[U-Boot] [PATCH] ARM: omap3: evm: Do not relocate FDT address

Derald Woods woods.technical at gmail.com
Mon Jan 1 18:38:53 UTC 2018


On Mon, Jan 1, 2018 at 7:29 AM, Tom Rini <trini at konsulko.com> wrote:

> On Thu, Dec 28, 2017 at 05:17:13PM -0600, Derald D. Woods wrote:
> > On Thu, Dec 28, 2017 at 02:24:08PM -0500, Tom Rini wrote:
> > > On Thu, Dec 28, 2017 at 01:21:05PM -0600, Derald D. Woods wrote:
> > > > On Thu, Dec 28, 2017 at 01:43:43PM -0500, Tom Rini wrote:
> > > > > On Thu, Dec 28, 2017 at 11:48:29AM -0600, Derald D. Woods wrote:
> > > > > > On Thu, Dec 28, 2017 at 10:37:18AM -0500, Tom Rini wrote:
> > > > > > > On Thu, Dec 28, 2017 at 01:25:43AM -0600, Derald D. Woods
> wrote:
> > > > > > >
> > > > > > > > This commit keeps the 'fdtaddr' as provided by
> DEFAULT_LINUX_BOOT_ENV.
> > > > > > > >
> > > > > > > > Signed-off-by: Derald D. Woods <woods.technical at gmail.com>
> > > > > > > > ---
> > > > > > > >  include/configs/omap3_evm.h | 1 +
> > > > > > > >  1 file changed, 1 insertion(+)
> > > > > > > >
> > > > > > > > diff --git a/include/configs/omap3_evm.h
> b/include/configs/omap3_evm.h
> > > > > > > > index 629d60b961..d95ccdf035 100644
> > > > > > > > --- a/include/configs/omap3_evm.h
> > > > > > > > +++ b/include/configs/omap3_evm.h
> > > > > > > > @@ -127,6 +127,7 @@
> > > > > > > >   "fdtfile=" CONFIG_DEFAULT_FDT_FILE "\0" \
> > > > > > > >   "mtdids=" CONFIG_MTDIDS_DEFAULT "\0" \
> > > > > > > >   "mtdparts=" CONFIG_MTDPARTS_DEFAULT "\0" \
> > > > > > > > + "fdt_high=0xffffffff\0" \
> > > > > > > >   "bootenv=uEnv.txt\0" \
> > > > > > > >   "optargs=\0" \
> > > > > > > >   "mmcdev=0\0" \
> > > > > > >
> > > > > > > What's the problem this solves, and how much memory is on the
> system in
> > > > > > > question?  bootm_size should be ensuring we never relocate it
> outside of
> > > > > > > the first 256MB of memory.  Thanks!
> > > > > > >
> > > > > >
> > > > > > The logic within "common/image-fdt.c" lead me down this path. The
> > > > > > addition of this fairly common environment variable allowed my
> > > > > > OMAP34XX board to boot the kernel without an appended devicetree.
> > > > > > When I use the variable to trigger 'disable_relocation', as the
> code
> > > > > > indicates, the kernel boot behavior is what I expect. I spent a
> few
> > > > > > hours following the code path to a single line edition that
> works.
> > > > > >
> > > > > >
> > > > > > Without 'fdt_high'
> > > > > > ==================
> > > > > >
> > > > > > U-Boot SPL 2018.01-rc2-00028-gaf9da945cd-dirty (Dec 28 2017 -
> 11:16:52)
> > > > > > Trying to boot from MMC1
> > > > > > reading u-boot.img
> > > > > > reading u-boot.img
> > > > > >
> > > > > >
> > > > > > U-Boot 2018.01-rc2-00028-gaf9da945cd-dirty (Dec 28 2017 -
> 11:16:52 -0600)
> > > > > >
> > > > > > OMAP3530-GP ES3.1, CPU-OPP2, L3-165MHz, Max CPU Clock 720 MHz
> > > > > > Model: TI OMAP35XX EVM (TMDSEVM3530)
> > > > > > OMAP3 EVM board + LPDDR/NAND
> > > > > > I2C:   ready
> > > > > > DRAM:  256 MiB
> > > > > > NAND:  256 MiB
> > > > > > MMC:   OMAP SD/MMC: 0
> > > > > > Read back SMSC id 0x92200000
> > > > > > OMAP die ID: 265a002400000000040365fa1801b01f
> > > > > > Net:   smc911x-0
> > > > > > starting USB...
