diff options
| -rw-r--r-- | gas/ChangeLog | 5 | ||||
| -rw-r--r-- | gas/doc/as.texinfo | 53 |
2 files changed, 53 insertions, 5 deletions
diff --git a/gas/ChangeLog b/gas/ChangeLog index 263d0b2..cfd6326 100644 --- a/gas/ChangeLog +++ b/gas/ChangeLog @@ -1,3 +1,8 @@ +2016-04-19 Martin Galvan <martin.galvan@tallertechnologies.com> + + * doc/as.texinfo (.cfi_remember_state, .cfi_restore_state): Improve + documentation. + 2016-04-17 Andrew Burgess <andrew.burgess@embecosm.com> Revert prevous change. diff --git a/gas/doc/as.texinfo b/gas/doc/as.texinfo index 7b36931..e2e744e 100644 --- a/gas/doc/as.texinfo +++ b/gas/doc/as.texinfo @@ -4816,11 +4816,54 @@ From now on the previous value of @var{register} can't be restored anymore. Current value of @var{register} is the same like in the previous frame, i.e. no restoration needed. -@subsection @code{.cfi_remember_state}, -First save all current rules for all registers by @code{.cfi_remember_state}, -then totally screw them up by subsequent @code{.cfi_*} directives and when -everything is hopelessly bad, use @code{.cfi_restore_state} to restore -the previous saved state. +@subsection @code{.cfi_remember_state} and @code{.cfi_restore_state} +@code{.cfi_remember_state} pushes the set of rules for every register onto an +implicit stack, while @code{.cfi_restore_state} pops them off the stack and +places them in the current row. This is useful for situations where you have +multiple @code{.cfi_*} directives that need to be undone due to the control +flow of the program. For example, we could have something like this (assuming +the CFA is the value of @code{rbp}): + +@smallexample + je label + popq %rbx + .cfi_restore %rbx + popq %r12 + .cfi_restore %r12 + popq %rbp + .cfi_restore %rbp + .cfi_def_cfa %rsp, 8 + ret +label: + /* Do something else */ +@end smallexample + +Here, we want the @code{.cfi} directives to affect only the rows corresponding +to the instructions before @code{label}. This means we'd have to add multiple +@code{.cfi} directives after @code{label} to recreate the original save +locations of the registers, as well as setting the CFA back to the value of +@code{rbp}. This would be clumsy, and result in a larger binary size. Instead, +we can write: + +@smallexample + je label + popq %rbx + .cfi_remember_state + .cfi_restore %rbx + popq %r12 + .cfi_restore %r12 + popq %rbp + .cfi_restore %rbp + .cfi_def_cfa %rsp, 8 + ret +label: + .cfi_restore_state + /* Do something else */ +@end smallexample + +That way, the rules for the instructions after @code{label} will be the same +as before the first @code{.cfi_restore} without having to use multiple +@code{.cfi} directives. @subsection @code{.cfi_return_column @var{register}} Change return column @var{register}, i.e. the return address is either |