On 08/08/2019 05.16, Tom Rini wrote:
On Fri, Aug 02, 2019 at 03:52:28PM -0700, Julius Werner wrote:
The Linux ramdisk should always be decompressed by the kernel itself, not by U-Boot. Therefore, the 'compression' node in the FIT image should always be set to "none" for ramdisk images, since the only point of using that node is if you want U-Boot to do the decompression itself.
Yet some systems populate the node to the compression algorithm used by the kernel instead. This used to be ignored, but now that we support decompression of all image types it becomes a problem. Since ramdisks should never be decompressed by U-Boot anyway, this patch adds a special exception for them to avoid these issues. Still, setting the 'compression' node like that is wrong in the first place, so we still want to print out a warning so that third-party distributions doing this can notice and fix it.
Signed-off-by: Julius Werner jwerner@chromium.org Reviewed-by: Heiko Schocher hs@denx.de Tested-by: Heiko Schocher hs@denx.de Reviewed-by: Simon Goldschmidt simon.k.r.goldschmidt@gmail.com
This part
+ if (image_type == IH_TYPE_RAMDISK && comp != IH_COMP_NONE) + puts("WARNING: 'compression' nodes for ramdisks are deprecated," + " please fix your .its file!\n"); +
ends up being a little confusing, because when one dutifully removes the compression = "foo" property, the warning is still there (because comp ends up being (u8)-1) - the only way to silence it is by actually _having_ a 'compression = "none"' property. (It also says node instead of property).
So, what is the intention? Should ramdisk images not have a compression property at all, or must it be present but set to "none", or are either acceptable?
Rasmus