diff mbox series

[v2,2/2] serial: 8250: Add new 8250-core based Broadcom STB driver

Message ID 20210115211543.33563-3-alcooperx@gmail.com (mailing list archive)
State Superseded
Headers show
Series serial: 8250: Add driver for Broadcom UART | expand

Commit Message

Alan Cooper Jan. 15, 2021, 9:15 p.m. UTC
Add a UART driver for the new Broadcom 8250 based STB UART. The new
UART is backward compatible with the standard 8250, but has some
additional features. The new features include a high accuracy baud
rate clock system and DMA support.

The driver will use the new optional BAUD MUX clock to select the best
one of the four master clocks (81MHz, 108MHz, 64MHz and 48MHz) to feed
the baud rate selection logic for any requested baud rate.  This allows
for more accurate BAUD rates when high speed baud rates are selected.

The driver will use the new UART DMA hardware if the UART DMA registers
are specified in Device Tree "reg" property. The DMA functionality can
be disabled on kernel boot with the argument:
"8250_bcm7271.disable_dma=Y".

The driver also set the UPSTAT_AUTOCTS flag when hardware flow control
is enabled. This flag is needed for UARTs that don't assert a CTS
changed interrupt when CTS changes and AFE (Hardware Flow Control) is
enabled.

The driver also contains a workaround for a bug in the Synopsis 8250
core. The problem is that at high baud rates, the RX partial FIFO
timeout interrupt can occur but there is no RX data (DR not set in
the LSR register). In this case the driver will not read the Receive
Buffer Register, which clears the interrupt, and the system will get
continuous UART interrupts until the next RX character arrives. The
fix originally suggested by Synopsis was to read the Receive Buffer
Register and discard the character when the DR bit in the LSR was
not set, to clear the interrupt. The problem was that occasionally
a character would arrive just after the DR bit check and a valid
character would be discarded. The fix that was added will clear
receive interrupts to stop the interrupt, deassert RTS to insure
that no new data can arrive, wait for 1.5 character times for the
sender to react to RTS and then check for data and either do a dummy
read or a valid read. Sysfs error counters were also added and were
used to help create test software that would cause the error condition.
The counters can be found at:
/sys/devices/platform/rdb/*serial/rx_bad_timeout_late_char
/sys/devices/platform/rdb/*serial/rx_bad_timeout_no_char

Signed-off-by: Al Cooper <alcooperx@gmail.com>
---
 MAINTAINERS                            |    8 +
 drivers/tty/serial/8250/8250_bcm7271.c | 1131 ++++++++++++++++++++++++
 drivers/tty/serial/8250/Kconfig        |   11 +
 drivers/tty/serial/8250/Makefile       |    1 +
 drivers/tty/serial/8250/bcm7271_uart.h |  158 ++++
 5 files changed, 1309 insertions(+)
 create mode 100644 drivers/tty/serial/8250/8250_bcm7271.c
 create mode 100644 drivers/tty/serial/8250/bcm7271_uart.h

Comments

Greg Kroah-Hartman Jan. 18, 2021, 5:45 p.m. UTC | #1
On Fri, Jan 15, 2021 at 04:15:43PM -0500, Al Cooper wrote:
> Add a UART driver for the new Broadcom 8250 based STB UART. The new
> UART is backward compatible with the standard 8250, but has some
> additional features. The new features include a high accuracy baud
> rate clock system and DMA support.
> 
> The driver will use the new optional BAUD MUX clock to select the best
> one of the four master clocks (81MHz, 108MHz, 64MHz and 48MHz) to feed
> the baud rate selection logic for any requested baud rate.  This allows
> for more accurate BAUD rates when high speed baud rates are selected.
> 
> The driver will use the new UART DMA hardware if the UART DMA registers
> are specified in Device Tree "reg" property. The DMA functionality can
> be disabled on kernel boot with the argument:
> "8250_bcm7271.disable_dma=Y".

Shouldn't that be on a per-device basis, and not a per-driver basis?

And why would you want to disable this, if you have support for this in
the DT?  Why not just rely on the DT setting?

thanks,

greg k-h
Al Cooper Jan. 18, 2021, 8:32 p.m. UTC | #2
On Mon, Jan 18, 2021 at 12:45 PM Greg Kroah-Hartman
<gregkh@linuxfoundation.org> wrote:
>
> On Fri, Jan 15, 2021 at 04:15:43PM -0500, Al Cooper wrote:
> > Add a UART driver for the new Broadcom 8250 based STB UART. The new
> > UART is backward compatible with the standard 8250, but has some
> > additional features. The new features include a high accuracy baud
> > rate clock system and DMA support.
> >
> > The driver will use the new optional BAUD MUX clock to select the best
> > one of the four master clocks (81MHz, 108MHz, 64MHz and 48MHz) to feed
> > the baud rate selection logic for any requested baud rate.  This allows
> > for more accurate BAUD rates when high speed baud rates are selected.
> >
> > The driver will use the new UART DMA hardware if the UART DMA registers
> > are specified in Device Tree "reg" property. The DMA functionality can
> > be disabled on kernel boot with the argument:
> > "8250_bcm7271.disable_dma=Y".
>
> Shouldn't that be on a per-device basis, and not a per-driver basis?

There is only one instance of the UART DMA hardware and it gets muxed
to just one of the possible UARTS.

>
> And why would you want to disable this, if you have support for this in
> the DT?  Why not just rely on the DT setting?

The DMA feature is used when the UART is connected to a Bluetooth
controller and the BAUD rate is typically 2-3Mbs. The ability to
easily disable DMA is very useful when debugging BT communication
problems in the field. DT settings could also be used to disable DMA,
but knowing the correct modifications to the "reg" and "reg-names"
properties is a lot more complicated.

Thanks
Al

>
> thanks,
>
> greg k-h
Greg Kroah-Hartman Jan. 19, 2021, 12:33 p.m. UTC | #3
On Mon, Jan 18, 2021 at 03:32:57PM -0500, Al Cooper wrote:
> On Mon, Jan 18, 2021 at 12:45 PM Greg Kroah-Hartman
> <gregkh@linuxfoundation.org> wrote:
> >
> > On Fri, Jan 15, 2021 at 04:15:43PM -0500, Al Cooper wrote:
> > > Add a UART driver for the new Broadcom 8250 based STB UART. The new
> > > UART is backward compatible with the standard 8250, but has some
> > > additional features. The new features include a high accuracy baud
> > > rate clock system and DMA support.
> > >
> > > The driver will use the new optional BAUD MUX clock to select the best
> > > one of the four master clocks (81MHz, 108MHz, 64MHz and 48MHz) to feed
> > > the baud rate selection logic for any requested baud rate.  This allows
> > > for more accurate BAUD rates when high speed baud rates are selected.
> > >
> > > The driver will use the new UART DMA hardware if the UART DMA registers
> > > are specified in Device Tree "reg" property. The DMA functionality can
> > > be disabled on kernel boot with the argument:
> > > "8250_bcm7271.disable_dma=Y".
> >
> > Shouldn't that be on a per-device basis, and not a per-driver basis?
> 
> There is only one instance of the UART DMA hardware and it gets muxed
> to just one of the possible UARTS.

But the driver doesn't know/care about that, it binds to any device that
matches it.  per-module/driver flags are not a good idea.

> > And why would you want to disable this, if you have support for this in
> > the DT?  Why not just rely on the DT setting?
> 
> The DMA feature is used when the UART is connected to a Bluetooth
> controller and the BAUD rate is typically 2-3Mbs. The ability to
> easily disable DMA is very useful when debugging BT communication
> problems in the field. DT settings could also be used to disable DMA,
> but knowing the correct modifications to the "reg" and "reg-names"
> properties is a lot more complicated.

So this is a debug-only option?  If so, why not just make it a debugfs
file then?  No need to clutter up a module parameter for this mess.

thanks,

greg k-h
Andy Shevchenko Jan. 19, 2021, 3:21 p.m. UTC | #4
On Fri, Jan 15, 2021 at 11:19 PM Al Cooper <alcooperx@gmail.com> wrote:
>
> Add a UART driver for the new Broadcom 8250 based STB UART. The new
> UART is backward compatible with the standard 8250, but has some
> additional features. The new features include a high accuracy baud
> rate clock system and DMA support.
>
> The driver will use the new optional BAUD MUX clock to select the best
> one of the four master clocks (81MHz, 108MHz, 64MHz and 48MHz) to feed
> the baud rate selection logic for any requested baud rate.  This allows
> for more accurate BAUD rates when high speed baud rates are selected.
>
> The driver will use the new UART DMA hardware if the UART DMA registers
> are specified in Device Tree "reg" property. The DMA functionality can
> be disabled on kernel boot with the argument:
> "8250_bcm7271.disable_dma=Y".
>
> The driver also set the UPSTAT_AUTOCTS flag when hardware flow control
> is enabled. This flag is needed for UARTs that don't assert a CTS
> changed interrupt when CTS changes and AFE (Hardware Flow Control) is
> enabled.
>
> The driver also contains a workaround for a bug in the Synopsis 8250
> core. The problem is that at high baud rates, the RX partial FIFO
> timeout interrupt can occur but there is no RX data (DR not set in
> the LSR register). In this case the driver will not read the Receive
> Buffer Register, which clears the interrupt, and the system will get
> continuous UART interrupts until the next RX character arrives. The
> fix originally suggested by Synopsis was to read the Receive Buffer
> Register and discard the character when the DR bit in the LSR was
> not set, to clear the interrupt. The problem was that occasionally
> a character would arrive just after the DR bit check and a valid
> character would be discarded. The fix that was added will clear
> receive interrupts to stop the interrupt, deassert RTS to insure
> that no new data can arrive, wait for 1.5 character times for the
> sender to react to RTS and then check for data and either do a dummy
> read or a valid read. Sysfs error counters were also added and were
> used to help create test software that would cause the error condition.
> The counters can be found at:
> /sys/devices/platform/rdb/*serial/rx_bad_timeout_late_char
> /sys/devices/platform/rdb/*serial/rx_bad_timeout_no_char

Brief looking into the code raises several questions:
 - is it driver from the last decade?
 - why it's not using what kernel provides?
 - we have a lot of nice helpers:
   - DMA Engine API
   - BIT() and GENMASK() macros
   - tons of different helpers like regmap API (if you wish to dump
registers via debugfs)

