public inbox for devel@edk2.groups.io
 help / color / mirror / Atom feed
From: Leif Lindholm <leif.lindholm@linaro.org>
To: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Marcin Wojtas <mw@semihalf.com>,
	edk2-devel-01 <edk2-devel@lists.01.org>,
	"Tian, Feng" <feng.tian@intel.com>,
	"Kinney, Michael D" <michael.d.kinney@intel.com>,
	"Gao, Liming" <liming.gao@intel.com>,
	Nadav Haklai <nadavh@marvell.com>,
	"jsd@semihalf.com" <jsd@semihalf.com>
Subject: Re: [PATCH] ArmPkg/GenericWatchdogDxe: Split 64bit register write to 2x32bit
Date: Thu, 2 Aug 2018 21:14:19 +0100	[thread overview]
Message-ID: <20180802201419.ewh3j6x2kqdq5ccn@bivouac.eciton.net> (raw)
In-Reply-To: <CAKv+Gu8cRwizi76n6HCw00a-JnQmVWnb+cP4wpVLnmqd5cE3FA@mail.gmail.com>

On Thu, Aug 02, 2018 at 08:32:18PM +0200, Ard Biesheuvel wrote:
> On 2 August 2018 at 17:12, Marcin Wojtas <mw@semihalf.com> wrote:
> > czw., 2 sie 2018 o 16:59 Ard Biesheuvel <ard.biesheuvel@linaro.org> napisał(a):
> >>
> >> On 2 August 2018 at 16:49, Marcin Wojtas <mw@semihalf.com> wrote:
> >> > EDK2 code uses a single 64bit write to update SBSA watchdog
> >> > compare registers, however an access to mmio registers should
> >> > be 32bit for some SoCs. Current usage of MmioWrite64 resulted
> >> > in an unpredicted behavior. Fix this by modifying
> >> > WatchdogWriteCompareRegister routine to use two consecutive
> >> > 32bit writes to Watchdog Compare Register Low and High.
> >> >
> >>
> >> You describe it as if this is generally the case, but this is just a
> >> silicon bug, right?
> >
> > Not sure if it's a bug, or simply SoC characterisctics to place SoC
> > registers to allow only mmio32 access to 32-bit registers. In any way,
> > even updated routine should be fine also for the ones capable of
> > mmio64 registers access. Do you have strong objections to the change?
> >
> 
> To be fair, the SBSA v5.0 does describe this register as
> 
> 0x010-0x013  WCV[31:0]   Watchdog compare value. Read/write registers
> 0x014-0x017  WCV[63:32]  containing the current value in the watchdog
>                          compare register.
> 
> so I guess it is implied that any implementation should tolerate this
> value being written as 2 32-bit quantities.
> 
> Leif?

Yep, the way the SBSA describes it, these are two separate 32-bit
registers, making accessing them with a single transaction ends up
being implementation defined behaviour.

We should maybe add separate _HIGH and _LOW #defines for the two rather
than calculating register addresses inline?

But apart from that:
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>

> >
> >>
> >> > Contributed-under: TianoCore Contribution Agreement 1.1
> >> > Signed-off-by: Marcin Wojtas <mw@semihalf.com>
> >> > ---
> >> >  ArmPkg/Drivers/GenericWatchdogDxe/GenericWatchdogDxe.c | 6 +++++-
> >> >  1 file changed, 5 insertions(+), 1 deletion(-)
> >> >
> >> > diff --git a/ArmPkg/Drivers/GenericWatchdogDxe/GenericWatchdogDxe.c b/ArmPkg/Drivers/GenericWatchdogDxe/GenericWatchdogDxe.c
> >> > index 3180f01..b25d210 100644
> >> > --- a/ArmPkg/Drivers/GenericWatchdogDxe/GenericWatchdogDxe.c
> >> > +++ b/ArmPkg/Drivers/GenericWatchdogDxe/GenericWatchdogDxe.c
> >> > @@ -56,7 +56,11 @@ WatchdogWriteCompareRegister (
> >> >    UINT64  Value
> >> >    )
> >> >  {
> >> > -  MmioWrite64 (GENERIC_WDOG_COMPARE_VALUE_REG, Value);
> >> > +  MmioWrite32 (GENERIC_WDOG_COMPARE_VALUE_REG, Value & MAX_UINT32);
> >> > +  MmioWrite32 (
> >> > +    GENERIC_WDOG_COMPARE_VALUE_REG + sizeof (UINT32),
> >> > +    (Value >> 32) & MAX_UINT32
> >> > +    );
> >> >  }
> >> >
> >> >  VOID
> >> > --
> >> > 2.7.4
> >> >


      reply	other threads:[~2018-08-02 20:14 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-08-02 14:49 [PATCH] ArmPkg/GenericWatchdogDxe: Split 64bit register write to 2x32bit Marcin Wojtas
2018-08-02 14:59 ` Ard Biesheuvel
2018-08-02 15:12   ` Marcin Wojtas
2018-08-02 18:32     ` Ard Biesheuvel
2018-08-02 20:14       ` Leif Lindholm [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-list from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20180802201419.ewh3j6x2kqdq5ccn@bivouac.eciton.net \
    --to=devel@edk2.groups.io \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox