grub_machine_get_bootlocation and partitions

grub_machine_get_bootlocation and partitions

[Andrei Borzenkov]

If partition information is actually present, it results in bogus and
invalid device name like

hd0,1

This device cannot be opened as it requires partmap name.

One obvious case when partition *is* passed to GRUB is multiboot. I
suppose nobody ever noticed it because standard grub-install generates
fixed partition part.

We probably can fallback to probing partmap on current level in this
case, except it means more code again. But current situation is rather
useless as is.

Re: grub_machine_get_bootlocation and partitions

[Vladimir 'phcoder' Serbinenko]

[-- Attachment #1: Type: text/plain, Size: 861 bytes --]

Do you have a failure scenario? hd0,1 is not recommended but it's still
valid and so should still work
Le 1 déc. 2015 5:19 PM, "Andrei Borzenkov" <arvidjaar@gmail.com> a écrit :

> If partition information is actually present, it results in bogus and
> invalid device name like
>
> hd0,1
>
> This device cannot be opened as it requires partmap name.
>
> One obvious case when partition *is* passed to GRUB is multiboot. I
> suppose nobody ever noticed it because standard grub-install generates
> fixed partition part.
>
> We probably can fallback to probing partmap on current level in this
> case, except it means more code again. But current situation is rather
> useless as is.
>
> _______________________________________________
> Grub-devel mailing list
> Grub-devel@gnu.org
> https://lists.gnu.org/mailman/listinfo/grub-devel
>

[-- Attachment #2: Type: text/html, Size: 1272 bytes --]

Re: grub_machine_get_bootlocation and partitions

[Andrei Borzenkov]

01.12.2015 20:45, Vladimir 'phcoder' Serbinenko пишет:
> Do you have a failure scenario? hd0,1 is not recommended but it's still
> valid and so should still work

Yes, I misunderstood what I have seen. Sorry.

> Le 1 déc. 2015 5:19 PM, "Andrei Borzenkov" <arvidjaar@gmail.com> a écrit :
> 
>> If partition information is actually present, it results in bogus and
>> invalid device name like
>>
>> hd0,1
>>
>> This device cannot be opened as it requires partmap name.
>>
>> One obvious case when partition *is* passed to GRUB is multiboot. I
>> suppose nobody ever noticed it because standard grub-install generates
>> fixed partition part.
>>
>> We probably can fallback to probing partmap on current level in this
>> case, except it means more code again. But current situation is rather
>> useless as is.
>>
>> _______________________________________________
>> Grub-devel mailing list
>> Grub-devel@gnu.org
>> https://lists.gnu.org/mailman/listinfo/grub-devel
>>
> 
> 
> 
> _______________________________________________
> Grub-devel mailing list
> Grub-devel@gnu.org
> https://lists.gnu.org/mailman/listinfo/grub-devel
>