Can you shrink this driver by 20-30% (I truly believe it's possible)
and split DMA driver to drivers/dma (which may already have something
similar there)?
Florian Fainelli Jan. 19, 2021, 6:16 p.m. UTC | #5
On 1/19/2021 7:21 AM, Andy Shevchenko wrote:
> On Fri, Jan 15, 2021 at 11:19 PM Al Cooper <alcooperx@gmail.com> wrote:
>>
>> Add a UART driver for the new Broadcom 8250 based STB UART. The new
>> UART is backward compatible with the standard 8250, but has some
>> additional features. The new features include a high accuracy baud
>> rate clock system and DMA support.
>>
>> The driver will use the new optional BAUD MUX clock to select the best
>> one of the four master clocks (81MHz, 108MHz, 64MHz and 48MHz) to feed
>> the baud rate selection logic for any requested baud rate.  This allows
>> for more accurate BAUD rates when high speed baud rates are selected.
>>
>> The driver will use the new UART DMA hardware if the UART DMA registers
>> are specified in Device Tree "reg" property. The DMA functionality can
>> be disabled on kernel boot with the argument:
>> "8250_bcm7271.disable_dma=Y".
>>
>> The driver also set the UPSTAT_AUTOCTS flag when hardware flow control
>> is enabled. This flag is needed for UARTs that don't assert a CTS
>> changed interrupt when CTS changes and AFE (Hardware Flow Control) is
>> enabled.
>>
>> The driver also contains a workaround for a bug in the Synopsis 8250
>> core. The problem is that at high baud rates, the RX partial FIFO
>> timeout interrupt can occur but there is no RX data (DR not set in
>> the LSR register). In this case the driver will not read the Receive
>> Buffer Register, which clears the interrupt, and the system will get
>> continuous UART interrupts until the next RX character arrives. The
>> fix originally suggested by Synopsis was to read the Receive Buffer
>> Register and discard the character when the DR bit in the LSR was
>> not set, to clear the interrupt. The problem was that occasionally
>> a character would arrive just after the DR bit check and a valid
>> character would be discarded. The fix that was added will clear
>> receive interrupts to stop the interrupt, deassert RTS to insure
>> that no new data can arrive, wait for 1.5 character times for the
>> sender to react to RTS and then check for data and either do a dummy
>> read or a valid read. Sysfs error counters were also added and were
>> used to help create test software that would cause the error condition.
>> The counters can be found at:
>> /sys/devices/platform/rdb/*serial/rx_bad_timeout_late_char
>> /sys/devices/platform/rdb/*serial/rx_bad_timeout_no_char
> 
> Brief looking into the code raises several questions:
>  - is it driver from the last decade?

Work on this driver started back in 2018, that was indeed the last decade.

>  - why it's not using what kernel provides?
>  - we have a lot of nice helpers:
>    - DMA Engine API

Not sure this makes sense, given that the DMA hardware that was added to
this UART block is only used by the UART block and no other pieces of HW
in the system, nor will they ever be. Not sure it makes sense to pay the
cost of an extra indirection and subsystem unless there are at least two
consumers of that DMA hardware to warrant modeling it after a dmaengine
driver. I also remember that Al researched before whether 8250_dma.c
could work, and came to the conclusion that it would not, but I will let
him comment on the specifics.


>    - BIT() and GENMASK() macros
>    - tons of different helpers like regmap API (if you wish to dump
> registers via debugfs)
> 
> Can you shrink this driver by 20-30% (I truly believe it's possible)
> and split DMA driver to drivers/dma (which may already have something
> similar there)?

See previous response.
Andy Shevchenko Jan. 20, 2021, 4:47 p.m. UTC | #6
On Tue, Jan 19, 2021 at 8:16 PM Florian Fainelli <f.fainelli@gmail.com> wrote:
> On 1/19/2021 7:21 AM, Andy Shevchenko wrote:
> > On Fri, Jan 15, 2021 at 11:19 PM Al Cooper <alcooperx@gmail.com> wrote:
> >>
> >> Add a UART driver for the new Broadcom 8250 based STB UART. The new
> >> UART is backward compatible with the standard 8250, but has some
> >> additional features. The new features include a high accuracy baud
> >> rate clock system and DMA support.
> >>
> >> The driver will use the new optional BAUD MUX clock to select the best
> >> one of the four master clocks (81MHz, 108MHz, 64MHz and 48MHz) to feed
> >> the baud rate selection logic for any requested baud rate.  This allows
> >> for more accurate BAUD rates when high speed baud rates are selected.
> >>
> >> The driver will use the new UART DMA hardware if the UART DMA registers
> >> are specified in Device Tree "reg" property. The DMA functionality can
> >> be disabled on kernel boot with the argument:
> >> "8250_bcm7271.disable_dma=Y".
> >>
> >> The driver also set the UPSTAT_AUTOCTS flag when hardware flow control
> >> is enabled. This flag is needed for UARTs that don't assert a CTS
> >> changed interrupt when CTS changes and AFE (Hardware Flow Control) is
> >> enabled.
> >>
> >> The driver also contains a workaround for a bug in the Synopsis 8250
> >> core. The problem is that at high baud rates, the RX partial FIFO
> >> timeout interrupt can occur but there is no RX data (DR not set in
> >> the LSR register). In this case the driver will not read the Receive
> >> Buffer Register, which clears the interrupt, and the system will get
> >> continuous UART interrupts until the next RX character arrives. The
> >> fix originally suggested by Synopsis was to read the Receive Buffer
> >> Register and discard the character when the DR bit in the LSR was
> >> not set, to clear the interrupt. The problem was that occasionally
> >> a character would arrive just after the DR bit check and a valid
> >> character would be discarded. The fix that was added will clear
> >> receive interrupts to stop the interrupt, deassert RTS to insure
> >> that no new data can arrive, wait for 1.5 character times for the
> >> sender to react to RTS and then check for data and either do a dummy
> >> read or a valid read. Sysfs error counters were also added and were
> >> used to help create test software that would cause the error condition.
> >> The counters can be found at:
> >> /sys/devices/platform/rdb/*serial/rx_bad_timeout_late_char
> >> /sys/devices/platform/rdb/*serial/rx_bad_timeout_no_char
> >
> > Brief looking into the code raises several questions:
> >  - is it driver from the last decade?
>
> Work on this driver started back in 2018, that was indeed the last decade.
>
> >  - why it's not using what kernel provides?
> >  - we have a lot of nice helpers:
> >    - DMA Engine API
>
> Not sure this makes sense, given that the DMA hardware that was added to
> this UART block is only used by the UART block and no other pieces of HW
> in the system, nor will they ever be. Not sure it makes sense to pay the
> cost of an extra indirection and subsystem unless there are at least two
> consumers of that DMA hardware to warrant modeling it after a dmaengine
> driver. I also remember that Al researched before whether 8250_dma.c
> could work, and came to the conclusion that it would not, but I will let
> him comment on the specifics.

I see. In any case I still believe that the driver can be shrinked by
a notable amount of lines.

> >    - BIT() and GENMASK() macros
> >    - tons of different helpers like regmap API (if you wish to dump
> > registers via debugfs)
> >
> > Can you shrink this driver by 20-30% (I truly believe it's possible)
> > and split DMA driver to drivers/dma (which may already have something
> > similar there)?
>
> See previous response.
Greg Kroah-Hartman Jan. 20, 2021, 5:05 p.m. UTC | #7
On Wed, Jan 20, 2021 at 06:47:52PM +0200, Andy Shevchenko wrote:
> On Tue, Jan 19, 2021 at 8:16 PM Florian Fainelli <f.fainelli@gmail.com> wrote:
> > On 1/19/2021 7:21 AM, Andy Shevchenko wrote:
> > > On Fri, Jan 15, 2021 at 11:19 PM Al Cooper <alcooperx@gmail.com> wrote:
> > >>
> > >> Add a UART driver for the new Broadcom 8250 based STB UART. The new
> > >> UART is backward compatible with the standard 8250, but has some
> > >> additional features. The new features include a high accuracy baud
> > >> rate clock system and DMA support.
> > >>
> > >> The driver will use the new optional BAUD MUX clock to select the best
> > >> one of the four master clocks (81MHz, 108MHz, 64MHz and 48MHz) to feed
> > >> the baud rate selection logic for any requested baud rate.  This allows
> > >> for more accurate BAUD rates when high speed baud rates are selected.
> > >>
> > >> The driver will use the new UART DMA hardware if the UART DMA registers
> > >> are specified in Device Tree "reg" property. The DMA functionality can
> > >> be disabled on kernel boot with the argument:
> > >> "8250_bcm7271.disable_dma=Y".
> > >>
> > >> The driver also set the UPSTAT_AUTOCTS flag when hardware flow control
> > >> is enabled. This flag is needed for UARTs that don't assert a CTS
> > >> changed interrupt when CTS changes and AFE (Hardware Flow Control) is
> > >> enabled.
> > >>
> > >> The driver also contains a workaround for a bug in the Synopsis 8250
> > >> core. The problem is that at high baud rates, the RX partial FIFO
> > >> timeout interrupt can occur but there is no RX data (DR not set in
> > >> the LSR register). In this case the driver will not read the Receive
> > >> Buffer Register, which clears the interrupt, and the system will get
> > >> continuous UART interrupts until the next RX character arrives. The
> > >> fix originally suggested by Synopsis was to read the Receive Buffer
> > >> Register and discard the character when the DR bit in the LSR was
> > >> not set, to clear the interrupt. The problem was that occasionally
> > >> a character would arrive just after the DR bit check and a valid
> > >> character would be discarded. The fix that was added will clear
> > >> receive interrupts to stop the interrupt, deassert RTS to insure
> > >> that no new data can arrive, wait for 1.5 character times for the
> > >> sender to react to RTS and then check for data and either do a dummy
> > >> read or a valid read. Sysfs error counters were also added and were
> > >> used to help create test software that would cause the error condition.
> > >> The counters can be found at:
> > >> /sys/devices/platform/rdb/*serial/rx_bad_timeout_late_char
> > >> /sys/devices/platform/rdb/*serial/rx_bad_timeout_no_char
> > >
> > > Brief looking into the code raises several questions:
> > >  - is it driver from the last decade?
> >
> > Work on this driver started back in 2018, that was indeed the last decade.
> >
> > >  - why it's not using what kernel provides?
> > >  - we have a lot of nice helpers:
> > >    - DMA Engine API
> >
> > Not sure this makes sense, given that the DMA hardware that was added to
> > this UART block is only used by the UART block and no other pieces of HW
> > in the system, nor will they ever be. Not sure it makes sense to pay the
> > cost of an extra indirection and subsystem unless there are at least two
> > consumers of that DMA hardware to warrant modeling it after a dmaengine
> > driver. I also remember that Al researched before whether 8250_dma.c
> > could work, and came to the conclusion that it would not, but I will let
> > him comment on the specifics.
> 
> I see. In any case I still believe that the driver can be shrinked by
> a notable amount of lines.

Patches always gladly accepted :)
Andy Shevchenko Jan. 20, 2021, 5:09 p.m. UTC | #8
On Wed, Jan 20, 2021 at 7:05 PM Greg Kroah-Hartman
<gregkh@linuxfoundation.org> wrote:
> On Wed, Jan 20, 2021 at 06:47:52PM +0200, Andy Shevchenko wrote:
> > On Tue, Jan 19, 2021 at 8:16 PM Florian Fainelli <f.fainelli@gmail.com> wrote:
> > > On 1/19/2021 7:21 AM, Andy Shevchenko wrote:
> > > > On Fri, Jan 15, 2021 at 11:19 PM Al Cooper <alcooperx@gmail.com> wrote:

...

