linux-kernel - Re: [PATCH v2 00/28] drivers: Consolidate device lookup helpers

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 for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <c6813d08-b70b-f153-a651-94d1bcaa5522@arm.com>
Date:   Mon, 17 Jun 2019 11:07:29 +0100
From:   Suzuki K Poulose <suzuki.poulose@....com>
To:     joe@...ches.com
Cc:     devicetree@...r.kernel.org, linux-kernel@...r.kernel.org,
        gregkh@...uxfoundation.org, rafael@...nel.org
Subject: Re: [PATCH v2 00/28] drivers: Consolidate device lookup helpers

Cc: Greg, Rafael

Hi Joe,

On 14/06/2019 19:24, Joe Perches wrote:
> (dropping the very long cc list just cc'ing LKML and devicetree)
> 
> On Fri, 2019-06-14 at 18:53 +0100, Suzuki K Poulose wrote:
>> We have device iterators to find a particular device matching a criteria
>> for a given bus/class/driver. i.e, {bus,class,driver}_find_device() APIs.
>> The matching criteria is a function pointer for the APIs. Often the lookup
>> is based on a generic property of a device (e.g, name, fwnode, of node pointer
>> or device type) rather than a driver specific information. However, each driver
>> writes up its own "match" function, spilling the similar match functions all
>> over the driver subsystems.
>>
>> Additionally the prototype for the "match" functions accepted by the above APIs
>> have a minute difference which prevents us otherwise sharing the match functions.
>> i.e,
>> 	int (*match)(struct device *dev, void *data) for {bus/driver}_find_device()
>> 	  vs
>> 	int (*match)(struct device *dev, const void *) for class_find_device()
>>
> 
> As you are doing treewide conversions, perhaps using
> 
> 	bool (*match)(...)
> 
> is a more sensible api.

I agree that it is more suitable api. However, that would need much more
conversions for the existing "class_find_device()" , which are not touched by
the series and would make this series even more bigger. With that said, I
am fine with implementing the suggestion if Greg/Rafael thinks that is fine.

>> Changes since v1:
>>   - Drop start parameter for *_find_device_by_devt().
>>   - Fix build warnings for s390
>>   - Add *_find_device_by_acpi_dev() wrappers.
>>   - Group wrappers and the consumers into single patch, reducing
>>     the total patches to 28 from 57. (Rafael).
>>   - Better description for acpi cleanup patch.
>>   - Added tags from v1.
> 
> Below this is a _very_ long list of cc:'s.

Unfortunately, yes.

> If the list is generated using scripts/get_maintainer.pl
> perhaps it is more sensible to add --nogit --nogit-fallback

Yes, and trimmed manually a bit to remove the "commit-signers".
I have tried to keep only the maintainers/reviewers/supporters.
Thanks for the option, I will give that a try.

> to its arguments to cc actual maintainers and avoid people
> that have submitted cleanup style patches to various files.
> 

Cheers
Suzuki