qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Qemu-devel] [PATCH] Make -kernel flag optional on ARM.


From: Grant Likely
Subject: Re: [Qemu-devel] [PATCH] Make -kernel flag optional on ARM.
Date: Wed, 10 Jul 2013 09:17:48 +0100

On Wed, Jul 10, 2013 at 8:44 AM, Peter Maydell <address@hidden> wrote:
> On 10 July 2013 08:33, Grant Likely <address@hidden> wrote:
>> Sometimes we want to boot the system via firmware instead of loading a
>> kernel into ram with the -kernel parameter. This patch makes the -kernel
>> parameter optional so that a bios image provided by the -pflash flag
>> will be executed.
>>
>> For example:
>> qemu-system-arm -M vexpress-a15 -pflash <filename>
>>
>> Note: Currently the file must be at least the size of the emulated flash
>> device (64M) otherwise QEMU will silently not use the data. This will be
>> fixed in a separate patch
>>
>> This is part of the larger solution. For this to work it requires the
>> platform to have flash mapped or aliased to address 0 so that there is
>> some code for the emulator to execute.
>
> That patch is already on list :-)

Heh, I wrote the above before grabbing and testing your patch.

>> diff --git a/hw/arm/boot.c b/hw/arm/boot.c
>> index 7c0090f..3b9e108 100644
>> --- a/hw/arm/boot.c
>> +++ b/hw/arm/boot.c
>> @@ -361,12 +361,6 @@ void arm_load_kernel(ARMCPU *cpu, struct arm_boot_info 
>> *info)
>>      int big_endian;
>>      QemuOpts *machine_opts;
>>
>> -    /* Load the kernel.  */
>> -    if (!info->kernel_filename) {
>> -        fprintf(stderr, "Kernel image must be specified\n");
>> -        exit(1);
>> -    }
>
> Why can't we just return from the function here? If we don't
> have a kernel we can assume we don't have a dtb or an initrd
> either. I don't think we want to do any of the other weird
> stuff with setting up a reset handler or boot pen for the
> secondary CPUs either -- we should just assume the bootrom
> image in the flash handles this.

I've just tested this, and you're right. I assumed that some of the
other stuff would be necessary, but it does work just fine to return.
Second version of the patch posted now.

g.



reply via email to

[Prev in Thread] Current Thread [Next in Thread]