> > > Not sure this makes sense, given that the DMA hardware that was added to
> > > this UART block is only used by the UART block and no other pieces of HW
> > > in the system, nor will they ever be. Not sure it makes sense to pay the
> > > cost of an extra indirection and subsystem unless there are at least two
> > > consumers of that DMA hardware to warrant modeling it after a dmaengine
> > > driver. I also remember that Al researched before whether 8250_dma.c
> > > could work, and came to the conclusion that it would not, but I will let
> > > him comment on the specifics.
> >
> > I see. In any case I still believe that the driver can be shrinked by
> > a notable amount of lines.
>
> Patches always gladly accepted :)

Or a good review... :-)
Greg Kroah-Hartman Jan. 20, 2021, 5:15 p.m. UTC | #9
On Wed, Jan 20, 2021 at 07:09:46PM +0200, Andy Shevchenko wrote:
> On Wed, Jan 20, 2021 at 7:05 PM Greg Kroah-Hartman
> <gregkh@linuxfoundation.org> wrote:
> > On Wed, Jan 20, 2021 at 06:47:52PM +0200, Andy Shevchenko wrote:
> > > On Tue, Jan 19, 2021 at 8:16 PM Florian Fainelli <f.fainelli@gmail.com> wrote:
> > > > On 1/19/2021 7:21 AM, Andy Shevchenko wrote:
> > > > > On Fri, Jan 15, 2021 at 11:19 PM Al Cooper <alcooperx@gmail.com> wrote:
> 
> ...
> 
> > > > Not sure this makes sense, given that the DMA hardware that was added to
> > > > this UART block is only used by the UART block and no other pieces of HW
> > > > in the system, nor will they ever be. Not sure it makes sense to pay the
> > > > cost of an extra indirection and subsystem unless there are at least two
> > > > consumers of that DMA hardware to warrant modeling it after a dmaengine
> > > > driver. I also remember that Al researched before whether 8250_dma.c
> > > > could work, and came to the conclusion that it would not, but I will let
> > > > him comment on the specifics.
> > >
> > > I see. In any case I still believe that the driver can be shrinked by
> > > a notable amount of lines.
> >
> > Patches always gladly accepted :)
> 
> Or a good review... :-)

Please do so!
Al Cooper Jan. 20, 2021, 5:30 p.m. UTC | #10
I took another look at using the DMA framework for this and I still
don't think it fits. The problem is that the DMA hardware is so
tightly coupled to the 8250 UART hardware that a generic DMA model
where a DMA engine will read the data from the UART data register when
Data Ready is asserted and transfer it to memory, doesn't fit.  For
example:
- The flush timeout, parity error, framing error and overrun errors
are kept separately in the DMA hardware.
- The DMA hardware expects to be enabled before receiving input data
and it will continually write the data to ping pong data buffers as
long as a buffer empty bit is set on the next buffer.
- When DMA is enabled it bypasses some of the normal 8250 control and
data registers.

I think that trying to force this hardware into the current 8250 DMA
model using the DMA framework would not be a good approach.
I'll look at the other suggestions.

Thanks
Al



On Wed, Jan 20, 2021 at 11:47 AM Andy Shevchenko
<andy.shevchenko@gmail.com> wrote:
>
> On Tue, Jan 19, 2021 at 8:16 PM Florian Fainelli <f.fainelli@gmail.com> wrote:
> > On 1/19/2021 7:21 AM, Andy Shevchenko wrote:
> > > On Fri, Jan 15, 2021 at 11:19 PM Al Cooper <alcooperx@gmail.com> wrote:
> > >>
> > >> Add a UART driver for the new Broadcom 8250 based STB UART. The new
> > >> UART is backward compatible with the standard 8250, but has some
> > >> additional features. The new features include a high accuracy baud
> > >> rate clock system and DMA support.
> > >>
> > >> The driver will use the new optional BAUD MUX clock to select the best
> > >> one of the four master clocks (81MHz, 108MHz, 64MHz and 48MHz) to feed
> > >> the baud rate selection logic for any requested baud rate.  This allows
> > >> for more accurate BAUD rates when high speed baud rates are selected.
> > >>
> > >> The driver will use the new UART DMA hardware if the UART DMA registers
> > >> are specified in Device Tree "reg" property. The DMA functionality can
> > >> be disabled on kernel boot with the argument:
> > >> "8250_bcm7271.disable_dma=Y".
> > >>
> > >> The driver also set the UPSTAT_AUTOCTS flag when hardware flow control
> > >> is enabled. This flag is needed for UARTs that don't assert a CTS
> > >> changed interrupt when CTS changes and AFE (Hardware Flow Control) is
> > >> enabled.
> > >>
> > >> The driver also contains a workaround for a bug in the Synopsis 8250
> > >> core. The problem is that at high baud rates, the RX partial FIFO
> > >> timeout interrupt can occur but there is no RX data (DR not set in
> > >> the LSR register). In this case the driver will not read the Receive
> > >> Buffer Register, which clears the interrupt, and the system will get
> > >> continuous UART interrupts until the next RX character arrives. The
> > >> fix originally suggested by Synopsis was to read the Receive Buffer
> > >> Register and discard the character when the DR bit in the LSR was
> > >> not set, to clear the interrupt. The problem was that occasionally
> > >> a character would arrive just after the DR bit check and a valid
> > >> character would be discarded. The fix that was added will clear
> > >> receive interrupts to stop the interrupt, deassert RTS to insure
> > >> that no new data can arrive, wait for 1.5 character times for the
> > >> sender to react to RTS and then check for data and either do a dummy
> > >> read or a valid read. Sysfs error counters were also added and were
> > >> used to help create test software that would cause the error condition.
> > >> The counters can be found at:
> > >> /sys/devices/platform/rdb/*serial/rx_bad_timeout_late_char
> > >> /sys/devices/platform/rdb/*serial/rx_bad_timeout_no_char
> > >
> > > Brief looking into the code raises several questions:
> > >  - is it driver from the last decade?
> >
> > Work on this driver started back in 2018, that was indeed the last decade.
> >
> > >  - why it's not using what kernel provides?
> > >  - we have a lot of nice helpers:
> > >    - DMA Engine API
> >
> > Not sure this makes sense, given that the DMA hardware that was added to
> > this UART block is only used by the UART block and no other pieces of HW
> > in the system, nor will they ever be. Not sure it makes sense to pay the
> > cost of an extra indirection and subsystem unless there are at least two
> > consumers of that DMA hardware to warrant modeling it after a dmaengine
> > driver. I also remember that Al researched before whether 8250_dma.c
> > could work, and came to the conclusion that it would not, but I will let
> > him comment on the specifics.
>
> I see. In any case I still believe that the driver can be shrinked by
> a notable amount of lines.
>
> > >    - BIT() and GENMASK() macros
> > >    - tons of different helpers like regmap API (if you wish to dump
> > > registers via debugfs)
> > >
> > > Can you shrink this driver by 20-30% (I truly believe it's possible)
> > > and split DMA driver to drivers/dma (which may already have something
> > > similar there)?
> >
> > See previous response.
>
>
> --
> With Best Regards,
> Andy Shevchenko
diff mbox series

Patch

diff --git a/MAINTAINERS b/MAINTAINERS
index 79b400c97059..0ff3a33a3834 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -3581,6 +3581,14 @@  S:	Supported
 F:	Documentation/devicetree/bindings/i2c/brcm,brcmstb-i2c.yaml
 F:	drivers/i2c/busses/i2c-brcmstb.c
 
