Re: [PATCH v5 2/2] lib: stackdepot: Add support to disable stack depot

[Vijayanand Jitta]

On 1/22/2021 5:55 AM, Minchan Kim wrote:
> On Mon, Jan 18, 2021 at 03:26:42PM +0530, vjitta@xxxxxxxxxxxxxx wrote:
>> From: Vijayanand Jitta <vjitta@xxxxxxxxxxxxxx>
>>
>> Add a kernel parameter stack_depot_disable to disable
>> stack depot. So that stack hash table doesn't consume
>> any memory when stack depot is disabled.
> 
> The usecase is CONFIG_PAGE_OWNER without page_owner=on.
> Without this patch, stackdepot will consume the memory
> for the hashtable. By default, it's 8M which is never trivial.
> 
> With this option, in CONFIG_PAGE_OWNER configured system,
> page_owner=off, stack_depot_disable in kernel command line,
> we could save the wasted memory for the hashtable.
> 

Sure, will update the commit text with above details.
>>
>> Signed-off-by: Vinayak Menon <vinmenon@xxxxxxxxxxxxxx>
>> Signed-off-by: Vijayanand Jitta <vjitta@xxxxxxxxxxxxxx>
> 
> Please also update kernel-parameters.txt.
> 

Sure.
>> ---
>>  include/linux/stackdepot.h |  1 +
>>  init/main.c                |  2 ++
>>  lib/stackdepot.c           | 33 +++++++++++++++++++++++++++++----
>>  3 files changed, 32 insertions(+), 4 deletions(-)
>>
>> diff --git a/include/linux/stackdepot.h b/include/linux/stackdepot.h
>> index 24d49c7..eafd9aa 100644
>> --- a/include/linux/stackdepot.h
>> +++ b/include/linux/stackdepot.h
>> @@ -21,4 +21,5 @@ unsigned int stack_depot_fetch(depot_stack_handle_t handle,
>>  
>>  unsigned int filter_irq_stacks(unsigned long *entries, unsigned int nr_entries);
>>  
>> +int stack_depot_init(void);
>>  #endif
>> diff --git a/init/main.c b/init/main.c
>> index 32b2a8a..8fcf9bb 100644
>> --- a/init/main.c
>> +++ b/init/main.c
>> @@ -98,6 +98,7 @@
>>  #include <linux/mem_encrypt.h>
>>  #include <linux/kcsan.h>
>>  #include <linux/init_syscalls.h>
>> +#include <linux/stackdepot.h>
>>  
>>  #include <asm/io.h>
>>  #include <asm/bugs.h>
>> @@ -827,6 +828,7 @@ static void __init mm_init(void)
>>  	page_ext_init_flatmem();
>>  	init_debug_pagealloc();
>>  	report_meminit();
>> +	stack_depot_init();
>>  	mem_init();
>>  	kmem_cache_init();
>>  	kmemleak_init();
>> diff --git a/lib/stackdepot.c b/lib/stackdepot.c
>> index dff8521..d20e6fd 100644
>> --- a/lib/stackdepot.c
>> +++ b/lib/stackdepot.c
>> @@ -31,6 +31,8 @@
>>  #include <linux/stackdepot.h>
>>  #include <linux/string.h>
>>  #include <linux/types.h>
>> +#include <linux/vmalloc.h>
> 
> Why do we need vmalloc?
> 

Its not needed, will remove it.

> Otherwise, looks good to me.
> Thank you!
> 

Thanks for the review.

Thanks,
Vijay
-- 
QUALCOMM INDIA, on behalf of Qualcomm Innovation Center, Inc. is a
member of Code Aurora Forum, hosted by The Linux Foundation