> > > > > > USB0:   USB EHCI 1.00
> > > > > > scanning bus 0 for devices... 1 USB Device(s) found
> > > > > > Hit any key to stop autoboot:  0
> > > > > > switch to partitions #0, OK
> > > > > > mmc0 is current device
> > > > > > Scanning mmc 0:1...
> > > > > > switch to partitions #0, OK
> > > > > > mmc0 is current device
> > > > > > ** Unable to read file uEnv.txt **
> > > > > > reading zImage
> > > > > > 4637344 bytes read in 407 ms (10.9 MiB/s)
> > > > > > reading omap3-evm.dtb
> > > > > > 62832 bytes read in 10 ms (6 MiB/s)
> > > > > > Booting zImage from mmc ...
> > > > > > *  fdt: cmdline image address = 0x88000000
> > > > > > ## Checking for 'FDT'/'FDT Image' at 88000000
> > > > > > *  fdt: raw FDT blob
> > > > > > ## Flattened Device Tree blob at 88000000
> > > > > >    Booting using the fdt blob at 0x88000000
> > > > > >    of_flat_tree at 0x88000000 size 0x0000f570
> > > > > > ## device tree at 88000000 ... 8800f56f (len=75120 [0x12570])
> > > > > >    Loading Device Tree to 8ffed000, end 8ffff56f ... OK
> > > > > >
> > > > > > Starting kernel ...
> > > > > >
> > > > > > [HANG]
> > > > > >
> > > > > >
> > > > > > Make note of the "of_flat_tree" and "Loading Device Tree" lines.
> > > > >
> > > > > Ah, hmmm.  Can you please try setting bootm_size, as an
> experiment, to
> > > > > 0x0a000000 ?  I wonder if, since you have 256MB of memory we're not
> > > > > putting the DTB into where U-Boot is and we stomp on it a bit,
> causing
> > > > > Linux to boot, see an invalid DTB and not have a console available
> to
> > > > > tell us.  Thanks!
> > > > >
> > > >
> > > >
> > > > That seemed to do it.
> > >
> > > Ah, good!
> > >
> > > > ---8<-------------------------------------------------------
> ---------
> > > >
> > > > OMAP3_EVM # echo $bootm_size
> > > > 0x0a000000
> > > > OMAP3_EVM # reset
> > > > resetting ...
> > > > O
> > > > U-Boot SPL 2018.01-rc2-00028-gaf9da945cd-dirty (Dec 28 2017 -
> 13:03:42)
> > > > Trying to boot from MMC1
> > > > reading u-boot.img
> > > > reading u-boot.img
> > > >
> > > >
> > > > U-Boot 2018.01-rc2-00028-gaf9da945cd-dirty (Dec 28 2017 - 13:03:42
> -0600)
> > > >
> > > > OMAP3530-GP ES3.1, CPU-OPP2, L3-165MHz, Max CPU Clock 720 MHz
> > > > Model: TI OMAP35XX EVM (TMDSEVM3530)
> > > > OMAP3 EVM board + LPDDR/NAND
> > > > I2C:   ready
> > > > DRAM:  256 MiB
> > > > NAND:  256 MiB
> > > > MMC:   OMAP SD/MMC: 0
> > > > Read back SMSC id 0x92200000
> > > > OMAP die ID: 265a002400000000040365fa1801b01f
> > > > Net:   smc911x-0
> > > > starting USB...
