linux-kernel - Re: [PATCH v2] staging: r8188eu: handle rtw_init_netdev

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:   Mon, 24 Jan 2022 09:49:18 +0300
From:   Dan Carpenter <dan.carpenter@...cle.com>
To:     Vihas Mak <makvihas@...il.com>
Cc:     Larry Finger <Larry.Finger@...inger.net>,
        Phillip Potter <phil@...lpotter.co.uk>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        Michael Straube <straube.linux@...il.com>,
        Martin Kaiser <martin@...ser.cx>,
        linux-staging@...ts.linux.dev, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2] staging: r8188eu: handle rtw_init_netdev_name()
 failure appropriately

On Sun, Jan 23, 2022 at 11:47:35PM +0530, Vihas Mak wrote:
> rtw_init_netdev_name() calls dev_alloc_name() which allocates the name
> for the device as per the given name format.
> The name format is specified by the module parameter "ifname".
> It returns a negative err code if the format is invalid. Handle this
> error appropriately.
> Cancel the timers ininitliazed by rtw_init_drv_sw() before calling
> rtw_free_drv_sw() and then proceed to free the adapter.
> 
> Also, if register_netdev() fails then goto free_drv_sw instead of
> goto handle_dualmac.
> 
> Signed-off-by: Vihas Mak <makvihas@...il.com>
> ---
> v1->v2:
>     free the adapter and netdev instead of warning the user about 
>     allocation failure.

Thanks!

Reviewed-by: Dan Carpenter <dan.carpenter@...cle.com>

Of course, all this code is staging code and terrible.  This function
is needlessly difficult to read/review.

TODO: re-write probe error handling

Step 1: Keep the success path and error path separate.

-	status = _SUCCESS;
+	return padapter;

Step 2: Eliminate do-nothing-gotos.  s/goto exit/return NULL/

Step 3: Delete the vfree(pnpi->priv); from rtw_free_netdev() and call
        vfree(pnpi->priv); from probe and rtw_usb_if1_deinit() instead.
        Avoid a layering violation.

Step 4: Every allocation function needs a matching free function.  Move
        the rtw_cancel_all_timer() into the rtw_free_drv_sw() function.
        Open coding it is a layering violation.

Step 5: Get rid of the rtw_handle_dualmac() function.  It has a bad
        name and a global variable.  What is the point of this function?

But that stuff is for later patches.

regards,
dan carpenter