From: james.morse@arm•com (James Morse)
To: linux-arm-kernel@lists•infradead.org
Subject: [PATCH v24 5/9] arm64: kdump: add kdump support
Date: Fri, 19 Aug 2016 14:28:06 +0100 [thread overview]
Message-ID: <57B70966.2000105@arm.com> (raw)
In-Reply-To: <20160819012651.GE20080@linaro.org>
On 19/08/16 02:26, AKASHI Takahiro wrote:
> Can you try the following change?
> If it fixes your problem, I will post it as a patch.
Almost! This causes booting with acpi=on to fail for the familiar
alignment-fault reasons[2],
details and a suggested fix below.
I think we should have this change as it matches x86's use of acpi, and means we
don't rely on re-parsing the efi memory map to learn which areas of memory
shouldn't be in the vmcore.
> ===8<===
> From 740563e4a437f0d6ecf6e421c91433f9b8f19041 Mon Sep 17 00:00:00 2001
> From: AKASHI Takahiro <takahiro.akashi@linaro•org>
> Date: Fri, 19 Aug 2016 09:57:52 +0900
> Subject: [PATCH] arm64: mark reserved memblock regions explicitly
>
> ---
> arch/arm64/kernel/setup.c | 9 +++++++--
> 1 file changed, 7 insertions(+), 2 deletions(-)
>
> diff --git a/arch/arm64/kernel/setup.c b/arch/arm64/kernel/setup.c
> index 38eda13..38589b5 100644
> --- a/arch/arm64/kernel/setup.c
> +++ b/arch/arm64/kernel/setup.c
> @@ -205,10 +205,15 @@ static void __init request_standard_resources(void)
>
> for_each_memblock(memory, region) {
> res = alloc_bootmem_low(sizeof(*res));
> - res->name = "System RAM";
> + if (memblock_is_nomap(region)) {
> + res->name = "reserved";
> + res->flags = IORESOURCE_MEM | IORESOURCE_BUSY;
This causes acpica to choke. arch/arm64/include/asm/acpi.h:acpi_os_ioremap()
calls page_is_ram(), which expects IORESOURCE_SYSTEM_RAM. From kernel/resource.c:
> /*
> * This generic page_is_ram() returns true if specified address is
> * registered as System RAM in iomem_resource list.
> */
> int __weak page_is_ram(unsigned long pfn)
We are trying to infer information about the EFI memory map by looking through
iomem_resource list generated from memblock.
drivers/firmware/efi/arm-init.c:reserve_regions() adds memory with the WB
attribute to memblock via early_init_dt_add_memory_arch(), so changing
page_is_ram() for memblock_is_memory() is one step closer to checking the
attributes in the efi memory map (which turns out to tricky).
With your v24 on v4.8-rc1 and 'mark reserved memblock regions explicitly', and
this extra hack [0], I can boot, kdump, extract the vmcore (with read() and
mmap()), and pull things out of it with crash.
Thanks,
James
> + } else {
> + res->name = "System RAM";
> + res->flags = IORESOURCE_SYSTEM_RAM | IORESOURCE_BUSY;
> + }
> res->start = __pfn_to_phys(memblock_region_memory_base_pfn(region));
> res->end = __pfn_to_phys(memblock_region_memory_end_pfn(region)) - 1;
> - res->flags = IORESOURCE_SYSTEM_RAM | IORESOURCE_BUSY;
>
> request_resource(&iomem_resource, res);
[0] whitespace damaged hack for acpi_os_ioremap().
------------------------------------%<------------------------------------
+++ b/arch/arm64/include/asm/acpi.h
@@ -12,6 +12,7 @@
#ifndef _ASM_ACPI_H
#define _ASM_ACPI_H
+#include <linux/memblock.h>
#include <linux/mm.h>
#include <linux/psci.h>
@@ -32,7 +33,11 @@
static inline void __iomem *acpi_os_ioremap(acpi_physical_address phys,
acpi_size size)
{
- if (!page_is_ram(phys >> PAGE_SHIFT))
+ /*
+ * EFI's reserve_regions() call adds memory with the WB attribute
+ * to memblock via early_init_dt_add_memory_arch().
