[PATCH] const initdata.

[PATCH] const initdata.

[Dave Jones]

As defined in in Rusty's kernel-hacking doc, __initdata must not be
marked as const.  Patch below does this for the PCI subsystem.

*nb*, This kind of patch needs to be done in quite a few other
places too.

regards,

Dave.

-- 
| Dave Jones.                    http://www.codemonkey.org.uk
| SuSE Labs .

diff -urN --exclude-from=/home/davej/.exclude linux-ac/drivers/pci/names.c linux-dj/drivers/pci/names.c
--- linux-ac/drivers/pci/names.c	Mon Oct  2 20:00:16 2000
+++ linux-dj/drivers/pci/names.c	Fri Aug 24 00:48:59 2001
@@ -32,9 +32,9 @@
  * real memory.. Parse the same file multiple times
  * to get all the info.
  */
-#define VENDOR( vendor, name )		static const char __vendorstr_##vendor[] __initdata = name;
+#define VENDOR( vendor, name )		static char __vendorstr_##vendor[] __initdata = name;
 #define ENDVENDOR()
-#define DEVICE( vendor, device, name ) 	static const char __devicestr_##vendor##device[] __initdata = name;
+#define DEVICE( vendor, device, name ) 	static char __devicestr_##vendor##device[] __initdata = name;
 #include "devlist.h"


@@ -43,7 +43,7 @@
 #define DEVICE( vendor, device, name )	{ 0x##device, 0, __devicestr_##vendor##device },
 #include "devlist.h"

-static const struct pci_vendor_info __initdata pci_vendor_list[] = {
+static struct pci_vendor_info __initdata pci_vendor_list[] = {
 #define VENDOR( vendor, name )		{ 0x##vendor, sizeof(__devices_##vendor) / sizeof(struct pci_device_info), __vendorstr_##vendor, __devices_##vendor },
 #define ENDVENDOR()
 #define DEVICE( vendor, device, name )

Re: [PATCH] const initdata.

[Anders Peter Fugmann]

How "bad" is it to have __initdata declared static?

A quick scan through the kernel tree shows
that 27 .h and 355 .c files needa patching.
(a total of 845 places)

Numbers are from 2.4.9 kernel.

It should be very easy to correct through a script.

Regards
Anders Fugmann


Dave Jones wrote:
> As defined in in Rusty's kernel-hacking doc, __initdata must not be
> marked as const.  Patch below does this for the PCI subsystem.
> 
> *nb*, This kind of patch needs to be done in quite a few other
> places too.
> 
> regards,
> 
> Dave.
> 
> 

Re: [PATCH] const initdata.

[Alan Cox]

> How "bad" is it to have __initdata declared static?

static is fine as it just changes name scoping, const however can try and
put data in other places such as code segments. That breaks stuff when
compiling with certain gcc versions

Re: [PATCH] const initdata.

[Anders Peter Fugmann]

Ah.. I misread the first posting.

Then we are down to zero .h and 43 .c files that are affected.
(a total of 121 lines containing 'const' before '__initdata')

Can these be fixed without testing (ie. through a script).
If so, I'm willing to post a patch.

Regards
Anders Fugmann





Alan Cox wrote:
>>How "bad" is it to have __initdata declared static?
>>
> 
> static is fine as it just changes name scoping, const however can try and
> put data in other places such as code segments. That breaks stuff when
> compiling with certain gcc versions
> 
>