linux-kernel - Re: [PATCH v2 3/3] iommu: rockchip: Allocate tables from all available memory

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]

Message-ID: <0feaf393-21ca-f6ff-bc5d-d179d6169a70@kwiboo.se>
Date:   Thu, 15 Jun 2023 22:26:16 +0000 (UTC)
From:   Jonas Karlman <jonas@...boo.se>
To:     Robin Murphy <robin.murphy@....com>,
        Joerg Roedel <joro@...tes.org>, Will Deacon <will@...nel.org>,
        Heiko Stuebner <heiko@...ech.de>
Cc:     Benjamin Gaignard <benjamin.gaignard@...labora.com>,
        iommu@...ts.linux.dev, linux-arm-kernel@...ts.infradead.org,
        linux-rockchip@...ts.infradead.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2 3/3] iommu: rockchip: Allocate tables from all
 available memory

On 2023-06-15 23:25, Robin Murphy wrote:
> On 2023-06-15 21:10, Jonas Karlman wrote:
>> Newer Rockchip SoCs, RK356x and RK3588, support more then 4GB of memory.
> 
> Older ones supposedly did too, e.g. commit 79db45be2b8b ("ARM: dts: 
> rockchip: convert rk3288 device tree files to 64 bits"). Are we certain 
> that nobody actually has a system with IOMMUv1 and more than 4GB of RAM?

In IOMMU v1 bit 11:0 read back as 0 from MMU_DTE_ADDR reg, so I expect
that the old limit for v1 is 4GB. I will reword this to focus on IOMMU
v1 vs v2 instead of SoCs in v3.

> 
>> However, the RK IOMMU driver is using the GFP_DMA32 flag to limit
>> allocation of the discovery and page tables into memory below 4GB.
>>
>> Let's remove this limitation now that the discovery table address is
> 
> Nit: s/discovery/directory/g again

Will fix in v3 :-)

Regards,
Jonas

> 
> Thanks,
> Robin.
> 
>> correctly configured for addresses above 4GB.
>>
>> Signed-off-by: Jonas Karlman <jonas@...boo.se>
>> ---
>> v2:
>> - no change
>>
>>   drivers/iommu/rockchip-iommu.c | 4 ++--
>>   1 file changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/iommu/rockchip-iommu.c b/drivers/iommu/rockchip-iommu.c
>> index 62be9bf42390..46498fc382ee 100644
>> --- a/drivers/iommu/rockchip-iommu.c
>> +++ b/drivers/iommu/rockchip-iommu.c
>> @@ -727,7 +727,7 @@ static u32 *rk_dte_get_page_table(struct rk_iommu_domain *rk_domain,
>>   	if (rk_dte_is_pt_valid(dte))
>>   		goto done;
>>   
>> -	page_table = (u32 *)get_zeroed_page(GFP_ATOMIC | GFP_DMA32);
>> +	page_table = (u32 *)get_zeroed_page(GFP_ATOMIC);
>>   	if (!page_table)
>>   		return ERR_PTR(-ENOMEM);
>>   
>> @@ -1076,7 +1076,7 @@ static struct iommu_domain *rk_iommu_domain_alloc(unsigned type)
>>   	 * Each level1 (dt) and level2 (pt) table has 1024 4-byte entries.
>>   	 * Allocate one 4 KiB page for each table.
>>   	 */
>> -	rk_domain->dt = (u32 *)get_zeroed_page(GFP_KERNEL | GFP_DMA32);
>> +	rk_domain->dt = (u32 *)get_zeroed_page(GFP_KERNEL);
>>   	if (!rk_domain->dt)
>>   		goto err_free_domain;
>>