> > > > USB0:   USB EHCI 1.00
> > > > scanning bus 0 for devices... 1 USB Device(s) found
> > > > Hit any key to stop autoboot:  0
> > > > switch to partitions #0, OK
> > > > mmc0 is current device
> > > > Scanning mmc 0:1...
> > > > switch to partitions #0, OK
> > > > mmc0 is current device
> > > > ** Unable to read file uEnv.txt **
> > > > reading zImage
> > > > 4637344 bytes read in 407 ms (10.9 MiB/s)
> > > > reading omap3-evm.dtb
> > > > 62832 bytes read in 10 ms (6 MiB/s)
> > > > Booting zImage from mmc ...
> > > > *  fdt: cmdline image address = 0x88000000
> > > > ## Checking for 'FDT'/'FDT Image' at 88000000
> > > > *  fdt: raw FDT blob
> > > > ## Flattened Device Tree blob at 88000000
> > > >    Booting using the fdt blob at 0x88000000
> > > >    of_flat_tree at 0x88000000 size 0x0000f570
> > > > ## device tree at 88000000 ... 8800f56f (len=75120 [0x12570])
> > > >    Loading Device Tree to 89fed000, end 89fff56f ... OK
> > > >
> > > > Starting kernel ...
> > > >
> > > > [    0.000000] Booting Linux on physical CPU 0x0
> > > > [    0.000000] Linux version 4.15.0-rc5-1-gdc7de7cf5f08
> (ddwoods at ethiopia) (gcc version 7.2.0 (crosstool-NG
> crosstool-ng-1.23.0-269-ge832b9b2 - Linux 4.14.y)) #17 PREEMPT Wed Dec 27
> 22:59:37 CST 2017
> > > > [    0.000000] CPU: ARMv7 Processor [411fc083] revision 3 (ARMv7),
> cr=10c5387d
> > > > [    0.000000] CPU: PIPT / VIPT nonaliasing data cache, VIPT
> nonaliasing instruction cache
> > > > [    0.000000] OF: fdt: Machine model: TI OMAP35XX EVM (TMDSEVM3530)
> > > > [    0.000000] Memory policy: Data cache writeback
> > > > [    0.000000] cma: Reserved 16 MiB at 0x8e800000
> > > > [    0.000000] CPU: All CPU(s) started in SVC mode.
> > > > [    0.000000] OMAP3430/3530 ES3.1 (l2cache iva sgx neon isp)
> > > > [    0.000000] random: fast init done
> > > > [    0.000000] Built 1 zonelists, mobility grouping on.  Total
> pages: 64706
> > > > [    0.000000] Kernel command line: console=ttyO0,115200n8
> mtdparts=omap2-nand.0:512k(spl),1792k(u-boot),128k(dtb),
> 128k(u-boot-env),6m(kernel),-(rootfs) root=/dev/mmcblk0p2 rw
> rootfstype=ext4 rootwait
> > > > [    0.000000] Dentry cache hash table entries: 32768 (order: 5,
> 131072 bytes)
> > > > [    0.000000] Inode-cache hash table entries: 16384 (order: 4,
> 65536 bytes)
> > > > [    0.000000] Memory: 218956K/261120K available (9216K kernel code,
> 742K rwdata, 3004K rodata, 1024K init, 7808K bss, 25780K reserved, 16384K
> cma-reserved, 0K highmem)
> > > > [    0.000000] Virtual kernel memory layout:
> > > > [    0.000000]     vector  : 0xffff0000 - 0xffff1000   (   4 kB)
> > > > [    0.000000]     fixmap  : 0xffc00000 - 0xfff00000   (3072 kB)
> > > > [    0.000000]     vmalloc : 0xd0000000 - 0xff800000   ( 760 MB)
> > > > [    0.000000]     lowmem  : 0xc0000000 - 0xcff00000   ( 255 MB)
> > > > [    0.000000]     pkmap   : 0xbfe00000 - 0xc0000000   (   2 MB)
> > > > [    0.000000]     modules : 0xbf000000 - 0xbfe00000   (  14 MB)
> > > > [    0.000000]       .text : 0x(ptrval) - 0x(ptrval)   (10208 kB)
> > > > [    0.000000]       .init : 0x(ptrval) - 0x(ptrval)   (1024 kB)
> > > > [    0.000000]       .data : 0x(ptrval) - 0x(ptrval)   ( 743 kB)
> > > > [    0.000000]        .bss : 0x(ptrval) - 0x(ptrval)   (7809 kB)
> > > > [    0.000000] SLUB: HWalign=64, Order=0-3, MinObjects=0, CPUs=1,
> Nodes=1
> > > > [    0.000000] ftrace: allocating 29168 entries in 86 pages
> > > > [    0.000000] Running RCU self tests
> > > > [    0.000000] Preemptible hierarchical RCU implementation.
