Re: [PATCH v4 RFT 1/2] powerpc/fadump: reduce memory consumption for capture kernel

From: Hari Bathini
Date: Wed May 03 2017 - 04:50:36 EST

On Wednesday 03 May 2017 12:43 PM, Hari Bathini wrote:

On Tuesday 02 May 2017 09:26 PM, Michal Suchanek wrote:
With fadump (dump capture) kernel booting like a regular kernel, it almost
needs the same amount of memory to boot as the production kernel, which is
unwarranted for a dump capture kernel. But with no option to disable some
of the unnecessary subsystems in fadump kernel, that much memory is wasted
on fadump, depriving the production kernel of that memory.

Introduce kernel parameter 'fadump_append=' that would take regular kernel
parameters to be appended when fadump is active.
This 'fadump_append=' parameter can be leveraged to pass parameters like
nr_cpus=1, cgroup_disable=memory and numa=off, to disable unwarranted

Also, ensure the log "Firmware-assisted dump is active" is printed early
in the boot process to put the subsequent fadump messages in context.

Suggested-by: Michael Ellerman <mpe@xxxxxxxxxxxxxx>
Signed-off-by: Hari Bathini <hbathini@xxxxxxxxxxxxxxxxxx>
Signed-off-by: Michal Suchanek <msuchanek@xxxxxxx>
- use space separated arguments instead of comma separated
- do not append parameters when fadummp is disabled
arch/powerpc/kernel/fadump.c | 27 ++++++++++++++++++++++++---
1 file changed, 24 insertions(+), 3 deletions(-)

diff --git a/arch/powerpc/kernel/fadump.c b/arch/powerpc/kernel/fadump.c
index ebf2e9c..e0c728a 100644
--- a/arch/powerpc/kernel/fadump.c
+++ b/arch/powerpc/kernel/fadump.c
@@ -79,8 +79,10 @@ int __init early_init_dt_scan_fw_dump(unsigned long node,
* dump data waiting for us.
fdm_active = of_get_flat_dt_prop(node, "ibm,kernel-dump", NULL);
- if (fdm_active)
+ if (fdm_active) {
+ pr_info("Firmware-assisted dump is active.\n");
fw_dump.dump_active = 1;
+ }

/* Get the sizes required to store dump data for the firmware provided
* dump sections.
@@ -263,8 +265,12 @@ int __init fadump_reserve_mem(void)
unsigned long base, size, memory_boundary;

- if (!fw_dump.fadump_enabled)
+ if (!fw_dump.fadump_enabled) {
+ if (fw_dump.dump_active)
+ pr_warn("Firmware-assisted dump was active but kernel"
+ " booted with fadump disabled!\n");
return 0;
+ }

if (!fw_dump.fadump_supported) {
printk(KERN_INFO "Firmware-assisted dump is not supported on"
@@ -304,7 +310,6 @@ int __init fadump_reserve_mem(void)
memory_boundary = memblock_end_of_DRAM();

if (fw_dump.dump_active) {
- printk(KERN_INFO "Firmware-assisted dump is active.\n");
* If last boot has crashed then reserve all the memory
* above boot_memory_size so that we don't touch it until
@@ -363,6 +368,22 @@ unsigned long __init arch_reserved_kernel_pages(void)
return memblock_reserved_size() / PAGE_SIZE;

+/* Look for fadump_append= cmdline option. */
+static int __init early_fadump_append_param(char *p)
+ if (!p)
+ return 1;
+ if (fw_dump.fadump_enabled && fw_dump.dump_active) {
+ pr_info("enforcing additional parameters (%s) passed through "
+ "'fadump_append=' parameter\n", p);
+ parse_early_options(p);

While testing this on a system with yaboot bootloader, it seems that parsing
a parameter with syntax like fadump_append="nr_cpus numa=off" is not

fadump_append="nr_cpus numa=off" should have read fadump_append="nr_cpus=1 numa=off".
Nonetheless, considering different environments this is supposed to work, I was trying to convey
comma-separated append list maybe better over space-separated one..