Bug 5329 - LinuxHost does not deal with a missing DT_DEBUG -> r_debug (ie points to 0x0)
Summary: LinuxHost does not deal with a missing DT_DEBUG -> r_debug (ie points to 0x0)
Status: NEW
Alias: None
Product: frysk
Classification: Unclassified
Component: general (show other bugs)
Version: unspecified
: P2 normal
Target Milestone: ---
Assignee: Unassigned
URL:
Keywords:
Depends on:
Blocks: 1595
  Show dependency tree / graph
 
Reported: 2007-11-15 09:34 UTC by Phil Muldoon
Modified: 2007-11-15 15:19 UTC (History)
0 users

See Also:
Host:
Target:
Build:
Last reconfirmed:

Attachments
Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Phil Muldoon 2007-11-15 09:34:48 UTC 
Theory: If a process is blocked at entry-point, as DaemonBlockedAtEntry() does,
the dynamic linker has not built this information yet. DT_DEBUG points to 0x0,
not to r_debug, which does not allow LinuxHost to find the Linkmap table to
build the metadata.

If a process is specified, and rich metadata cannot be built, should the
LinuxHost revert to basic metadata, or just completely abort?
Comment 1 Phil Muldoon 2007-11-15 09:37:35 UTC 
"If a process is blocked at entry-point, and a core file is created from a
process at that time, as DaemonBlockedAtEntry() .... "

That's how it should read
Comment 2 Andrew Cagney 2007-11-15 15:19:40 UTC 
Nice catch; for a traditional core file, it's a "you loose" situtation.

Does/can fcore dump the maps table to a note section so that, for frysk it isn't
so much of a loose-loose?

Not dumping the maps page can be added to not dumping the full exe/args bugs.