ath0 performence issues "ar9300_Stub_GetCTSTimeout" "ar9300_Stub_GetCTSTimeout

[Adrian Chadd]

Ah, add ATH_DEBUG, AH_DEBUG and AH_DIAGAPI to your kernel and recompile!


-a


On 19 March 2015 at 19:32, Miguel Clara <miguelmclara at gmail.com> wrote:
>
> On Fri, Mar 20, 2015 at 2:11 AM, Miguel Clara <miguelmclara at gmail.com>
> wrote:
>>
>>
>> On Fri, Mar 20, 2015 at 1:55 AM, Miguel Clara <miguelmclara at gmail.com>
>> wrote:
>>>
>>>
>>> On Fri, Mar 20, 2015 at 1:18 AM, Miguel Clara <miguelmclara at gmail.com>
>>> wrote:
>>>>
>>>>
>>>> On Thu, Mar 19, 2015 at 9:49 PM, Adrian Chadd <adrian at freebsd.org>
>>>> wrote:
>>>>>
>>>>> Whilst running this, can you run 'athstats 1' on the freebsd box?
>>>>>
>>>>> i wonder what the signal level and retransmit rates are.
>>>>
>>>>
>>>> Not sure hwo to do that... do I need ath_debug?
>>>>
>>>> I did found '/usr/src/tools/tools/ath/athstats' and tried to "make"
>>>>
>>>> clang -O2 -pipe   -DATH_DEFAULT='"ath0"'
>>>> -I/usr/src/tools/tools/ath/athstats
>>>> -I/usr/src/tools/tools/ath/athstats/../common
>>>> -I/usr/src/tools/tools/ath/athstats/../../../../sys
>>>> -I/usr/src/tools/tools/ath/athstats/../../../../sys/dev/ath
>>>> -I/usr/src/tools/tools/ath/athstats/../../../../sys/dev/ath/ath_hal
>>>> -I/usr/src/tools/tools/ath/athstats/../../../../sys/contrib/dev/ath/ath_hal
>>>> -I/usr/src/tools/tools/ath/athstats -DATH_SUPPORT_ANI -DATH_SUPPORT_TDMA
>>>> -DATH_DEFAULT='"ath0"' -I/usr/src/tools/tools/ath/athstats
>>>> -I/usr/src/tools/tools/ath/athstats/../common
>>>> -I/usr/src/tools/tools/ath/athstats/../../../../sys
>>>> -I/usr/src/tools/tools/ath/athstats/../../../../sys/dev/ath
>>>> -I/usr/src/tools/tools/ath/athstats/../../../../sys/dev/ath/ath_hal
>>>> -I/usr/src/tools/tools/ath/athstats/../../../../sys/contrib/dev/ath/ath_hal
>>>> -I/usr/src/tools/tools/ath/athstats -std=gnu99 -fstack-protector
>>>> -fbracket-depth=512 -Qunused-arguments  -rpath /usr/lib/private -o athstats
>>>> main.o athstats.o -L/usr/src/lib/libbsdstat -lbsdstat
>>>> /usr/bin/ld: cannot find -lbsdstat
>>>> clang: error: linker command failed with exit code 1 (use -v to see
>>>> invocation)
>>>> *** Error code 1
>>>>
>>>> Stop.
>>>> make: stopped in /usr/src/tools/tools/ath/athstats
>>>>
>>>> I installed bsdstats but after make clean;make  I still get the same.
>>>
>>>
>>> Oh wait:
>>> -L/usr/src/lib/libbsdstat -lbsdstat
>>>
>>> /usr/bin/ld: cannot find -lbsdstat
>>>
>>> Should it look in /usr/private?
>>>
>>> Or maybe not... I saw this:
>>> ######################
>>> commit 050d4409aa855fa1f82f6ca0dbb7e0c865c05e32
>>> Author: bapt <bapt at FreeBSD.org>
>>> Date:   Tue Nov 25 22:37:27 2014 +0000
>>>
>>>     Convert to LIBADD
>>>
>>> Notes:
>>>     svn path=/head/; revision=275084
>>>
>>> diff --git a/tools/tools/ath/athstats/Makefile
>>> b/tools/tools/ath/athstats/Makefile
>>> index 8d1bfa6..a7ff43b 100644
>>> --- a/tools/tools/ath/athstats/Makefile
>>> +++ b/tools/tools/ath/athstats/Makefile
>>> @@ -23,9 +23,7 @@ CFLAGS+=-DATH_SUPPORT_TDMA
>>>
>>>  CFLAGS.clang+= -fbracket-depth=512
>>>
>>> -USEPRIVATELIB= bsdstat
>>> -
>>> -LDADD= ${LDBSDSTAT}
>>> +LIBADD=        bsdstat
>>>
>>>  opt_ah.h:
>>>         echo "#define AH_DEBUG 1" > opt_ah.h
>>> #################
>>>
>>> reverting gives me this:
>>>
>>> [...]ib/private -rpath /usr/lib/private -o athstats main.o athstats.o
>>> athstats.o: In function `athstats_new':
>>> athstats.c:(.text+0x3e): undefined reference to `bsdstat_init'
>>> clang: error: linker command failed with exit code 1 (use -v to see
>>> invocation)
>>>
>>> If I can get around this I'd love to post those stats
>>>
>>
>> Sadly I'm no C expert... I'm not even a coder but "bsdstat_init" defiantly
>> exists and the include bsdstat.h is in athstats.h...
>> makes no sense (to me) :|
>>
>
> In the mean time I checked out a version before bsdstat and I was able to
> complie, output follows
> % sudo athstats
> athstats: ioctl: ath0: Invalid argument <------------------ Is this
> expected?
> athstats: ioctl: ath0: Invalid argument
> 990785       data frames received
> 559194       data frames transmit
> 7602         tx frames with an alternate rate
> 27634        short on-chip tx retries
> 100075       long on-chip tx retries
> 374          tx failed 'cuz too many retries
> 24           mib overflow interrupts
> 3            tx linearized to cluster
> MCS3         current transmit rate
> 17           beacon miss interrupts
> 31173        rx failed 'cuz of bad CRC
> 114          rx failed 'cuz of PHY err
>     5            OFDM restart
>     109          CCK restart
> 247          periodic calibrations
> -0/+0        TDMA slot adjust (usecs, smoothed)
> 28           rssi of last ack
> 40           avg recv rssi
> -96          rx noise floor
> 614          tx frames through raw api
> 601749       A-MPDU sub-frames received
> 187766       Half-GI frames received
> 443941       40MHz frames received
> 12776        CRC errors for non-last A-MPDU subframes
> 198          CRC errors for last subframe in an A-MPDU
> 558554       Frames transmitted with HT Protection
> 169          TX Timeout
> Antenna profile:
> [0] tx   558817 rx    36629
> [1] tx        0 rx   772652
> [2] tx        0 rx    26737
> [3] tx        0 rx    30276
> [4] tx        0 rx    28833
> [5] tx        0 rx    25606
> [6] tx        0 rx    30300
> [7] tx        0 rx    39752
>
>
>
>>
>>
>>>
>>>>
>>>>> Thanks,
>>>>>
>>>>>
>>>>> -a
>>>>
>>>>
>>>
>>
>