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: <70ce6ecd-a5a0-42a7-9063-2982f18078cc@broadcom.com>
Date: Mon, 12 Aug 2024 22:00:35 -0700
From: Florian Fainelli <florian.fainelli@...adcom.com>
To: Cristian Marussi <cristian.marussi@....com>
Cc: linux-arm-kernel@...ts.infradead.org, Rob Herring <robh@...nel.org>,
 Krzysztof Kozlowski <krzk+dt@...nel.org>, Conor Dooley
 <conor+dt@...nel.org>, Sudeep Holla <sudeep.holla@....com>,
 "open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS"
 <devicetree@...r.kernel.org>, open list <linux-kernel@...r.kernel.org>,
 arm-scmi@...r.kernel.org, james.quinlan@...adcom.com,
 justin.chen@...adcom.com, kapil.hali@...adcom.com,
 bcm-kernel-feedback-list@...adcom.com
Subject: Re: [PATCH 2/2] firmware: arm_scmi: Support 'reg-io-width' property
 for shared memory



On 8/12/2024 10:46 AM, Florian Fainelli wrote:
[snip]
>> what about these (and other) header reads if reg-io-width is defined 
>> as < 32 ?
>> Should not these accesses be size-wise too ? or I am missing smth ...
> 
> Good question, I suppose it depends whether 'reg-io-width' means that 
> this must be the strict access width we use, or if this is the minimum 
> access width supported. If the former, then yes, we do have to make a 
> whole lot of changes to support the only access width being supported, 
> if the latter, then we ought to be OK, because doing a 32-bit access 
> should drive more byte enables at the bus level, yet still return the 
> expected data.
> 
> A minimum or only supported access width of 64-bit would be quite 
> interesting, and not somewhat compatible with how SCMI is defined, so 
> maybe that one should not be supported at all, even if this is how 
> memcpy_{to,from}_io() decides to operate on parts of the memory that are 
> 8bytes aligned.

I am inclined to dropping support for doing 1 and 2 byte accesses, and 
support only 4-byte accesses, since the existing SCMI code makes use of 
io{read,write}32 in many places, unless you feel strongly about it.

1 and 2 byte accesses only do not quite make sense for a SRAM IMHO, that 
is, if you can support 1 byte, then you must support 4 byte, too.
-- 
Florian


Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