netdev - Re: [PATCH net-next v3 2/3] net: airoha: npu: Add airoha_npu_soc

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: <20251014134609.GA3239414@horms.kernel.org>
Date: Tue, 14 Oct 2025 14:46:09 +0100
From: Simon Horman <horms@...nel.org>
To: Lorenzo Bianconi <lorenzo@...nel.org>
Cc: Andrew Lunn <andrew+netdev@...n.ch>,
	"David S. Miller" <davem@...emloft.net>,
	Eric Dumazet <edumazet@...gle.com>,
	Jakub Kicinski <kuba@...nel.org>, Paolo Abeni <pabeni@...hat.com>,
	Rob Herring <robh@...nel.org>,
	Krzysztof Kozlowski <krzk+dt@...nel.org>,
	Conor Dooley <conor+dt@...nel.org>, netdev@...r.kernel.org,
	devicetree@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
	linux-mediatek@...ts.infradead.org
Subject: Re: [PATCH net-next v3 2/3] net: airoha: npu: Add
 airoha_npu_soc_data struct

On Tue, Oct 14, 2025 at 11:23:37AM +0200, Lorenzo Bianconi wrote:
> > On Mon, Oct 13, 2025 at 03:58:50PM +0200, Lorenzo Bianconi wrote:
> > 
> > ...
> > 
> > > @@ -182,49 +192,53 @@ static int airoha_npu_send_msg(struct airoha_npu *npu, int func_id,
> > >  	return ret;
> > >  }
> > >  
> > > -static int airoha_npu_run_firmware(struct device *dev, void __iomem *base,
> > > -				   struct resource *res)
> > > +static int airoha_npu_load_firmware(struct device *dev, void __iomem *addr,
> > > +				    const struct airoha_npu_fw *fw_info)
> > >  {
> > >  	const struct firmware *fw;
> > > -	void __iomem *addr;
> > >  	int ret;
> > >  
> > > -	ret = request_firmware(&fw, NPU_EN7581_FIRMWARE_RV32, dev);
> > > +	ret = request_firmware(&fw, fw_info->name, dev);
> > >  	if (ret)
> > >  		return ret == -ENOENT ? -EPROBE_DEFER : ret;
> > >  
> > > -	if (fw->size > NPU_EN7581_FIRMWARE_RV32_MAX_SIZE) {
> > > +	if (fw->size > fw_info->max_size) {
> > >  		dev_err(dev, "%s: fw size too overlimit (%zu)\n",
> > > -			NPU_EN7581_FIRMWARE_RV32, fw->size);
> > > +			fw_info->name, fw->size);
> > >  		ret = -E2BIG;
> > >  		goto out;
> > >  	}
> > >  
> > > -	addr = devm_ioremap_resource(dev, res);
> > > -	if (IS_ERR(addr)) {
> > > -		ret = PTR_ERR(addr);
> > > -		goto out;
> > > -	}
> > > -
> > >  	memcpy_toio(addr, fw->data, fw->size);
> > > +out:
> > >  	release_firmware(fw);
> > >  
> > > -	ret = request_firmware(&fw, NPU_EN7581_FIRMWARE_DATA, dev);
> > > -	if (ret)
> > > -		return ret == -ENOENT ? -EPROBE_DEFER : ret;
> > > +	return ret;
> > > +}
> > >  
> > > -	if (fw->size > NPU_EN7581_FIRMWARE_DATA_MAX_SIZE) {
> > > -		dev_err(dev, "%s: fw size too overlimit (%zu)\n",
> > > -			NPU_EN7581_FIRMWARE_DATA, fw->size);
> > > -		ret = -E2BIG;
> > > -		goto out;
> > > -	}
> > > +static int airoha_npu_run_firmware(struct device *dev, void __iomem *base,
> > > +				   struct resource *res)
> > > +{
> > > +	const struct airoha_npu_soc_data *soc;
> > > +	void __iomem *addr;
> > > +	int ret;
> > >  
> > > -	memcpy_toio(base + REG_NPU_LOCAL_SRAM, fw->data, fw->size);
> > > -out:
> > > -	release_firmware(fw);
> > > +	soc = of_device_get_match_data(dev);
> > > +	if (!soc)
> > > +		return -EINVAL;
> > >  
> > > -	return ret;
> > > +	addr = devm_ioremap_resource(dev, res);
> > > +	if (IS_ERR(addr))
> > > +		return PTR_ERR(addr);
> > > +
> > > +	/* Load rv32 npu firmware */
> > > +	ret = airoha_npu_load_firmware(dev, addr, &soc->fw_rv32);
> > > +	if (ret)
> > > +		return ret;
> > > +
> > > +	/* Load data npu firmware */
> > > +	return airoha_npu_load_firmware(dev, base + REG_NPU_LOCAL_SRAM,
> > > +					&soc->fw_data);
> > 
> > Hi Lorenzo,
> 
> Hi Simon,
> 
> > 
> > There are two calls to airoha_npu_load_firmware() above.
> > And, internally, airoha_npu_load_firmware() will call release_firmware()
> > if an error is encountered.
> > 
> > But should release_firmware() be called for the firmware requested
> > by the first call to airoha_npu_load_firmware() if the second call fails?
> > Such clean-up seems to have been the case prior to this patch.
> 
> release_firmware() is intended to release the resources allocated by the
> corresponding call to request_firmware() in airoha_npu_load_firmware().
> According to my understanding we always run release_firmware() in
> airoha_npu_load_firmware() before returning to the caller. Even before this
> patch we run release_firmware() on the 'first' firmware image before requesting
> the second one. Am I missing something?
> 
> > 
> > Also, not strictly related. Should release_firmware() be called (twice)
> > when the driver is removed?
> 
> For the above reasons, it is not important to call release_firmware() removing
> the module. Agree?

Thanks, agreed.

For some reason I missed that release_firmware() is called in
airoha_npu_load_firmware() regardless of error - I thought it was only
in error paths for some reason.

So I agree that the firmware is always released by the time
airoha_npu_load_firmware() is returned. As thus there is never
a need to release it afterwards.

Reviewed-by: Simon Horman <horms@...nel.org>