Hi Richard,

On Wed, 2022-11-30 at 09:07 +0100, Alexander Kanavin wrote:

On Tue, 29 Nov 2022 at 16:45, Stephen Jolley
<sjolley.yp.pm@...> wrote:

We’d welcome a proposal/series on how to move forward with the
Y2038 work for 32 bit platforms.

I have the following proposal:

1. A branch is made where:
a. "-D_TIME_BITS=64 -D_FILE_OFFSET_BITS=64" is enabled globally.
b. qemu is always started with "-rtc base=2040-01-01", simulating
Y2038 actually occurring.
c. an additional runtime test verifies that both RTC clock and
system clock report 2040.

2. This branch is run through a-full on the autobuilder. Any
uncovered issues are filed as bugs.

3. Once *all* of the bugs are addressed, repeat point 2.

4. Once there are no more open bugs, 1a is merged into master.

Any fatal flaws in the plan?

Others have made some good comments. My thoughts:

* We need to add some runtime tests to oeqa for this (in addition to
the ptests)

* We need to have a 32 bit ptest run on the autobuilder (qemux86
should work, not sure we can make qemuarm fast). Whether this is
manually triggered, not sure. We could have a smaller set of ptests
to run for it?

Y2038 ptests maybe?

Here is the list of integrated tests to ptests:
https://github.com/lmajewski/y2038-tests

* Could we optionally disable some of the glibc 32 bit function calls
to ensure they're not being used?

Could you be more specific here? Would you like to disable some
syscalls?

We don't really want to diverge from
upstream glibc much though.

Could you be more specific here? The glibc now supports the whole set
of syscalls as of 2.34 version?

To enable them one needs to pass -D_TIME_BITS=64 flag when compiling
programs.

This is now the official glibc ABI.

* We need to work out how to communicate this change happened and have
people "buy in" to it.

Ok.

The reason for that is that if someone has
existing binaries, there could be problems using them after the
change.

The binary shall work without issues on glibc 2.34+ and 5.10+ kernel
without issues.

The only problem happens when new binaries with 64 bit time support are
run on glibc or kernel not supporting 64 bit time.

We therefore need to be sure they are aware of it.

Cheers,

Richard

Best regards,

Lukasz Majewski

--

DENX Software Engineering GmbH, Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-59 Fax: (+49)-8142-66989-80 Email: lukma@...

On Wed, 2022-11-30 at 14:36 +0100, Lukasz Majewski wrote:

On Wed, 2022-11-30 at 09:07 +0100, Alexander Kanavin wrote:

On Tue, 29 Nov 2022 at 16:45, Stephen Jolley
<sjolley.yp.pm@...> wrote:

We’d welcome a proposal/series on how to move forward with the
Y2038 work for 32 bit platforms.

I have the following proposal:

1. A branch is made where:
a. "-D_TIME_BITS=64 -D_FILE_OFFSET_BITS=64" is enabled globally.
b. qemu is always started with "-rtc base=2040-01-01", simulating
Y2038 actually occurring.
c. an additional runtime test verifies that both RTC clock and
system clock report 2040.

2. This branch is run through a-full on the autobuilder. Any
uncovered issues are filed as bugs.

3. Once *all* of the bugs are addressed, repeat point 2.

4. Once there are no more open bugs, 1a is merged into master.

Any fatal flaws in the plan?

Others have made some good comments. My thoughts:

* We need to add some runtime tests to oeqa for this (in addition to
the ptests)

* We need to have a 32 bit ptest run on the autobuilder (qemux86
should work, not sure we can make qemuarm fast). Whether this is
manually triggered, not sure. We could have a smaller set of ptests
to run for it?

Y2038 ptests maybe?

Here is the list of integrated tests to ptests:
https://github.com/lmajewski/y2038-tests

Perhaps, yes.

* Could we optionally disable some of the glibc 32 bit function calls
to ensure they're not being used?

Could you be more specific here? Would you like to disable some
syscalls?

I'm meaning disabling the 32 bit glibc time functions.

We don't really want to diverge from
upstream glibc much though.

Could you be more specific here? The glibc now supports the whole set
of syscalls as of 2.34 version?

To enable them one needs to pass -D_TIME_BITS=64 flag when compiling
programs.

This is now the official glibc ABI.

Right, but the 32 bit time functions/symbols are still available for
older binaries. My point is that anything using those older functions
is likely in need of attention so for Yocto Project/OE usage,
identifying those would be helpful. If we were to disable them, that
would make such usage very obvious.

The reason for that is that if someone has
existing binaries, there could be problems using them after the
change.

The binary shall work without issues on glibc 2.34+ and 5.10+ kernel
without issues.

Not necessarily. If it were a binary library, compiled with 32 bit
time_t, new binaries using it would use a different sized field.

The only problem happens when new binaries with 64 bit time support are
run on glibc or kernel not supporting 64 bit time.

That is definitely not the only problem. Some of the problems are
unlikely but we do need to consider them.

Cheers,

Richard