Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > linux.kernel > #1540239

Re: [PATCH 1/1] arm/module: maximum utilization of module area.

From Vaneet Narang <v.narang@samsung.com>
Newsgroups linux.kernel
Subject Re: [PATCH 1/1] arm/module: maximum utilization of module area.
Date 2016-12-12 11:20 +0100
Message-ID <sNvSb-341-61@gated-at.bofh.it> (permalink)
References <sLjBL-8aU-19@gated-at.bofh.it> <sNvSb-341-63@gated-at.bofh.it>
Organization linux.* mail to news gateway

Show all headers | View raw


[Multipart message — attachments visible in raw view] - view raw

Hi Ard,

>> +#ifndef CONFIG_THUMB2_KERNEL
>> +#define MODULES_START  ALIGN((unsigned long)_etext - SZ_32M, PAGE_SIZE)
>
>On a multi_v7_defconfig kernel, the text size is >8 MB, which means
>you are only adding ~7 MB to the module area, while consuming 16 MB of
>additional address space.

I am not sure if 16MB virtual address space will make any difference on embedded 
systems where physical memory is already less than virtual address space.
if required address space wastage can be reduced by keeping TASK_SIZE as 
PAGE_OFFSET - 24MB like below

#define MODULES_VADDR          (PAGE_OFFSET - SZ_24M)
#define TASK_SIZE              (UL(CONFIG_PAGE_OFFSET) - UL(SZ_24M))


> Given that 20 MB modules are very uncommon,

Size of all modules can be 20MB, this seems to be normal scenario.

>I think it is better to enable CONFIG_ARM_MODULE_PLTS instead. That

CONFIG_ARM_MODULE_PLTS has function call overhead as it refers PLT table
while calling kernel functions. Also size of modules will also gets increased a bit.
So using short calls from modules to kernel will be faster. 
These changes trying to utilize best available space for kernel modules for
making short calls. 
So CONFIG_ARM_MODULE_PLTS is not required when modules
can be accomdated within 20MB.

>way, there is no need to update these defaults for everyone.
>

>> +#else
>> +#define MODULES_START MODULES_VADDR

Thanks & Regards,
Vaneet Narang

Back to linux.kernel | Previous | NextPrevious in thread | Next in thread | Find similar | Unroll thread


Thread

[PATCH 1/1] arm/module: maximum utilization of module area. Maninder Singh <maninder1.s@samsung.com> - 2016-12-06 09:50 +0100
  Re: [PATCH 1/1] arm/module: maximum utilization of module area. Ard Biesheuvel <ard.biesheuvel@linaro.org> - 2016-12-06 10:20 +0100
  Re: [PATCH 1/1] arm/module: maximum utilization of module area. Nicolas Pitre <nicolas.pitre@linaro.org> - 2016-12-06 17:00 +0100
  Re: [PATCH 1/1] arm/module: maximum utilization of module area. Vaneet Narang <v.narang@samsung.com> - 2016-12-12 11:20 +0100
  Re: [PATCH 1/1] arm/module: maximum utilization of module area. Russell King - ARM Linux <linux@armlinux.org.uk> - 2016-12-12 11:30 +0100
  RE: Re: [PATCH 1/1] arm/module: maximum utilization of module area. Vaneet Narang <v.narang@samsung.com> - 2016-12-12 13:10 +0100
    Re: Re: [PATCH 1/1] arm/module: maximum utilization of module area. Russell King - ARM Linux <linux@armlinux.org.uk> - 2016-12-12 14:00 +0100
    RE: Re: [PATCH 1/1] arm/module: maximum utilization of module  area. Nicolas Pitre <nicolas.pitre@linaro.org> - 2016-12-12 16:30 +0100
      Re: Re: [PATCH 1/1] arm/module: maximum utilization of module area. Ard Biesheuvel <ard.biesheuvel@linaro.org> - 2016-12-12 18:00 +0100

csiph-web