+BROADCOM BRCMSTB UART DRIVER
+M:	Al Cooper <alcooperx@gmail.com>
+L:	linux-usb@vger.kernel.org
+L:	bcm-kernel-feedback-list@broadcom.com
+S:	Maintained
+F:	Documentation/devicetree/bindings/serial/brcm,bcm7271-uart.yaml
+F:	drivers/tty/serial/8250/8250_bcm7271.c
+
 BROADCOM BRCMSTB USB EHCI DRIVER
 M:	Al Cooper <alcooperx@gmail.com>
 L:	linux-usb@vger.kernel.org
diff --git a/drivers/tty/serial/8250/8250_bcm7271.c b/drivers/tty/serial/8250/8250_bcm7271.c
new file mode 100644
index 000000000000..194519a3a158
--- /dev/null
+++ b/drivers/tty/serial/8250/8250_bcm7271.c
@@ -0,0 +1,1131 @@ 
+// SPDX-License-Identifier: GPL-2.0-only
+/* Copyright (c) 2020, Broadcom */
+/*
+ * 8250-core based driver for Broadcom ns16550a UARTs
+ *
+ * This driver uses the standard 8250 driver core but adds additional
+ * optional features including the ability to use a baud rate clock
+ * mux for more accurate high speed baud rate selection and also
+ * an optional DMA engine.
+ *
+ */
+
+#include <linux/module.h>
+#include <linux/types.h>
+#include <linux/tty.h>
+#include <linux/errno.h>
+#include <linux/device.h>
+#include <linux/io.h>
+#include <linux/of.h>
+#include <linux/dma-mapping.h>
+#include <linux/tty_flip.h>
+#include <linux/delay.h>
+#include <linux/clk.h>
+
+#include "8250.h"
+#include "bcm7271_uart.h"
+
+#define REGS_8250 0
+#define REGS_DMA_RX 1
+#define REGS_DMA_TX 2
+#define REGS_DMA_ISR 3
+#define REGS_DMA_ARB 4
+#define REGS_MAX 5
+
+#define TX_BUF_SIZE 4096
+#define RX_BUF_SIZE 4096
+#define RX_BUFS_COUNT 2
+#define KHZ    1000
+#define MHZ(x) ((x) * KHZ * KHZ)
+
+static const u32 brcmstb_rate_table[] = {
+	MHZ(81),
+	MHZ(108),
+	MHZ(64),		/* Actually 64285715 for some chips */
+	MHZ(48),
+};
+
+static const u32 brcmstb_rate_table_7278[] = {
+	MHZ(81),
+	MHZ(108),
+	0,
+	MHZ(48),
+};
+
+static bool disable_dma;
+
+struct brcmuart_priv {
+	int		line;
+	struct clk	*baud_mux_clk;
+	unsigned long	default_mux_rate;
+	u32		real_rates[ARRAY_SIZE(brcmstb_rate_table)];
+	const u32	*rate_table;
+	ktime_t		char_wait;
+	struct uart_port *up;
+	struct hrtimer	hrt;
+	u32		flags;
+#define BRCMUART_PRIV_FLAGS_SHUTDOWN 1
+	bool		dma_enabled;
+	struct uart_8250_dma dma;
+	void __iomem	*regs[REGS_MAX];
+	dma_addr_t	rx_addr;
+	void		*rx_bufs;
+	size_t		rx_size;
+	int		rx_next_buf;
+	dma_addr_t	tx_addr;
+	void		*tx_buf;
+	size_t		tx_size;
+	bool		tx_running;
+	bool		rx_running;
+
+	/* sysfs attributes */
+	u64		dma_rx_partial_buf;
+	u64		dma_rx_full_buf;
+	u32		rx_bad_timeout_late_char;
+	u32		rx_bad_timeout_no_char;
+	u32		rx_missing_close_timeout;
+
+};
+
+/*
+ * Register access routines
+ */
+static u32 udma_readl(struct brcmuart_priv *priv,
+		int reg_type, int offset)
+{
+	return readl(priv->regs[reg_type] + offset);
+}
+
+static void udma_writel(struct brcmuart_priv *priv,
+			int reg_type, int offset, u32 value)
+{
+	writel(value, priv->regs[reg_type] + offset);
+}
+
+static void udma_set(struct brcmuart_priv *priv,
+		int reg_type, int offset, u32 bits)
+{
+	void __iomem *reg = priv->regs[reg_type] + offset;
+	u32 value;
+
+	value = readl(reg);
+	value |= bits;
+	writel(value, reg);
+}
+
+static void udma_unset(struct brcmuart_priv *priv,
+		int reg_type, int offset, u32 bits)
+{
+	void __iomem *reg = priv->regs[reg_type] + offset;
+	u32 value;
+
+	value = readl(reg);
+	value &= ~bits;
+	writel(value, reg);
+}
+
+#define UDMA_DUMP_RX_REG(x)	\
+	dev_dbg(dev, (#x ": 0x%x\n"),			\
+		udma_readl(priv, REGS_DMA_RX, x))
+
+static void __maybe_unused udma_dump_rx_regs(struct device *dev,
+					     struct brcmuart_priv *priv)
+{
+	UDMA_DUMP_RX_REG(UDMA_RX_REVISION);
+	UDMA_DUMP_RX_REG(UDMA_RX_CTRL);
+	UDMA_DUMP_RX_REG(UDMA_RX_STATUS);
+	UDMA_DUMP_RX_REG(UDMA_RX_TRANSFER_LEN);
+	UDMA_DUMP_RX_REG(UDMA_RX_TRANSFER_TOTAL);
+	UDMA_DUMP_RX_REG(UDMA_RX_BUFFER_SIZE);
+	UDMA_DUMP_RX_REG(UDMA_RX_SRC_ADDR);
+	UDMA_DUMP_RX_REG(UDMA_RX_TIMEOUT);
+	UDMA_DUMP_RX_REG(UDMA_RX_BUFFER_CLOSE);
+	UDMA_DUMP_RX_REG(UDMA_RX_BUF0_PTR_LO);
+	UDMA_DUMP_RX_REG(UDMA_RX_BUF0_PTR_HI);
+	UDMA_DUMP_RX_REG(UDMA_RX_BUF0_STATUS);
+	UDMA_DUMP_RX_REG(UDMA_RX_BUF0_DATA_LEN);
+	UDMA_DUMP_RX_REG(UDMA_RX_BUF1_PTR_LO);
+	UDMA_DUMP_RX_REG(UDMA_RX_BUF1_PTR_HI);
+	UDMA_DUMP_RX_REG(UDMA_RX_BUF1_STATUS);
+	UDMA_DUMP_RX_REG(UDMA_RX_BUF1_DATA_LEN);
+}
+
+/*
+ * The UART DMA engine hardware can be used by multiple UARTS, but
+ * only one at a time. Sharing is not currently supported so
+ * the first UART to request the DMA engine will get it and any
+ * subsequent requests by other UARTS will fail.
+ */
+static int brcmuart_arbitration(struct brcmuart_priv *priv, bool acquire)
+{
+	u32 rx_grant;
+	u32 tx_grant;
+	int waits;
+	int ret = 0;
+
+	if (acquire) {
+		udma_set(priv, REGS_DMA_ARB, UDMA_ARB_RX, UDMA_ARB_REQ);
+		udma_set(priv, REGS_DMA_ARB, UDMA_ARB_TX, UDMA_ARB_REQ);
+
+		waits = 1;
+		while (1) {
+			rx_grant = udma_readl(priv, REGS_DMA_ARB, UDMA_ARB_RX);
+			tx_grant = udma_readl(priv, REGS_DMA_ARB, UDMA_ARB_TX);
+			if (rx_grant & tx_grant & UDMA_ARB_GRANT)
+				return 0;
+			if (waits-- == 0)
+				break;
+			msleep(1);
+		}
+		ret = 1;
+	}
+
+	udma_unset(priv, REGS_DMA_ARB, UDMA_ARB_RX, UDMA_ARB_REQ);
+	udma_unset(priv, REGS_DMA_ARB, UDMA_ARB_TX, UDMA_ARB_REQ);
+	return ret;
+}
+
+static void brcmuart_init_dma_hardware(struct brcmuart_priv *priv)
+{
+	u32 daddr;
+	u32 value;
+	int x;
+
+	/* Start with all interrupts disabled */
+	udma_writel(priv, REGS_DMA_ISR, UDMA_INTR_MASK_SET, 0xffffffff);
+
+	udma_writel(priv, REGS_DMA_RX, UDMA_RX_BUFFER_SIZE, RX_BUF_SIZE);
+
+	/*
+	 * Setup buffer close to happen when 32 character times have
+	 * elapsed since the last character was received.
+	 */
+	udma_writel(priv, REGS_DMA_RX, UDMA_RX_BUFFER_CLOSE, 16*10*32);
+	value = (RX_BUFS_COUNT << UDMA_RX_CTRL_NUM_BUF_USED_SHIFT)
+		| UDMA_RX_CTRL_BUF_CLOSE_MODE
+		| UDMA_RX_CTRL_BUF_CLOSE_ENA;
+	udma_writel(priv, REGS_DMA_RX, UDMA_RX_CTRL, value);
+
+	udma_writel(priv, REGS_DMA_RX, UDMA_RX_BLOCKOUT_COUNTER, 0);
+	daddr = priv->rx_addr;
+	for (x = 0; x < RX_BUFS_COUNT; x++) {
+
+		/* Set RX transfer length to 0 for unknown */
+		udma_writel(priv, REGS_DMA_RX, UDMA_RX_TRANSFER_LEN, 0);
+
+		udma_writel(priv, REGS_DMA_RX, UDMA_RX_BUFx_PTR_LO(x),
+			    lower_32_bits(daddr));
+		udma_writel(priv, REGS_DMA_RX, UDMA_RX_BUFx_PTR_HI(x),
+			    upper_32_bits(daddr));
+		daddr += RX_BUF_SIZE;
+	}
+
+	daddr = priv->tx_addr;
+	udma_writel(priv, REGS_DMA_TX, UDMA_TX_BUFx_PTR_LO(0),
+		    lower_32_bits(daddr));
+	udma_writel(priv, REGS_DMA_TX, UDMA_TX_BUFx_PTR_HI(0),
+		    upper_32_bits(daddr));
+	udma_writel(priv, REGS_DMA_TX, UDMA_TX_CTRL,
+		    UDMA_TX_CTRL_NUM_BUF_USED_1);
+
+	/* clear all interrupts then enable them */
+	udma_writel(priv, REGS_DMA_ISR, UDMA_INTR_CLEAR, 0xffffffff);
+	udma_writel(priv, REGS_DMA_ISR, UDMA_INTR_MASK_CLEAR,
+		UDMA_RX_INTERRUPTS | UDMA_TX_INTERRUPTS);
+
+}
+
+static int start_rx_dma(struct uart_8250_port *p)
+{
+	struct brcmuart_priv *priv = p->port.private_data;
+	struct device *dev = p->port.dev;
+	int x;
+
+	dev_dbg(dev, "%s\n", __func__);
+
+	udma_unset(priv, REGS_DMA_RX, UDMA_RX_CTRL, UDMA_RX_CTRL_ENA);
+
+	/* Clear the RX ready bit for all buffers */
+	for (x = 0; x < RX_BUFS_COUNT; x++)
+		udma_unset(priv, REGS_DMA_RX, UDMA_RX_BUFx_STATUS(x),
+			UDMA_RX_BUFX_STATUS_DATA_RDY);
+
+	/* always start with buffer 0 */
+	udma_unset(priv, REGS_DMA_RX, UDMA_RX_STATUS,
+		   UDMA_RX_STATUS_ACTIVE_BUF_MASK);
+	priv->rx_next_buf = 0;
+
+	udma_set(priv, REGS_DMA_RX, UDMA_RX_CTRL, UDMA_RX_CTRL_ENA);
+	priv->rx_running = true;
+	return 0;
+}
+
+static int stop_rx_dma(struct uart_8250_port *p)
+{
+	struct brcmuart_priv *priv = p->port.private_data;
+	struct device *dev = p->port.dev;
+
+	dev_dbg(dev, "%s\n", __func__);
+
+	/* If RX is running, set the RX ABORT */
+	if (priv->rx_running)
+		udma_set(priv, REGS_DMA_RX, UDMA_RX_CTRL, UDMA_RX_CTRL_ABORT);
+	return 0;
+}
+
+static int stop_tx_dma(struct uart_8250_port *p)
+{
+	struct brcmuart_priv *priv = p->port.private_data;
+	struct device *dev = p->port.dev;
+	u32 value;
+
+	dev_dbg(dev, "%s\n", __func__);
+
+	/* If TX is running, set the TX ABORT */
+	value = udma_readl(priv, REGS_DMA_TX, UDMA_TX_CTRL);
+	if (value & UDMA_TX_CTRL_ENA)
+		udma_set(priv, REGS_DMA_TX, UDMA_TX_CTRL, UDMA_TX_CTRL_ABORT);
+	priv->tx_running = false;
+	return 0;
+}
+
+/*
+ * NOTE: printk's in this routine will hang the system if this is
+ * the console tty
+ */
+static int brcmuart_tx_dma(struct uart_8250_port *p)
+{
+	struct brcmuart_priv *priv = p->port.private_data;
+	struct circ_buf *xmit = &p->port.state->xmit;
+	struct device *dev = p->port.dev;
+	u32 tx_size;
+
+	if (uart_tx_stopped(&p->port) || priv->tx_running ||
+		uart_circ_empty(xmit)) {
+		return 0;
+	}
+	tx_size = CIRC_CNT_TO_END(xmit->head, xmit->tail, UART_XMIT_SIZE);
+
+	dev_dbg(dev, "%s: size: %d\n", __func__, tx_size);
+
+	priv->dma.tx_err = 0;
+	memcpy(priv->tx_buf, &xmit->buf[xmit->tail], tx_size);
+	xmit->tail += tx_size;
+	xmit->tail &= UART_XMIT_SIZE - 1;
+	p->port.icount.tx += tx_size;
+
+	if (uart_circ_chars_pending(xmit) < WAKEUP_CHARS)
+		uart_write_wakeup(&p->port);
+
+	udma_writel(priv, REGS_DMA_TX, UDMA_TX_TRANSFER_LEN, tx_size);
+	udma_writel(priv, REGS_DMA_TX, UDMA_TX_BUF0_DATA_LEN, tx_size);
+	udma_unset(priv, REGS_DMA_TX, UDMA_TX_BUF0_STATUS, UDMA_TX_BUFX_EMPTY);
+	udma_set(priv, REGS_DMA_TX, UDMA_TX_CTRL, UDMA_TX_CTRL_ENA);
+	priv->tx_running = true;
+
+	return 0;
+}
+
+static void brcmuart_rx_buf_done_isr(struct uart_port *up, int index)
+{
+	struct brcmuart_priv *priv = up->private_data;
+	struct tty_port *tty_port = &up->state->port;
+	u32 status;
+	u32 length;
+	u32 copied;
+
+	/* Make sure we're still in sync with the hardware */
+	status = udma_readl(priv, REGS_DMA_RX, UDMA_RX_BUFx_STATUS(index));
+	length = udma_readl(priv, REGS_DMA_RX, UDMA_RX_BUFx_DATA_LEN(index));
+
+	if ((status & UDMA_RX_BUFX_STATUS_DATA_RDY) == 0) {
+		dev_err(up->dev, "RX done interrupt but DATA_RDY not found\n");
+		return;
+	}
+	if (status & (UDMA_RX_BUFX_STATUS_OVERRUN_ERR |
+		      UDMA_RX_BUFX_STATUS_FRAME_ERR |
+		      UDMA_RX_BUFX_STATUS_PARITY_ERR)) {
+		if (status & UDMA_RX_BUFX_STATUS_OVERRUN_ERR) {
+			up->icount.overrun++;
+			dev_warn(up->dev, "RX OVERRUN Error\n");
+		}
+		if (status & UDMA_RX_BUFX_STATUS_FRAME_ERR) {
+			up->icount.frame++;
+			dev_warn(up->dev, "RX FRAMING Error\n");
+		}
+		if (status & UDMA_RX_BUFX_STATUS_PARITY_ERR) {
+			up->icount.parity++;
+			dev_warn(up->dev, "RX PARITY Error\n");
+		}
+	}
+	copied = (u32)tty_insert_flip_string(
+		tty_port,
+		priv->rx_bufs + (index * RX_BUF_SIZE),
+		length);
+	if (copied != length) {
+		dev_warn(up->dev, "Flip buffer rverrun of %d bytes\n",
+			 length - copied);
+		up->icount.overrun += length - copied;
+	}
+	up->icount.rx += length;
+	if (status & UDMA_RX_BUFX_STATUS_CLOSE_EXPIRED)
