linux-kernel - Re: [PATCH] serial: msm: Configure correct working mode before starting earlycon

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: <Z_Zp0yCXrYM9zG83@linaro.org>
Date: Wed, 9 Apr 2025 14:36:35 +0200
From: Stephan Gerhold <stephan.gerhold@...aro.org>
To: Mukesh Kumar Savaliya <quic_msavaliy@...cinc.com>
Cc: Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
	Jiri Slaby <jirislaby@...nel.org>, Stephen Boyd <sboyd@...nel.org>,
	linux-arm-msm@...r.kernel.org, linux-kernel@...r.kernel.org,
	linux-serial@...r.kernel.org, Sam Day <me@...cday.com>,
	Chandana Kishori Chiluveru <cchiluve@....qualcomm.com>
Subject: Re: [PATCH] serial: msm: Configure correct working mode before
 starting earlycon

On Wed, Apr 09, 2025 at 03:30:02PM +0530, Mukesh Kumar Savaliya wrote:
> On 4/8/2025 10:52 PM, Stephan Gerhold wrote:
> > The MSM UART DM controller supports different working modes, e.g. DMA or
> > the "single-character mode", where all reads/writes operate on a single
> > character rather than 4 chars (32-bit) at once. When using earlycon,
> > __msm_console_write() always writes 4 characters at a time, but we don't
> > know which mode the bootloader was using and we don't set the mode either.
> > 
> Looks surprising. I haven't seen pre-kernel UART console ever works in DMA
> mode. It was always fixed to work in FIFO/PIO mode. From what i know.

I'm sure you're right, since it doesn't really make sense to implement
DMA mode for the UART console in earlier firmware or the bootloader.

DMA is just the side note here though. As I wrote in the patch
description, the real problem this patch fixes is the "single-character
mode". This is still FIFO/PIO mode, except that the register will
take/return just a single character rather than 4 chars at once.

This mode is used by firmware on some platforms and I'm planning to make
use of this mode in U-Boot to fix some edge cases. It's much more simple
to implement reliably for something minimal like U-Boot. With that
change in U-Boot, I get the garbled output shown below in the earlycon
serial console. This patch fixes it.

> > This causes garbled output if the bootloader was using the single-character
> > mode, because only every 4th character appears in the serial console, e.g.
> > 
> >    "[ 00oni pi  000xf0[ 00i s 5rm9(l)l s 1  1 SPMTA 7:C 5[ 00A ade k d[
> >     00ano:ameoi .Q1B[ 00ac _idaM00080oo'"
> > 
> > If the bootloader was using the DMA ("DM") mode, output would likely fail
> > entirely. Later, when the full serial driver probes, the port is
> > re-initialized and output works as expected.
> > 
> AFAIR, console UART was always configured in FIFO/PIO mode. For non Console
> application e.g. Bluetooth it works in DMA mode.

I think console UART in msm_serial does use DMA at least for receiving
at the moment, since we don't differentiate between "console" or
"non-console" use case in the upstream driver (as far as I can tell).
My patch doesn't change anything about this though, it just ensures the
UART controller is in the expected mode before starting earlycon.

Thanks,
Stephan