Received-SPF: pass (google.com: domain of linux-kernel+bounces-187782-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.80.249 as permitted sender) client-ip=147.75.80.249;
Date: Thu, 23 May 2024 12:30:52 -0400
From: Alan Stern <stern@rowland.harvard.edu>
To: shichao lai <shichaorai@gmail.com>
Cc: gregkh@linuxfoundation.org, oneukum@suse.com, linux-usb@vger.kernel.org,
  usb-storage@lists.one-eyed-alien.net, linux-kernel@vger.kernel.org,
  xingwei lee <xrivendell7@gmail.com>, yue sun <samsun1006219@gmail.com>
Subject: Re: [PATCHv2] Check whether divisor is non-zero before division
Message-ID: <eb995078-1923-43d5-a20f-9d4a7edee719@rowland.harvard.edu>
References: <20240523092608.874986-1-shichaorai@gmail.com>
 <a0afa88b-f84c-4b45-a265-2e6bcbb84b35@rowland.harvard.edu>
 <bb581989-4ac5-4ffe-9f80-01b5f993146f@rowland.harvard.edu>
 <CACjpba5iJ5dC=rB_Ckaqe4BKesrAN2VmsDCPZJ=frufNgA96Uw@mail.gmail.com>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <CACjpba5iJ5dC=rB_Ckaqe4BKesrAN2VmsDCPZJ=frufNgA96Uw@mail.gmail.com>

On Thu, May 23, 2024 at 11:13:08PM +0800, shichao lai wrote:
> Thanks for your comprehensive analysis.
> I added some pr_info() to check the workflow, and I found that the
> uzonesize was not initialized in fact!
> 
> The workflow is shown as below.
> Before alauda_read_data(), there are in fact many alauda_check_media(),
> but none of them enter the branch of alauda_init_media(), where
> uzonesize is set to nonzero value.
> The key branch condition is "status[0] & 0x08", which is always
> unsatisfied in this repro.
> 
> ```
> alauda_transport
>     alauda_check_media
>         if (status[0] & 0x08) // not satisfied
>             alauda_init_media()
>                 // initialize uzonesize
>     alauda_read_data
> ```

Good work!  So the problem is that the driver believes the status[0] & 
0x08 test.  

The way to fix this is to add an "initialized" flag to the alauda_info 
structure.  Then alauda_check_media() should call alauda_init_media() if 
the 0x08 bit is set in status[0] _or_ if info->initialized is 0.  And of 
course, alauda_check_media() should then set info->initialized to 1 if 
the alauda_init_media() call succeeds.

Would you like to write and test a patch that does this?

Alan Stern