linux-kernel - Re: [PATCH] USB: realtek_cr: fix return check for dma functions

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] [day] [month] [year] [list]

Message-ID: <20200811194319.GB344152@rowland.harvard.edu>
Date:   Tue, 11 Aug 2020 15:43:19 -0400
From:   Alan Stern <stern@...land.harvard.edu>
To:     Tom Rix <trix@...hat.com>
Cc:     gregkh@...uxfoundation.org, acozzette@...hmc.edu,
        linux-usb@...r.kernel.org, usb-storage@...ts.one-eyed-alien.net,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH] USB: realtek_cr: fix return check for dma functions

On Tue, Aug 11, 2020 at 11:54:28AM -0700, Tom Rix wrote:
> 
> On 8/11/20 10:53 AM, Alan Stern wrote:

> >>> Instead of changing all these call sites, wouldn't it be a lot easier 
> >>> just to change rts51x_read_mem() to make it always return a negative 
> >>> value (such as -EIO) when there's an error?
> >>>
> >>> Alan Stern
> >> I thought about that but there was already existing (retval != 
> >> STATUS_SUCCESS) checks for these calls.
> > The only values that routine currently returns are 
> > USB_STOR_TRANSPORT_ERROR, -EIO, and 0.  None of the callers distinguish 
> > between the first two values, so you can just change the first to the 
> > second.
> >
> > Note that STATUS_SUCCESS is simply 0.
> 
> Yes, i noted all of these already. My change is consistent with the 
> existing correct checks.  consistency is important.  returning a neg 
> value to reuse the exiting check should mean the STATUS_SUCCESS != 0 
> checks are changed to neg check.

Do you mean the "retval == STATUS_SUCCESS" checks?  Those checks would 
end up doing exactly the same thing as they do now, since 
USB_STOR_TRANSPORT_ERROR and -EIO are both different from 0.

Yes, it is true that consistency is important.  But correctness is more 
important than consistency.

>  i can do this larger change if 
> required.

Let me put it this way: Suppose you changed the USB_STOR_TRANSPORT_ERROR 
in rts51x_read_mem() to -EIO, without changing anything else.  Wouldn't 
that fix the problem reported by the clang static analysis?  If not, why 
not?

Alan Stern