linux-kernel - Re: [PATCH RESEND v2 0/2] driver core: bus: Fix issues related to bus_rescan_devices

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: <beebf034-f2a6-4451-b68c-2f23f8dc92ec@icloud.com>
Date: Wed, 13 Nov 2024 21:43:53 +0800
From: Zijun Hu <zijun_hu@...oud.com>
To: Greg Kroah-Hartman <gregkh@...uxfoundation.org>
Cc: "Rafael J. Wysocki" <rafael@...nel.org>, linux-kernel@...r.kernel.org,
 Zijun Hu <quic_zijuhu@...cinc.com>
Subject: Re: [PATCH RESEND v2 0/2] driver core: bus: Fix issues related to
 bus_rescan_devices_helper()

On 2024/11/12 19:48, Greg Kroah-Hartman wrote:
> On Tue, Oct 22, 2024 at 07:18:00PM +0800, Zijun Hu wrote:
>> This patch series is to fix issues related to bus_rescan_devices_helper().
>>
>> The function is improperly used for 2 incompatible scenarios as
>> explained below:
>>
>> Scenario A: scan drivers for a single device user specify
>>  - user may care about precise synchronous scanning result, so the
>>    function can not collapse error codes.
> 
> I do not understand this, what is wrong that this is fixing?
> 

for scanning drivers for single device, it gives user wrong scanning
result(success or failure).

patch 1/2 is a concrete example.

>> Scenario B: scan drivers for all devices of a bus
>>  - user may need to scan drivers for a bus's devices as many as
>>    possible, so the function needs to ignore inconsequential error
>>    codes for a device in order to continue to scan for next device.
> 
> How often is that needed?  And why can't that still work with the
> existing code?
>

1) API bus_rescan_devices() invokes it and can't achieve its design
purpose due to error described above.

2) as shown by recent mainlined commit, usage of API
bus_rescan_devices() have been dropped due to some bugs.
Commit: 3d6ebf16438d ("cxl/port: Fix cxl_bus_rescan() vs
bus_rescan_devices()")

3) there are only 2 usages for the API now.
// does not do what the comments say
https://elixir.bootlin.com/linux/v6.11/source/drivers/pcmcia/ds.c#L725
// do multi repeated iterating, can be simplified if fix the API bugs
https://elixir.bootlin.com/linux/v6.11/source/drivers/hid/hid-core.c#L2967

4) i have more reply in below link about the API's bugs.
https://lore.kernel.org/all/20240913-bus_match_unlikely-v2-2-5c0c3bfda2f6@quicinc.com/

we may go to patch 2/2 to continue discussion.

> thanks,
> 
> greg k-h