linux-kernel - Re: [PATCH 1/4] staging: r8188eu: add error handling of rtw

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]

Date:   Thu, 19 May 2022 08:43:23 +0300
From:   Pavel Skripkin <paskripkin@...il.com>
To:     Dan Carpenter <dan.carpenter@...cle.com>
Cc:     gregkh@...uxfoundation.org, Larry.Finger@...inger.net,
        phil@...lpotter.co.uk, straube.linux@...il.com,
        fmdefrancesco@...il.com, linux-kernel@...r.kernel.org,
        linux-staging@...ts.linux.dev
Subject: Re: [PATCH 1/4] staging: r8188eu: add error handling of rtw_read8

Hi Dan,

On 5/19/22 07:33, Dan Carpenter wrote:
> On Thu, May 19, 2022 at 01:11:51AM +0300, Pavel Skripkin wrote:
>> @@ -240,12 +259,14 @@ int rtl8188e_firmware_download(struct adapter *padapter)
>>  {
>>  	int ret = _SUCCESS;
>>  	u8 write_fw_retry = 0;
>> +	u8 reg;
>>  	unsigned long fwdl_timeout;
>>  	struct dvobj_priv *dvobj = adapter_to_dvobj(padapter);
>>  	struct device *device = dvobj_to_dev(dvobj);
>>  	struct rt_firmware_hdr *fwhdr = NULL;
>>  	u8 *fw_data;
>>  	u32 fw_size;
>> +	int res;
>>  
>>  	if (!dvobj->firmware.data)
>>  		ret = load_firmware(&dvobj->firmware, device);
>> @@ -269,7 +290,11 @@ int rtl8188e_firmware_download(struct adapter *padapter)
>>  
>>  	/*  Suggested by Filen. If 8051 is running in RAM code, driver should inform Fw to reset by itself, */
>>  	/*  or it will cause download Fw fail. 2010.02.01. by tynli. */
>> -	if (rtw_read8(padapter, REG_MCUFWDL) & RAM_DL_SEL) { /* 8051 RAM code */
>> +	res = rtw_read8(padapter, REG_MCUFWDL, &reg);
>> +	if (res)
>> +		goto exit;
> 
> You didn't introduce this bug, but this path needs to have an error code
> set.  Also we really need to get rid of the _FAIL garbage.  When I saw
> this, I got "ret" and "res" mixed up so I thought we were returning
> negative error codes instead of _FAIL.  That would   But then I saw we
> are returning success.
> 

I see now, that 'res' and 'ret' got mixed up in my mind too. Will fix up


>> +
>> +	if (reg & RAM_DL_SEL) { /* 8051 RAM code */
>>  		rtw_write8(padapter, REG_MCUFWDL, 0x00);
>>  		rtw_reset_8051(padapter);
>>  	}
>> @@ -278,7 +303,14 @@ int rtl8188e_firmware_download(struct adapter *padapter)
>>  	fwdl_timeout = jiffies + msecs_to_jiffies(500);
>>  	while (1) {
>>  		/* reset the FWDL chksum */
>> -		rtw_write8(padapter, REG_MCUFWDL, rtw_read8(padapter, REG_MCUFWDL) | FWDL_CHKSUM_RPT);
>> +		res = rtw_read8(padapter, REG_MCUFWDL, &reg);
>> +		if (res == -ENODEV)
>> +			break;
>> +
>> +		if (res)
>> +			continue;
> 
> This continue is wrong.  If res = -EPERM then it's a forever loop.
> Let's just break for every error.
> 

I was trying to avoid strict breaking the loop on any error, since I am 
afraid this might break the driver.

What about:

	do {
		/* reset the FWDL chksum */
		ret = rtw_read8(padapter, REG_MCUFWDL, &reg);
		if (ret == -ENODEV || ret == -EPERM)
			break;

		if (ret) {
			ret == _FAIL;
			continue;
		}

		rtw_write8(padapter, REG_MCUFWDL, reg | FWDL_CHKSUM_RPT);

		ret = write_fw(padapter, fw_data, fw_size);
	} while (!(ret == _SUCCESS ||
		    (time_after(jiffies, fwdl_timeout) && write_fw_retry++ >= 3)))

The idea is to break only on fatal errors to make things less strict




With regards,
Pavel Skripkin

Download attachment "OpenPGP_signature" of type "application/pgp-signature" (841 bytes)