[PATCH] 2.5.5-pre1, allow RAM disk to be build

[PATCH] 2.5.5-pre1, allow RAM disk to be build

[Rudmer van Dijk]

Hi,

since 2.5.x (can't remember version correctly) the setup for RAM disk changed and I could not build it anymore into the kernel due to compilation errors.
In the latest prepatch (2.5.5-pre1) this is still not fixed. I am using the following patch to fix this and it works from 2.5.x upto 2.5.5-pre1 without any problems.

please apply,

Rudmer

PS. CC me as I am not on the list

--
# diff -uN linux-2.5.5-pre1/arch/i386/kernel/setup.c.orig linux-2.5.5-pre1/arch/i386/kernel/setup.c
--- linux-2.5.5-pre1/arch/i386/kernel/setup.c.orig   Thu Feb 14 20:39:00 2002
+++ linux-2.5.5-pre1/arch/i386/kernel/setup.c   Thu Feb 14 20:40:17 2002
@@ -661,7 +661,9 @@
        unsigned long bootmap_size, low_mem_size;
        unsigned long start_pfn, max_low_pfn;
        int i;
-
+#ifdef CONFIG_BLK_DEV_RAM
+       extern int rd_image_start, rd_prompt, rd_doload;
+#endif
 #ifdef CONFIG_VISWS
        visws_get_board_type_and_rev();
 #endif

-- 

Get your free email from www.linuxmail.org 


Powered by Outblaze

Re: [PATCH] 2.5.5-pre1, allow RAM disk to be build

[Dave Jones]

On Sun, Feb 17, 2002 at 02:07:35AM +0800, Rudmer van Dijk wrote:
 > since 2.5.x (can't remember version correctly) the setup for RAM disk changed and I could not build it anymore into the kernel due to compilation errors.
 > In the latest prepatch (2.5.5-pre1) this is still not fixed. I am using the following patch to fix this and it works from 2.5.x upto 2.5.5-pre1 without any problems.

The real fix I think is to fix include/linux/blk.h
setup.c uses it inside CONFIG_BLK_DEV_RAM, but blk.h
only defines it with CONFIG_BLK_DEV_INITRD
 
-- 
| Dave Jones.        http://www.codemonkey.org.uk
| SuSE Labs

Re: [PATCH] 2.5.5-pre1, allow RAM disk to be build

[Rudmer van Dijk]

From: Dave Jones <davej@suse.de>

> On Sun, Feb 17, 2002 at 02:07:35AM +0800, Rudmer van Dijk wrote:
>  > since 2.5.x (can't remember version correctly) the setup for RAM disk changed and I could not build it anymore into the kernel due to compilation errors.
>  > In the latest prepatch (2.5.5-pre1) this is still not fixed. I am using the following patch to fix this and it works from 2.5.x upto 2.5.5-pre1 without any problems.
> 
> The real fix I think is to fix include/linux/blk.h
> setup.c uses it inside CONFIG_BLK_DEV_RAM, but blk.h
> only defines it with CONFIG_BLK_DEV_INITRD
>
It was a quick fix because I thought someone else was going to fix it...

How about the following patch?
kernel builds and boots with patch (with CONFIG_BLK_DEV_INITRD disabled).

I think this is unrelated but I suddenly got the following messages while trying to build another kernel:
Feb 18 00:02:51 frodo kernel: hdc: lost interrupt
Feb 18 00:03:31 frodo last message repeated 4 times
Feb 18 00:04:31 frodo last message repeated 6 times
Feb 18 00:05:31 frodo last message repeated 6 times
Feb 18 00:06:31 frodo last message repeated 6 times

now my some files (including gcc) are screwed and I cannot test this patch with CONFIG_BLK_DEV_INITRD enabled...

Rudmer

--
frodo:/home/rudmer/linux-2.5.5-pre1 # diff -uN include/linux/blk.h.orig include/linux/blk.h
--- include/linux/blk.h.orig    Sun Feb 17 20:27:28 2002
+++ include/linux/blk.h Sun Feb 17 19:49:27 2002
@@ -16,18 +16,24 @@
 extern void set_device_ro(kdev_t dev,int flag);
 extern void add_blkdev_randomness(int major);

+#ifdef CONFIG_BLK_DEV_RAM
+
+extern int rd_doload;          /* 1 = load ramdisk, 0 = don't load */
+extern int rd_prompt;          /* 1 = prompt for ramdisk, 0 = don't prompt */
+extern int rd_image_start;     /* starting block # of image */
+
 #ifdef CONFIG_BLK_DEV_INITRD

 #define INITRD_MINOR 250 /* shouldn't collide with /dev/ram* too soon ... */

 extern unsigned long initrd_start,initrd_end;
 extern int initrd_below_start_ok; /* 1 if it is not an error if initrd_start < memory_start */
-extern int rd_doload;          /* 1 = load ramdisk, 0 = don't load */
-extern int rd_prompt;          /* 1 = prompt for ramdisk, 0 = don't prompt */
-extern int rd_image_start;     /* starting block # of image */
 void initrd_init(void);

+#endif /* CONFIG_BLK_DEV_INITRD */
+
 #endif
+
 /*
  * end_request() and friends. Must be called with the request queue spinlock
  * acquired. All functions called within end_request() _must_be_ atomic.
-- 

Get your free email from www.linuxmail.org 


Powered by Outblaze