From: "Laszlo Ersek" <lersek@redhat.com>
To: devel@edk2.groups.io, mw@semihalf.com
Cc: leif@nuviainc.com, ard.biesheuvel@arm.com, jaz@semihalf.com,
kostap@marvell.com
Subject: Re: [edk2-devel] [edk2-platforms PATCH] Marvell/RealTimeClockLib: Allow only Unix Epoch in LibSetWakeupTime
Date: Mon, 21 Dec 2020 21:32:29 +0100 [thread overview]
Message-ID: <1191b863-3779-2751-e47e-deef8adc73f5@redhat.com> (raw)
In-Reply-To: <4f9497c7-0097-3a92-6b52-ece4a6defa9f@redhat.com>
On 12/21/20 21:09, Laszlo Ersek wrote:
> Hi Marcin,
>
> On 12/21/20 17:37, Marcin Wojtas wrote:
>> Because the Armada RTC uses a 32-bit counter for seconds,
>> the maximum time span is just over 136 years.
>> Time is stored in Unix Epoch format, so it starts in 1970,
>> Therefore it can not exceed the year 2106.
>>
>> The issue emerged during ACS test case, which does not pass
>> Unix Epoch-relative time and caused EfiTimeToEpoch to assert.
>>
>> Signed-off-by: Marcin Wojtas <mw@semihalf.com>
>> ---
>> Silicon/Marvell/Armada7k8k/Library/RealTimeClockLib/RealTimeClockLib.c | 10 ++++++++++
>> 1 file changed, 10 insertions(+)
>>
>> diff --git a/Silicon/Marvell/Armada7k8k/Library/RealTimeClockLib/RealTimeClockLib.c b/Silicon/Marvell/Armada7k8k/Library/RealTimeClockLib/RealTimeClockLib.c
>> index a811fd368e..40ab01ed41 100644
>> --- a/Silicon/Marvell/Armada7k8k/Library/RealTimeClockLib/RealTimeClockLib.c
>> +++ b/Silicon/Marvell/Armada7k8k/Library/RealTimeClockLib/RealTimeClockLib.c
>> @@ -179,6 +179,16 @@ LibSetWakeupTime (
>> {
>> UINTN WakeupSeconds;
>>
>> + //
>> + // Because the Armada RTC uses a 32-bit counter for seconds,
>> + // the maximum time span is just over 136 years.
>> + // Time is stored in Unix Epoch format, so it starts in 1970,
>> + // Therefore it can not exceed the year 2106.
>> + //
>> + if ((Time->Year < 1970) || (Time->Year >= 2106)) {
>> + return EFI_UNSUPPORTED;
>> + }
>> +
>> // Convert time to raw seconds
>> WakeupSeconds = EfiTimeToEpoch (Time);
>> if (WakeupSeconds > MAX_UINT32) {
>>
>
> please see:
>
> - edk2-platforms commit fbdfe8c4100d ("Silicon/Marvell/RealTimeClockLib:
> make EpochSeconds, WakeupSeconds UINTN", 2020-12-21) -- part of this is
> already visible in the context above,
>
> - edk2 commit c06635ea3f4b ("EmbeddedPkg/TimeBaseLib: remove useless
> truncation to 32-bit", 2020-12-21).
>
> If you advance the edk2 submodule reference in edk2-platforms to or past
> c06635ea3f4b, then the "Time->Year >= 2106" condition will be covered
> already, by the (WakeupSeconds > MAX_UINT32) check.
(Oops, sorry, I forgot that edk2-platforms consumes edk2 *only* through PACKAGES_PATH, and never through a submodule; so please adjust the above submodule statement accordingly.)
BTW, is this library instance ever built for 32-bit ARM? Because in that case (== UINTN meaning UINT32), checking Year against 2106 makes sense as well. While build-testing the above-noted patches, I tried to build them for 32-bit ARM too, but that arch did not seem applicable: while the following command works:
build \
-a AARCH64 \
-b NOOPT \
-p Platform/SolidRun/Armada80x0McBin/Armada80x0McBin.dsc \
-t GCC5 \
-m EmbeddedPkg/RealTimeClockRuntimeDxe/RealTimeClockRuntimeDxe.inf
the same with "-a ARM" fails, with the following error message:
edk2-platforms/Platform/SolidRun/Armada80x0McBin/Armada80x0McBin.dsc(...): error 4000: Instance of library class [ArmSoftFloatLib] is not found
in [edk2/CryptoPkg/Library/OpensslLib/OpensslLib.inf] [ARM]
consumed by module [edk2/SignedCapsulePkg/Universal/SystemFirmwareUpdate/SystemFirmwareReportDxe.inf]
which implies that "Armada80x0McBin.dsc" is never actually built for ARM.
... Hm, the following platforms:
- Platform/Marvell/Armada80x0Db/Armada80x0Db.dsc
- Platform/Marvell/Armada70x0Db/Armada70x0Db.dsc
do build the library for ARM. So this patch seems justified after all, as-posted (i.e., including the year 2106 check).
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Thanks
Laszlo
>
> Preventing ASSERT (JulianDate >= EPOCH_JULIAN_DATE) from firing in
> EfiGetEpochDays() still makes sense, so the (Time->Year < 1970) check
> can still be added usefully, I think.
>
> Thanks
> Laszlo
>
next prev parent reply other threads:[~2020-12-21 20:32 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-12-21 16:37 [edk2-platforms PATCH] Marvell/RealTimeClockLib: Allow only Unix Epoch in LibSetWakeupTime Marcin Wojtas
2020-12-21 20:09 ` [edk2-devel] " Laszlo Ersek
2020-12-21 20:32 ` Laszlo Ersek [this message]
2021-01-04 17:14 ` 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=1191b863-3779-2751-e47e-deef8adc73f5@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