Re: [PATCH 12/15] tty: serial: switch from circ_buf to kfifo
From: Anders Roxell
Date: Wed Apr 17 2024 - 07:19:39 EST
On Wed, 17 Apr 2024 at 12:20, Marek Szyprowski <m.szyprowski@xxxxxxxxxxx> wrote:
>
> On 17.04.2024 12:08, Anders Roxell wrote:
> > On 2024-04-15 15:28, Jiri Slaby wrote:
> >> On 15. 04. 24, 14:58, Marek Szyprowski wrote:
> >>> On 05.04.2024 08:08, Jiri Slaby (SUSE) wrote:
> >>>> Switch from struct circ_buf to proper kfifo. kfifo provides much better
> >>>> API, esp. when wrap-around of the buffer needs to be taken into account.
> >>>> Look at pl011_dma_tx_refill() or cpm_uart_tx_pump() changes for example.
> >>>>
> >>>> Kfifo API can also fill in scatter-gather DMA structures, so it easier
> >>>> for that use case too. Look at lpuart_dma_tx() for example. Note that
> >>>> not all drivers can be converted to that (like atmel_serial), they
> >>>> handle DMA specially.
> >>>>
> >>>> Note that usb-serial uses kfifo for TX for ages.
> >>>>
> >>>> omap needed a bit more care as it needs to put a char into FIFO to start
> >>>> the DMA transfer when OMAP_DMA_TX_KICK is set. In that case, we have to
> >>>> do kfifo_dma_out_prepare twice: once to find out the tx_size (to find
> >>>> out if it is worths to do DMA at all -- size >= 4), the second time for
> >>>> the actual transfer.
> >>>>
> >>>> All traces of circ_buf are removed from serial_core.h (and its struct
> >>>> uart_state).
> >>>>
> >>>> Signed-off-by: Jiri Slaby (SUSE) <jirislaby@xxxxxxxxxx>
> >>>> ...
> >>> This patch landed in linux-next as commit 1788cf6a91d9 ("tty: serial:
> >>> switch from circ_buf to kfifo"). Unfortunately it breaks UART operation
> >>> on thr Amlogic Meson based boards (drivers/tty/serial/meson_uart.c
> >>> driver) and Qualcomm RB5 board (drivers/tty/serial/qcom_geni_serial.c).
> >>> Once the init process is started, a complete garbage is printed to the
> >>> serial console. Here is an example how it looks:
> >> Oh my!
> >>
> >> Both drivers move the tail using both kfifo and uart_xmit_advance()
> >> interfaces. Bah. Does it help to remove that uart_xmit_advance() for both of
> >> them? (TX stats will be broken.)
> >>
> >> Users of uart_port_tx() are not affected.
> >>
> >> This is my fault when merging uart_xmit_advance() with this series.
> >>
> > I'm trying to run on two dragonboard devices db410c and db845c and both
> > fails to boot see the boot failure from db845c [1], linux-next tag: next-20240415.
> > I tried to apply the patch [2] (that you proposed in this thread) ontop of next-20240415. However, that didn't
> > help bootlog on db845c [3].
>
> This is a different issue, which I've reported 2 days ago. See the
> following thread:
>
> https://lore.kernel.org/all/d3eb9f21-f3e1-43ec-bf41-984c6aa5cfc8@xxxxxxxxxxx/
Oh ok, I did the bisection on db845v, and that led me to this
patch 1788cf6a91d9 ("tty: serial: switch from circ_buf to kfifo")
Cheers,
Anders