Quoting Robert Millan <[EMAIL PROTECTED]>:
On Thu, Jun 26, 2008 at 12:54:54PM -0400, Pavel Roskin wrote:
+ dos_part = grub_cpu_to_le32 (root_dev->disk->partition->index);
+ bsd_part = grub_cpu_to_le32 (-1);
[...]
grub_util_info ("dos partition is %d, bsd partition is
On Thu, Jun 26, 2008 at 12:54:54PM -0400, Pavel Roskin wrote:
> + dos_part = grub_cpu_to_le32 (root_dev->disk->partition->index);
> + bsd_part = grub_cpu_to_le32 (-1);
> [...]
>grub_util_info ("dos partition is %d, bsd partition is %d",
> - grub_le_to_cpu32 (*i
On Thu, 2008-06-26 at 10:48 -0400, Pavel Roskin wrote:
> Anyway, I think there is a much simpler approach. Changes to the memory
> image of core.img should be applied after the image is tested for
> readability. Until then, they should be in temporary variables.
Here's the patch. It's an alter
On Thu, 2008-06-26 at 16:20 +0200, Robert Millan wrote:
> On Wed, Jun 25, 2008 at 05:48:34PM -0400, Pavel Roskin wrote:
> >
> > I'm also surprised that the code alternately uses dir and
> > DEFAULT_DIRECTORY to calculate core_path. core_path is calculated 3
> > times in one function! If dir and
On Wed, Jun 25, 2008 at 05:48:34PM -0400, Pavel Roskin wrote:
>
> I'm also surprised that the code alternately uses dir and
> DEFAULT_DIRECTORY to calculate core_path. core_path is calculated 3
> times in one function! If dir and DEFAULT_DIRECTORY are used correctly,
> I suggest that two differe
On Wed, 2008-06-25 at 17:48 -0400, Pavel Roskin wrote:
> It turns out following happens in grub-setup:
>
> - core.img is read by the OS facilities
> - the memory image is modified
> - the memory image is compared to core.img read by GRUB FS code and fails
> - the memory image is modified again
>
On Thu, 2008-06-26 at 03:02 +0200, Javier Martín wrote:
> El mié, 25-06-2008 a las 17:48 -0400, Pavel Roskin escribió:
> > I'm also surprised that the code alternately uses dir and
> > DEFAULT_DIRECTORY to calculate core_path. core_path is calculated 3
> > times in one function! If dir and DEFAUL
El mié, 25-06-2008 a las 17:48 -0400, Pavel Roskin escribió:
> I'm also surprised that the code alternately uses dir and
> DEFAULT_DIRECTORY to calculate core_path. core_path is calculated 3
> times in one function! If dir and DEFAULT_DIRECTORY are used correctly,
> I suggest that two different v
Hello!
I have a system with a hard drive with a geometry that doesn't leave
space for GRUB to be embedded:
Disk /dev/sda: 30.0 GB, 30020272128 bytes
64 heads, 32 sectors/track, 28629 cylinders
Units = cylinders of 2048 * 512 = 1048576 bytes
Disk identifier: 0xfafa98d2
Device Boot Start