+		priv->dma_rx_partial_buf++;
+	else if (length != RX_BUF_SIZE)
+		/*
+		 * This is a bug in the controller that doesn't cause
+		 * any problems but will be fixed in the future.
+		 */
+		priv->rx_missing_close_timeout++;
+	else
+		priv->dma_rx_full_buf++;
+
+	tty_flip_buffer_push(tty_port);
+
+	dev_dbg(up->dev, "RX: %d chars, chan %d, stat: 0x%x, 0x%x\n",
+		length, index, status,
+		(unsigned int)((unsigned char *)priv->rx_bufs)[index *
+							       RX_BUF_SIZE]);
+}
+
+static void brcmuart_rx_isr(struct uart_port *up, u32 rx_isr)
+{
+	struct brcmuart_priv *priv = up->private_data;
+	struct device *dev = up->dev;
+	u32 rx_done_isr;
+	u32 check_isr;
+	char seq_err[] = "RX buffer ready out of sequence, restarting RX DMA\n";
+
+	rx_done_isr = (rx_isr & UDMA_INTR_RX_READY_MASK);
+	dev_dbg(dev, "Got RX interrupt: 0x%x, next: %d\n", rx_done_isr,
+		priv->rx_next_buf);
+	while (rx_done_isr) {
+		check_isr = UDMA_INTR_RX_READY_BUF0 << priv->rx_next_buf;
+		if (check_isr & rx_done_isr) {
+			brcmuart_rx_buf_done_isr(up, priv->rx_next_buf);
+		} else {
+			dev_err(dev, seq_err);
+			start_rx_dma(up_to_u8250p(up));
+			break;
+		}
+		if (rx_isr & UDMA_RX_ERR_INTERRUPTS) {
+			if (rx_isr & UDMA_INTR_RX_ERROR)
+				dev_dbg(dev, "RX FRAME/OVERRUN/PARITY Error\n");
+			if (rx_isr & UDMA_INTR_RX_TIMEOUT)
+				dev_err(dev, "RX TIMEOUT Error\n");
+			if (rx_isr & UDMA_INTR_RX_ABORT)
+				dev_dbg(dev, "RX ABORT\n");
+			priv->rx_running = false;
+		}
+		/* If not ABORT, re-enable RX buffer */
+		if (!(rx_isr & UDMA_INTR_RX_ABORT))
+			udma_unset(priv, REGS_DMA_RX,
+				   UDMA_RX_BUFx_STATUS(priv->rx_next_buf),
+				   UDMA_RX_BUFX_STATUS_DATA_RDY);
+		dev_dbg(dev, "RX interrupt DONE for channel %d\n",
+			priv->rx_next_buf);
+		rx_done_isr &= ~check_isr;
+		priv->rx_next_buf++;
+		if (priv->rx_next_buf == RX_BUFS_COUNT)
+			priv->rx_next_buf = 0;
+	}
+}
+
+static void brcmuart_tx_isr(struct uart_port *up, u32 isr)
+{
+	struct brcmuart_priv *priv = up->private_data;
+	struct device *dev = up->dev;
+	struct uart_8250_port *port_8250 = up_to_u8250p(up);
+	struct circ_buf	*xmit = &port_8250->port.state->xmit;
+	int length;
+
+	length = udma_readl(priv, REGS_DMA_TX, UDMA_TX_TRANSFER_TOTAL);
+	dev_dbg(dev, "Got TX interrupt, length %d\n", length);
+
+	if (isr & UDMA_INTR_TX_ABORT) {
+		if (priv->tx_running)
+			dev_err(dev, "Unexpected TX_ABORT interrupt\n");
+		return;
+	}
+	priv->tx_running = false;
+	if (!uart_circ_empty(xmit) && !uart_tx_stopped(up))
+		brcmuart_tx_dma(port_8250);
+}
+
+static irqreturn_t brcmuart_isr(int irq, void *dev_id)
+{
+	struct uart_port *up = dev_id;
+	struct device *dev = up->dev;
+	struct brcmuart_priv *priv = up->private_data;
+	unsigned long flags;
+	u32 interrupts;
+	u32 rval;
+	u32 tval;
+
+	interrupts = udma_readl(priv, REGS_DMA_ISR, UDMA_INTR_STATUS);
+	dev_dbg(up->dev, "ISR: interrupts: 0x%x\n", interrupts);
+	if (interrupts == 0)
+		return IRQ_NONE;
+
+	spin_lock_irqsave(&up->lock, flags);
+
+	/* Clear all interrupts */
+	udma_writel(priv, REGS_DMA_ISR, UDMA_INTR_CLEAR, interrupts);
+
+	rval = UDMA_IS_RX_INTERRUPT(interrupts);
+	if (rval)
+		brcmuart_rx_isr(up, rval);
+	tval = UDMA_IS_TX_INTERRUPT(interrupts);
+	if (tval)
+		brcmuart_tx_isr(up, tval);
+	if ((rval | tval) == 0)
+		dev_warn(dev, "Spurious interrupt: 0x%x\n", interrupts);
+
+	spin_unlock_irqrestore(&up->lock, flags);
+	return IRQ_HANDLED;
+}
+
+static int brcmuart_startup(struct uart_port *port)
+{
+	int res;
+	struct uart_8250_port *up = up_to_u8250p(port);
+	struct brcmuart_priv *priv = up->port.private_data;
+	struct device *dev = port->dev;
+
+	dev_dbg(dev, "%s\n", __func__);
+
+	priv->flags &= ~BRCMUART_PRIV_FLAGS_SHUTDOWN;
+
+	/*
+	 * prevent serial8250_do_startup() from allocating non-existent
+	 * DMA resources
+	 */
+	up->dma = NULL;
+
+	res = serial8250_do_startup(port);
+	if (!priv->dma_enabled)
+		return res;
+	/*
+	 * Disable the Receive Data Interrupt because the DMA engine
+	 * will handle this.
+	 */
+	up->ier &= ~UART_IER_RDI;
+	serial_port_out(port, UART_IER, up->ier);
+
+	priv->tx_running = false;
+	priv->dma.rx_dma = NULL;
+	priv->dma.tx_dma = brcmuart_tx_dma;
+	up->dma = &priv->dma;
+
+	brcmuart_init_dma_hardware(priv);
+	start_rx_dma(up);
+	return res;
+}
+
+static void brcmuart_shutdown(struct uart_port *port)
+{
+	struct uart_8250_port *up = up_to_u8250p(port);
+	struct brcmuart_priv *priv = up->port.private_data;
+	struct device *dev = port->dev;
+	unsigned long flags;
+
+	dev_dbg(dev, "%s\n", __func__);
+
+	spin_lock_irqsave(&port->lock, flags);
+	priv->flags |= BRCMUART_PRIV_FLAGS_SHUTDOWN;
+	if (priv->dma_enabled) {
+		stop_rx_dma(up);
+		stop_tx_dma(up);
+		/* disable all interrupts */
+		udma_writel(priv, REGS_DMA_ISR, UDMA_INTR_MASK_SET,
+			UDMA_RX_INTERRUPTS | UDMA_TX_INTERRUPTS);
+	}
+
+	/*
+	 * prevent serial8250_do_shutdown() from trying to free
+	 * DMA resources that we never alloc'd for this driver.
+	 */
+	up->dma = NULL;
+
+	spin_unlock_irqrestore(&port->lock, flags);
+	serial8250_do_shutdown(port);
+}
+
+/*
+ * Not all clocks run at the exact specified rate, so set each requested
+ * rate and then get the actual rate.
+ */
+static void init_real_clk_rates(struct device *dev, struct brcmuart_priv *priv)
+{
+	int x;
+	int rc;
+
+	priv->default_mux_rate = clk_get_rate(priv->baud_mux_clk);
+	dev_dbg(dev, "Default BAUD MUX Clock rate is %lu\n",
+		priv->default_mux_rate);
+
+	for (x = 0; x < ARRAY_SIZE(priv->real_rates); x++) {
+		if (priv->rate_table[x] == 0) {
+			priv->real_rates[x] = 0;
+			continue;
+		}
+		rc = clk_set_rate(priv->baud_mux_clk, priv->rate_table[x]);
+		if (rc) {
+			dev_err(dev, "Error selecting BAUD MUX clock for %u\n",
+				priv->rate_table[x]);
+			priv->real_rates[x] = priv->rate_table[x];
+		} else {
+			priv->real_rates[x] = clk_get_rate(priv->baud_mux_clk);
+		}
+	}
+	 clk_set_rate(priv->baud_mux_clk, priv->default_mux_rate);
+}
+
+static void set_clock_mux(struct uart_port *up, struct brcmuart_priv *priv,
+			u32 baud)
+{
+	u32 percent;
+	u32 best_percent = UINT_MAX;
+	u32 quot;
+	u32 best_quot = 1;
+	u32 rate;
+	int best_index = -1;
+	u64 hires_rate;
+	u64 hires_baud;
+	u64 hires_err;
+	int rc;
+	int i;
+	int real_baud;
+
+	/* If the Baud Mux Clock was not specified, just return */
+	if (priv->baud_mux_clk == NULL)
+		return;
+
+	/* Find the closest match for specified baud */
+	for (i = 0; i < ARRAY_SIZE(priv->real_rates); i++) {
+		if (priv->real_rates[i] == 0)
+			continue;
+		rate = priv->real_rates[i] / 16;
+		quot = DIV_ROUND_CLOSEST(rate, baud);
+		if (!quot)
+			continue;
+
+		/* increase resolution to get xx.xx percent */
+		hires_rate = (u64)rate * 10000;
+		hires_baud = (u64)baud * 10000;
+
+		hires_err = div_u64(hires_rate, (u64)quot);
+
+		/* get the delta */
+		if (hires_err > hires_baud)
+			hires_err = (hires_err - hires_baud);
+		else
+			hires_err = (hires_baud - hires_err);
+
+		percent = (unsigned long)DIV_ROUND_CLOSEST_ULL(hires_err, baud);
+		dev_dbg(up->dev,
+			"Baud rate: %u, MUX Clk: %u, Error: %u.%u%%\n",
+			baud, priv->real_rates[i], percent / 100,
+			percent % 100);
+		if (percent < best_percent) {
+			best_percent = percent;
+			best_index = i;
+			best_quot = quot;
+		}
+	}
+	if (best_index == -1) {
+		dev_err(up->dev, "Error, %d BAUD rate is too fast.\n", baud);
+		return;
+	}
+	rate = priv->real_rates[best_index];
+	rc = clk_set_rate(priv->baud_mux_clk, rate);
+	if (rc)
+		dev_err(up->dev, "Error selecting BAUD MUX clock\n");
+
+	/* Error over 3 percent will cause data errors */
+	if (best_percent > 300)
+		dev_err(up->dev, "Error, baud: %d has %u.%u%% error\n",
+			baud, percent / 100, percent % 100);
+
+	real_baud = rate / 16 / best_quot;
+	dev_dbg(up->dev, "Selecting BAUD MUX rate: %u\n", rate);
+	dev_dbg(up->dev, "Requested baud: %u, Actual baud: %u\n",
+		baud, real_baud);
+
+	/* calc nanoseconds for 1.5 characters time at the given baud rate */
+	i = 1000000000 / real_baud / 10;
+	i += (i / 2);
+	priv->char_wait = ns_to_ktime(i);
+
+	up->uartclk = rate;
+}
+
+static void brcmstb_set_termios(struct uart_port *up,
+				struct ktermios *termios,
+				struct ktermios *old)
+{
+	struct uart_8250_port *p8250 = up_to_u8250p(up);
+	struct brcmuart_priv *priv = up->private_data;
+
+	if (priv->dma_enabled)
+		stop_rx_dma(p8250);
+	set_clock_mux(up, priv, tty_termios_baud_rate(termios));
+	serial8250_do_set_termios(up, termios, old);
+	if (p8250->mcr & UART_MCR_AFE)
+		p8250->port.status |= UPSTAT_AUTOCTS;
+	if (priv->dma_enabled)
+		start_rx_dma(p8250);
+}
+
+static int brcmuart_handle_irq(struct uart_port *p)
+{
+	unsigned int iir = serial_port_in(p, UART_IIR);
+	struct brcmuart_priv *priv = p->private_data;
+	struct uart_8250_port *up = up_to_u8250p(p);
+	unsigned int status;
+	unsigned long flags;
+	unsigned int ier;
+	unsigned int mcr;
+	int handled = 0;
+
+	/*
+	 * There's a bug in some 8250 cores where we get a timeout
+	 * interrupt but there is no data ready.
+	 */
+	if (((iir & UART_IIR_ID) == UART_IIR_RX_TIMEOUT) &&
+	    ((priv->flags & BRCMUART_PRIV_FLAGS_SHUTDOWN) == 0)) {
+		spin_lock_irqsave(&p->lock, flags);
+		status = serial_port_in(p, UART_LSR);
+		if ((status & UART_LSR_DR) == 0) {
+
+			ier = serial_port_in(p, UART_IER);
+			/*
+			 * if Receive Data Interrupt is enabled and
+			 * we're uing hardware flow control, deassert
+			 * RTS and wait for any chars in the pipline to
+			 * arrive and then check for DR again.
+			 */
+			if ((ier & UART_IER_RDI) && (up->mcr & UART_MCR_AFE)) {
+				ier &= ~(UART_IER_RLSI | UART_IER_RDI);
+				serial_port_out(p, UART_IER, ier);
+				mcr = serial_port_in(p, UART_MCR);
+				mcr &= ~UART_MCR_RTS;
+				serial_port_out(p, UART_MCR, mcr);
+				hrtimer_start(&priv->hrt, priv->char_wait,
+					      HRTIMER_MODE_REL);
+			} else {
+				serial_port_in(p, UART_RX);
+			}
+
+			handled = 1;
+		}
+		spin_unlock_irqrestore(&p->lock, flags);
+		if (handled)
+			return 1;
+	}
+	return serial8250_handle_irq(p, iir);
+}
+
+static enum hrtimer_restart brcmuart_hrtimer_func(struct hrtimer *t)
+{
+	struct brcmuart_priv *priv = container_of(t, struct brcmuart_priv, hrt);
+	struct uart_port *p = priv->up;
+	struct uart_8250_port *up = up_to_u8250p(p);
+	unsigned int status;
+	unsigned long flags;
+
+	if (priv->flags & BRCMUART_PRIV_FLAGS_SHUTDOWN)
+		return HRTIMER_NORESTART;
+
+	spin_lock_irqsave(&p->lock, flags);
+	status = serial_port_in(p, UART_LSR);
+
+	/*
+	 * If a character did not arrive after the timeout, clear the false
+	 * receive timeout.
+	 */
+	if ((status & UART_LSR_DR) == 0) {
+		serial_port_in(p, UART_RX);
+		priv->rx_bad_timeout_no_char++;
+	} else {
+		priv->rx_bad_timeout_late_char++;
+	}
+
+	/* re-enable receive unless upper layer has disabled it */
+	if ((up->ier & (UART_IER_RLSI | UART_IER_RDI)) ==
+	    (UART_IER_RLSI | UART_IER_RDI)) {
