Hi, I am working on a ARM imx6ull board with custom kernel. I have some freeze issue. I would like to be able to save across reboot the kernel panic log thanks to ramoops feature. I dont have physical access to the board during the kernel panic. Only way for me to understand and fix the issue is to save and store the kernel panic log to send the log to a server later To enable ramoops, I update my kernel config :
[arthur] cat output/build/linux-4.1.15_2.1.07/.config | grep PSTORE
CONFIG_PSTORE=y CONFIG_PSTORE_CONSOLE=y # CONFIG_PSTORE_PMS
I configure the ramoops platform data in the board config file of the kernel tree :
arthur * 4.1.15_2.0.0_nano] git diff arch/arm
diff --git a/arch/arm/mach-imx/mach-imx6ul.c b/arch/arm/mach-imx/mach-imx6ul.c
index dcabfb1..5b1d74c 100644
--- a/arch/arm/mach-imx/mach-imx6ul.c
+++ b/arch/arm/mach-imx/mach-imx6ul.c
@@ -23,6 +23,22 @@
#include "common.h"
#include "cpuidle.h"
#include "hardware.h"
+#include <linux/pstore_ram.h>
+#include <linux/memblock.h>
+static struct ramoops_platform_data ramoops_data = {
+ .mem_address = 0x80000000,
+ .mem_size = 0x00005000, // 5 Mb ??
+ .record_size = 0x00001000, // 1 Mb
+ .dump_oops = 1,
+};
+
+static struct platform_device ramoops_dev = {
+ .name = "ramoops",
+ .dev = {
+ .platform_data = &ramoops_data,
+ },
+};
+
static void __init imx6ul_enet_clk_init(void)
{
@@ -169,6 +185,15 @@ static inline void imx6ul_enet_init(void)
static void __init imx6ul_init_machine(void)
{
struct device *parent;
+ int ret;
+
+ ret = platform_device_register(&ramoops_dev);
+ if (ret) {
+ printk(KERN_ERR "unable to register platform device\n");
+ return;
+ }
+ /* Do I need that ???? */
+ //memblock_reserve(ramoops_data.mem_address, ramoops_data.mem_size);
parent = imx_soc_device_init();
if (parent == NULL)
Then I tried to mount the pstore partition :
# mount -t pstore -o kmsg_bytes=1000 - /sys/fs/pstore TUTU size : -335633131 ------------[ cut here ]------------ WARNING: CPU: 0 PID: 547 at mm/page_alloc.c:2668 __alloc_pages_nodemask+0x19c/0x838() Modules linked in: bcmdhd CPU: 0 PID: 547 Comm: mount Not tainted 4.1.15 #27 Hardware name: Freescale i.MX6 Ultralite (Device Tree) [<800144bc>] (unwind_backtrace) from [<800121ec>] (show_stack+0x10/0x14) [<800121ec>] (show_stack) from [<80034444>] (warn_slowpath_common+0x80/0xac) [<80034444>] (warn_slowpath_common) from [<8003450c>] (warn_slowpath_null+0x1c/0x24) [<8003450c>] (warn_slowpath_null) from [<800a2704>] (__alloc_pages_nodemask+0x19c/0x838) [<800a2704>] (__alloc_pages_nodemask) from [<800b5a60>] (kmalloc_order+0x10/0x20) [<800b5a60>] (kmalloc_order) from [<80253fa8>] (persistent_ram_save_old+0x118/0x134) [<80253fa8>] (persistent_ram_save_old) from [<80253154>] (ramoops_get_next_prz+0x6c/0x78) [<80253154>] (ramoops_get_next_prz) from [<802531a0>] (ramoops_pstore_read+0x40/0x254) [<802531a0>] (ramoops_pstore_read) from [<80252990>] (pstore_get_records+0xa8/0x2e0) [<80252990>] (pstore_get_records) from [<80251da8>] (pstore_fill_super+0xa8/0xbc) [<80251da8>] (pstore_fill_super) from [<800db540>] (mount_single+0x88/0xac) [<800db540>] (mount_single) from [<800db5d4>] (mount_fs+0x14/0xa4) [<800db5d4>] (mount_fs) from [<800f4688>] (vfs_kern_mount+0x48/0x114) [<800f4688>] (vfs_kern_mount) from [<800f7744>] (do_mount+0x198/0xc28) [<800f7744>] (do_mount) from [<800f84fc>] (SyS_mount+0x74/0xa0) [<800f84fc>] (SyS_mount) from [<8000f2a0>] (ret_fast_syscall+0x0/0x3c) persistent_ram: failed to allocate buffer
In the function persistent_ram_save_old for unknow reason the size is negative. Do I miss
something ? When I try to use other value for ramoops configuration, most of the time my
kernel is not able to boot anymore.
Hi Arthur
for freeze issues one can check ddr memory
i.MX6/7 DDR Stress Test Tool V2.80
and further debug it using AN4553 Using Open Source Debugging Tools for Linux on i.MX Processors
https://www.nxp.com/docs/en/application-note/AN4553.pdf
issues with custom kernels may be also posted on meta-fsl-arm mailing list:
https://lists.yoctoproject.org/listinfo/meta-freescale
Best regards
igor
-----------------------------------------------------------------------------------------------------------------------
Note: If this post answers your question, please click the Correct Answer button. Thank you!
-----------------------------------------------------------------------------------------------------------------------
Thanks for your answer Igor. I also send an email to the kernel mailing list. They suggest me to use kdump/kexec instead.
Working perfectly with 4.9 fsl community kernel.