Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Tue, 6 Sep 2022 06:52:14 +0300
From:   Jarkko Sakkinen <jarkko@kernel.org>
To:     Dave Hansen <dave.hansen@intel.com>
Cc:     linux-sgx@vger.kernel.org,
        Haitao Huang <haitao.huang@linux.intel.com>,
        Vijay Dhanraj <vijay.dhanraj@intel.com>,
        Reinette Chatre <reinette.chatre@intel.com>,
        Dave Hansen <dave.hansen@linux.intel.com>,
        Kai Huang <kai.huang@intel.com>,
        Thomas Gleixner <tglx@linutronix.de>,
        Ingo Molnar <mingo@redhat.com>, Borislav Petkov <bp@alien8.de>,
        "maintainer:X86 ARCHITECTURE (32-BIT AND 64-BIT)" <x86@kernel.org>,
        "H. Peter Anvin" <hpa@zytor.com>,
        "open list:X86 ARCHITECTURE (32-BIT AND 64-BIT)" 
        <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH RFC] x86/sgx: Use a heap allocated list head for
 unsanitized pages
Message-ID: <YxbD7oWO1D6JmyER@kernel.org>
References: <20220906031230.107108-1-jarkko@kernel.org>
 <c387b67e-7545-db72-25f7-7e368dddd75d@intel.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <c387b67e-7545-db72-25f7-7e368dddd75d@intel.com>
Precedence: bulk

On Mon, Sep 05, 2022 at 08:20:18PM -0700, Dave Hansen wrote:
> On 9/5/22 20:12, Jarkko Sakkinen wrote:
> > Allocate the list head for the unsanitized pages from heap, and transfer
> > its to ownership to ksgxd, which takes care of destroying it. Remove
> > sgx_dirty_page_list, as a global list is no longer required.
> > 
> > Signed-off-by: Jarkko Sakkinen <jarkko@kernel.org>
> > ---
> > Depends on https://lore.kernel.org/linux-sgx/20220906000221.34286-1-jarkko@kernel.org/T/#t
> > Would this be plausible?
> > ---
> >  arch/x86/kernel/cpu/sgx/main.c | 44 ++++++++++++++++++++++++----------
> >  1 file changed, 31 insertions(+), 13 deletions(-)
> 
> What's the benefit?  I guess it saves 16 bytes of kernel image for ~20
> lines of code.  Does it add more than 16 bytes to kernel text? ;)

That's why I asked, and added RFC. It was so small step that I thought to
try it out.

I agree that it does not save memory for any significant amount but it
does take a bit way from driver global data, which is IMHO most of the
time +1 (simplifies state). Not saying that enough to pick this but
perhaps it will find its place as part of a patch set. By itself it
does not stand, I agree.

BR, Jarkko