+		status = serial_port_in(p, UART_IER);
+		status |= (UART_IER_RLSI | UART_IER_RDI);
+		serial_port_out(p, UART_IER, status);
+		status = serial_port_in(p, UART_MCR);
+		status |= UART_MCR_RTS;
+		serial_port_out(p, UART_MCR, status);
+	}
+	spin_unlock_irqrestore(&p->lock, flags);
+	return HRTIMER_NORESTART;
+}
+
+static ssize_t rx_bad_timeout_no_char_show(struct device *dev,
+					   struct device_attribute *attr,
+					   char *buf)
+{
+	struct brcmuart_priv *priv = dev_get_drvdata(dev);
+
+	return sprintf(buf, "%d\n", priv->rx_bad_timeout_no_char);
+}
+static DEVICE_ATTR_RO(rx_bad_timeout_no_char);
+
+static ssize_t rx_bad_timeout_late_char_show(struct device *dev,
+					     struct device_attribute *attr,
+					     char *buf)
+{
+	struct brcmuart_priv *priv = dev_get_drvdata(dev);
+
+	return sprintf(buf, "%d\n", priv->rx_bad_timeout_late_char);
+}
+static DEVICE_ATTR_RO(rx_bad_timeout_late_char);
+
+static ssize_t rx_missing_close_timeout_show(struct device *dev,
+					     struct device_attribute *attr,
+					     char *buf)
+{
+	struct brcmuart_priv *priv = dev_get_drvdata(dev);
+
+	return sprintf(buf, "%d\n", priv->rx_missing_close_timeout);
+}
+static DEVICE_ATTR_RO(rx_missing_close_timeout);
+
+static ssize_t dma_rx_partial_buf_show(struct device *dev,
+				       struct device_attribute *attr,
+				       char *buf)
+{
+	struct brcmuart_priv *priv = dev_get_drvdata(dev);
+
+	return sprintf(buf, "%lld\n", priv->dma_rx_partial_buf);
+}
+static DEVICE_ATTR_RO(dma_rx_partial_buf);
+
+static ssize_t dma_rx_full_buf_show(struct device *dev,
+				    struct device_attribute *attr,
+				    char *buf)
+{
+	struct brcmuart_priv *priv = dev_get_drvdata(dev);
+
+	return sprintf(buf, "%lld\n", priv->dma_rx_full_buf);
+}
+static DEVICE_ATTR_RO(dma_rx_full_buf);
+
+static struct attribute *brcmuart_dev_attrs[] = {
+	&dev_attr_rx_bad_timeout_no_char.attr,
+	&dev_attr_rx_bad_timeout_late_char.attr,
+	&dev_attr_rx_missing_close_timeout.attr,
+	&dev_attr_dma_rx_partial_buf.attr,
+	&dev_attr_dma_rx_full_buf.attr,
+	NULL,
+};
+
+static const struct attribute_group brcmuart_dev_attr_group = {
+	.attrs = brcmuart_dev_attrs,
+};
+
+static const struct of_device_id brcmuart_dt_ids[] = {
+	{
+		.compatible = "brcm,bcm7278-uart",
+		.data = brcmstb_rate_table_7278,
+	},
+	{
+		.compatible = "brcm,bcm7271-uart",
+		.data = brcmstb_rate_table,
+	},
+	{},
+};
+
+MODULE_DEVICE_TABLE(of, brcmuart_dt_ids);
+
+static void brcmuart_free_bufs(struct device *dev, struct brcmuart_priv *priv)
+{
+	if (priv->rx_bufs)
+		dma_free_coherent(dev, priv->rx_size, priv->rx_bufs,
+				  priv->rx_addr);
+	if (priv->tx_buf)
+		dma_free_coherent(dev, priv->tx_size, priv->tx_buf,
+				  priv->tx_addr);
+}
+
+static void brcmuart_throttle(struct uart_port *port)
+{
+	struct brcmuart_priv *priv = port->private_data;
+
+	dev_dbg(port->dev, "%s\n", __func__);
+
+	udma_writel(priv, REGS_DMA_ISR, UDMA_INTR_MASK_SET, UDMA_RX_INTERRUPTS);
+}
+
+static void brcmuart_unthrottle(struct uart_port *port)
+{
+	struct brcmuart_priv *priv = port->private_data;
+
+	dev_dbg(port->dev, "%s\n", __func__);
+
+	udma_writel(priv, REGS_DMA_ISR, UDMA_INTR_MASK_CLEAR,
+		    UDMA_RX_INTERRUPTS);
+}
+
+module_param(disable_dma, bool, 0644);
+MODULE_PARM_DESC(disable_dma, "Disable UART DMA capability");
+
+static int brcmuart_probe(struct platform_device *pdev)
+{
+	struct resource *regs;
+	struct device_node *np = pdev->dev.of_node;
+	const struct of_device_id *of_id = NULL;
+	struct uart_8250_port *new_port;
+	struct device *dev = &pdev->dev;
+	struct brcmuart_priv *priv;
+	struct clk *baud_mux_clk;
+	struct uart_8250_port up;
+	struct resource *irq;
+	void __iomem *membase = 0;
+	resource_size_t mapbase = 0;
+	u32 clk_rate = 0;
+	int ret;
+	int x;
+	int dma_irq;
+	static const char * const reg_names[REGS_MAX] = {
+		"uart", "dma_rx", "dma_tx", "dma_intr2", "dma_arb"
+	};
+
+	irq = platform_get_resource(pdev, IORESOURCE_IRQ, 0);
+	if (!irq) {
+		dev_err(dev, "missing irq\n");
+		return -EINVAL;
+	}
+	priv = devm_kzalloc(dev, sizeof(struct brcmuart_priv),
+			GFP_KERNEL);
+	if (!priv)
+		return -ENOMEM;
+
+	of_id = of_match_node(brcmuart_dt_ids, np);
+	if (!of_id || !of_id->data)
+		priv->rate_table = brcmstb_rate_table;
+	else
+		priv->rate_table = of_id->data;
+
+	for (x = 0; x < REGS_MAX; x++) {
+		regs = platform_get_resource_byname(pdev, IORESOURCE_MEM,
+						reg_names[x]);
+		if (!regs)
+			break;
+		priv->regs[x] =	devm_ioremap(dev, regs->start,
+					     resource_size(regs));
+		if (IS_ERR(priv->regs[x]))
+			return PTR_ERR(priv->regs[x]);
+		if (x == REGS_8250) {
+			mapbase = regs->start;
+			membase = priv->regs[x];
+		}
+	}
+
+	/* We should have just the uart base registers or all the registers */
+	if (x != 1 && x != REGS_MAX) {
+		dev_warn(dev, "%s registers not specified\n", reg_names[x]);
+		return -EINVAL;
+	}
+
+	/* if the DMA registers were specified, try to enable DMA */
+	if ((x > REGS_DMA_RX) && !disable_dma) {
+		if (brcmuart_arbitration(priv, 1) == 0) {
+			u32 txrev = 0;
+			u32 rxrev = 0;
+
+			txrev = udma_readl(priv, REGS_DMA_RX, UDMA_RX_REVISION);
+			rxrev = udma_readl(priv, REGS_DMA_TX, UDMA_TX_REVISION);
+			if ((txrev >= UDMA_TX_REVISION_REQUIRED) &&
+				(rxrev >= UDMA_RX_REVISION_REQUIRED)) {
+
+				/* Enable the use of the DMA hardware */
+				priv->dma_enabled = true;
+			} else {
+				brcmuart_arbitration(priv, 0);
+				dev_err(dev,
+					"Unsupported DMA Hardware Revision\n");
+			}
+		} else {
+			dev_err(dev,
+				"Timeout arbitrating for UART DMA hardware\n");
+		}
+	}
+
+	of_property_read_u32(np, "clock-frequency", &clk_rate);
+
+	/* See if a Baud clock has been specified */
+	baud_mux_clk = of_clk_get_by_name(np, "sw_baud");
+	if (IS_ERR(baud_mux_clk)) {
+		if (PTR_ERR(baud_mux_clk) == -EPROBE_DEFER)
+			return -EPROBE_DEFER;
+		dev_info(dev, "BAUD MUX clock not specified\n");
+	} else {
+		dev_info(dev, "BAUD MUX clock found\n");
+		ret = clk_prepare_enable(baud_mux_clk);
+		if (ret)
+			return ret;
+		priv->baud_mux_clk = baud_mux_clk;
+		init_real_clk_rates(dev, priv);
+		clk_rate = priv->default_mux_rate;
+	}
+
+	if (clk_rate == 0) {
+		dev_err(dev, "clock-frequency or clk not defined\n");
+		return -EINVAL;
+	}
+
+	dev_info(dev, "DMA is %senabled\n", priv->dma_enabled ? "" : "not ");
+
+	memset(&up, 0, sizeof(up));
+	up.port.type = PORT_16550A;
+	up.port.uartclk = clk_rate;
+	up.port.dev = dev;
+	up.port.mapbase = mapbase;
+	up.port.membase = membase;
+	up.port.irq = irq->start;
+	up.port.handle_irq = brcmuart_handle_irq;
+	up.port.regshift = 2;
+	up.port.iotype = of_device_is_big_endian(np) ?
+		UPIO_MEM32BE : UPIO_MEM32;
+	up.port.flags = UPF_SHARE_IRQ | UPF_BOOT_AUTOCONF
+		| UPF_FIXED_PORT | UPF_FIXED_TYPE;
+	up.port.dev = dev;
+	up.port.private_data = priv;
+	up.capabilities = UART_CAP_FIFO | UART_CAP_AFE;
+	up.port.fifosize = 32;
+
+	/* Check for a fixed line number */
+	ret = of_alias_get_id(np, "serial");
+	if (ret >= 0)
+		up.port.line = ret;
+
+	/* setup HR timer */
+	hrtimer_init(&priv->hrt, CLOCK_MONOTONIC, HRTIMER_MODE_ABS);
+	priv->hrt.function = brcmuart_hrtimer_func;
+
+	up.port.shutdown = brcmuart_shutdown;
+	up.port.startup = brcmuart_startup;
+	up.port.throttle = brcmuart_throttle;
+	up.port.unthrottle = brcmuart_unthrottle;
+	up.port.set_termios = brcmstb_set_termios;
+
+	if (priv->dma_enabled) {
+		priv->rx_size = RX_BUF_SIZE * RX_BUFS_COUNT;
+		priv->rx_bufs = dma_alloc_coherent(dev,
+						   priv->rx_size,
+						   &priv->rx_addr, GFP_KERNEL);
+		if (!priv->rx_bufs)
+			goto err;
+		priv->tx_size = UART_XMIT_SIZE;
+		priv->tx_buf = dma_alloc_coherent(dev,
+						  priv->tx_size,
+						  &priv->tx_addr, GFP_KERNEL);
+		if (!priv->tx_buf)
+			goto err;
+	}
+
+	ret = serial8250_register_8250_port(&up);
+	if (ret < 0) {
+		dev_err(dev, "unable to register 8250 port\n");
+		goto err;
+	}
+	priv->line = ret;
+	new_port = serial8250_get_port(ret);
+	priv->up = &new_port->port;
+	if (priv->dma_enabled) {
+		dma_irq = platform_get_irq_byname(pdev,  "dma");
+		if (dma_irq < 0) {
+			dev_err(dev, "no IRQ resource info\n");
+			goto err1;
+		}
+		ret = devm_request_irq(dev, dma_irq, brcmuart_isr,
+				IRQF_SHARED, "uart DMA irq", &new_port->port);
+		if (ret) {
+			dev_err(dev, "unable to register IRQ handler\n");
+			goto err1;
+		}
+	}
+	platform_set_drvdata(pdev, priv);
+	ret = sysfs_create_group(&dev->kobj, &brcmuart_dev_attr_group);
+	if (ret)
+		dev_warn(dev, "Error creating sysfs attributes\n");
+
+	return 0;
+
+err1:
+	serial8250_unregister_port(priv->line);
+err:
+	brcmuart_free_bufs(dev, priv);
+	brcmuart_arbitration(priv, 0);
+	return -ENODEV;
+}
+
+static int brcmuart_remove(struct platform_device *pdev)
+{
+	struct brcmuart_priv *priv = platform_get_drvdata(pdev);
+
+	sysfs_remove_group(&pdev->dev.kobj, &brcmuart_dev_attr_group);
+	hrtimer_cancel(&priv->hrt);
+	serial8250_unregister_port(priv->line);
+	brcmuart_free_bufs(&pdev->dev, priv);
+	brcmuart_arbitration(priv, 0);
+	return 0;
+}
+
+static int __maybe_unused brcmuart_suspend(struct device *dev)
+{
+	struct brcmuart_priv *priv = dev_get_drvdata(dev);
+
+	serial8250_suspend_port(priv->line);
+	clk_disable_unprepare(priv->baud_mux_clk);
+
+	return 0;
+}
+
+static int __maybe_unused brcmuart_resume(struct device *dev)
+{
+	struct brcmuart_priv *priv = dev_get_drvdata(dev);
+	int ret;
+
+	ret = clk_prepare_enable(priv->baud_mux_clk);
+	if (ret)
+		dev_err(dev, "Error enabling BAUD MUX clock\n");
+
+	/*
+	 * The hardware goes back to it's default after suspend
+	 * so get the "clk" back in sync.
+	 */
+	ret = clk_set_rate(priv->baud_mux_clk, priv->default_mux_rate);
+	if (ret)
+		dev_err(dev, "Error restoring default BAUD MUX clock\n");
+	if (priv->dma_enabled) {
+		brcmuart_arbitration(priv, 1);
+		brcmuart_init_dma_hardware(priv);
+		start_rx_dma(serial8250_get_port(priv->line));
+	}
+	serial8250_resume_port(priv->line);
+	return 0;
+}
+
+static const struct dev_pm_ops brcmuart_dev_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(brcmuart_suspend, brcmuart_resume)
+};
+
+static struct platform_driver brcmuart_platform_driver = {
+	.driver = {
+		.name	= "bcm7271-uart",
+		.pm		= &brcmuart_dev_pm_ops,
+		.of_match_table = brcmuart_dt_ids,
+	},
+	.probe		= brcmuart_probe,
+	.remove		= brcmuart_remove,
+};
+module_platform_driver(brcmuart_platform_driver);
+
+MODULE_AUTHOR("Al Cooper");
+MODULE_DESCRIPTION("Broadcom NS16550A compatible serial port driver");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/tty/serial/8250/Kconfig b/drivers/tty/serial/8250/Kconfig
index 603137da4736..436a686bf7fd 100644
--- a/drivers/tty/serial/8250/Kconfig
+++ b/drivers/tty/serial/8250/Kconfig
@@ -501,6 +501,7 @@  config SERIAL_8250_PXA
 	  applicable to both devicetree and legacy boards, and early console is
 	  part of its support.
 