> > > > [    0.000000]    RCU event tracing is enabled.
> > > > [    0.000000]    RCU lockdep checking is enabled.
> > > > [    0.000000]    Tasks RCU enabled.
> > > > [    0.000000] NR_IRQS: 16, nr_irqs: 16, preallocated irqs: 16
> > > >
> > > > [...]
> > > > ---8<-------------------------------------------------------
> ---------
> > > >
> > > >
> > > > This is the patch that I used.
> > > >
> > > > ---8<-------------------------------------------------------
> ---------
> > > > diff --git a/include/configs/ti_armv7_common.h
> b/include/configs/ti_armv7_common.h
> > > > index 91e139853c..23d2ef840e 100644
> > > > --- a/include/configs/ti_armv7_common.h
> > > > +++ b/include/configs/ti_armv7_common.h
> > > > @@ -48,7 +48,7 @@
> > > >         "ramdisk_addr_r=0x88080000\0" \
> > > >         "scriptaddr=0x80000000\0" \
> > > >         "pxefile_addr_r=0x80100000\0" \
> > > > -       "bootm_size=0x10000000\0" \
> > > > +       "bootm_size=0x0a000000\0" \
> > > >         "boot_fdt=try\0"
> > > >
> > > >  #define DEFAULT_FIT_TI_ARGS \
> > > > ---8<-------------------------------------------------------
> ---------
> > > >
> > > > This, I believe, would be of help for many OMAP34XX boards which have
> > > > the same memory layout as the OMAP3 EVM. Do you think submitting a
> patch
> > > > for the above would be of value? Or should it really be determined at
> > > > the individual board level?
> > >
> > > I'm not quite sure about the best way forward here.  Can you do a
> > > 'bdinfo' on your EVM and see how much memory U-Boot is taking up while
> > > running?   Thanks!
> > >
> >
> > Is there a reason disabling FDT relocation is not a valid path? It would
> > at least keep things localized to a given board. I also like knowing
> > that 'fdtaddr' is the same throughout the startup process. I just want
> > to know if there is some rule that is being broken with
> > 'fdt_high=0xffffffff'. It appears to exist by design.
>
> The downside to stopping relocation of the fdt (and of the ramdisk via
> initrd_high=0xffffffff) is that you introduce the chance (more or less,
> based on where you load everything else) that the device tree gets
> stomped on by the kernel, esp on ARM (not aarch64 where this has been
> addressed) because we don't know the size of the kernel BSS.  For this
> release, OK, lets go and do what you've got now, stopping fdt
> relocation.  What I would like to see is some logic added to the boot
> code to see what the top of bootm_size runs into where U-Boot itself is,
> and then rounding bootm_size down (and warning the user).
>
>
​Thanks for the clear explanation of the issue with respect to ARM.

Derald​


More information about the U-Boot mailing list