[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CA+CK2bBgUH8v_bYEyJKPsLZFDxse6xYRwGR8KN=SzgHnrR9yhA@mail.gmail.com>
Date: Fri, 6 Sep 2019 11:39:58 -0400
From: Pavel Tatashin <pasha.tatashin@...een.com>
To: James Morse <james.morse@....com>
Cc: James Morris <jmorris@...ei.org>, Sasha Levin <sashal@...nel.org>,
"Eric W. Biederman" <ebiederm@...ssion.com>,
kexec mailing list <kexec@...ts.infradead.org>,
LKML <linux-kernel@...r.kernel.org>,
Jonathan Corbet <corbet@....net>,
Catalin Marinas <catalin.marinas@....com>, will@...nel.org,
Linux ARM <linux-arm-kernel@...ts.infradead.org>,
Marc Zyngier <marc.zyngier@....com>,
Vladimir Murzin <vladimir.murzin@....com>,
Matthias Brugger <matthias.bgg@...il.com>,
Bhupesh Sharma <bhsharma@...hat.com>,
linux-mm <linux-mm@...ck.org>,
Mark Rutland <mark.rutland@....com>
Subject: Re: [PATCH v3 02/17] arm64, hibernate: use get_safe_page directly
On Fri, Sep 6, 2019 at 11:17 AM James Morse <james.morse@....com> wrote:
>
> Hi Pavel,
>
> Nit: The pattern for the subject prefix should be "arm64: hibernate:"..
> Its usually possible to spot the pattern from "git log --oneline $file".
Sure, I will change here and in other places to "arm64: hibernate:"
>
> On 21/08/2019 19:31, Pavel Tatashin wrote:
> > create_safe_exec_page is a local function that uses the
> > get_safe_page() to allocate page table and pages and one pages
> > that is getting mapped.
>
> I can't parse this.
>
> create_safe_exec_page() uses hibernate's allocator to create a set of page table to map a
> single page that will contain the relocation code.
Thanks I will rephrase it with your suggestion.
>
>
> > Remove the allocator related arguments, and use get_safe_page
> > directly, as it is done in other local functions in this
> > file.
> ... because kexec can't use this as it doesn't have a working allocator.
> Removing this function pointer makes it easier to refactor the code later.
Thanks, I will add it to the description.
>
> (this thing is only a function pointer so kexec could use it too ... It looks like you're
> creating extra work. Patch 7 moves these new calls out to a new file... presumably so
> another patch can remove them again)
>
> As stand-alone cleanup the patch looks fine, but you probably don't need to do this.
Without this clean-up moving to common code becomes messier. So, I
would like to keep this change.
Thank you,
Pasha
Powered by blists - more mailing lists