[meta-ti] [master/thud][PATCH v2 3/3] jailhouse: use ramfs for linux-demo
Jacob Stiffler
j-stiffler at ti.com
Fri Aug 30 09:30:19 PDT 2019
On 8/30/2019 12:29 PM, Denys Dmytriyenko wrote:
> On Fri, Aug 30, 2019 at 12:23:43PM -0400, Jacob Stiffler wrote:
>> On 8/30/2019 12:13 PM, Denys Dmytriyenko wrote:
>>> On Tue, Aug 20, 2019 at 02:11:12PM -0400, Jacob Stiffler wrote:
>>>> * Configure the linux-demo to use a ramfs for the linux inmate.
>>>> * Using JH_RAMFS_IMAGE, this can be set in some global conf.
>>>>
>>>> Signed-off-by: Jacob Stiffler <j-stiffler at ti.com>
>>>> ---
>>>> recipes-kernel/jailhouse/jailhouse_git.bb | 21 +++++++++++++++++----
>>>> 1 file changed, 17 insertions(+), 4 deletions(-)
>>>>
>>>> diff --git a/recipes-kernel/jailhouse/jailhouse_git.bb b/recipes-kernel/jailhouse/jailhouse_git.bb
>>>> index 5d572d0..ba6772f 100644
>>>> --- a/recipes-kernel/jailhouse/jailhouse_git.bb
>>>> +++ b/recipes-kernel/jailhouse/jailhouse_git.bb
>>>> @@ -58,8 +58,10 @@ JH_SYSCONFIG_CELL ?= ""
>>>> JH_SYSCONFIG_CELL_am65xx ?= "k3-am654-idk.cell"
>>>> JH_SYSCONFIG_CELL_j7-evm ?= "k3-j721e-evm.cell"
>>>> +JH_RAMFS_IMAGE ?= "core-image-minimal"
>>> Ok, after thinking about it a bit, I will agree to JH_RAMFS_IMAGE variable,
>>> but I don't want to loose the default INITRAMFS_IMAGE usage. So, how about
>>> changing the default here to something like this:
>>>
>>> JH_RAMFS_IMAGE ?= "${INITRAMFS_IMAGE}"
>>>
>>> That way either of the variables can be set my machine config or local.conf
>>> and the default use is also preserved. Thoughts?
>>
>> I like that idea better as well.
>>
>>
>>>> JH_CMDLINE ?= ""
>>>> -JH_CMDLINE_k3 ?= "console=ttyS1,115200n8 root=/dev/mmcblk0p2 rw rootfstype=ext4 rootwait"
>>>> +JH_CMDLINE_k3 ?= "console=ttyS1,115200n8"
>>>> do_configure() {
>>>> if [ -d ${STAGING_DIR_HOST}/${CELLCONF_DIR} ];
>>>> @@ -101,6 +103,16 @@ do_install() {
>>>> install -m 0644 ${B}/inmates/demos/${JH_ARCH}/*.bin ${D}${INMATES_DIR}
>>>> install -d ${D}/boot
>>>> + if [ -n "${JH_RAMFS_IMAGE}" ]
>>>> + then
>>>> + if [ -f ${DEPLOY_DIR_IMAGE}/${JH_RAMFS_IMAGE}-${MACHINE}.cpio ]
>>>> + then
>>>> + install -m 0644 ${DEPLOY_DIR_IMAGE}/${JH_RAMFS_IMAGE}-${MACHINE}.cpio ${D}/boot
>>>> + else
>>>> + bbwarn "Could find find JH_RAMFS_IMAGE (${JH_RAMFS_IMAGE}-${MACHINE}.cpio)!"
>>> Typo in the message. Also, should it be an error instead of a warning?
>>
>> I did not want this recipe to force the requirement for an ramfs. But
>> thinking about it now, making it an error is better in case the linux-demo
>> is desired. If it is not, then JH_RAMFS_IMAGE can be unset in some conf.
> Thanks. Can you do a v4 of all 3 patches quickly to address the comments in
> 1/3 and 3/3, so I can merge them in for rc4? Thanks again.
Sure. Just got to make sure they still build.
>
>>>> + bbwarn "Please make sure that \"cpio\" is in IMAGE_FSTYPES."
>>>> + fi
>>>> + fi
>>>> if [ -n "${JH_INMATE_DTB}" -a -n "${JH_LINUX_DEMO_CELL}" ]; then
>>>> cd ${TOOLS_SRC_DIR}
>>>> @@ -110,6 +122,7 @@ do_install() {
>>>> ./jailhouse-cell-linux -w ${D}${JH_DATADIR}/${JH_INMATE_DTB} \
>>>> -a ${JH_ARCH} -c "${JH_CMDLINE}" \
>>>> -d ../configs/${JH_ARCH}/dts/${JH_INMATE_DTB} \
>>>> + -i ${D}/boot/${JH_RAMFS_IMAGE}-${MACHINE}.cpio \
>>>> ${D}${CELL_DIR}/${JH_LINUX_DEMO_CELL} \
>>>> ${DEPLOY_DIR_IMAGE}/Image \
>>>> | tr -cd '\11\12\15\40-\176' \
>>>> @@ -143,9 +156,9 @@ CELLS = ""
>>>> python __anonymous () {
>>>> d.appendVarFlag('do_install', 'depends', ' virtual/kernel:do_deploy')
>>>> - initrd = d.getVar('INITRAMFS_IMAGE', True)
>>>> - if initrd:
>>>> - d.appendVarFlag('do_install', 'depends', ' ${INITRAMFS_IMAGE}:do_image_complete')
>>>> + ramfs = d.getVar('JH_RAMFS_IMAGE', True)
>>>> + if ramfs:
>>>> + d.appendVarFlag('do_install', 'depends', ' ${JH_RAMFS_IMAGE}:do_image_complete')
>>>> # Setup DEPENDS and RDEPENDS to included cells
>>>> cells = d.getVar('CELLS', True) or ""
>>>> --
>>>> 2.7.4
>>>>
>>>> --
>>>> _______________________________________________
>>>> meta-ti mailing list
>>>> meta-ti at yoctoproject.org
>>>> https://lists.yoctoproject.org/listinfo/meta-ti
More information about the meta-ti
mailing list