[mythtv-users] Enabling experimental ATRPMS
Cory Papenfuss
papenfuss at juneau.me.vt.edu
Mon Sep 27 15:35:33 EDT 2004
DOH! I saw your post yesterday but was confused as to the intent.
The compat.h applied, but not the ivtv-firmware.c patch. Did you patch against
a vanilla ivtv-0.1.10-pre2-ck100z?
[mythtv at localhost driver]$ pwd
/home/mythtv/ivtv-0.1.10-pre2-ck100z/driver
[mythtv at localhost driver]$ patch -p2 < ivtv-firmware.c.patch
patching file ivtv-firmware.c
Hunk #1 FAILED at 95.
1 out of 1 hunk FAILED -- saving rejects to file ivtv-firmware.c.rej
[mythtv at localhost driver]$ head ivtv-firmware.c.rej
***************
*** 95,136 ****
return 0;
}
static int load_fw_direct(const char *fn, char *mem)
{
- int fd;
long l;
- mm_segment_t fs = get_fs();
I don't speak "diff-ese" yet... :)
?
*************************************************************************
* Cory Papenfuss *
* Electrical Engineering candidate Ph.D. graduate student *
* Virginia Polytechnic Institute and State University *
*************************************************************************
On Mon, 27 Sep 2004, Jarod Wilson wrote:
> On Monday 27 September 2004 05:41, Cory Papenfuss wrote:
>>> That should provide you with the needed tuner, compile a patched ck100z
>>> ivtv driver, and still use rpms for everything else.
>>
>> I went and got my old patched one, but it appears to have issues. Last
>> time I fried it up, I was running a vanilla 2.6.8.1 kernel (with tuner.c,
>> tuner.h patch). Now I've got the FC2 kernel source tree and the compile
>> gives me:
>>
>> make[1]: Entering directory `/lib/modules/2.6.8-1.521smp/build'
>> Building modules, stage 2.
>> MODPOST
>> *** Warning: "close"
>> [/home/mythtv/ivtv-0.1.10-pre2-ck100z_patched/driver/ivtv.ko] undefined!
>> *** Warning: "open"
>> [/home/mythtv/ivtv-0.1.10-pre2-ck100z_patched/driver/ivtv.ko] undefined!
>> *** Warning: "read"
>> [/home/mythtv/ivtv-0.1.10-pre2-ck100z_patched/driver/ivtv.ko] undefined!
>> *** Warning: "lseek"
>> [/home/mythtv/ivtv-0.1.10-pre2-ck100z_patched/driver/ivtv.ko] undefined!
>>
>> ... with the associated
>> ivtv: Unknown symbol sys_close
>> ivtv: Unknown symbol sys_lseek
>> ivtv: Unknown symbol sys_read
>> ivtv: Unknown symbol sys_open
>>
>> when trying to load the modules. Last time I copied compat.h from the
>> newer drivers and all was well... This time, notsomuch. I'm thinking that
>> the #if LINUX_VERSION_CODE > KERNEL_VERSION(2, 6, 7)
>> in compat.h is getting confused with a non-vanilla kernel source.
>
> Nope, it works fine here w/the FC2 2.6.8-1.521 kernel. RH disallowed the use
> of sys_* calls entirely, which isn't being done (yet) in the kernel.org
> sources, so there's an additional change you need to make in ivtv-firmware.c.
>
>> Did you
>> say you'd gotten a workaround for this?
>
> Yup. I posted a patch to the ivtv mailing list just yesterday. Essentially,
> all you need to do is take the load_fw_direct function from a newer version
> and substitute it for the one in 100z.
>
>> I'm also wondering if 108m might
>> be a better choice for all this. Good stability?
>
> Both 108k and 108m kernel panic my EPIA/PVR-350 system and 112t segfaults. The
> patched 100z loads without a problem.
>
> --
> Jarod C. Wilson, RHCE
> jcw at wilsonet.com
>
> Got a question? Read this first...
> http://catb.org/~esr/faqs/smart-questions.html
> MythTV, Fedora Core & ATrpms documentation:
> http://wilsonet.com/mythtv/
> MythTV Searchable Mailing List Archive
> http://www.gossamer-threads.com/lists/mythtv/
>
More information about the mythtv-users
mailing list