From mboxrd@z Thu Jan  1 00:00:00 1970
Subject: Re: [edk2-devel] [PATCH 1/1] ArmPlatformPkg/PrePeiCore: Print the firmware version early in boot
To: Ard Biesheuvel <ardb@kernel.org>,devel@edk2.groups.io
From: "Sami Mujawar" <sami.mujawar@arm.com>
X-Originating-Location: Cambridge, England, GB (217.140.106.51)
X-Originating-Platform: Windows Firefox 102
User-Agent: GROUPS.IO Web Poster
MIME-Version: 1.0
Date: Thu, 21 Jul 2022 09:23:13 -0700
References: <CAMj1kXHk5wLwXYCM-Kp=XmQ+yDvW6a1FGZArocBBfVzTM8rn=w@mail.gmail.com>
In-Reply-To: <CAMj1kXHk5wLwXYCM-Kp=XmQ+yDvW6a1FGZArocBBfVzTM8rn=w@mail.gmail.com>
Message-ID: <15026.1658420593869880980@groups.io>
Content-Type: multipart/alternative; boundary="JnzTVq38E3reU2EcyriC"

--JnzTVq38E3reU2EcyriC
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: quoted-printable

Hi Rebecca, Ard,

On Tue, May 3, 2022 at 12:43 AM, Ard Biesheuvel wrote:

>=20
>=20
>> // Data Cache enabled on Primary core when MMU is enabled.
>> ArmDisableDataCache ();
>> // Invalidate instruction cache
>> @@ -90,6 +95,18 @@ CEntryPoint (
>>=20
>> // If not primary Jump to Secondary Main
>> if (ArmPlatformIsPrimaryCore (MpId)) {
>> + // Initialize the Serial Port
>> + SerialPortInitialize ();
>> + CharCount =3D AsciiSPrint (
>> + Buffer,
>> + sizeof (Buffer),
>> + "UEFI firmware (version %s built at %a on %a)\n\r",
>> + (CHAR16 *)PcdGetPtr (PcdFirmwareVersionString),
>=20
> Given that this runs before any library constructors are called, I'd
> be inclined to use FixedPcdGetPtr() here (and [FixedPcd] in the .INFs
> below), to ensure that we don't dereference garbage.
>=20
> I also wonder whether we need a call to
> ProcessLibraryConstructorList() somewhere.

[SAMI] There is a patch at https://edk2.groups.io/g/devel/message/91113 whi=
ch aims to call ProcessLibraryConstructorList(). The idea being that the De=
bugLib constructor would be called which initializes the serial port.
So, maybe these patches need adjustment so that we avoid calling SerialPort=
Initialize () twice.
[/SAMI]

regards,

Sami Mujawar

>=20
>=20
>=20
>> + __TIME__,
>> + __DATE__
>> + );
>> + SerialPortWrite ((UINT8 *)Buffer, CharCount);
>=20
>

--JnzTVq38E3reU2EcyriC
Content-Type: text/html; charset="utf-8"
Content-Transfer-Encoding: quoted-printable

Hi Rebecca, Ard,<br /><br />On Tue, May 3, 2022 at 12:43 AM, Ard Biesheuvel=
 wrote:<br />
<blockquote>
<blockquote>// Data Cache enabled on Primary core when MMU is enabled.<br /=
>ArmDisableDataCache ();<br />// Invalidate instruction cache<br />@@ -90,6=
 +95,18 @@ CEntryPoint (<br /><br />// If not primary Jump to Secondary Mai=
n<br />if (ArmPlatformIsPrimaryCore (MpId)) {<br />+ // Initialize the Seri=
al Port<br />+ SerialPortInitialize ();<br />+ CharCount =3D AsciiSPrint (<=
br />+ Buffer,<br />+ sizeof (Buffer),<br />+ "UEFI firmware (version %s bu=
ilt at %a on %a)\n\r",<br />+ (CHAR16 *)PcdGetPtr (PcdFirmwareVersionString=
),</blockquote>
Given that this runs before any library constructors are called, I'd<br />b=
e inclined to use FixedPcdGetPtr() here (and [FixedPcd] in the .INFs<br />b=
elow), to ensure that we don't dereference garbage.<br /><br />I also wonde=
r whether we need a call to<br />ProcessLibraryConstructorList() somewhere.=
</blockquote>
[SAMI] There is a patch at <a href=3D"https://edk2.groups.io/g/devel/messag=
e/91113" target=3D"_blank" rel=3D"noopener">https://edk2.groups.io/g/devel/=
message/91113</a> which aims to call ProcessLibraryConstructorList(). The i=
dea being that the DebugLib constructor would be called which initializes t=
he serial port.<br />So, maybe these patches need adjustment so that we avo=
id calling SerialPortInitialize () twice.<br />[/SAMI]<br /><br />regards,<=
br /><br />Sami Mujawar<br />
<blockquote><br />
<blockquote>+ __TIME__,<br />+ __DATE__<br />+ );<br />+ SerialPortWrite ((=
UINT8 *)Buffer, CharCount);</blockquote>
</blockquote>

--JnzTVq38E3reU2EcyriC--