Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > linux.kernel > #1507493 > unrolled thread

Re: [PATCH 2/6] sparc64: Add ATU (new IOMMU) support

Started byDavid Miller <davem@davemloft.net>
First post2016-10-24 19:50 +0200
Last post2016-10-24 22:50 +0200
Articles 2 — 2 participants

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.


Contents

  Re: [PATCH 2/6] sparc64: Add ATU (new IOMMU) support David Miller <davem@davemloft.net> - 2016-10-24 19:50 +0200
    Re: [PATCH 2/6] sparc64: Add ATU (new IOMMU) support tndave <tushar.n.dave@oracle.com> - 2016-10-24 22:50 +0200

#1507493 — Re: [PATCH 2/6] sparc64: Add ATU (new IOMMU) support

FromDavid Miller <davem@davemloft.net>
Date2016-10-24 19:50 +0200
SubjectRe: [PATCH 2/6] sparc64: Add ATU (new IOMMU) support
Message-ID<svRxM-2mA-13@gated-at.bofh.it>
From: Tushar Dave <tushar.n.dave@oracle.com>
Date: Mon, 10 Oct 2016 11:12:03 -0700

> +	/* calculate size of IOTSB */
> +	table_size = (atu->size / IO_PAGE_SIZE) * 8;
> +	table = kzalloc(table_size, GFP_KERNEL);
> +	if (!table) {
> +		err = -ENOMEM;
> +		goto table_failed;
> +	}
> +	table = (void *)IO_PAGE_ALIGN((u64)table);

If you use the page allocator directly you can avoid all of these
funny alignment calculations.

> @@ -931,8 +1046,10 @@ static int pci_sun4v_probe(struct platform_device *op)
>  	struct pci_pbm_info *pbm;
>  	struct device_node *dp;
>  	struct iommu *iommu;
> +	struct atu *atu;
>  	u32 devhandle;
>  	int i, err = -ENODEV;
> +	static bool hv_atu = 1;

Please use true/false with boolean variables.

> +			hv_atu = 0;

Likewise.

[toc] | [next] | [standalone]


#1507719

Fromtndave <tushar.n.dave@oracle.com>
Date2016-10-24 22:50 +0200
Message-ID<svUlY-4eX-35@gated-at.bofh.it>
In reply to#1507493

On 10/24/2016 10:46 AM, David Miller wrote:
> From: Tushar Dave <tushar.n.dave@oracle.com>
> Date: Mon, 10 Oct 2016 11:12:03 -0700
>
>> +	/* calculate size of IOTSB */
>> +	table_size = (atu->size / IO_PAGE_SIZE) * 8;
>> +	table = kzalloc(table_size, GFP_KERNEL);
>> +	if (!table) {
>> +		err = -ENOMEM;
>> +		goto table_failed;
>> +	}
>> +	table = (void *)IO_PAGE_ALIGN((u64)table);
>
> If you use the page allocator directly you can avoid all of these
> funny alignment calculations.
Okay. I will look into this.
>
>> @@ -931,8 +1046,10 @@ static int pci_sun4v_probe(struct platform_device *op)
>>  	struct pci_pbm_info *pbm;
>>  	struct device_node *dp;
>>  	struct iommu *iommu;
>> +	struct atu *atu;
>>  	u32 devhandle;
>>  	int i, err = -ENODEV;
>> +	static bool hv_atu = 1;
>
> Please use true/false with boolean variables.
Sure.
>
>> +			hv_atu = 0;
>
> Likewise.
>
Thanks.
-Tushar

[toc] | [prev] | [standalone]


Back to top | Article view | linux.kernel


csiph-web