This is the mail archive of the libc-hacker@sources.redhat.com mailing list for the glibc project.
Note that libc-hacker is a closed list. You may look at the archives of this list, but subscription and posting are not open.
Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
---|---|---|
Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |
Other format: | [Raw text] |
On Tue, Jan 07, 2003 at 03:08:05PM +0100, Martin Schwidefsky wrote: > Hi, > we - that is Ulrich Weigand my myself - have started working on the > TLS support for s390/s390x. I've done a preliminary binutils patch > which seems to work. This version uses the standard IA-32 calling > convention to __tls_get_addr. Now Ulrich came up with an idea > how to improve the instruction sequences but this requires two > changes to __tls_get_addr: > 1) __tls_get_addr would be called with the got offset of a tls_index > structure located in the got instead of a pointer to the tls_index > structure. > 2) __tls_get_addr is to return the offset to the TCB of the variable > the tls_index refers to instead of the address. > With these two changes we could improve the code quality for > TLS accesses quite a bit. > The question now is: can we have a s390/s390x version of > __tls_get_addr with these changes without breaking anything? The question is if it should be __tls_get_addr and not say __tls_get_offset on s390*... Would you be passing %r12 to that routine as another argument or would it have some assembly wrapper which would do that? Jakub
Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
---|---|---|
Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |