From: Laszlo Ersek <lersek@redhat.com>
To: Haojian Zhuang <haojian.zhuang@linaro.org>, edk2-devel@lists.01.org
Cc: Leif Lindholm <leif.lindholm@linaro.org>,
Ard Biesheuvel <ard.biesheuvel@linaro.org>
Subject: Re: [PATCH v3 2/2] ArmPkg/PlatformBootManagerLib: load platform boot options
Date: Fri, 20 Apr 2018 14:55:03 +0200 [thread overview]
Message-ID: <b5dd6853-a67b-8963-d6ca-348c3ff16967@redhat.com> (raw)
In-Reply-To: <1524134560-1245-3-git-send-email-haojian.zhuang@linaro.org>
On 04/19/18 12:42, Haojian Zhuang wrote:
> Make platform driver to create predefined boot options and related
> hot keys.
>
> Cc: Laszlo Ersek <lersek@redhat.com>
> Cc: Leif Lindholm <leif.lindholm@linaro.org>
> Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
> Contributed-under: TianoCore Contribution Agreement 1.1
> Signed-off-by: Haojian Zhuang <haojian.zhuang@linaro.org>
> ---
> ArmPkg/Library/PlatformBootManagerLib/PlatformBm.c | 83 ++++++++++++++++++++++
> .../PlatformBootManagerLib.inf | 2 +
> 2 files changed, 85 insertions(+)
>
> diff --git a/ArmPkg/Library/PlatformBootManagerLib/PlatformBm.c b/ArmPkg/Library/PlatformBootManagerLib/PlatformBm.c
> index 61ab61ccc780..3a40592b58dd 100644
> --- a/ArmPkg/Library/PlatformBootManagerLib/PlatformBm.c
> +++ b/ArmPkg/Library/PlatformBootManagerLib/PlatformBm.c
> @@ -30,6 +30,7 @@
> #include <Protocol/LoadedImage.h>
> #include <Protocol/PciIo.h>
> #include <Protocol/PciRootBridgeIo.h>
> +#include <Protocol/PlatformBootManager.h>
> #include <Guid/EventGroup.h>
> #include <Guid/TtyTerm.h>
>
> @@ -392,6 +393,86 @@ PlatformRegisterFvBootOption (
>
> STATIC
> VOID
> +GetPlatformOptions (
> + VOID
> + )
> +{
> + EFI_STATUS Status;
> + EFI_BOOT_MANAGER_LOAD_OPTION *CurrentBootOptions;
> + EFI_BOOT_MANAGER_LOAD_OPTION *PlatformBootOptionArray;
> + EFI_INPUT_KEY *PlatformKeyArray;
> + PLATFORM_BOOT_MANAGER_PROTOCOL *PlatformBootManager;
> + UINTN CurrentBootOptionCount;
> + UINTN OptionIndex;
> + UINTN Index;
> +
> + Status = gBS->LocateProtocol (&gPlatformBootManagerProtocolGuid, NULL,
> + (VOID **)&PlatformBootManager);
> + if (EFI_ERROR (Status)) {
> + return;
> + }
> + Status = PlatformBootManager->GetPlatformBootOptionsAndKeys (
> + &PlatformBootOptionArray,
> + &PlatformKeyArray
> + );
The indentation is not correct. Please use:
Status = PlatformBootManager->GetPlatformBootOptionsAndKeys (
&PlatformBootOptionArray,
&PlatformKeyArray
);
(I don't insist that the patch be resubmitted just because of this.)
> + if (EFI_ERROR (Status)) {
> + return;
> + }
> + CurrentBootOptions = EfiBootManagerGetLoadOptions (
> + &CurrentBootOptionCount, LoadOptionTypeBoot
> + );
The coding style is not correct. Choose one of the folowing two:
- canonical:
CurrentBootOptions = EfiBootManagerGetLoadOptions (
&CurrentBootOptionCount,
LoadOptionTypeBoot
);
- less then canonical, but still acceptable:
CurrentBootOptions = EfiBootManagerGetLoadOptions (&CurrentBootOptionCount,
LoadOptionTypeBoot);
Anyway it's up to the ArmPkg maintainers to choose a preference; I shouldn't
comment more on indentation.
> + OptionIndex = EfiBootManagerFindLoadOption (
> + &PlatformBootOptionArray[0],
> + CurrentBootOptions,
> + CurrentBootOptionCount
> + );
EfiBootManagerFindLoadOption() returns an INTN value (so that it can
return -1), but OptionIndex is of type UINTN. Is it your intent to store
MAX_UINTN if the retval is -1?
> + if (OptionIndex == -1) {
> + OptionIndex = 1;
> + } else {
> + OptionIndex = OptionIndex + 1;
> + }
> + Index = 0;
> + //
> + // Last entries of PlatformBootOptionArray and PlatformKeyArray are empty.
> + //
> + while (PlatformBootOptionArray[Index].Description != NULL) {
> + if (OptionIndex == 1) {
> + //
> + // Append the BootLoadOption
> + //
> + Status = EfiBootManagerAddLoadOptionVariable (
> + &PlatformBootOptionArray[Index],
> + OptionIndex + Index
> + );
> + ASSERT_EFI_ERROR (Status);
> + }
> + //
> + // If UnicodeChar isn't empty, there's a hot key.
> + //
> + if (PlatformKeyArray[Index].UnicodeChar) {
> + //
> + // The index of Boot Options counts from 1.
> + // The last index equals to the count of total Boot Options.
> + //
> + Status = EfiBootManagerAddKeyOptionVariable (
> + NULL, OptionIndex + Index, 0,
> + PlatformKeyArray[Index], NULL
> + );
> + ASSERT_EFI_ERROR (Status);
> + }
> + Index++;
> + }
I'm sorry, I don't understand the logic. It seems that you determine the
insertion point in the boot order by looking up the first
platform-specific boot option in the current option list. If there is no
match, you insert the platform-specific boot options at positions 1, 2,
3, ..., leaving the preexistent boot option at position 0 undisturbed,
and pushing the rest of the preexistent boot options back. Why is this
useful?
Instead, I would imagine: iterate over all the platform-specific boot
options, and in the loop body, check whether each one is already in
CurrentBootOptions() or not. If it's already there (somewhere --
anywhere), just skip it. If it's not there yet, then append it, by
passing MAX_UINTN as Position to EfiBootManagerAddLoadOptionVariable().
That will leave the user's BootOrder undisturbed, just make sure that
the platform-specific boot options are always present -- in the worst
case, at the very end of the boot order.
Let me show you what I have in mind (untested, clearly):
> diff --git a/ArmPkg/Library/PlatformBootManagerLib/PlatformBm.c b/ArmPkg/Library/PlatformBootManagerLib/PlatformBm.c
> index 61ab61ccc780..3daab129043b 100644
> --- a/ArmPkg/Library/PlatformBootManagerLib/PlatformBm.c
> +++ b/ArmPkg/Library/PlatformBootManagerLib/PlatformBm.c
> @@ -30,6 +30,7 @@
> #include <Protocol/LoadedImage.h>
> #include <Protocol/PciIo.h>
> #include <Protocol/PciRootBridgeIo.h>
> +#include <Protocol/PlatformBootManager.h>
> #include <Guid/EventGroup.h>
> #include <Guid/TtyTerm.h>
>
> @@ -389,6 +390,100 @@ PlatformRegisterFvBootOption (
> EfiBootManagerFreeLoadOptions (BootOptions, BootOptionCount);
> }
>
> +STATIC
> +VOID
> +RegisterPlatformOptions (
> + VOID
> + )
> +{
> + EFI_STATUS Status;
> + PLATFORM_BOOT_MANAGER_PROTOCOL *PlatformBootManager;
> + UINTN BootCount;
> + EFI_BOOT_MANAGER_LOAD_OPTION *BootOptions;
> + EFI_INPUT_KEY *BootKeys;
> + UINTN CurrentBootCount;
> + EFI_BOOT_MANAGER_LOAD_OPTION *CurrentBootOptions;
> + UINTN Index;
> +
> + //
> + // Get the platform's boot options.
