[PATCH v2 1/3] mm: vmalloc: Let user to control huge vmalloc default behavior

Kefeng Wang wangkefeng.wang at huawei.com
Wed Jan 19 23:57:58 AEDT 2022


On 2022/1/18 10:52, Nicholas Piggin wrote:
> Excerpts from Kefeng Wang's message of December 28, 2021 12:59 am:
>> Introduce HUGE_VMALLOC_DEFAULT_ENABLED and make it default y, this
>> let user to choose whether or not enable huge vmalloc mappings by
>> default.
>>
>> Meanwhile, add new hugevmalloc=on/off parameter to enable or disable
>> this feature at boot time, nohugevmalloc is still supported and
>> equivalent to hugevmalloc=off.
> Runtime options are bad enough, Kconfig and boot options are even worse.

nohugevmalloc is like blacklists, on the other hand, Add a 
HUGE_VMALLOC_DEFAULT_ENABLED

to close hugevmalloc default and enable it only via hugevmalloc=on is 
whiteList.


Only parts of our products wants this feature,  we add some interfaces 
which only

alloc hugevmalloc for them, eg, 
vmap_hugepage/vmalloc_hugepage/remap_vmalloc_hugepage_range..

for our products, but it's not the choice of most products, also add 
nohugevmalloc

for most products is expensive, so this is the reason for adding the patch.

more config/cmdline are more flexible for test/products,

>
> The 'nohugevmalloc' option mirrors 'nohugeiomap' and is not expected to
> ever be understood by an administrator unless a kernel developer is
> working with them to hunt down a regression.
>
> IMO there should be no new options. You could switch it off for
> CONFIG_BASE_SMALL perhaps, and otherwise just try to work on heuristics
> first. Bring in new options once it's proven they're needed.

but yes, this patch is optional, could others give some more comments 
about this way?

Thanks.

> Aside from that, thanks for working on these ports, great work.
>
> Thanks,
> Nick
> .


More information about the Linuxppc-dev mailing list