On Wed, 7 Sep 2022, Geert Uytterhoeven wrote: > On Wed, Sep 7, 2022 at 4:04 AM Stephen Walsh <vk3...@vk3heg.net> wrote: > > On Wed, 7 Sep 2022 11:59:17 +1000 (AEST) > > Finn Thain <fth...@linux-m68k.org> wrote: > > > > > > Here are three debugs..... mixing the kernel/intrid's... > > > > > > What happens if you boot the 5.19 kernel/initrd pair with > > > "root=/dev/sda2" omitted from the kernel parameters? > > > > > > It looks like the oops from the 5.18 build disappeared in the 5.19 > > > build. > > It disappeared because the (too large?) initrd is no longer loaded... > > > > That makes me think that the 5.18 build would no longer oops when > > > used with the 5.19 initrd (?) > > > > All boot attempts have ran with the following amiboot line (Changing > > the kernel/initrd of course) > > > > amiboot -k vmlinux-5.15.0-2-m68k -r initrd.img-5.15.0-2-m68k > > root=/dev/sda2 fb=false debug=mem > > Are you sure that is what you really used? > > I missed this before, but all kernels that failed to mount their root > file systems, and printed no "initrd: <start> - <end>" lines, had the > "-r <initrd>" parameter in the kernel command line part: > > Kernel command line: ignore_loglevel -r initrd.img-5.19.0-1-m68k > root=/dev/sda2 fb=false debug=mem BOOT_IMAGE=vmlinux-5.19.0-1-m68k > > Hence Amiboot didn't provide the initrd... >
Okay, so the first couple of logs Stephen sent used the correct command line. From those we can see, Success: [ 0.000000] initrd: 0783283d - 08000000 Failure: [ 0.000000] initrd: 0f7f81c2 - 10000000 Stephen, can you please confirm that (given the correct bootloader command) 5.15 does indeed oops when presented with an initrd that's 8 MiB (8388608 bytes) or larger? And conversely, 5.18 does not oops when presented with an initrd that's smaller than 8 MiB?
