| commit | b344e78c65330d3b2d4f767e91349f2a6004e960 | |
| author | simonb <simonb@chromium.org> | |
| Thu, 18 Sep 2014 12:31:17 +0000 (18 05:31 -0700) | ||
| committer | Commit bot <commit-bot@chromium.org> | |
| Thu, 18 Sep 2014 12:31:26 +0000 (18 12:31 +0000) | ||
| tree | 7ac702cea1e7c43d53a843607a4d78c2bb7f3a0b | treesnapshot (tar.gz zip) |
| parent | 694cd937163f3eb45a2e52dd7da7899c2c9db146 | commitdiff |
Set spliced LOAD segment's memsz equal to filesz.
Current packing sets the spliced LOAD segment's memsz to larger than
its filesz. This causes the runtime linker to create a 'zero' mapping
to pad from filesz to memsz.
This isn't a problem -- the mapping sits on already reserved address
space and uses no real memory -- but 'null' mappings like this appear
in /proc/pid/maps and can disrupt or confuse tools which read and
parse it, for example Breakpad.
This change makes the spliced LOAD segment's memsz and filesz equal.
It also tidies code a little so that it is now more obvious when
filesz is equal to memsz, and vaddr is equal to paddr.
Also, rebuild test data to match.
BUG=385553
Review URL: https://codereview.chromium.org/571333003
Cr-Commit-Position: refs/heads/master@{#295459}
Current packing sets the spliced LOAD segment's memsz to larger than
its filesz. This causes the runtime linker to create a 'zero' mapping
to pad from filesz to memsz.
This isn't a problem -- the mapping sits on already reserved address
space and uses no real memory -- but 'null' mappings like this appear
in /proc/pid/maps and can disrupt or confuse tools which read and
parse it, for example Breakpad.
This change makes the spliced LOAD segment's memsz and filesz equal.
It also tidies code a little so that it is now more obvious when
filesz is equal to memsz, and vaddr is equal to paddr.
Also, rebuild test data to match.
BUG=385553
Review URL: https://codereview.chromium.org/571333003
Cr-Commit-Position: refs/heads/master@{#295459}