From: "Jeremy Linton" <jeremy.linton@arm.com>
To: Ard Biesheuvel <ardb@kernel.org>
Cc: edk2-devel-groups-io <devel@edk2.groups.io>,
Peter Batard <pete@akeo.ie>,
Ard Biesheuvel <ardb+tianocore@kernel.org>,
Leif Lindholm <leif@nuviainc.com>,
Andrei Warkentin <awarkentin@vmware.com>,
Sunny Wang <Sunny.Wang@arm.com>,
Samer El-Haj-Mahmoud <samer.el-haj-mahmoud@arm.com>
Subject: Re: [PATCH 5/5] Platform/RaspberryPi: Disconnect/shutdown all drivers before reboot
Date: Tue, 5 Oct 2021 16:24:35 -0500 [thread overview]
Message-ID: <452e1938-3d6d-97f0-9922-35a7d573c621@arm.com> (raw)
In-Reply-To: <CAMj1kXHQ0UStu5pr6JBTiS1B0qWP9kRHbRRpYwQeH1V=QChvpg@mail.gmail.com>
Hi,
On 10/5/21 5:11 AM, Ard Biesheuvel wrote:
> On Sat, 2 Oct 2021 at 02:52, Jeremy Linton <jeremy.linton@arm.com> wrote:
>>
>> In theory we should be properly cleaning up all the device drivers before
>> pulling the big switch. Particularly the partition mgr will issue
>> flush commands to attached disks as it goes down. This assures that
>> devices running in WB mode (that correctly handle flush/sync/etc) commands
>> are persisted to physical media before we hit reset.
>>
>> Without this, there are definitly cases where the relevant specifications
>> don't guarantee persistence of data in their buffers in the face of
>> reset conditions. We can't really do anything about the many
>> devices that don't honor persistance requests but we can start here.
>>
>> Signed-off-by: Jeremy Linton <jeremy.linton@arm.com>
>> ---
>> Platform/RaspberryPi/Library/ResetLib/ResetLib.c | 44 ++++++++++++++++++++++++
>> 1 file changed, 44 insertions(+)
>>
>> diff --git a/Platform/RaspberryPi/Library/ResetLib/ResetLib.c b/Platform/RaspberryPi/Library/ResetLib/ResetLib.c
>> index a70eee485d..036f619cb5 100644
>> --- a/Platform/RaspberryPi/Library/ResetLib/ResetLib.c
>> +++ b/Platform/RaspberryPi/Library/ResetLib/ResetLib.c
>> @@ -19,11 +19,54 @@
>> #include <Library/TimerLib.h>
>> #include <Library/EfiResetSystemLib.h>
>> #include <Library/ArmSmcLib.h>
>> +#include <Library/UefiBootServicesTableLib.h>
>> #include <Library/UefiLib.h>
>> #include <Library/UefiRuntimeLib.h>
>>
>> #include <IndustryStandard/ArmStdSmc.h>
>>
>> +
>> +/**
>> + Disconnect everything.
>> + Modified from the UEFI 2.3 spec (May 2009 version)
>> +
>> + @retval EFI_SUCCESS The operation was successful.
>> +
>> +**/
>
> STATIC
>
>> +EFI_STATUS
>> +DisconnectAll(
>
> Space before (
>
>> + VOID
>> + )
>> +{
>> + EFI_STATUS Status;
>> + UINTN HandleCount;
>> + EFI_HANDLE *HandleBuffer;
>> + UINTN HandleIndex;
>> +
>> + //
>> + // Retrieve the list of all handles from the handle database
>> + //
>> + Status = gBS->LocateHandleBuffer (
>> + AllHandles,
>> + NULL,
>> + NULL,
>> + &HandleCount,
>> + &HandleBuffer
>> + );
>> + if (!EFI_ERROR (Status)) {
>
> I understand that this code is copy/pasted but I'd still prefer to
> avoid the 'success handling' anti pattern here.
Sure.
>
> if (EFI_ERROR (Status)) {
> return Status;
> }
>
>> + for (HandleIndex = 0; HandleIndex < HandleCount; HandleIndex++) {
>> + Status = gBS->DisconnectController (
>> + HandleBuffer[HandleIndex],
>> + NULL,
>> + NULL
>> + );
>> + }
>> + gBS->FreePool(HandleBuffer);
>> + }
>> + return (EFI_SUCCESS);
>
> No need for ()
Yup
>
>> +}
>> +
>> +
>> /**
>> Resets the entire platform.
>>
>> @@ -57,6 +100,7 @@ LibResetSystem (
>> if (Delay != 0) {
>> DEBUG ((DEBUG_INFO, "Platform will be reset in %d.%d seconds...\n",
>> Delay / 1000000, (Delay % 1000000) / 100000));
>> + DisconnectAll ();
>
> Capture Status here and ASSERT_EFI_ERROR() ??
>
> Maybe it is overkill, and maybe DisconnectController() fails
> spuriously, so I am not entirely sure, but adding a local function
> that returns a value and then ignore it seems slightly sloppy to me.
Which makes the above bits about failure returns sorta redundant as I
should probably just make DisconnectAll() void. There isn't really
anything to do with a failed return other than print a message and
ignore it.
>
>> MicroSecondDelay (Delay);
>> }
>> }
>> --
>> 2.13.7
>>
next prev parent reply other threads:[~2021-10-05 21:25 UTC|newest]
Thread overview: 19+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-10-02 0:52 [PATCH 0/5] Platform/Rpi: Various cleanups Jeremy Linton
2021-10-02 0:52 ` [PATCH 1/5] Platform/RaspberryPi: Fix vfr warning caused by two defaults Jeremy Linton
2021-10-02 1:12 ` Andrei Warkentin
2021-10-02 0:52 ` [PATCH 2/5] Platform/RaspberryPi: Expand locking to cover return data Jeremy Linton
2021-10-02 1:17 ` Andrei Warkentin
2021-10-05 10:12 ` Ard Biesheuvel
2021-10-05 21:19 ` Jeremy Linton
2021-10-06 15:31 ` Ard Biesheuvel
2021-10-02 0:52 ` [PATCH 3/5] Platform/RaspberryPi: Update Linux quirk name Jeremy Linton
2021-10-02 1:13 ` Andrei Warkentin
2021-10-14 21:22 ` Jeremy Linton
2021-10-02 0:52 ` [PATCH 4/5] Platform/RaspberryPi: Normal memory should not be marked as uncached Jeremy Linton
2021-10-02 1:14 ` Andrei Warkentin
2021-10-05 10:05 ` Ard Biesheuvel
2021-10-02 0:52 ` [PATCH 5/5] Platform/RaspberryPi: Disconnect/shutdown all drivers before reboot Jeremy Linton
2021-10-02 1:16 ` Andrei Warkentin
2021-10-05 10:11 ` Ard Biesheuvel
2021-10-05 21:24 ` Jeremy Linton [this message]
2021-10-05 21:35 ` Ard Biesheuvel
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=452e1938-3d6d-97f0-9922-35a7d573c621@arm.com \
--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