Opened 12 years ago

Closed 12 years ago

Last modified 12 years ago

#2670 closed defect (worksforme)

binary dies on sh4 stapi

Reported by: japie Owned by:
Priority: blocker Component: DVBApi
Severity: high Keywords:
Cc: Sensitive: no

Description

Revision

7209 onwards

Issue Description

binary dies when starting without leaving anything in console or log file neither any in syslog. (dmesg)
When using "oscam -h" help text is shown fine.
Last (known) working version 7209 but only tried 7230 and 7232

When the issue occurs

every time

How the issue is reproducable

starting it...

Change History (13)

comment:1 by tossa-fan, 12 years ago

r7228 is working fine here.

comment:2 by japie, 12 years ago

Have been building some different revisions and r7209 is the last working version on my box, all newer seem to die without notification.
I refreshed my source and rebuild r7209 to verify that my toolchain was still working with the same results, everything later dies.

The strange thing is that I have recent svn running on my x86_64 box without problems, so it is platform or box specific.

Icecrypt S4000 & Octagon 1018 & Atevio 7000
Builds made with STlinux-2.3 (which was used to build the firmware on the box)

Linux hmp7109 2.6.17.14_stm22_0039-HMP_7109-STFAE #3 PREEMPT Tue Feb 15 10:44:24 KST 2011 sh4 GNU/Linux

Unix starttime: 1344148003
Starttime:      05.08.2012 06:26:43
Version:        1.20-unstable_svn  Rev. 7209
Max PID:        unlimited


Active modules:
Web interface support:      yes
SSL support:                no
DVB API support:            yes
DVB API with STAPI support: yes
Anti-cascading support:     yes
Irdeto guessing:            yes
Debug mode:                 no
Monitor:                    no
Loadbalancing support:      yes
LCD support:                no
IPv6 support:               no
Cache exchange support:     yes
camd 3.3x:                  no
camd 3.5 UDP:               no
camd 3.5 TCP:               no
newcamd:                    yes
CCcam:                      yes
Pandora:                    no
gbox:                       no
radegast:                   no
serial:                     no
constant CW:                no
Cardreader:                 yes
Nagra:                      yes
Irdeto:                     yes
Conax:                      yes
Cryptoworks:                yes
Seca:                       yes
Viaccess:                   yes
NDS Videoguard:             yes
DRE Crypt:                  yes
TONGFANG:                   yes
Bulcrypt:                   yes
Last edited 12 years ago by japie (previous) (diff)

comment:3 by Deas, 12 years ago

wrong toolchain when the info you posted is correct.

Linux hmp7109 2.6.17.14_stm22

Builds made with STlinux-2.3

comment:4 by japie, 12 years ago

Your completely right Deas, I used 2.2 not 2.3... (symlink confused me)

comment:5 by Deas, 12 years ago

i just installed a self compiled stm22 binary (r7232) on a ufs910 - works like a charm.

so either you use the wrong toolchain or you forgot to enable binary mode in your ftp program which destroys your binary. but the problem you try to explain here does only exist on your side. nobody else here or in the forum reported such a problem in the last days.

comment:6 by japie, 12 years ago

That is clear Deas, if it doesn't effect anyone else you can close the ticket but I do want to make clear that the toolchain I use I have been using for over a year now and until r7210 without any problems.

I started using my own builds due to other builds segfaulting on the Icecrypt and found that this one seems to create the most stable binaries, if I rebuild 7209 it runs and 7210 doesn't, binairy mode is enabled... (otherwise 7209 wouldn't function also and "oscam -h" wouldn't execute either ;)

I will try r7232 you provided in the bin forum thread as soon as my wife stops watching tv...

comment:7 by japie, 12 years ago

Deas, your build runs fine...
I don't understand why mine stopped working at certain revision, maybe because -Os is causing the compiler to haywire on code after 7209

Anyway thanks for your patience and help...

comment:8 by japie, 12 years ago

Resolution: fixed
Status: newclosed

comment:9 by japie, 12 years ago

Resolution: fixed
Status: closedreopened

comment:10 by japie, 12 years ago

Deas, your build segfaults after a couple of hours on this box like all other builds (yours and others) did in the past.
Can't tell for sure if this is the same as I experienced in the past or due to something new/buggy in recent code.

Don't know what to do with this ticket so I reopened it, please advice. (still having problems believing that my toolchain is to blame, will do some more testing and post beneath)

comment:11 by japie, 12 years ago

Indeed -Os is messing up, keeping compiler options blank creates a working binary, sort for the headaches I caused...

comment:12 by japie, 12 years ago

Resolution: worksforme
Status: reopenedclosed

comment:13 by Deas, 12 years ago

for what reason did you set different compiler options?

Note: See TracTickets for help on using tickets.