Re: [PATCH 1/2 v3] image_types: Add elf image type

[Raymond Danks <ray.danks@se-eng.com>]

On 07/01/2012 10:22 PM, Saul Wold wrote:
> On 06/29/2012 01:19 PM, Raymond Danks wrote:
>> On x86, an ELF image file may be stored as a coreboot payload.
>> The image file is constructed, using the mkelfimage utility,
>> from a kernel and an initrd.
>>
>> Signed-off-by: Raymond Danks<ray.danks@se-eng.com>
>> ---
>> This was originally submitted to the openembedded project:
>> http://patches.openembedded.org/patch/7689/
>>
>> v3:
>>   Include DEPLOY_DIR_IMAGE and KERNEL_IMAGETYPE for location of 
>> ELF_KERNEL
>>
>>   meta/classes/image_types.bbclass |   18 +++++++++++++++++-
>>   1 files changed, 17 insertions(+), 1 deletions(-)
>>
>> diff --git a/meta/classes/image_types.bbclass 
>> b/meta/classes/image_types.bbclass
>> index 55f122e..64fa14a 100644
>> --- a/meta/classes/image_types.bbclass
>> +++ b/meta/classes/image_types.bbclass
>> @@ -7,6 +7,12 @@ def get_imagecmds(d):
>>       ctypes = d.getVar('COMPRESSIONTYPES', True).split()
>>       cimages = {}
>>
>> +    if "elf" in alltypes:
>> +        alltypes.remove("elf")
>> +        if "cpio.gz" not in alltypes:
>> +                alltypes.append("cpio.gz")
>> +        alltypes.append("elf")
>> +
>>       # Filter out all the compressed images from types
>>       for type in alltypes:
>>           basetype = None
>> @@ -173,6 +179,14 @@ IMAGE_CMD_cpio () {
>>       cd ${IMAGE_ROOTFS}&&  (find . | cpio -o -H 
>> newc>${DEPLOY_DIR_IMAGE}/${IMAGE_NAME}.rootfs.cpio)
>>   }
>>
>> +ELF_KERNEL ?= ${DEPLOY_DIR_IMAGE}/${KERNEL_IMAGETYPE}
>
> I guess this was not tested very well, as this failed right out of the 
> gate.
Thanks for giving this a shot Sau.  Could you be more specific regarding 
the failure you are seeing?
>
> Please let me know what kind of testing you have done.
This was tested rather exhaustively on an x86_64->x86_64 configuration 
prior to the latest changes to the above line.  I have only tested this 
using the bzImage kernel type.  I did feel that I had tested the above 
line as well, but it's possible I had something cached somewhere.  I'll 
kick off a clean build and see if anything surfaces.

>
> Thanks
>     Sau!
>> +ELF_APPEND ?= "ramdisk_size=32768 root=/dev/ram0 rw console="
>> +
>> +IMAGE_CMD_elf () {
>> +    test -f ${DEPLOY_DIR_IMAGE}/${IMAGE_NAME}.rootfs.elf&&  rm -f 
>> ${DEPLOY_DIR_IMAGE}/${IMAGE_NAME}.rootfs.elf
>> +    mkelfImage --kernel=${ELF_KERNEL} 
>> --initrd=${DEPLOY_DIR_IMAGE}/${IMAGE_NAME}.rootfs.cpio.gz 
>> --output=${DEPLOY_DIR_IMAGE}/${IMAGE_NAME}.rootfs.elf 
>> --append='${ELF_APPEND}' ${EXTRA_IMAGECMD}
>> +}
>> +
>>   UBI_VOLNAME ?= "${MACHINE}-rootfs"
>>
>>   IMAGE_CMD_ubi () {
>> @@ -199,6 +213,7 @@ EXTRA_IMAGECMD_ext2 ?= "-i 8192"
>>   EXTRA_IMAGECMD_ext3 ?= "-i 8192"
>>   EXTRA_IMAGECMD_ext4 ?= "-i 8192"
>>   EXTRA_IMAGECMD_btrfs ?= ""
>> +EXTRA_IMAGECMD_elf ?= ""
>>
>>   IMAGE_DEPENDS = ""
>>   IMAGE_DEPENDS_jffs2 = "mtd-utils-native"
>> @@ -210,11 +225,12 @@ IMAGE_DEPENDS_ext4 = "genext2fs-native 
>> e2fsprogs-native"
>>   IMAGE_DEPENDS_btrfs = "btrfs-tools-native"
>>   IMAGE_DEPENDS_squashfs = "squashfs-tools-native"
>>   IMAGE_DEPENDS_squashfs-lzma = "squashfs-lzma-tools-native"
>> +IMAGE_DEPENDS_elf = "virtual/kernel mkelfimage-native"
>>   IMAGE_DEPENDS_ubi = "mtd-utils-native"
>>   IMAGE_DEPENDS_ubifs = "mtd-utils-native"
>>
>>   # This variable is available to request which values are suitable 
>> for IMAGE_FSTYPES
>> -IMAGE_TYPES = "jffs2 sum.jffs2 cramfs ext2 ext2.gz ext2.bz2 ext3 
>> ext3.gz ext2.lzma btrfs live squashfs squashfs-lzma ubi tar tar.gz 
>> tar.bz2 tar.xz cpio cpio.gz cpio.xz cpio.lzma vmdk"
>> +IMAGE_TYPES = "jffs2 sum.jffs2 cramfs ext2 ext2.gz ext2.bz2 ext3 
>> ext3.gz ext2.lzma btrfs live squashfs squashfs-lzma ubi tar tar.gz 
>> tar.bz2 tar.xz cpio cpio.gz cpio.xz cpio.lzma vmdk elf"
>>
>>   COMPRESSIONTYPES = "gz bz2 lzma xz"
>>   COMPRESS_CMD_lzma = "lzma -k -f -7 ${IMAGE_NAME}.rootfs.${type}"