linux-kernel - Re: [PATCH v2] i2c: axxia: support sequence command mode

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: <20181217223122.ye5as5dn5f7rm2a4@ninjato>
Date:   Mon, 17 Dec 2018 23:31:22 +0100
From:   Wolfram Sang <wsa@...-dreams.de>
To:     "Adamski, Krzysztof (Nokia - PL/Wroclaw)" 
        <krzysztof.adamski@...ia.com>
Cc:     "linux-i2c@...r.kernel.org" <linux-i2c@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        Linus Walleij <linus.walleij@...aro.org>,
        Guenter Roeck <linux@...ck-us.net>,
        Tobias Jordan <Tobias.Jordan@...ktrobit.com>,
        Peter Rosin <peda@...ntia.se>,
        "Sverdlin, Alexander (Nokia - DE/Ulm)" <alexander.sverdlin@...ia.com>
Subject: Re: [PATCH v2] i2c: axxia: support sequence command mode

On Thu, Dec 13, 2018 at 12:09:38PM +0000, Adamski, Krzysztof (Nokia - PL/Wroclaw) wrote:
> In order to comply with SMBus specification, the Axxia I²C module will
> abort the multi message transfer if the delay between finishing sending
> one message and starting another is longer than 25ms. Unfortunately it
> isn't that hard to trigger this situation on a busy system. In order to
> fix this problem, we should make sure hardware does whole transaction
> without waiting for software to fill some data.
> 
> Fortunately, in addition to Manual mode that is currently used by the
> driver to perform I²C transfers, the module supports also so called
> Sequence mode. In this mode, the module automatically performs
> predefined sequence of operations - it sends a slave address, transmits
> specified number of bytes from the FIFO, changes transfer direction,
> resends the slave address and then reads specified number of bytes to
> FIFO. While very inflexible, this does fit a most common case of multi
> message transfer - the one where you first write a register number you
> want to read and then read it.
> 
> To use this mode effectively, a number of conditions must be met to
> ensure the transaction does fit the predefined sequence. In case this is
> not the case, a fallback to manual mode is used.
> 
> The initialization of this mode is very similar to Manual mode. The most
> notable difference is different bit in the Master Interrupt Status
> designating finishing of transaction. Also some of the errors, like TSS,
> cannot happen in this mode.
> 
> While it is possible to support transactions requesting a read of any
> size (RFL interrupt will be generated when FIFO size is not enough) the
> TFL interrupt is not available in this mode, thus the write part of the
> transaction cannot exceed FIFO_SIZE (8).
> 
> Note that in case of a NAK during transaction, the NA/ND status bits
> will be set before STOP command is generated, triggering an interrupt
> while the controller is still busy. Current solution for this problem is
> to actively wait for this command to stop before leaving xfer callback.
> 
> Signed-off-by: Krzysztof Adamski <krzysztof.adamski@...ia.com>

Fixed this 'checkpatch --strict' warning:

CHECK: braces {} should be used on all arms of this statement
#201: FILE: drivers/i2c/busses/i2c-axxia.c:422:

... and applied to for-next, thanks!


Download attachment "signature.asc" of type "application/pgp-signature" (834 bytes)