On Fri, Jan 21, 2005 at 01:44:58AM +0100, Lucian Muresan wrote:
> Hi folks,
>
> I updated DirctFB too latest CVS as of 20.01.2005, and rebuilt all the
> apps using it (most important mplayer and vdrplugin-softdevice at the
> moment) and I couldn't run any of them, they all had an output similar
> to that from dfbinfo pasted here at the end of the message.
> By searching after "Error reading from maven chip!" I found that the
> problem was introduced 4 days ago in matrox_maven.c. To verify, I kept
> the fresh CVS checkout but replaced matrox_maven.c/h with the version
> from Sunday, and everything seems to just work as expected.
I was fairly certain this change wouldn't break anything but apparently I
was wrong :(
Note that I don't actually have a DH add-on so I only tested it on a real
DH card. But I even tested it by swapping the tests so that DH add-on was
tested first and it still detected my DH card properly.
> Ville, if you need me to further test and diagnose this, just tell me
> what debugging/tracing I should turn on and where, or let me test your
> fixes.
Try the attached patch so we can see if the ioctl() fails or if the read
actually returns 0xFF for both addresses.
--
Ville Syrj�l�
[EMAIL PROTECTED]
http://www.sci.fi/~syrjala/
Index: matrox_maven.c
===================================================================
RCS file: /cvs/directfb/DirectFB/gfxdrivers/matrox/matrox_maven.c,v
retrieving revision 1.15
diff -u -r1.15 matrox_maven.c
--- matrox_maven.c 16 Jan 2005 20:50:51 -0000 1.15
+++ matrox_maven.c 21 Jan 2005 07:34:51 -0000
@@ -106,9 +106,12 @@
msgs, 2
};
- if (ioctl( fd, I2C_RDWR, &data ) < 0)
+ if (ioctl( fd, I2C_RDWR, &data ) < 0) {
+ D_PERROR( "DirectFB/Matrox/Maven: I2C_RDWR failed: addr = 0x%02X,
reg = 0x%02X!\n", addr, reg );
return 0xFF;
+ }
+ D_INFO( "DirectFB/Matrox/Maven: I2C_RDWR succeeded: addr = 0x%02X, reg =
0x%02X, val = 0x%02X\n", addr, reg, val );
return val;
}
