ThinkPad 850 Patches for 2.2.16
Hi,
Linar Yusupov did some patches for the Thinkpad 850 to make sound work on these machines. It fixes also some typos for /proc/cpuinfo.
I'm not sure if the CAROLINA_IRQ_EDGE_MASK_HI change breaks anything else.
The sound stuff should be moved out of arch/ppc into the driver cs4232, you can't have it as a module right now. xmms works fine, but I have many dropouts. Any ideas how to finetune that?
That sound-patch will proably break the other PReP machines, I will test it on the 43p 133 later. We need proably a _PREP_IBM_portable or so.
Gruss Olaf
--
$ man clone
BUGS
Main feature not yet implemented...
arch/ppc/kernel/prep_pci.c +2 -2
@@ -560,7 +560,7 @@* to either edge or level. */ #define CAROLINA_IRQ_EDGE_MASK_LO 0x00/* IRQ's 0-7 */ -#define CAROLINA_IRQ_EDGE_MASK_HI 0xA4/* IRQ's 8-15 [10,13,15] */ +#define CAROLINA_IRQ_EDGE_MASK_HI 0xA0/* IRQ's 8-15 [10,13,15] */ /* * 8259 edge/level control definitions
@@ -932,7 +932,7 @@switch (inb (0x0852 )) {case 0xff : - Motherboard_map_name ="IBM 850/860 Portable\n" ; + Motherboard_map_name ="IBM 850/860 Portable" ; Motherboard_map = Nobis_pci_IRQ_map; Motherboard_routes = Nobis_pci_IRQ_routes;break ;
arch/ppc/kernel/prep_setup.c +4 -4
@@ -140,7 +140,7 @@ } len +=sprintf (buffer+len,"%sKb," , (((*(unsigned char *)0x8000080d )>>2 )&1 )?"512" :"256" ); - len +=sprintf (buffer+len,"%sync\n" , + len +=sprintf (buffer+len,"%ssync\n" , ((*(unsigned char *)0x8000080d )>>7 ) ?"" :"a" );break ;case _PREP_Motorola:
@@ -329,9 +329,9 @@ }if ( _prep_type == _PREP_IBM ) { - snd_ptr->config.io_base =0x530 ; - snd_ptr->config.irq =5 ; - snd_ptr->config.dma = ppc_cs4232_dma =1 ; + snd_ptr->config.io_base =0x830 ; + snd_ptr->config.irq =10 ; + snd_ptr->config.dma = ppc_cs4232_dma =6 ;/* this is wrong - but leave it for now */ snd_ptr->config.dma2 = ppc_cs4232_dma2 =7 ; }
Quoted 6 lines
The sound stuff should be moved out of
arch/ppcinto the driver cs4232, you can't have it as a module right now. xmms works fine, but I have many dropouts. Any ideas how to finetune that?
That sound-patch will proably break the other PReP machines, I will test it on the 43p 133 later. We need proably a_PREP_IBM_portableor so.
No please ;)
I beleive you must have a way to probe for the machine type and do the right thing, don't you ?
Ben.
** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/
Quoted 2 lines · Benjamin Herrenschmidt
I beleive you must have a way to probe for the machine type and do the right thing, don't you ?
We need it, but not in arch/ppc :)
The cs4232.c driver has already some interrupt stuff that you can adjust in .config. But we know the needed values so we could hardcode it in the driver itself.
Gruss Olaf
--
$ man clone
BUGS
Main feature not yet implemented...
** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/
Quoted 7 lines · Olaf Hering
Quoted 2 lines · Benjamin Herrenschmidt
I beleive you must have a way to probe for the machine type and do the right thing, don't you ?
We need it, but not in
arch/ppc:)
Right; that auto-setup is already in prep_setup.c, and has been for a while.
Quoted 3 lines
The
cs4232.cdriver has already some interrupt stuff that you can adjust in .config. But we know the needed values so we could hardcode it in the driver itself.
Cort said if anyone can convince the cs4232 driver maintainer to accept the hack, then great.
Otherwise it seems the hack will live where it is, to allow the same binary kernel to run on Motorola and IBM PReP boxes with sound support.
-Hollis
** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/
Quoted 6 lines · Olaf Hering
Hi,
Linar Yusupov did some patches for the Thinkpad 850 to make sound work on these machines. It fixes also some typos for
/proc/cpuinfo.I'm not sure if the
CAROLINA_IRQ_EDGE_MASK_HIchange breaks anything else.
linux-pmac-stable works on the TP, but I need a common config. David Monroes carolina minidif is needed to access the SCSI disks.
I tried to remove that patch with CAROLINA_IRQ_EDGE_MASK_HI, but it hangs in the cs4231 detection.
I have another strange problem.
The "bootloader" in produces garbage output, it starts with a few random bytes and then it fills up the screen with more random bytes. These garbage goes into /proc/cmdline (its hard to copy&paste :):
olaf@tp:~ > hex /proc/cmdline
0000 72 6f 6f 74 3d 2f 64 65 76 2f 66 64 30 bc ad 73 0010 ff bc ad 73 ff bc ad 33 fe 3c ad 73 ff 0a
Any ideas how to fix that? Could be a termic problem, the machine runs here since a few days. And I cant remember that I got this garbage a few days ago. But typing here at the console works fine.
Gruss Olaf
--
$ man clone
BUGS
Main feature not yet implemented...
** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/