+ */
+ if (!memblock_is_memory(phys))
return ioremap(phys, size);
return ioremap_cache(phys, size);
------------------------------------%<------------------------------------
[1] cat /proc/iomem | tail -n 9
8000000000-8001e7ffff : reserved
8001e80000-83ff17ffff : System RAM
8002080000-8002b0ffff : Kernel code
8002d90000-8002f6ffff : Kernel data
80dfe00000-80ffdfffff : Crash kernel
83ff180000-83ff1cffff : reserved
83ff1d0000-83ff21ffff : System RAM
83ff220000-83ffe4ffff : reserved
83ffe50000-83ffffffff : System RAM
[2] Panic on boot without [0]
[ 0.037098] ACPI: Core revision 20160422
[ 0.041174] Unable to handle kernel paging request at virtual address
fffffc0008f831e7
[ 0.049155] pgd = fffffc0008f50000
[ 0.052576] [fffffc0008f831e7] *pgd=00000083fffe0003, *pud=00000083fffe0003,
*pmd=00000083fffe0003, *pte=00e80083ff1c0707
[ 0.063632] Internal error: Oops: 96000021 [#1] PREEMPT SMP
[ 0.069245] Modules linked in:
[ 0.072317] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 4.8.0-rc1+ #4892
[ 0.078891] Hardware name: AMD Seattle (Rev.B0) Development Board (Overdrive)
(DT)
[ 0.086518] task: fffffc0008dc8a80 task.stack: fffffc0008d90000
[ 0.092486] PC is at acpi_ns_lookup+0x238/0x378
[ 0.097049] LR is@acpi_ds_load1_begin_op+0x88/0x260
[ 0.102290] pc : [<fffffc000840fed8>] lr : [<fffffc0008405dcc>] pstate: 60000045
[ 0.109740] sp : fffffc0008d93bb0
[ 0.113071] x29: fffffc0008d93bb0 x28: 0000000000000001
[ 0.118420] x27: 0000000000000000 x26: 000000000000001b
[ 0.123768] x25: fffffc0008a813b7 x24: 000000000000001b
[ 0.129119] x23: 000000000000001b x22: 0000000000000001
[ 0.134466] x21: fffffc0008d93cb8 x20: 0000000000000001
[ 0.139814] x19: fffffc0008f831e7 x18: 0000000000000023
[ 0.145163] x17: ffffffffffffffff x16: 0000000000000000
[ 0.150511] x15: 000000000000038e x14: 0000000000007fff
[ 0.155859] x13: ffffffffffffff00 x12: 0000000000000008
[ 0.161208] x11: 0000000000000028 x10: 00000000ffffff76
[ 0.166556] x9 : 0000000000000000 x8 : fffffe03c00b0140
[ 0.171907] x7 : 0000000000000003 x6 : fffffc0008d93cb8
[ 0.177255] x5 : fffffe03c0060800 x4 : fffffc0008ee7fe8
[ 0.182603] x3 : fffffc0008ee8000 x2 : fffffc0008ee7fe8
[ 0.187951] x1 : fffffc0008f831e7 x0 : 0000000000000000
[ 0.193296]
[ 0.194789] Process swapper/0 (pid: 0, stack limit = 0xfffffc0008d90020)
[ 0.483605] Call trace:
[ 0.486062] Exception stack(0xfffffc0008d939e0 to 0xfffffc0008d93b10)
[ 0.568522] [<fffffc000840fed8>] acpi_ns_lookup+0x238/0x378
[ 0.574134] [<fffffc0008405dcc>] acpi_ds_load1_begin_op+0x88/0x260
[ 0.580358] [<fffffc0008415e3c>] acpi_ps_build_named_op+0xa8/0x170
[ 0.586582] [<fffffc0008416034>] acpi_ps_create_op+0x130/0x230
[ 0.592455] [<fffffc0008415988>] acpi_ps_parse_loop+0x174/0x580
[ 0.598419] [<fffffc00084168b4>] acpi_ps_parse_aml+0xa8/0x278
[ 0.604208] [<fffffc0008411efc>] acpi_ns_one_complete_parse+0x130/0x158
[ 0.610871] [<fffffc0008411f48>] acpi_ns_parse_table+0x24/0x44
[ 0.616745] [<fffffc00084116fc>] acpi_ns_load_table+0x50/0xe0
[ 0.622535] [<fffffc000841ad80>] acpi_tb_load_namespace+0xdc/0x25c
[ 0.628764] [<fffffc0008b383f4>] acpi_load_tables+0x3c/0xa8
[ 0.634377] [<fffffc0008b37548>] acpi_early_init+0x88/0xbc
[ 0.639903] [<fffffc0008b10b30>] start_kernel+0x330/0x394
[ 0.645340] [<fffffc0008b10210>] __primary_switched+0x7c/0x8c
[ 0.651127] Code: 2a1a03f7 2a0002d6 36380054 321902d6 (b9400260)
[ 0.657270] ---[ end trace 0000000000000000 ]---
[ 0.661922] Kernel panic - not syncing: Attempted to kill the idle task!
