* [gold patch] Incremental 20/22: Fix problem caused by realignment of data segment
@ 2011-05-02 23:16 Cary Coutant
2011-06-08 0:38 ` Ian Lance Taylor
0 siblings, 1 reply; 2+ messages in thread
From: Cary Coutant @ 2011-05-02 23:16 UTC (permalink / raw)
To: Ian Lance Taylor, Binutils
[-- Attachment #1: Type: text/plain, Size: 827 bytes --]
This patch fixes a problem with the TLS segment caused by realignment
of the data segment during the initial incremental-full link. When we
do a subsequent incremental-update, the initial attempt to assign
addresses to the data segment doesn't match the fixed layout, and we
end up with out-of-sync addresses. I avoid this issue simply by
keeping the original layout for --incremental-full. Part of the
problem was also some old code under a FIXME that ought to have been
removed in one of the earlier patches, but I didn't notice that when I
was rearranging all my git commits.
OK?
-cary
2011-05-02 Cary Coutant <ccoutant@google.com>
* layout.cc (Layout::set_segment_offsets): Don't adjust layout
for incremental links.
* output.cc (Output_segment::set_section_list_addresses): Remove
FIXME and test for TLS or BSS.
[-- Attachment #2: incr-patch-20.txt --]
[-- Type: text/plain, Size: 1455 bytes --]
Suppress realignment of data segment for incremental links, and
fix handling of TLS and BSS sections during incremental update.
2011-05-02 Cary Coutant <ccoutant@google.com>
* layout.cc (Layout::set_segment_offsets): Don't adjust layout
for incremental links.
* output.cc (Output_segment::set_section_list_addresses): Remove
FIXME and test for TLS or BSS.
diff --git a/gold/layout.cc b/gold/layout.cc
index 67ed8c3..537651a 100644
--- a/gold/layout.cc
+++ b/gold/layout.cc
@@ -2945,7 +2945,7 @@ Layout::set_segment_offsets(const Target* target, Output_segment* load_seg,
if (!are_addresses_set
&& !has_relro
&& aligned_addr != addr
- && !parameters->incremental_update())
+ && !parameters->incremental())
{
uint64_t first_off = (common_pagesize
- (aligned_addr
diff --git a/gold/output.cc b/gold/output.cc
index 832428e..8e81b07 100644
--- a/gold/output.cc
+++ b/gold/output.cc
@@ -4036,10 +4036,7 @@ Output_segment::set_section_list_addresses(Layout* layout, bool reset,
}
}
- // FIXME: Need to handle TLS and .bss with incremental update.
- if (!parameters->incremental_update()
- || (*p)->is_section_flag_set(elfcpp::SHF_TLS)
- || (*p)->is_section_type(elfcpp::SHT_NOBITS))
+ if (!parameters->incremental_update())
{
off = align_address(off, align);
(*p)->set_address_and_file_offset(addr + (off - startoff), off);
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: [gold patch] Incremental 20/22: Fix problem caused by realignment of data segment
2011-05-02 23:16 [gold patch] Incremental 20/22: Fix problem caused by realignment of data segment Cary Coutant
@ 2011-06-08 0:38 ` Ian Lance Taylor
0 siblings, 0 replies; 2+ messages in thread
From: Ian Lance Taylor @ 2011-06-08 0:38 UTC (permalink / raw)
To: Cary Coutant; +Cc: Binutils
Cary Coutant <ccoutant@google.com> writes:
> 2011-05-02 Cary Coutant <ccoutant@google.com>
>
> * layout.cc (Layout::set_segment_offsets): Don't adjust layout
> for incremental links.
> * output.cc (Output_segment::set_section_list_addresses): Remove
> FIXME and test for TLS or BSS.
This is OK.
Thanks.
Ian
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2011-06-08 0:38 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2011-05-02 23:16 [gold patch] Incremental 20/22: Fix problem caused by realignment of data segment Cary Coutant
2011-06-08 0:38 ` Ian Lance Taylor
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).