> + //
> + Status = gBS->LocateProtocol (&gPlatformBootManagerProtocolGuid, NULL,
> + (VOID **)&PlatformBootManager);
> + if (EFI_ERROR (Status)) {
> + return;
> + }
> + Status = PlatformBootManager->GetPlatformBootOptionsAndKeys (&BootCount,
> + &BootOptions, &BootKeys);
> + if (EFI_ERROR (Status)) {
> + return;
> + }
> + //
> + // Fetch the existent boot options. If there are none, CurrentBootCount will
> + // be zeroed.
> + //
> + CurrentBootOptions = EfiBootManagerGetLoadOptions (&CurrentBootCount,
> + LoadOptionTypeBoot);
> + //
> + // Process the platform boot options.
> + //
> + for (Index = 0; Index < BootCount; Index++) {
> + INTN Match;
> + UINTN BootOptionNumber;
> +
> + //
> + // If there are any preexistent boot options, and the subject platform boot
> + // option is already among them, then don't try to add it. Just get its
> + // assigned boot option number so we can associate a hotkey with it. Note
> + // that EfiBootManagerFindLoadOption() deals fine with (CurrentBootOptions
> + // == NULL) if (CurrentBootCount == 0).
> + //
> + Match = EfiBootManagerFindLoadOption (&BootOptions[Index],
> + CurrentBootOptions, CurrentBootCount);
> + if (Match >= 0) {
> + BootOptionNumber = CurrentBootOptions[Match].OptionNumber;
> + } else {
> + //
> + // Add the platform boot option as a new one, at the end of the boot
> + // order. Note that if the platform provided this boot option with an
> + // unassigned option number, then the below function call will assign a
> + // number.
> + //
> + Status = EfiBootManagerAddLoadOptionVariable (&BootOptions[Index],
> + MAX_UINTN);
> + if (EFI_ERROR (Status)) {
> + DEBUG ((DEBUG_ERROR, "%a: failed to register \"%s\": %r\n",
> + __FUNCTION__, BootOptions[Index].Description, Status));
> + continue;
> + }
> + BootOptionNumber = BootOptions[Index].OptionNumber;
> + }
> +
> + //
> + // Register a hotkey with the boot option, if requested.
> + //
> + if (BootKeys[Index].UnicodeChar == 'L\0') {
> + continue;
> + }
> + Status = EfiBootManagerAddKeyOptionVariable (
> + NULL, // AddedOption
> + BootOptionNumber,
> + 0 // Modifier
> + &BootKeys[Index],
> + NULL
> + );
> + if (EFI_ERROR (Status)) {
> + DEBUG ((DEBUG_ERROR, "%a: failed to register hotkey for \"%s\": %r\n",
> + __FUNCTION__, BootOptions[Index].Description, Status));
> + }
> + }
> + //
> + // (CurrentBootOptions == NULL) is handled gracefully by the below.
> + //
> + EfiBootManagerFreeLoadOptions (CurrentBootOptions, CurrentBootCount);
> + EfiBootManagerFreeLoadOptions (BootOptions, BootCount);
> + FreePool (BootKeys);
> +}
>
> STATIC
> VOID
> @@ -402,6 +497,8 @@ PlatformRegisterOptionsAndKeys (
> EFI_INPUT_KEY Esc;
> EFI_BOOT_MANAGER_LOAD_OPTION BootOption;
>
> + RegisterPlatformOptions ();
> +
> //
> // Register ENTER as CONTINUE key
> //
I apologize if I misunderstood you.
Thanks
Laszlo
prev parent reply other threads:[~2018-04-20 12:55 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-04-19 10:42 [PATCH v3 0/2] add platform boot manager protocol Haojian Zhuang
2018-04-19 10:42 ` [PATCH v3 1/2] EmbeddedPkg: " Haojian Zhuang
2018-04-20 11:33 ` Laszlo Ersek
2018-04-19 10:42 ` [PATCH v3 2/2] ArmPkg/PlatformBootManagerLib: load platform boot options Haojian Zhuang
2018-04-20 12:55 ` Laszlo Ersek [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=b5dd6853-a67b-8963-d6ca-348c3ff16967@redhat.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