| Message ID | 20240703175727.446205-1-festevam@gmail.com |
|---|---|
| State | Changes Requested |
| Delegated to: | Tom Rini |
| Headers | show |
| Series | msc_sm2s_imx8mp: Adjust the initrd_addr location | expand |
On Wed, Jul 03, 2024 at 02:57:27PM -0300, Fabio Estevam wrote: > From: Fabio Estevam <festevam@denx.de> > > Booting an initramfs with the current address may lead to initramfs > corruption and boot failure. > > Move it 0x48000000 which avoids initramfs overlap and allows a > successful initramfs boot. > > Signed-off-by: Fabio Estevam <festevam@denx.de> > --- > include/configs/msc_sm2s_imx8mp.h | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/include/configs/msc_sm2s_imx8mp.h b/include/configs/msc_sm2s_imx8mp.h > index 3c7d96cb3c0a..2a52455ace34 100644 > --- a/include/configs/msc_sm2s_imx8mp.h > +++ b/include/configs/msc_sm2s_imx8mp.h > @@ -39,7 +39,7 @@ > "fdt_addr_r=0x43000000\0" \ > "boot_fdt=try\0" \ > "fdtfile=" CONFIG_DEFAULT_FDT_FILE "\0" \ > - "initrd_addr=0x43800000\0" \ > + "initrd_addr=0x48000000\0" \ > "bootm_size=0x10000000\0" \ > "mmcpart=1\0" \ > "mmcroot=/dev/mmcblk1p2 rootwait rw\0" \ Shouldn't fdt_addr_r change too? This is making me suspect the whole map is too small for safety, how much memory does the platform have at minimum?
Hi Tom, On Wed, Jul 3, 2024 at 3:03 PM Tom Rini <trini@konsulko.com> wrote: > Shouldn't fdt_addr_r change too? This is making me suspect the whole map > is too small for safety, how much memory does the platform have at > minimum? It has 2GB of RAM. Changing only initrd_addr and keeping fdt_addr_r unchanged allows a successful initramfs boot.
On Wed, Jul 03, 2024 at 03:07:47PM -0300, Fabio Estevam wrote: > Hi Tom, > > On Wed, Jul 3, 2024 at 3:03 PM Tom Rini <trini@konsulko.com> wrote: > > > Shouldn't fdt_addr_r change too? This is making me suspect the whole map > > is too small for safety, how much memory does the platform have at > > minimum? > > It has 2GB of RAM. > > Changing only initrd_addr and keeping fdt_addr_r unchanged allows a > successful initramfs boot. How is the fdt at 0x43000000 not being clobbered if the initramfs at 0x43800000 is? Is it because it's just enough of a bad spot that we relocate the fdt out of the way? With the kernel at 0x40480000 everything else is too low. kernel address + 128MiB for fdt address, then +512KiB (or a bit more, for nicer numbers) for initrd.
Hi Tom, On Wed, Jul 3, 2024 at 3:22 PM Tom Rini <trini@konsulko.com> wrote: > How is the fdt at 0x43000000 not being clobbered if the initramfs at > 0x43800000 is? Is it because it's just enough of a bad spot that we > relocate the fdt out of the way? With the kernel at 0x40480000 > everything else is too low. kernel address + 128MiB for fdt address, then > +512KiB (or a bit more, for nicer numbers) for initrd. Thanks for the suggestion, it works. I sent a v2 with that proposal. Thanks
diff --git a/include/configs/msc_sm2s_imx8mp.h b/include/configs/msc_sm2s_imx8mp.h index 3c7d96cb3c0a..2a52455ace34 100644 --- a/include/configs/msc_sm2s_imx8mp.h +++ b/include/configs/msc_sm2s_imx8mp.h @@ -39,7 +39,7 @@ "fdt_addr_r=0x43000000\0" \ "boot_fdt=try\0" \ "fdtfile=" CONFIG_DEFAULT_FDT_FILE "\0" \ - "initrd_addr=0x43800000\0" \ + "initrd_addr=0x48000000\0" \ "bootm_size=0x10000000\0" \ "mmcpart=1\0" \ "mmcroot=/dev/mmcblk1p2 rootwait rw\0" \