Re: [PATCH net-next v3 01/10] net: wwan: tmi: Add PCIe core

From: Jakub Kicinski <kuba@kernel.org>
To: "Yanchao Yang (杨彦超)" <Yanchao.Yang@mediatek.com>
Cc: "Chris Feng (冯保林)" <Chris.Feng@mediatek.com>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	"linux-mediatek@lists.infradead.org"
	<linux-mediatek@lists.infradead.org>,
	"Mingliang Xu (徐明亮)" <mingliang.xu@mediatek.com>,
	"Min Dong (董敏)" <min.dong@mediatek.com>,
	"linuxwwan@intel.com" <linuxwwan@intel.com>,
	"m.chetan.kumar@intel.com" <m.chetan.kumar@intel.com>,
	"Liang Lu (吕亮)" <liang.lu@mediatek.com>,
	"Haijun Liu (刘海军)" <haijun.liu@mediatek.com>,
	"Haozhe Chang (常浩哲)" <Haozhe.Chang@mediatek.com>,
	"Hua Yang (杨华)" <Hua.Yang@mediatek.com>,
	"ryazanov.s.a@gmail.com" <ryazanov.s.a@gmail.com>,
	"Xiayu Zhang (张夏宇)" <Xiayu.Zhang@mediatek.com>,
	"loic.poulain@linaro.org" <loic.poulain@linaro.org>,
	"pabeni@redhat.com" <pabeni@redhat.com>,
	"edumazet@google.com" <edumazet@google.com>,
	"Ting Wang (王挺)" <ting.wang@mediatek.com>,
	"johannes@sipsolutions.net" <johannes@sipsolutions.net>,
	"netdev@vger.kernel.org" <netdev@vger.kernel.org>,
	"Aiden Wang ( 王咏麒)" <Aiden.Wang@mediatek.com>,
	"Felix Chen ( 陈非)" <Felix.Chen@mediatek.com>,
	"Lambert Wang ( 王伟)" <Lambert.Wang@mediatek.com>,
	"davem@davemloft.net" <davem@davemloft.net>,
	"Mingchuang Qiao (乔明闯)" <Mingchuang.Qiao@mediatek.com>,
	"Guohao Zhang (张国豪)" <Guohao.Zhang@mediatek.com>
Subject: Re: [PATCH net-next v3 01/10] net: wwan: tmi: Add PCIe core
Date: Thu, 16 Feb 2023 10:10:34 -0800	[thread overview]
Message-ID: <20230216101034.3f742834@kernel.org> (raw)
In-Reply-To: <2e518c17bf54298a2108de75fcd35aaf2b3397d3.camel@mediatek.com>

On Thu, 16 Feb 2023 12:50:44 +0000 Yanchao Yang (杨彦超) wrote:
> > > +	pci_clear_master(pdev);
> > > +	mtk_mhccif_exit(mdev);
> > > +	mtk_pci_free_irq(mdev);
> > > +	mtk_pci_bar_exit(mdev);
> > > +	pci_disable_device(pdev);
> > > +	pci_load_and_free_saved_state(pdev, &priv->saved_state);
> > > +	devm_kfree(dev, priv);
> > > +	devm_kfree(dev, mdev);  
> > 
> > Why are you using devm_ if you call kfree explicitly anyway?
> > You can save some memory by using kfree() directly.  
> devm_kzalloc(), devm_ioremap_resource(), devm_request_irq(),
> devm_gpio_request(), devm_clk_get()…..
> They will be freed automatically
> when corresponding device is freed, so you don’t have to free them
> explicitly. This also make probe error easier to handle. Is it ok?

Yes.