linux-kernel - Re: [RFC] dmaengine: Add metadata_ops for dma_async_tx

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] [day] [month] [year] [list]

Message-ID: <38aa48ba-5cf9-da12-bef8-b75e22f8b822@ti.com>
Date:   Fri, 17 Aug 2018 09:30:30 +0300
From:   Peter Ujfalusi <peter.ujfalusi@...com>
To:     Radhey Shyam Pandey <radheys@...inx.com>,
        "dan.j.williams@...el.com" <dan.j.williams@...el.com>,
        "vkoul@...nel.org" <vkoul@...nel.org>
CC:     "dmaengine@...r.kernel.org" <dmaengine@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "lars@...afoo.de" <lars@...afoo.de>
Subject: Re: [RFC] dmaengine: Add metadata_ops for dma_async_tx_descriptor

Radhey,

On 2018-08-16 16:29, Radhey Shyam Pandey wrote:
>> +/**
>> + * enum dma_desc_metadata_mode - per descriptor metadata mode types
>> supported
>> + * @DESC_METADATA_CLIENT - the metadata buffer is allocated/provided by
>> the
>> + *  client driver and it is attached (via the dmaengine_desc_attach_metadata()
>> + *  helper) to the descriptor.
>> + * @DESC_METADATA_EMBEDDED - the metadata buffer is
>> allocated/managed by the DMA
> Just a thought - We can rename it to DESC_METADATA_ENGINE? 
> i.e metadata allocation place - >  dma client/engine.

Sounds good.

> 
>> + *  driver. The client driver can ask for the pointer, maximum size and the
>> + *  currently used size of the metadata and can directly updata or read it.
> /s/updata/update
> 
>> + *  dmaengine_desc_get_metadata_ptr() and
>> dmaengine_desc_set_metadata_len() is
>> + *  provided as helper functions.
> It will be helpful if we add description for both DESC_METADATA_EMBEDDED 
> modes i.e DMA_DEV_TO_MEM and MEM_TO_DEV types. I think in DEV_TO_MEM
> we don't need to set_metadata_len(). Length will provided by DMA engine.

I agree, it is better to extend the explanation on the two modes.

...

>> +static inline int _desc_check_and_set_metadata_mode(
>> +	struct dma_async_tx_descriptor *desc, enum
>> dma_desc_metadata_mode mode)
>> +{
>> +	/* Make sure that the metadata mode is not mixed */
>> +	if (!desc->desc_metadata_mode) {
> Minor nit - we can refactor this code to have failure path early.

I don't think it would help readability, to move the failure case first:

if (desc->desc_metadata_mode &&
    (desc->desc_metadata_mode != mode)) {
	return -EINVAL;
} else if (!desc->desc_metadata_mode) {
	if (dmaengine_is_metadata_mode_supported(desc->chan, mode))
			desc->desc_metadata_mode = mode;
		else
			return -ENOTSUPP;
}

return 0;

> 
>> +		if (dmaengine_is_metadata_mode_supported(desc->chan,
>> mode))
>> +			desc->desc_metadata_mode = mode;
>> +		else
>> +			return -ENOTSUPP;
>> +	} else if (desc->desc_metadata_mode != mode) {
>> +		return -EINVAL;
>> +	}
>> +
>> +	return 0;
>> +}

- Péter

Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki.
Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki