This is the mail archive of the libc-alpha@sourceware.org mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Fifth draft of the Y2038 design document


On Wed, 1 Mar 2017, Zack Weinberg wrote:

> On Wed, Mar 1, 2017 at 2:11 AM, Albert ARIBAUD <albert.aribaud@3adev.fr> wrote:
> > GLIBC needs to provide 64-bit-time clock_gettime to 32-bit code in
> > various run-time cases:
> >
> > - above an 'old', only 32-bit-time, kernel;
> > - above a 'new', mixed 64- and 32-bit-time, kernel;
> > - above a 'newer', only 64-bit time, kernel.
> 
> Is it really necessary to support 'old' kernels?  That seems like it
> would add a great deal of complexity with no actual benefit.

It will be several years before glibc can require kernels newer than 
currently exist (only when all current kernels are no longer maintained as 
kernel.org releases, and maybe not even then if more issues like the one 
with 2.6.32 containers arise).  We want 64-bit time support in glibc 
before all old kernels cease to be supported, and we want glibc-internal 
uses of time-related interfaces to move over to using the 64-bit time 
interfaces (without depending on new kernel support).

Obviously 64-bit glibc interfaces won't work with 32-bit kernel interfaces 
after 2038, but the obsoletion of kernels lacking the new interfaces 
should come long before that.  And we may be able to depend on new kernel 
headers to simplify things, just not on a new kernel at runtime.

-- 
Joseph S. Myers
joseph@codesourcery.com


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]