driver is developed by guys who also have a real life, you are welcome to switch to another (probably commercial) driver with better support.

09/05/06 07:23:05 changed by mrenzmann

  • summary changed from Compression is not working in ahdemo mode. to Compression is not working in v0.9.1.

09/05/06 14:11:39 changed by *s*t*a*y*d*

If you don't want to contribute and don't like the fact that this (free) driver is >developed by guys who also have a real life, you are welcome to switch to another >(probably commercial) driver with better support.

Madwifi project and you, guys who also have a real life, are okay.

I know that it is beta, but it was announced and it is real problem.

I am agree with your opinon, but I am not in one thing.

People get used to compression does not working and me to, until the announcment.

You said: it seems that the feature either isn't used very often (which would lower the priority of this ticket compared to other open issues)

Of course, feature isnĀ“t used very often, because until 0.9.0 and it is 3 mounths and some days when it was released, it did not work at all.

People do not try some feature which was and is broken for years.

This ticket deserves major priority and your interest.

09/05/06 21:02:43 changed by mrenzmann

Your suggestion about the ticket priority has been noted. But please leave the decision about that to the developers. To be frank, there are a bunch of other issues that are (from my personal point of view) much more pressing, since they affect a wider range of people.

Be asured that this issue will be taken care of, but it will not happen with major priority at this time.

03/27/07 03:17:22 changed by *s*t*a*y*d*

Hi all,

I would like to ask you if something new on this ticket ? Or compression is stayed in uninteresting feature.

Thanks

05/01/07 23:16:42 changed by dhskhoo@yahoo.com

As of release 0.9.3, you can enable it by compiling with ATH_SUPERG_COMP(as in -DATH_SUPERG_COMP). Compression will be enabled by default with the define. Testing using iperf shows about 40Mbps with compression vs about 33Mbps with compression turned off (via iwpriv ath0 compression 0). Downloading a file already compressed (ie mpeg) yields about 20Mbps, which seems to be worse off than without ATH_SUPERG_COMP which can be as high as 32Mbps(with fast frame and bursting enabled).

05/02/07 00:38:02 changed by mentor

  • version changed from v0.9.1 to v0.9.3.
  • type changed from defect to enhancement.
  • description changed.
  • summary changed from Compression is not working in v0.9.1 to ATH_SUPERG_COMP not enabled by default.

(follow-up: ↓ 14 ) 05/07/07 14:25:19 changed by rozteck@interia.pl

I have enabled the ATH_SUPERG_COMP and made some tests with and without the compression set. And I got smaller transfer rate with compression enabled than when it was disabled. Do any one have similar problems with that? ps. I used two devices based on ixp425-platform (one with 266MHz processor and the other with 532MHz) running kernel 2.6.18-rt7.

05/07/07 15:14:03 changed by mentor

Read up a couple of comments :p - so, I guess we know why it isn't on by default...

(in reply to: ↑ 12 ) 05/07/07 15:16:31 changed by mrenzmann

Replying to rozteck@interia.pl:

And I got smaller transfer rate with compression enabled than when it was disabled. Do any one have similar problems with that?

Obviously yes, dhskhoo reported the same issue above for data that already has been compressed (mpeg in his case). I think this is not too surprising, as the chipset has to take care of additional work which has no effect other than adding some useless overhead.