[PATCH] Several patches agains linuxppc_devel-2.4.19-pre8

Hello


The attached patch fixes the following issues:

    - arch/ppc/kernel/ocp.c
       wrong comment and export of ocp_get_max() for module support.

    - arch/ppc/kernel/ocp_proc.c
       compiler warnings fixed in various seq_printf().
       i don't know if this also works correctly on PPC440
       with its 36-bit phys_addr_t?

    - arch/ppc/kernel/ocp_uart.c
       various fixes to opc_uart_init().
       initialize curr_uart, free previously allocated driver
       object and fix return value.

    - arch/ppc/kernel/ppc4xx_setup
       missing early_uart_init() prototype.

    - drivers/char/ibm_ocp_gpio.c
       stop modprobe complaining about tainted kernel ("GPL license").
       initialize curr_gpio and fix return value.

    - drivers/i2c/i2c-adap-ibm_ocp.c
       fixed call to ocp_get_max().
       initialize curr_iic.
       remove module parameters "base" and "irq".

    - drivers/i2c/i2c-algo-ibm_ocp.c
       remove module paramter "iic_test".


Dave

On Wed, May 15, 2002 at 01:35:55PM +0200, "David M?ller (ELSOFT AG)" wrote:

The attached patch fixes the following issues:

Looks good, except:
Don't #ifdef externs, if they aren't used it's OK.
Since this code only has to work in recent'ish kernels (in reality only
'current' kernels) don't bother with the
#ifdef MODULE_LICENSE
...
#endif

Just use it.

--
Tom Rini (TR1265)
http://gate.crashing.org/~trini/

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Tom Rini schrieb:

On Wed, May 15, 2002 at 01:35:55PM +0200, "David M?ller (ELSOFT AG)" wrote:

quoted

The attached patch fixes the following issues:

Looks good, except:
Don't #ifdef externs, if they aren't used it's OK.
Since this code only has to work in recent'ish kernels (in reality only
'current' kernels) don't bother with the
#ifdef MODULE_LICENSE
...
#endif

Just use it.

I have copied the MODULE_LICENSE stuff from another driver.
Does anybody knows if this feature is tagged to be a remove candiate in
future kernels and/or in the modutils?


Another question:

I've noticed that (at least) the following files have their executable
bit set, although they are simple C source files.
./include/asm-ppc/ppc405_dma.h
./arch/ppc/kernel/ppc405_dma.c
./arch/ppc/kernel/ppc405_pci.c
./arch/ppc/kernel/ppc4xx_setup.c
./arch/ppc/platforms/walnut.c
./arch/ppc/platforms/cpc700_pic.c

On the other hand, the PPCBoot kernel packer script in
./arch/ppc/boot/utils/mkimage.wrapper have the exec bit cleared.

Is it a bit task to fix this?


Dave


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

On Thu, May 16, 2002 at 11:26:33AM +0200, "David M?ller (ELSOFT AG)" wrote:

Tom Rini schrieb:

quoted

On Wed, May 15, 2002 at 01:35:55PM +0200, "David M?ller (ELSOFT AG)" wrote:

quoted

The attached patch fixes the following issues:

Looks good, except:
Don't #ifdef externs, if they aren't used it's OK.
Since this code only has to work in recent'ish kernels (in reality only
'current' kernels) don't bother with the
#ifdef MODULE_LICENSE
...
#endif

Just use it.

I have copied the MODULE_LICENSE stuff from another driver.
Does anybody knows if this feature is tagged to be a remove candidate in
future kernels and/or in the modutils?

Unfortunately, no.  I think other drivers do it for backwards
compatibility, which we don't have to worry about.

Another question:

I've noticed that (at least) the following files have their executable
bit set, although they are simple C source files.
./include/asm-ppc/ppc405_dma.h
./arch/ppc/kernel/ppc405_dma.c
./arch/ppc/kernel/ppc405_pci.c
./arch/ppc/kernel/ppc4xx_setup.c
./arch/ppc/platforms/walnut.c
./arch/ppc/platforms/cpc700_pic.c

On the other hand, the PPCBoot kernel packer script in
./arch/ppc/boot/utils/mkimage.wrapper have the exec bit cleared.

Is it a bit task to fix this?

I assume you ment big.  And no, I've just fixed these.  Thanks.

--
Tom Rini (TR1265)
http://gate.crashing.org/~trini/

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Tom Rini schrieb:

On Thu, May 16, 2002 at 11:26:33AM +0200, "David M?ller (ELSOFT AG)" wrote:

[old stuff deleted]

quoted

Another question:

I've noticed that (at least) the following files have their executable
bit set, although they are simple C source files.
./include/asm-ppc/ppc405_dma.h
./arch/ppc/kernel/ppc405_dma.c
./arch/ppc/kernel/ppc405_pci.c
./arch/ppc/kernel/ppc4xx_setup.c
./arch/ppc/platforms/walnut.c
./arch/ppc/platforms/cpc700_pic.c

On the other hand, the PPCBoot kernel packer script in
./arch/ppc/boot/utils/mkimage.wrapper have the exec bit cleared.

Is it a bit task to fix this?

I assume you ment big.  And no, I've just fixed these.  Thanks.

Yes, of course. Here's the list of exec bit set files for 2.5.15:
./arch/ppc/kernel/ppc405_dma.c
./arch/ppc/kernel/ppc405_pci.c
./arch/ppc/kernel/ppc4xx_serial.c
./arch/ppc/kernel/ppc4xx_setup.c
./arch/ppc/platforms/walnut.c
./include/asm-ppc/ppc405_dma.h


Dave


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Hello

Two new patches:

    - arch/ppc/kernel/ocp.c
       add default case in ocp_unregister().
       GCC 3.1 is not happy without it ;-)

    - drivers/net/ibm_ocp/ibm_ocp_debug.c
       ibm_ocp_debug.c seems not to be updated by the latest 4xx ethernet
       driver patches.


Dave