linux-kernel - Re: [usb-storage] Re: [PATCH] usb: storage: realtek_cr: Simplify rts51x_bulk

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: <49559415-4493-4b90-a9bc-c6b9074a61f6@rowland.harvard.edu>
Date: Tue, 12 Aug 2025 21:38:20 -0400
From: Alan Stern <stern@...land.harvard.edu>
To: Thorsten Blum <thorsten.blum@...ux.dev>
Cc: Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
	linux-usb@...r.kernel.org, usb-storage@...ts.one-eyed-alien.net,
	linux-kernel@...r.kernel.org
Subject: Re: [usb-storage] Re: [PATCH] usb: storage: realtek_cr: Simplify
 rts51x_bulk_transport()

On Tue, Aug 12, 2025 at 11:28:56PM +0200, Thorsten Blum wrote:
> Hi Alan,
> 
> On 12. Aug 2025, at 22:06, Alan Stern wrote:
> > On Tue, Aug 12, 2025 at 04:43:58PM +0200, Thorsten Blum wrote:
> >> Change the function parameter 'buf_len' from 'int' to 'unsigned int' and
> >> only update the local variable 'residue' if needed.
> >> 
> >> Update the rts51x_read_status() function signature accordingly.
> > 
> > That last part isn't really necessary, is it?  It doesn't make the code 
> > any clearer, less buggy, or quicker to execute.
> 
> It's mostly for consistency because the parameter 'len' is used to call
> rts51x_bulk_transport() which now expects an unsigned integer. I'd still
> argue that it makes the code and the function signature a bit clearer
> because now the type communicates that 'len' cannot be negative.
> 
> >> 	- if (residue)
> >> 		- residue = residue < buf_len ? residue : buf_len;
> >> 	+ if (residue > buf_len)
> >> 		+ residue = buf_len;
> > 
> > This really has nothing at all to do with whether buf_len is a signed 
> > quantity -- it should never be negative.  (And I have no idea why the 
> > original code includes that test for residue being nonzero.)
> 
> I agree with "it should never be negative" and ideally the type should
> reflect this if possible.
> 
> It's also easier to reason about the code when comparing two unsigned
> integers than having to think about implicit type conversion.
> 
> > Much more serious is something you didn't change: Just above these lines 
> > it says:
> > 
> > 	residue = bcs->Residue;
> > 
> > It should say:
> > 
> > 	residue = le32_to_cpu(bcs->Residue);
> 
> That should probably be another patch, no?

So we're really talking about three separate things:

	Making buf_len and len unsigned;

	Simplifying the calculation of residue;

	Using the correct byte order for bcs->Residue.

The last one fixes a real bug; the other two are very minor by 
comparison.  Regardless, they should be in three separate patches.

If you would like to submit three new patches, please do.

Alan Stern