Re: [stable-kernel][5.15.y][PATCH 0/5] Fix a regression on sc16is7xx

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Thu, 12 Dec 2024 22:00:00 +0800
Hui Wang <hui.wang@xxxxxxxxxxxxx> wrote:

> 
> On 12/12/24 21:44, Greg KH wrote:
> > On Wed, Dec 11, 2024 at 12:25:39PM +0800, Hui Wang wrote:
> >> Recently we found the fifo_read() and fifo_write() are broken in our
> >> 5.15 kernel after rebase to the latest 5.15.y, the 5.15.y integrated
> >> the commit e635f652696e ("serial: sc16is7xx: convert from _raw_ to
> >> _noinc_ regmap functions for FIFO"), but it forgot to integrate a
> >> prerequisite commit 3837a0379533 ("serial: sc16is7xx: improve regmap
> >> debugfs by using one regmap per port").
> >>
> >> And about the prerequisite commit, there are also 4 commits to fix it,
> >> So in total, I backported 5 patches to 5.15.y to fix this regression.
> >>
> >> 0002-xxx and 0004-xxx could be cleanly applied to 5.15.y, the remaining
> >> 3 patches need to resolve some conflict.
> >>
> >> Hugo Villeneuve (5):
> >>    serial: sc16is7xx: improve regmap debugfs by using one regmap per port
> >>    serial: sc16is7xx: remove wasteful static buffer in
> >>      sc16is7xx_regmap_name()
> >>    serial: sc16is7xx: remove global regmap from struct sc16is7xx_port
> >>    serial: sc16is7xx: remove unused line structure member
> >>    serial: sc16is7xx: change EFR lock to operate on each channels
> >>
> >>   drivers/tty/serial/sc16is7xx.c | 185 +++++++++++++++++++--------------
> >>   1 file changed, 107 insertions(+), 78 deletions(-)
> > How well did you test this series?  It seems you forgot about commit
> > 133f4c00b8b2 ("serial: sc16is7xx: fix TX fifo corruption"), right?
> >
> > Please do better testing and resend a working set of patches.
> 
> Okay, got it.

Hi Hui / Greg,
I am testing these changes on my RS-485 board, and I
found out that this patch is required:

commit b4a778303ea0 ("serial: sc16is7xx: add missing support for rs485
devicetree properties")

With it, it now works (basic loopback test) on 5.15 branch and with my
hardware.

As per Greg's suggestion, I have also tested (and reworked) commit
133f4c00b8b2 ("serial: sc16is7xx: fix TX fifo corruption"), with a
prerequisite patch for it to apply more easily: 53a8c50802745 ("serial:
sc16is7xx: refactor FIFO access functions to increase commonality").

And finally I have added commit c41698d1a04cb ("serial: sc16is7xx: fix
invalid FIFO access with special register set").

I will submit these 4 patches to stable soon.

-- 
Hugo Villeneuve




[Index of Archives]     [Linux Kernel]     [Kernel Development Newbies]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite Hiking]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux