linux-kernel - Re: [PATCH] serial: 8250_of: replace kzalloc with devm

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: <2025082908-charbroil-saline-ef5f@gregkh>
Date: Fri, 29 Aug 2025 14:44:14 +0200
From: Greg KH <gregkh@...uxfoundation.org>
To: Osama Abdelkader <osama.abdelkader@...il.com>
Cc: jirislaby@...nel.org, andriy.shevchenko@...ux.intel.com,
	elder@...cstar.com, benjamin.larsson@...exis.eu,
	u.kleine-koenig@...libre.com, linux-kernel@...r.kernel.org,
	linux-serial@...r.kernel.org
Subject: Re: [PATCH] serial: 8250_of: replace kzalloc with devm_kzalloc

On Fri, Aug 29, 2025 at 11:49:41AM +0200, Osama Abdelkader wrote:
> 
> On 8/28/25 7:51 AM, Greg KH wrote:
> > On Thu, Aug 28, 2025 at 01:11:05AM +0200, Osama Abdelkader wrote:
> >> Use devm_kzalloc for automatic memory cleanup.
> > Why?
> >
> > I do not see a good reason here as to how this makes anything better
> > overall?  How was it tested?
> >
> > thanks,
> >
> > greg k-h
> 
> Hi Greg,
> 
> Thanks for the feedback, the change to devm_kzalloc ensures the allocated
> memory is tied to the device's lifetime. This removed the need for explicit
> kfree() calls in the remove path and avoids potential leaks in probe error 
> paths.

But there are no existing errors, so why change working code?

> It also aligns the driver with others in the 8250 subsystem which 
> already use devm-managed resources.

This code is older than the devm api :)

> For testing, I built the kernel and booted it on QEMU riscv with of_serial
> enabled. The driver probed successfully and the serial console worked as
> expected, also tested unbinding/rebinding the driver via sysfs to confirm
> no leaks or errors occur.

But did you test the error paths?  That is what you changed here.

And changes like this, for old, working, code, is usually not needed
unless you are fixing a bug somewhere.

thanks,

greg k-h