• source navigation • diff markup • identifier search • freetext search •
Version: 2.0.40 2.2.26 2.4.37 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 3.10 3.11 3.12 3.13 3.14 3.15 3.16 3.17 3.18
1 Tagged virtual addresses in AArch64 Linux 2 ========================================= 3 4 Author: Will Deacon <will.deacon@arm.com> 5 Date : 12 June 2013 6 7 This document briefly describes the provision of tagged virtual 8 addresses in the AArch64 translation system and their potential uses 9 in AArch64 Linux. 10 11 The kernel configures the translation tables so that translations made 12 via TTBR0 (i.e. userspace mappings) have the top byte (bits 63:56) of 13 the virtual address ignored by the translation hardware. This frees up 14 this byte for application use, with the following caveats: 15 16 (1) The kernel requires that all user addresses passed to EL1 17 are tagged with tag 0x00. This means that any syscall 18 parameters containing user virtual addresses *must* have 19 their top byte cleared before trapping to the kernel. 20 21 (2) Non-zero tags are not preserved when delivering signals. 22 This means that signal handlers in applications making use 23 of tags cannot rely on the tag information for user virtual 24 addresses being maintained for fields inside siginfo_t. 25 One exception to this rule is for signals raised in response 26 to watchpoint debug exceptions, where the tag information 27 will be preserved. 28 29 (3) Special care should be taken when using tagged pointers, 30 since it is likely that C compilers will not hazard two 31 virtual addresses differing only in the upper byte. 32 33 The architecture prevents the use of a tagged PC, so the upper byte will 34 be set to a sign-extension of bit 55 on exception return.
This page was automatically generated by LXR 0.3.1 (source). • Linux is a registered trademark of Linus Torvalds • Contact us