Open Hub
Open Hub

 Settings |  Report Duplicate

0
I Use This!
Activity Not Available

Commits : Listings

Analyzed over 1 year ago. based on code collected over 3 years ago.
Jan 17, 2021 — Jan 17, 2022
Commit Message Contributor Files Modified Lines Added Lines Removed Code Location Date
Implement support for "essid any": this works, but hopefully the split between essid and essid_found won't have some strange side effects (due to using essid whereas we should use the current authoritative essid_found somewhere in the code or so...). Improve station registration logging (infrastructure type, speed, ...). Merge separate packet logging lines, improve output format.
andim2
More... about 22 years ago
v 0.1h
andim2
More... about 22 years ago
Add card statistics.
andim2
More... about 22 years ago
Dan Aloni: IPv6 fix (add code to recognize special IPv6 broadcast address)
andim2
More... about 22 years ago
Readd stop/wake net queue to protect driver from ping flood attack. Log tx ring buf status. Loop over completed entries in dmaTxDataISR(). Now operation seems to be fairly reliable. Further testing planned.
andim2
More... about 22 years ago
Rework acx100_interrupt() to ACK interrupt items faster (I believe we need to be fast here). Update and cleanup dmaRxXfrISR to check for full descriptors within a loop: we were missing IRQ invocations probably due to not being fast enough in ACKing (and handling) IRQs, but it seems to be a general problem with this ACK register stuff; so let's handle Rx descs (and everything else, too!) in a loop. Code should also be cleaner now. pool_idx2 was being accessed outside the lock (urks!). Why disable irqs? spin_unlock should be ok. But OTOH we're working towards handling everything in a loop, so we should probably indeed disable IRQs while doing so (not that it hurts, but it wastes CPU). Hmm. I now have reliable pings in ~3ms range, with a very small amount of DUPs (argh, still DUPs, but no idea why!) and no problematic behaviour. (OK, floodping still kills the connection, but that's because we don't stop the net queue in time before it's too late)
andim2
More... about 22 years ago
angelbane: added missing dma_tx_data to transmit_disassoc twiddled with locks in dma_tx_data/get_tx_desc probably broke some stuff in acx100_ioctl_unkown11
andim2
More... about 22 years ago
angelbane: removed extraneous get_tx_desc() from transmit deauthen
andim2
More... about 22 years ago
angelbane: added locking around dmaRxXfrISR, dma_tx_data
andim2
More... about 22 years ago
Print more card details, some code cleanup. Rename acx100_download_fw() to acx100_upload_fw(), since it's an *upload* from the PC host, NOT a download.
andim2
More... about 22 years ago
Linux 2.6.x fix: remove the bloody tqueue.h!
andim2
More... about 22 years ago
Mention good debug setting. *sigh* :-)
andim2
More... about 22 years ago
bschrauw: time of packet got printed as signed long
andim2
More... about 22 years ago
bschrauw: removing some comments referring to V1 driver
andim2
More... about 22 years ago
bschrauw: Cleaning up the code a bit
andim2
More... about 22 years ago
bschrauw: Reenabled waking up queue after dmaTxDataISR bacause this caused my card to stop working after a few secs when doing a floodping (not with normal operation though).
andim2
More... about 22 years ago
Update EEPROM ID check and radio type printing for DrayTek Vigor 520.
andim2
More... about 22 years ago
v 0.1g
andim2
More... about 22 years ago
Version 0.1g
andim2
More... about 22 years ago
Added WEP patch as discussed in the forum. Hopefully it's ok...
andim2
More... about 22 years ago
bschrauw: removed all DRIVER_VERSION stuff
andim2
More... about 22 years ago
bschrauw: fixed bug introduced by removed DRIVER_VERSION that caused driver to stop working
andim2
More... about 22 years ago
Patch Parade: - improve EEPROM card ID check to include empty DWL-650+ ID - replace several CPU burning loops by proper scheduling code, and use TASK_UNINTERRUPTIBLE instead of the incorrect TASK_INTERRUPTIBLE - fix net queue problems when not associated yet (this is not the final, correct net queue handling yet!) - move processing of Rx packets in front of Tx descriptor clearing in IRQ handler, to make that faster - fix crash when trying to load 0 length firmware image - fix ifup/ifdown/ifup broken interrupt reinitialisation now state management works (maybe at least 80% or so) - don't reinit templates when reinitialising the card
andim2
More... about 22 years ago
Print response status in human-readable form.
andim2
More... about 22 years ago
Add card removal anti-crash item.
andim2
More... about 22 years ago
Of course only *some* DWL-650+ use Maxim. Add hint for NFS use.
andim2
More... about 22 years ago
bschrauw: removed all DRIVER_VERSION stuff
andim2
More... about 22 years ago
bschrauw: Further implementation of compile time decidable debug mode
andim2
More... about 22 years ago
bschrauw: First attempt to provide a means of turning off all debug info without having all the debug code in place. This is supposed to be always used, except when users explicitly want a lot of debug info. It is NOT fully operational yet, now all logging is turned off. Basic driver messages should still be logged. I propose make a acx_msg that always gets logged, and a acx_debug that only gets logged when the user compiled the driver in debug mode (which makes it a lot slower).
andim2
More... about 22 years ago
Fix state management: the card should only be up after we did an ifconfig up. In other words, do NOT scan or enable Tx or do any other radio activity.
andim2
More... about 22 years ago