Re: [Xen-devel] [PATCH] xen: Add EFI_LOAD_OPTION support

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 03.01.18 at 17:04, <tamas@xxxxxxxxxxxxx> wrote:
> On Wed, Jan 3, 2018 at 4:20 AM, Jan Beulich <JBeulich@xxxxxxxx> wrote:
>>>>> On 02.01.18 at 16:56, <tamas@xxxxxxxxxxxxx> wrote:
>>> +        if ( elo->Attributes & LOAD_OPTION_ACTIVE )
>>
>> Without any other (earlier) check, how can you reliably tell this
>> being a pointer to EFI_LOAD_OPTION from it being the
>> currently used one to CHAR16? Is the distinction perhaps UEFI
>> version dependent? In the 2.3 spec I can't find any information
>> on the layout of what ->LoadOptions points to.
> 
> AFAICT there is no clear cut way to distinguish what is being passed
> here. When launched via UEFI Xen receives the EFI_LOAD_OPTION buffer
> here already, which it tries to parse as a string and fails. Take a
> look at the same problem in the shim:
> https://github.com/rhboot/shim/blob/master/shim.c#L2501. If there is a
> clearer way to distinguish what is being passed here or more checks
> that can be done I would be open for suggestions.

Well, first of all the code you point to tells me that this situation is
indeed as bad as it can be. However, the code also shows you
some heuristics you could re-use. What they don't seem to utilize
is the fact that as long as there are few enough bits defined for
the Attributes field, checking that one for being a printable
character (and the upper 16 bits to be non-null) might give a good
indication what form we're dealing with. In fact it might be that
when it's a string, the first character is always '\', but I'm afraid
that's not written down anywhere.

Jan


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel