crashkernel reservation failed - No suitable area found on a cortina/gemini SoC
Corentin Labbe
clabbe.montjoie at gmail.com
Fri Apr 9 16:06:27 BST 2021
Le Fri, Apr 09, 2021 at 04:53:13PM +0200, Corentin Labbe a écrit :
> Le Wed, Apr 07, 2021 at 07:59:27AM -0700, Bruce Mitchell a écrit :
> > On 4/7/2021 07:48, Corentin Labbe wrote:
> > > Le Wed, Apr 07, 2021 at 07:28:26AM -0700, Bruce Mitchell a écrit :
> > >> On 4/7/2021 07:23, Corentin Labbe wrote:
> > >>> Le Wed, Apr 07, 2021 at 07:13:04AM -0700, Bruce Mitchell a écrit :
> > >>>> On 4/7/2021 05:54, Corentin Labbe wrote:
> > >>>>> Hello
> > >>>>>
> > >>>>> I try to do kexec on a cortina/gemini SoC.
> > >>>>> On a "normal" boot, kexec fail to find memory so I added crashkernel=8M to cmdline. (kernel size is ~6M).
> > >>>>> But now, kernel fail to reserve memory:
> > >>>>> Load Kern image from 0x30020000 to 0x800000 size 7340032
> > >>>>> Booting Linux on physical CPU 0x0
> > >>>>> Linux version 5.12.0-rc5-next-20210401+ (compile at Red) (armv7a-unknown-linux-gnueabihf-gcc (Gentoo 9.3.0-r2 p4) 9.3.0, GNU ld (Gentoo 2.34 p6) 2.34.0) #98 PREEMPT Wed Apr 7 14:14:08 CEST 2021
> > >>>>> CPU: FA526 [66015261] revision 1 (ARMv4), cr=0000397f
> > >>>>> CPU: VIVT data cache, VIVT instruction cache
> > >>>>> OF: fdt: Machine model: Edimax NS-2502
> > >>>>> Memory policy: Data cache writeback
> > >>>>> Zone ranges:
> > >>>>> Normal [mem 0x0000000000000000-0x0000000007ffffff]
> > >>>>> HighMem empty
> > >>>>> Movable zone start for each node
> > >>>>> Early memory node ranges
> > >>>>> node 0: [mem 0x0000000000000000-0x0000000007ffffff]
> > >>>>> Initmem setup node 0 [mem 0x0000000000000000-0x0000000007ffffff]
> > >>>>> crashkernel reservation failed - No suitable area found.
> > >>>>> Built 1 zonelists, mobility grouping on. Total pages: 32512
> > >>>>> Kernel command line: console=ttyS0,19200n8 ip=dhcp crashkernel=8M
> > >>>>> Dentry cache hash table entries: 16384 (order: 4, 65536 bytes, linear)
> > >>>>> Inode-cache hash table entries: 8192 (order: 3, 32768 bytes, linear)
> > >>>>> mem auto-init: stack:off, heap alloc:off, heap free:off
> > >>>>> Memory: 119476K/131072K available (5034K kernel code, 579K rwdata, 1372K rodata, 3020K init, 210K bss, 11596K reserved, 0K cma-reserved, 0K highmem)
> > >>>>> SLUB: HWalign=32, Order=0-3, MinObjects=0, CPUs=1, Nodes=1
> > >>>>>
> > >>>>> What can I do ?
> > >>>>>
> > >>>>> Thanks
> > >>>>> Regards
> > >>>>>
> > >>>>> _______________________________________________
> > >>>>> kexec mailing list
> > >>>>> kexec at lists.infradead.org
> > >>>>> http://lists.infradead.org/mailman/listinfo/kexec
> > >>>>>
> > >>>>
> > >>>> Hello Corentin,
> > >>>>
> > >>>> I see much larger crashkernel=xxM being shown here
> > >>>> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/admin-guide/kdump/kdump.rst
> > >>>> and from many of my other searches.
> > >>>>
> > >>>> Here is an interesting article on kdump for ARM-32
> > >>>> https://kaiwantech.wordpress.com/2017/07/13/setting-up-kdump-and-crash-for-arm-32-an-ongoing-saga/
> > >>>>
> > >>>>
> > >>>> Here is the kernel command line reference
> > >>>> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/admin-guide/kernel-parameters.txt?h=v5.11#n732
> > >>>>
> > >>>> I feel your frustrations too.
> > >>>
> > >>> Hello
> > >>>
> > >>> Thanks but I have already read those documentation.
> > >>> I search to know why the kernel cannot find 8M of memory ouf of 128.
> > >>>
> > >>> Regards
> > >>>
> > >>
> > >> How much more memory does the kernel and initrd above and beyond just
> > >> their physical size? (heaps, stacks, buffers, virtual filesystems)
> > >
> > > The kernel size include a rootfs.cpio.lzma of 3MB and dtb is appended.
> > > The total kernel size is 7MB.
> > > The uncompressed size of the kernel is 13M (size of vmlinux)
> > > The uncompressed size of rootfs is 11M.
> > >
> > > cat /proc/meminfo
> > > MemTotal: 122496 kB
> > > MemFree: 103700 kB
> > > MemAvailable: 101936 kB
> > > Buffers: 0 kB
> > > Cached: 10904 kB
> > > SwapCached: 0 kB
> > > Active: 4304 kB
> > > Inactive: 8012 kB
> > > Active(anon): 4304 kB
> > > Inactive(anon): 8012 kB
> > > Active(file): 0 kB
> > > Inactive(file): 0 kB
> > > Unevictable: 0 kB
> > > Mlocked: 0 kB
> > > HighTotal: 0 kB
> > > HighFree: 0 kB
> > > LowTotal: 122496 kB
> > > LowFree: 103700 kB
> > > SwapTotal: 0 kB
> > > SwapFree: 0 kB
> > > Dirty: 0 kB
> > > Writeback: 0 kB
> > > AnonPages: 1428 kB
> > > Mapped: 3552 kB
> > > Shmem: 10904 kB
> > > KReclaimable: 608 kB
> > > Slab: 2960 kB
> > > SReclaimable: 608 kB
> > > SUnreclaim: 2352 kB
> > > KernelStack: 312 kB
> > > PageTables: 136 kB
> > > NFS_Unstable: 0 kB
> > > Bounce: 0 kB
> > > WritebackTmp: 0 kB
> > > CommitLimit: 61248 kB
> > > Committed_AS: 14336 kB
> > > VmallocTotal: 901120 kB
> > > VmallocUsed: 64 kB
> > > VmallocChunk: 0 kB
> > > Percpu: 32 kB
> > > CmaTotal: 0 kB
> > > CmaFree: 0 kB
> > >
> >
> > I believe you need space for all of that,
> > the smallest that would work for me was 20MB.
>
> I tried without any change.
>
> Anyway when trying to kexec I got:
> kexec --no-ifdown --command-line="console=ttyS0,19200n8" /tmp/kernel |
> Could not find a free area of memory of 0x668a8a bytes...
> Cannot load /tmp/kernel
>
> So reserving 8M is enough according to what kexec said.
>
> So anyone know why the kernel cannot reserve 8M ?
> Thanks
It seems to be related to:
arch/arm/kernel/setup.c:977 "The crash region must be aligned to 128MB to avoid"
Hacking CRASH_ALIGN to 64 permit Linux to reserve 8M at boot.
But kexec still fail after with the same reason. (Could not find a free area of memory of 0x668a8a bytes)
More information about the kexec
mailing list