[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20250806054053.1507456-1-liqiang01@kylinos.cn>
Date: Wed, 6 Aug 2025 13:40:53 +0800
From: Li Qiang <liqiang01@...inos.cn>
To: akpm@...ux-foundation.org,
david@...hat.com
Cc: linux-mm@...ck.org,
linux-kernel@...r.kernel.org,
lorenzo.stoakes@...cle.com,
Liam.Howlett@...cle.com,
vbabka@...e.cz,
rppt@...nel.org,
surenb@...gle.com,
mhocko@...e.com
Subject: Re: [PATCH] mm: memory: Force-inline PTE/PMD zapping functions
On 8/5/25 13:15 PM, Vlastimil Babka wrote:
> Is the before/after swapped here? This output suggests some functions
> became NOT inlined.
>
> If I'm right the output binary becomes slightly larger. But it doesn't
> matter.
Thank you for your careful review. You're absolutely right - the
before/after columns in my previous data table were accidentally
reversed. I've corrected this in the updated measurements below:
[root@...alhost linux_old1]# ./scripts/bloat-o-meter befor.vmlinux after.vmlinux
add/remove: 0/6 grow/shrink: 1/0 up/down: 4747/-4569 (178)
Function old new delta
unmap_page_range 1018 5765 +4747
__pfx_zap_pte_range 16 - -16
__pfx_zap_present_ptes.constprop 16 - -16
__pfx_zap_pmd_range.isra 16 - -16
zap_pmd_range.isra 589 - -589
zap_pte_range 1236 - -1236
zap_present_ptes.constprop 2696 - -2696
Total: Before=35379608, After=35379786, chg +0.00%
[root@...alhost linux_old1]#
Powered by blists - more mailing lists