Re: [PATCH] libelf: improve PVH elfnote parsing

[Jan Beulich <jbeulich@xxxxxxxx>]

On 14.05.2021 15:50, Roger Pau Monne wrote:
> @@ -426,7 +426,7 @@ static elf_errorstatus elf_xen_addr_calc_check(struct 
> elf_binary *elf,
>      }
>  
>      /* Initial guess for virt_base is 0 if it is not explicitly defined. */
> -    if ( parms->virt_base == UNSET_ADDR )
> +    if ( parms->virt_base == UNSET_ADDR || hvm )
>      {
>          parms->virt_base = 0;
>          elf_msg(elf, "ELF: VIRT_BASE unset, using %#" PRIx64 "\n",
> @@ -442,7 +442,7 @@ static elf_errorstatus elf_xen_addr_calc_check(struct 
> elf_binary *elf,
>       * If we are using the modern ELF notes interface then the default
>       * is 0.
>       */
> -    if ( parms->elf_paddr_offset == UNSET_ADDR )
> +    if ( parms->elf_paddr_offset == UNSET_ADDR || hvm )
>      {
>          if ( parms->elf_note_start )
>              parms->elf_paddr_offset = 0;

Both of these would want their respective comments also updated, I
think: There's no defaulting or guessing really in PVH mode, is
there?

> @@ -456,8 +456,13 @@ static elf_errorstatus elf_xen_addr_calc_check(struct 
> elf_binary *elf,
>      parms->virt_kstart = elf->pstart + virt_offset;
>      parms->virt_kend   = elf->pend   + virt_offset;
>  
> -    if ( parms->virt_entry == UNSET_ADDR )
> -        parms->virt_entry = elf_uval(elf, elf->ehdr, e_entry);
> +    if ( parms->virt_entry == UNSET_ADDR || hvm )
> +    {
> +        if ( parms->phys_entry != UNSET_ADDR32 )

Don't you need "&& hvm" here to prevent ...

> +            parms->virt_entry = parms->phys_entry;

... this taking effect for a kernel capable of running in both
PV and PVH modes, instead of ...

> +        else
> +            parms->virt_entry = elf_uval(elf, elf->ehdr, e_entry);

... this (when actually in PV mode)?

Jan