linux-kernel - Re: [PATCH -next v2 2/4] soc: ti: knav_dma: Use dev_err

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: <ruphb2pue6tylszwekw7lwzgoyrtdru56vhcwqpnr63wrh37pu@mzpgfi62tyik>
Date: Tue, 3 Sep 2024 15:40:55 +0200
From: Uwe Kleine-König <u.kleine-koenig@...libre.com>
To: Jinjie Ruan <ruanjinjie@...wei.com>
Cc: Nishanth Menon <nm@...com>, ssantosh@...nel.org, 
	linux-kernel@...r.kernel.org, linux-arm-kernel@...ts.infradead.org, krzk@...nel.org, 
	jic23@...nel.org
Subject: Re: [PATCH -next v2 2/4] soc: ti: knav_dma: Use dev_err_probe() to
 simplfy code

Hello,

On Sat, Aug 31, 2024 at 09:59:33AM +0800, Jinjie Ruan wrote:
> On 2024/8/30 18:31, Nishanth Menon wrote:
> > On 14:32-20240830, Jinjie Ruan wrote:
> >> Use the dev_err_probe() helper to simplify error handling
> >> during probe.
> >>
> >> Signed-off-by: Jinjie Ruan <ruanjinjie@...wei.com>
> >> ---
> >> v2:
> >> - Split into 2 patches.
> >> ---
> >>  drivers/soc/ti/knav_dma.c | 12 ++++--------
> >>  1 file changed, 4 insertions(+), 8 deletions(-)
> >>
> >> diff --git a/drivers/soc/ti/knav_dma.c b/drivers/soc/ti/knav_dma.c
> >> index 15e41d3a5e22..eeec422a46f0 100644
> >> --- a/drivers/soc/ti/knav_dma.c
> >> +++ b/drivers/soc/ti/knav_dma.c
> >> @@ -708,17 +708,13 @@ static int knav_dma_probe(struct platform_device *pdev)
> >>  	struct device_node *node = pdev->dev.of_node;
> >>  	int ret = 0;
> >>  
> >> -	if (!node) {
> >> -		dev_err(&pdev->dev, "could not find device info\n");
> >> -		return -EINVAL;
> >> -	}
> >> +	if (!node)
> >> +		return dev_err_probe(&pdev->dev, -EINVAL, "could not find device info\n");
> >>  
> >>  	kdev = devm_kzalloc(dev,
> >>  			sizeof(struct knav_dma_pool_device), GFP_KERNEL);
> >> -	if (!kdev) {
> >> -		dev_err(dev, "could not allocate driver mem\n");
> >> -		return -ENOMEM;
> >> -	}
> >> +	if (!kdev)
> >> +		return dev_err_probe(dev, -ENOMEM, "could not allocate driver mem\n");
> > 
> > These make no sense to me :( -> just using dev_err_probe when there is
> > no chance of -EPROBE_DEFER ?
> 
> I noticed a change in dev_err_probe() this year, which is described in
> this patch:
> 
> For an out-of-memory error there should be no additional output. Adapt
> dev_err_probe() to not emit the error message when err is -ENOMEM.
> This simplifies handling errors that might among others be -ENOMEM.

Notice this was carefully worded. Calling dev_err_probe() if you know
that the error is ENOMEM isn't helpful. The change was introduced to
simplify

	ret = some_function_that_might_return_ENOMEM_and_other_errors()
	if (ret == -ENOMEM)
		return ret;
	else if (ret)
		return dev_err_probe(dev, ret, "....");

to

	ret = some_function_that_might_return_ENOMEM_and_other_errors()
	if (ret)
		return dev_err_probe(dev, ret, "....");

But adding a dev_err_probe() if you know in the ret != 0 branch that ret
must be -ENOMEM, actively adding a dev_err_probe() gives very little
improvement. The main effect is to increase the size of the resulting
kernel image (or module).

Back when I made dev_err_probe() silent for -ENOMEM, it was suggested to
even make it fail to compile if ret is ENOMEM.

I wouldn't necessarily object to new code using dev_err_probe() to check
the return value of a function that can only return (success or)
-ENOMEM, but I agree to Krzysztof that changing existing code is little
helpful.

Best regards
Uwe

Download attachment "signature.asc" of type "application/pgp-signature" (489 bytes)