+
 config SERIAL_8250_TEGRA
 	tristate "8250 support for Tegra serial ports"
 	default SERIAL_8250
@@ -510,6 +511,16 @@  config SERIAL_8250_TEGRA
 	  Select this option if you have machine with an NVIDIA Tegra SoC and
 	  wish to enable 8250 serial driver for the Tegra serial interfaces.
 
+config SERIAL_8250_BCM7271
+	bool "Broadcom 8250 based serial port"
+	depends on SERIAL_8250 && (ARCH_BRCMSTB || COMPILE_TEST)
+	default ARCH_BRCMSTB
+	help
+	  If you have a Broadcom STB based board and want to use the
+	  enhanced features of the Broadcom 8250 based serial port,
+	  including DMA support and high accuracy BAUD rates, say
+	  Y to this option. If unsure, say N.
+
 config SERIAL_OF_PLATFORM
 	tristate "Devicetree based probing for 8250 ports"
 	depends on SERIAL_8250 && OF
diff --git a/drivers/tty/serial/8250/Makefile b/drivers/tty/serial/8250/Makefile
index a8bfb654d490..b9bcd73c8997 100644
--- a/drivers/tty/serial/8250/Makefile
+++ b/drivers/tty/serial/8250/Makefile
@@ -38,6 +38,7 @@  obj-$(CONFIG_SERIAL_8250_LPSS)		+= 8250_lpss.o
 obj-$(CONFIG_SERIAL_8250_MID)		+= 8250_mid.o
 obj-$(CONFIG_SERIAL_8250_PXA)		+= 8250_pxa.o
 obj-$(CONFIG_SERIAL_8250_TEGRA)		+= 8250_tegra.o
+obj-$(CONFIG_SERIAL_8250_BCM7271)	+= 8250_bcm7271.o
 obj-$(CONFIG_SERIAL_OF_PLATFORM)	+= 8250_of.o
 