[ 0.668673] ---[ end Kernel panic - not syncing: Attempted to kill the idle task!
next prev parent reply other threads:[~2016-08-19 13:28 UTC|newest]
Thread overview: 52+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-08-09 1:52 [PATCH v24 0/9] arm64: add kdump support AKASHI Takahiro
2016-08-09 1:52 ` [PATCH v24 1/9] arm64: kdump: reserve memory for crash dump kernel AKASHI Takahiro
2016-08-09 1:55 ` [PATCH v24 2/9] memblock: add memblock_cap_memory_range() AKASHI Takahiro
2016-08-10 16:26 ` James Morse
2016-08-09 1:56 ` [PATCH v24 3/9] arm64: limit memory regions based on DT property, usable-memory-range AKASHI Takahiro
2016-08-09 1:56 ` [PATCH v24 4/9] arm64: kdump: implement machine_crash_shutdown() AKASHI Takahiro
2016-08-09 1:56 ` [PATCH v24 5/9] arm64: kdump: add kdump support AKASHI Takahiro
2016-08-10 16:38 ` James Morse
2016-08-10 18:18 ` Pratyush Anand
2016-08-11 10:03 ` Pratyush Anand
2016-08-16 10:13 ` James Morse
2016-08-17 15:33 ` [PATCH] fixup! " James Morse
2016-08-18 7:32 ` AKASHI Takahiro
2016-08-19 8:00 ` Pratyush Anand
2016-08-19 13:34 ` James Morse
2016-08-19 15:19 ` Pratyush Anand
2016-08-18 7:15 ` [PATCH v24 5/9] " AKASHI Takahiro
2016-08-18 7:19 ` Dave Young
2016-08-19 1:26 ` AKASHI Takahiro
2016-08-19 11:22 ` Pratyush Anand
2016-08-22 1:29 ` AKASHI Takahiro
2016-08-22 7:07 ` Pratyush Anand
2016-08-22 13:47 ` James Morse
2016-08-23 0:38 ` AKASHI Takahiro
2016-08-23 11:23 ` Pratyush Anand
2016-08-24 8:04 ` Dave Young
2016-08-24 10:25 ` James Morse
2016-08-25 1:04 ` Dave Young
2016-08-19 13:28 ` James Morse [this message]
2016-08-22 1:23 ` AKASHI Takahiro
2016-08-24 14:44 ` Ard Biesheuvel
2016-08-26 6:22 ` AKASHI Takahiro
2016-08-09 1:56 ` [PATCH v24 6/9] arm64: kdump: add VMCOREINFO's for user-space coredump tools AKASHI Takahiro
2016-08-09 1:56 ` [PATCH v24 7/9] arm64: kdump: enable kdump in the arm64 defconfig AKASHI Takahiro
2016-08-09 1:56 ` [PATCH v24 8/9] arm64: kdump: update a kernel doc AKASHI Takahiro
2016-08-09 1:57 ` [PATCH v24 9/9] Documentation: dt: chosen properties for arm64 kdump AKASHI Takahiro
2016-08-19 13:26 ` Rob Herring
2016-08-22 4:28 ` AKASHI Takahiro
2016-08-30 16:34 ` Rob Herring
2016-08-30 23:45 ` AKASHI Takahiro
2016-08-31 5:02 ` AKASHI Takahiro
2016-09-02 10:11 ` AKASHI Takahiro
2016-09-27 23:39 ` Mark Rutland
2016-08-09 2:04 ` [PATCH v24 0/9] arm64: add kdump support AKASHI Takahiro
2016-08-31 3:41 ` Manish Jaggi
2016-08-31 5:31 ` AKASHI Takahiro
2016-09-02 12:53 ` Manish Jaggi
2016-09-05 8:15 ` AKASHI Takahiro
2016-09-05 12:42 ` Manish Jaggi
2016-09-06 15:33 ` Marc Zyngier
2016-09-06 16:15 ` Manish Jaggi
2016-09-06 16:42 ` Marc Zyngier
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=57B70966.2000105@arm.com \
--to=james.morse@arm$(echo .)com \
--cc=linux-arm-kernel@lists$(echo .)infradead.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox