On Wed, 19 Dec 2018, Vineet Gupta wrote: > I took a stab at this but not really happy with taking this approach. > > (1). Common code assumes disparate kernel and userland sigaction struct even > though there's no reason for a *new* port to be: its not like all glibc code > is > shared/common although I agree it is best to do so as much as possible > So this requires explicit copy over of 1 struct into other, when it could have > been avoided altogether for some cases atleast (!SA_RESTORER).
So make the generic code optimize those cases based on appropriate conditionals (making sure to verify those conditionals are right for every architecture in glibc). Any new architecture having much architecture-specific code for the kernel interface in glibc, beyond the basic definitions of how to call a syscall, is suspect, given that the kernel structures should be consistent across asm-generic architectures; we ought to make the defaults work so they are genuinely suitable as defaults for new architectures. This may require changes to the sysdeps/unix/sysv/linux/ code if it's currently "generic" to old architectures but not so good for asm-generic ones. -- Joseph S. Myers jos...@codesourcery.com _______________________________________________ linux-snps-arc mailing list linux-snps-arc@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-snps-arc
