lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date:   Thu, 9 Feb 2023 14:39:11 +0100
From:   AngeloGioacchino Del Regno 
        <angelogioacchino.delregno@...labora.com>
To:     Yong Wu <yong.wu@...iatek.com>, Joerg Roedel <joro@...tes.org>,
        Matthias Brugger <matthias.bgg@...il.com>,
        Mauro Carvalho Chehab <mchehab@...nel.org>,
        Rob Herring <robh+dt@...nel.org>
Cc:     Will Deacon <will@...nel.org>, Robin Murphy <robin.murphy@....com>,
        Krzysztof Kozlowski <krzysztof.kozlowski+dt@...aro.org>,
        Hans Verkuil <hverkuil@...all.nl>, nfraprado@...labora.com,
        linux-media@...r.kernel.org, devicetree@...r.kernel.org,
        linux-arm-kernel@...ts.infradead.org,
        linux-mediatek@...ts.infradead.org, linux-kernel@...r.kernel.org,
        iommu@...ts.linux.dev, mingyuan.ma@...iatek.com,
        yf.wang@...iatek.com, libo.kang@...iatek.com,
        Yunfei Dong <yunfei.dong@...iatek.com>,
        kyrie wu <kyrie.wu@...iatek.corp-partner.google.com>,
        chengci.xu@...iatek.com, youlin.pei@...iatek.com,
        anan.sun@...iatek.com
Subject: Re: [PATCH v2 03/10] iommu/mediatek: Get regionid from larb/port id

Il 08/02/23 06:36, Yong Wu ha scritto:
> After commit f1ad5338a4d5 ("of: Fix "dma-ranges" handling for bus
> controllers"), the dma-ranges is not allowed for dts leaf node.
> but we still would like to separate to different masters
> into different iova regions.
> 
> Thus we have to separate it by the HW larbid and portid. For example,
> larb1/2 are in region2 and larb3 is in region3. The problem is that
> some ports inside a larb are in region4 while some ports inside this
> larb are in region5. Therefore I define a "larb_region_msk" to help
> record the information for each a port. Take a example for a larb:
>   [1] = ~0: means all ports in this larb are in region1;
>   [2] = BIT(3) | BIT(4): means port3/4 in this larb are region2;
>   [3] = ~(BIT(3) | BIT(4)): means all the other ports except port3/4
>                             in this larb are region3.
> 
> This method also avoids the users forget/abuse the iova regions.
> 
> Signed-off-by: Yong Wu <yong.wu@...iatek.com>
> ---
>   drivers/iommu/mtk_iommu.c | 43 +++++++++++++++++++++------------------
>   1 file changed, 23 insertions(+), 20 deletions(-)
> 
> diff --git a/drivers/iommu/mtk_iommu.c b/drivers/iommu/mtk_iommu.c
> index d5a4955910ff..fc3d9be120a0 100644
> --- a/drivers/iommu/mtk_iommu.c
> +++ b/drivers/iommu/mtk_iommu.c
> @@ -8,7 +8,6 @@
>   #include <linux/clk.h>
>   #include <linux/component.h>
>   #include <linux/device.h>
> -#include <linux/dma-direct.h>
>   #include <linux/err.h>
>   #include <linux/interrupt.h>
>   #include <linux/io.h>
> @@ -194,6 +193,7 @@ struct mtk_iommu_plat_data {
>   	enum mtk_iommu_plat	m4u_plat;
>   	u32			flags;
>   	u32			inv_sel_reg;
> +	const u32		(*larb_region_msk)[32];

Can you please document this larb region mask in code, other than the commit
description?

I can see this being essential for the next person reading this driver's code
without digging through the commit history. At least some comment on top of
the pointer, or on top of the struct declaration... and perhaps also describe
briefly that the array is "indexed by region" (so 1 = region 1; 2 = region 2)
and that the region index corresponds to the same index as `mtk_iommu_iova_region`.


Before doing that, I'd like to check if anyone else has a better solution for
that... because when looking at data for one of the SoCs in here, it looks a bit
intimidating!

Copy-paste from patch [04/10] of this series for the reader's commodity:

static const unsigned int mt8195_larb_region_msk[][32] = {
	[0] = {~0, ~0, ~0, ~0},               /* Region0: all ports for larb0/1/2/3 */
	[1] = {0, 0, 0, 0, 0, 0, 0, 0,
	       0, 0, 0, 0, 0, 0, 0, 0,
	       0, 0, 0, ~0, ~0, ~0, ~0, ~0,   /* Region1: larb19/20/21/22/23/24 */
	       ~0},
	[2] = {0, 0, 0, 0, ~0, ~0, ~0, ~0,    /* Region2: the other larbs. */
	       ~0, ~0, ~0, ~0, ~0, ~0, ~0, ~0,
	       ~0, ~0, 0, 0, 0, 0, 0, 0,
	       0, ~0, ~0, ~0, ~0},
	[3] = {0},
	[4] = {[18] = BIT(0) | BIT(1)},       /* Only larb18 port0/1 */
	[5] = {[18] = BIT(2) | BIT(3)},       /* Only larb18 port2/3 */
};

^^^^ That's what I actually mean by "intimidating"... :-P

It's just looks though, there's nothing much complicated here.

Regards,
Angelo


Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