On 25/11/2019 19:49, Masayoshi Mizuma wrote:
> From: Masayoshi Mizuma <m.mizuma@xxxxxxxxxxxxxx>
>
> kexec reboot stops in early boot sequence because efi_config_parse_tables()
> refers garbage data. We can see the log with memblock=debug kernel option:
>
> efi: ACPI 2.0=0x9821790014 PROP=0x8757f5c0 SMBIOS 3.0=0x9820740000 MEMRESERVE=0x9820bfdc58
> memblock_reserve: [0x0000009820bfdc58-0x0000009820bfdc67] efi_config_parse_tables+0x228/0x278
> memblock_reserve: [0x0000000082760000-0x00000000324d07ff] efi_config_parse_tables+0x228/0x278
> memblock_reserve: [0xcc4f84ecc0511670-0x5f6e5214a7fd91f9] efi_config_parse_tables+0x244/0x278
> memblock_reserve: [0xd2fd4144b9af693d-0xad0c1db1086f40a2] efi_config_parse_tables+0x244/0x278
> memblock_reserve: [0x0c719bb159b1fadc-0x5aa6e62a1417ce12] efi_config_parse_tables+0x244/0x278
> ...
>
> That happens because 0x82760000, struct linux_efi_memreserve, is destroyed.
> 0x82760000 is pointed from efi.mem_reseve, and efi.mem_reserve points the
> head page of LPI pending table and LPI property table which are allocated by
> gic_reserve_range().
>
> The destroyer is kexec. kexec locates the initrd to the area:
>
> ]# kexec -d -l /boot/vmlinuz-5.4.0-rc7 /boot/initramfs-5.4.0-rc7.img --reuse-cmdline
> ...
> initrd: base 82290000, size 388dd8ah (59301258)
> ...
>
> From dynamic debug log. initrd is located in segment[1]:
> machine_kexec_prepare:70:
> kexec kimage info:
> type: 0
> start: 85b30680
> head: 0
> nr_segments: 4
> segment[0]: 0000000080480000 - 0000000082290000, 0x1e10000 bytes, 481 pages
> segment[1]: 0000000082290000 - 0000000085b20000, 0x3890000 bytes, 905 pages
> segment[2]: 0000000085b20000 - 0000000085b30000, 0x10000 bytes, 1 pages
> segment[3]: 0000000085b30000 - 0000000085b40000, 0x10000 bytes, 1 pages
>
> kexec searches the memory region to locate initrd through
> "System RAM" in /proc/iomem. The pending tables are included in
> "System RAM" because they are allocated by alloc_pages(), so kexec
> destroys the LPI pending tables.
>
Doesn't that mean that you haven't enough memory reserved so that you have to
fallback to allocate it via __get_free_page()?
> Introduce /sys/firmware/efi/memreserve to tell the pages pointed by
> efi.mem_reserve so that kexec can avoid the area to locate initrd.
>
Doesn't that need a patch for kexec-tools to actually take this into account?
> Signed-off-by: Masayoshi Mizuma <m.mizuma@xxxxxxxxxxxxxx>
> ---
> drivers/firmware/efi/efi.c | 45 +++++++++++++++++++++++++++++++++++++-
> 1 file changed, 44 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/firmware/efi/efi.c b/drivers/firmware/efi/efi.c
> index e98bbf8e5..0aa07cc09 100644
> --- a/drivers/firmware/efi/efi.c
> +++ b/drivers/firmware/efi/efi.c
> @@ -141,6 +141,47 @@ static ssize_t systab_show(struct kobject *kobj,
>
> static struct kobj_attribute efi_attr_systab = __ATTR_RO_MODE(systab, 0400);
>
> +static struct linux_efi_memreserve *efi_memreserve_root __ro_after_init;
> +#ifdef CONFIG_KEXEC
> +static ssize_t memreserve_show(struct kobject *kobj,
> + struct kobj_attribute *attr, char *buf)
> +{
> + struct linux_efi_memreserve *rsv;
> + phys_addr_t start, end;
> + unsigned long prsv;
> + char *str = buf;
> + int count, i;
> +
> + if (!kobj || !buf)
> + return -EINVAL;
> +
> + if ((efi_memreserve_root == (void *)ULONG_MAX) ||
> + (!efi_memreserve_root))
> + return -ENODEV;
> +
> + for (prsv = efi_memreserve_root->next; prsv; prsv = rsv->next) {
> + rsv = memremap(prsv, sizeof(*rsv), MEMREMAP_WB);
> + if (!rsv) {
> + pr_err("Could not map efi_memreserve\n");
> + return -ENOMEM;
> + }
> + count = atomic_read(&rsv->count);
> + for (i = 0; i < count; i++) {
> + start = rsv->entry[i].base;
> + end = start + rsv->entry[i].size - 1;
> +
> + str += sprintf(str, "%pa-%pa\n", &start, &end);
What happens if we provide a buf which is too small?
Regards,
Matthias
