[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-Id: <20210519000704.3661773-1-andrew@aj.id.au>
Date: Wed, 19 May 2021 09:37:02 +0930
From: Andrew Jeffery <andrew@...id.au>
To: linux-serial@...r.kernel.org
Cc: gregkh@...uxfoundation.org, jirislaby@...nel.org, joel@....id.au,
linux-kernel@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
linux-aspeed@...ts.ozlabs.org, openbmc@...ts.ozlabs.org,
jenmin_yuan@...eedtech.com, ryan_chen@...eedtech.com,
miltonm@...ibm.com
Subject: [PATCH v2 0/2] serial: 8250: Mitigate Tx stall risk for Aspeed VUARTs
Hello,
Briefly, the series works around a hardware race condition in the Tx path for
Aspeed virtual UARTs. A write burst to THR on the APB interface may provoke a
transfer stall where LSR[DR] on the LPC interface remains clear despite the
presence of data in the Rx FIFO.
For the work-around patch, v2 addresses the request for a comment about the use
of serial_in():
https://lore.kernel.org/lkml/d7918dcf-b938-498c-a012-3d93a748431b@www.fastmail.com/T/#md75702fbc3704bd4b375f1251a1415bcddea26a3
The second patch addresses the request for use of BIT() instead of an explicit
shift by converting all of the UART_{CAP,BUG}_* macros.
Please review!
Andrew
Andrew Jeffery (2):
serial: 8250: Add UART_BUG_TXRACE workaround for Aspeed VUART
serial: 8250: Use BIT(x) for UART_{CAP,BUG}_*
drivers/tty/serial/8250/8250.h | 32 +++++++++++----------
drivers/tty/serial/8250/8250_aspeed_vuart.c | 1 +
drivers/tty/serial/8250/8250_port.c | 10 +++++++
3 files changed, 28 insertions(+), 15 deletions(-)
--
2.30.2
Powered by blists - more mailing lists