qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
* [Qemu-devel] ARM bootloader boot blobbing.
@ 2012-09-26  6:27 Peter Crosthwaite
  2012-09-26 21:52 ` Edgar E. Iglesias
  0 siblings, 1 reply; 3+ messages in thread
From: Peter Crosthwaite @ 2012-09-26  6:27 UTC (permalink / raw)
  To: qemu-devel@nongnu.org Developers, Peter Maydell, Paul Brook,
	Andreas Färber
  Cc: Edgar E. Iglesias

Hi All,

Can anyone think of a reason why the arm primary bootloader cant be
done by just direct interaction with the CPU? Currently we have this
...

/* The worlds second smallest bootloader.  Set r0-r2, then jump to kernel.  */
static uint32_t bootloader[] = {
  0xe3a00000, /* mov     r0, #0 */
  0xe59f1004, /* ldr     r1, [pc, #4] */
  0xe59f2004, /* ldr     r2, [pc, #4] */
  0xe59ff004, /* ldr     pc, [pc, #4] */
  0, /* Board ID */
  0, /* Address of kernel args.  Set by integratorcp_init.  */
  0  /* Kernel entry point.  Set by integratorcp_init.  */
};

... which gets injected into RAM then we set the PC to this blob and
go. But couldnt we just set R0-2 directly from the bootloader and just
straight to the kernel entry point? Why do we have to blob in a
lightweight bootloader?

Regards,
Peter

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [Qemu-devel] ARM bootloader boot blobbing.
  2012-09-26  6:27 [Qemu-devel] ARM bootloader boot blobbing Peter Crosthwaite
@ 2012-09-26 21:52 ` Edgar E. Iglesias
  2012-09-27  1:14   ` Peter Crosthwaite
  0 siblings, 1 reply; 3+ messages in thread
From: Edgar E. Iglesias @ 2012-09-26 21:52 UTC (permalink / raw)
  To: Peter Crosthwaite
  Cc: Peter Maydell, qemu-devel@nongnu.org Developers,
	Andreas Färber, Paul Brook

On Wed, Sep 26, 2012 at 04:27:57PM +1000, Peter Crosthwaite wrote:
> Hi All,
> 
> Can anyone think of a reason why the arm primary bootloader cant be
> done by just direct interaction with the CPU? Currently we have this
> ...
> 
> /* The worlds second smallest bootloader.  Set r0-r2, then jump to kernel.  */
> static uint32_t bootloader[] = {
>   0xe3a00000, /* mov     r0, #0 */
>   0xe59f1004, /* ldr     r1, [pc, #4] */
>   0xe59f2004, /* ldr     r2, [pc, #4] */
>   0xe59ff004, /* ldr     pc, [pc, #4] */
>   0, /* Board ID */
>   0, /* Address of kernel args.  Set by integratorcp_init.  */
>   0  /* Kernel entry point.  Set by integratorcp_init.  */
> };
> 
> ... which gets injected into RAM then we set the PC to this blob and
> go. But couldnt we just set R0-2 directly from the bootloader and just
> straight to the kernel entry point? Why do we have to blob in a
> lightweight bootloader?

Hi Peter,

I can't speak for this specific case but I've used similar approaches
for other boards when needing more realistic emulation, e.g if emulating
possible boot loaders in rom and supporting warm jumps back to the rom etc.

Cheers

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [Qemu-devel] ARM bootloader boot blobbing.
  2012-09-26 21:52 ` Edgar E. Iglesias
@ 2012-09-27  1:14   ` Peter Crosthwaite
  0 siblings, 0 replies; 3+ messages in thread
From: Peter Crosthwaite @ 2012-09-27  1:14 UTC (permalink / raw)
  To: Edgar E. Iglesias
  Cc: Peter Maydell, qemu-devel@nongnu.org Developers,
	Andreas Färber, Paul Brook

On Thu, Sep 27, 2012 at 7:52 AM, Edgar E. Iglesias
<edgar.iglesias@gmail.com> wrote:
> On Wed, Sep 26, 2012 at 04:27:57PM +1000, Peter Crosthwaite wrote:
>> Hi All,
>>
>> Can anyone think of a reason why the arm primary bootloader cant be
>> done by just direct interaction with the CPU? Currently we have this
>> ...
>>
>> /* The worlds second smallest bootloader.  Set r0-r2, then jump to kernel.  */
>> static uint32_t bootloader[] = {
>>   0xe3a00000, /* mov     r0, #0 */
>>   0xe59f1004, /* ldr     r1, [pc, #4] */
>>   0xe59f2004, /* ldr     r2, [pc, #4] */
>>   0xe59ff004, /* ldr     pc, [pc, #4] */
>>   0, /* Board ID */
>>   0, /* Address of kernel args.  Set by integratorcp_init.  */
>>   0  /* Kernel entry point.  Set by integratorcp_init.  */
>> };
>>
>> ... which gets injected into RAM then we set the PC to this blob and
>> go. But couldnt we just set R0-2 directly from the bootloader and just
>> straight to the kernel entry point? Why do we have to blob in a
>> lightweight bootloader?
>
> Hi Peter,
>
> I can't speak for this specific case but I've used similar approaches
> for other boards when needing more realistic emulation, e.g if emulating
> possible boot loaders in rom and supporting warm jumps back to the rom etc.
>

Cant you just do the CPU register setting warm as well however? Im
looking at microblaze_boot.c and comparing it to ARM, here it is:

static void main_cpu_reset(void *opaque)
{
    MicroBlazeCPU *cpu = opaque;
    CPUMBState *env = &cpu->env;

    cpu_reset(CPU(cpu));
    env->regs[5] = boot_info.cmdline;
    env->regs[7] = boot_info.fdt;
    env->sregs[SR_PC] = boot_info.bootstrap_pc;
    if (boot_info.machine_cpu_reset) {
        boot_info.machine_cpu_reset(cpu);
    }
}

They seem to me to be feature equivalent with respect to register and
PC setting, yet microblaze is much simpler. Only broken case I can see
is if the guest explicitly wants to jump back to the blob without an
actual reset, however that is impossible in ARM as the bootloader does
not pass its location to the guest so the guest cant jump to the right
place. If we want to support more complex bootloaders that the guest
interacts with, then isnt that guest in itself?

Regards,
Peter

> Cheers

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2012-09-27  1:14 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2012-09-26  6:27 [Qemu-devel] ARM bootloader boot blobbing Peter Crosthwaite
2012-09-26 21:52 ` Edgar E. Iglesias
2012-09-27  1:14   ` Peter Crosthwaite

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).