 CFLAGS_8250_ingenic.o += -I$(srctree)/scripts/dtc/libfdt
diff --git a/drivers/tty/serial/8250/bcm7271_uart.h b/drivers/tty/serial/8250/bcm7271_uart.h
new file mode 100644
index 000000000000..78cce03f9e17
--- /dev/null
+++ b/drivers/tty/serial/8250/bcm7271_uart.h
@@ -0,0 +1,158 @@ 
+/* SPDX-License-Identifier: GPL-2.0-only */
+/* Copyright (c) 2020, Broadcom */
+#ifndef __BCM7271_UART_H__
+#define __BCM7271_UART_H__
+
+/* Register definitions for UART DMA block. Version 1.1 or later. */
+#define UDMA_ARB_RX		0x00
+#define UDMA_ARB_TX		0x04
+#define		UDMA_ARB_REQ				0x00000001
+#define		UDMA_ARB_GRANT				0x00000002
+
+#define UDMA_RX_REVISION	0x00
+#define UDMA_RX_REVISION_REQUIRED			0x00000101
+#define UDMA_RX_CTRL		0x04
+#define		UDMA_RX_CTRL_BUF_CLOSE_MODE		0x00010000
+#define		UDMA_RX_CTRL_MASK_WR_DONE		0x00008000
+#define		UDMA_RX_CTRL_ENDIAN_OVERRIDE		0x00004000
+#define		UDMA_RX_CTRL_ENDIAN			0x00002000
+#define		UDMA_RX_CTRL_OE_IS_ERR			0x00001000
+#define		UDMA_RX_CTRL_PE_IS_ERR			0x00000800
+#define		UDMA_RX_CTRL_FE_IS_ERR			0x00000400
+#define		UDMA_RX_CTRL_NUM_BUF_USED_MASK		0x000003c0
+#define		UDMA_RX_CTRL_NUM_BUF_USED_SHIFT	6
+#define		UDMA_RX_CTRL_BUF_CLOSE_CLK_SEL_SYS	0x00000020
+#define		UDMA_RX_CTRL_BUF_CLOSE_ENA		0x00000010
+#define		UDMA_RX_CTRL_TIMEOUT_CLK_SEL_SYS	0x00000008
+#define		UDMA_RX_CTRL_TIMEOUT_ENA		0x00000004
+#define		UDMA_RX_CTRL_ABORT			0x00000002
+#define		UDMA_RX_CTRL_ENA			0x00000001
+#define UDMA_RX_STATUS		0x08
+#define		UDMA_RX_STATUS_ACTIVE_BUF_MASK		0x0000000f
+#define UDMA_RX_TRANSFER_LEN	0x0c
+#define UDMA_RX_TRANSFER_TOTAL	0x10
+#define UDMA_RX_BUFFER_SIZE	0x14
+#define UDMA_RX_SRC_ADDR	0x18
+#define UDMA_RX_TIMEOUT		0x1c
+#define UDMA_RX_BUFFER_CLOSE	0x20
+#define UDMA_RX_BLOCKOUT_COUNTER 0x24
+#define UDMA_RX_BUF0_PTR_LO	0x28
+#define UDMA_RX_BUF0_PTR_HI	0x2c
+#define UDMA_RX_BUF0_STATUS	0x30
+#define		UDMA_RX_BUFX_STATUS_OVERRUN_ERR		0x00000010
+#define		UDMA_RX_BUFX_STATUS_FRAME_ERR		0x00000008
+#define		UDMA_RX_BUFX_STATUS_PARITY_ERR		0x00000004
+#define		UDMA_RX_BUFX_STATUS_CLOSE_EXPIRED	0x00000002
+#define		UDMA_RX_BUFX_STATUS_DATA_RDY		0x00000001
+#define UDMA_RX_BUF0_DATA_LEN	0x34
+#define UDMA_RX_BUF1_PTR_LO	0x38
+#define UDMA_RX_BUF1_PTR_HI	0x3c
+#define UDMA_RX_BUF1_STATUS	0x40
+#define UDMA_RX_BUF1_DATA_LEN	0x44
+
+#define UDMA_TX_REVISION	0x00
+#define UDMA_TX_REVISION_REQUIRED			0x00000101
+#define UDMA_TX_CTRL		0x04
+#define		UDMA_TX_CTRL_ENDIAN_OVERRIDE		0x00000080
+#define		UDMA_TX_CTRL_ENDIAN			0x00000040
+#define		UDMA_TX_CTRL_NUM_BUF_USED_MASK		0x00000030
+#define		UDMA_TX_CTRL_NUM_BUF_USED_1		0x00000010
+#define		UDMA_TX_CTRL_ABORT			0x00000002
+#define		UDMA_TX_CTRL_ENA			0x00000001
+#define UDMA_TX_DST_ADDR	0x08
+#define UDMA_TX_BLOCKOUT_COUNTER 0x10
+#define UDMA_TX_TRANSFER_LEN	0x14
+#define UDMA_TX_TRANSFER_TOTAL	0x18
+#define UDMA_TX_STATUS		0x20
+#define UDMA_TX_BUF0_PTR_LO	0x24
+#define UDMA_TX_BUF0_PTR_HI	0x28
+#define UDMA_TX_BUF0_STATUS	0x2c
+#define		UDMA_TX_BUFX_LAST			0x00000002
+#define		UDMA_TX_BUFX_EMPTY			0x00000001
+#define UDMA_TX_BUF0_DATA_LEN	0x30
+#define UDMA_TX_BUF0_DATA_SENT	0x34
+#define UDMA_TX_BUF1_PTR_LO	0x38
+
+#define UDMA_INTR_STATUS	0x00
+#define		UDMA_INTR_ARB_TX_GRANT			0x00040000
+#define		UDMA_INTR_ARB_RX_GRANT			0x00020000
+#define		UDMA_INTR_TX_ALL_EMPTY			0x00010000
+#define		UDMA_INTR_TX_EMPTY_BUF1			0x00008000
+#define		UDMA_INTR_TX_EMPTY_BUF0			0x00004000
+#define		UDMA_INTR_TX_ABORT			0x00002000
+#define		UDMA_INTR_TX_DONE			0x00001000
+#define		UDMA_INTR_RX_ERROR			0x00000800
+#define		UDMA_INTR_RX_TIMEOUT			0x00000400
+#define		UDMA_INTR_RX_READY_BUF7			0x00000200
+#define		UDMA_INTR_RX_READY_BUF6			0x00000100
+#define		UDMA_INTR_RX_READY_BUF5			0x00000080
+#define		UDMA_INTR_RX_READY_BUF4			0x00000040
+#define		UDMA_INTR_RX_READY_BUF3			0x00000020
+#define		UDMA_INTR_RX_READY_BUF2			0x00000010
+#define		UDMA_INTR_RX_READY_BUF1			0x00000008
+#define		UDMA_INTR_RX_READY_BUF0			0x00000004
+#define		UDMA_INTR_RX_READY_MASK			0x000003fc
+#define		UDMA_INTR_RX_READY_SHIFT		2
+#define		UDMA_INTR_RX_ABORT			0x00000002
+#define		UDMA_INTR_RX_DONE			0x00000001
+#define UDMA_INTR_SET		0x04
+#define UDMA_INTR_CLEAR		0x08
+#define UDMA_INTR_MASK_STATUS	0x0c
+#define UDMA_INTR_MASK_SET	0x10
+#define UDMA_INTR_MASK_CLEAR	0x14
+
+
+#define UDMA_RX_INTERRUPTS ( \
+	UDMA_INTR_RX_ERROR | \
+	UDMA_INTR_RX_TIMEOUT | \
+	UDMA_INTR_RX_READY_BUF0 | \
+	UDMA_INTR_RX_READY_BUF1 | \
+	UDMA_INTR_RX_READY_BUF2 | \
+	UDMA_INTR_RX_READY_BUF3 | \
+	UDMA_INTR_RX_READY_BUF4 | \
+	UDMA_INTR_RX_READY_BUF5 | \
+	UDMA_INTR_RX_READY_BUF6 | \
+	UDMA_INTR_RX_READY_BUF7 | \
+	UDMA_INTR_RX_ABORT | \
+	UDMA_INTR_RX_DONE)
+
+#define UDMA_RX_ERR_INTERRUPTS ( \
+	UDMA_INTR_RX_ERROR | \
+	UDMA_INTR_RX_TIMEOUT | \
+	UDMA_INTR_RX_ABORT | \
+	UDMA_INTR_RX_DONE)
+
+#define UDMA_TX_INTERRUPTS ( \
+	UDMA_INTR_TX_ABORT | \
+	UDMA_INTR_TX_DONE)
+
+#define UDMA_IS_RX_INTERRUPT(status) ((status) & UDMA_RX_INTERRUPTS)
+#define UDMA_IS_TX_INTERRUPT(status) ((status) & UDMA_TX_INTERRUPTS)
+
+
+/* Current devices have 8 sets of RX buffer registers */
+#define UDMA_RX_BUFS_COUNT	8
+#define UDMA_RX_BUFS_REG_OFFSET (UDMA_RX_BUF1_PTR_LO - UDMA_RX_BUF0_PTR_LO)
+#define UDMA_RX_BUFx_PTR_LO(x)	(UDMA_RX_BUF0_PTR_LO + \
+				 ((x) * UDMA_RX_BUFS_REG_OFFSET))
+#define UDMA_RX_BUFx_PTR_HI(x)	(UDMA_RX_BUF0_PTR_HI + \
+				 ((x) * UDMA_RX_BUFS_REG_OFFSET))
+#define UDMA_RX_BUFx_STATUS(x)	(UDMA_RX_BUF0_STATUS + \
+				 ((x) * UDMA_RX_BUFS_REG_OFFSET))
+#define UDMA_RX_BUFx_DATA_LEN(x) (UDMA_RX_BUF0_DATA_LEN + \
+				  ((x) * UDMA_RX_BUFS_REG_OFFSET))
+
+/* Current devices have 2 sets of TX buffer registers */
+#define UDMA_TX_BUFS_COUNT	2
+#define UDMA_TX_BUFS_REG_OFFSET (UDMA_TX_BUF1_PTR_LO - UDMA_TX_BUF0_PTR_LO)
+#define UDMA_TX_BUFx_PTR_LO(x)	(UDMA_TX_BUF0_PTR_LO + \
+				 ((x) * UDMA_TX_BUFS_REG_OFFSET))
+#define UDMA_TX_BUFx_PTR_HI(x)	(UDMA_TX_BUF0_PTR_HI + \
+				 ((x) * UDMA_TX_BUFS_REG_OFFSET))
+#define UDMA_TX_BUFx_STATUS(x)	(UDMA_TX_BUF0_STATUS + \
+				 ((x) * UDMA_TX_BUFS_REG_OFFSET))
+#define UDMA_TX_BUFx_DATA_LEN(x) (UDMA_TX_BUF0_DATA_LEN + \
+				  ((x) * UDMA_TX_BUFS_REG_OFFSET))
+#define UDMA_TX_BUFx_DATA_SENT(x) (UDMA_TX_BUF0_DATA_SENT + \
+				   ((x) * UDMA_TX_BUFS_REG_OFFSET))
+#endif /* __BCM7271_UART_H__ */