[Bug translator/2436] New: inline function context_switch in IA64 platfrom

[Bug translator/2436] New: inline function context_switch in IA64 platfrom

[bibo dot mao at intel dot com]

In IA64 platform, buildok/six.stp always failed to run. The error information 
is shown as follows:
Pass 1: parsed user script and 10 library script(s) in 71usr/3sys/75real ms.
parsed 'context_switch' -> func 'context_switch'
pattern 'kernel' matches module 'kernel'
focused on module 'kernel' = [a000000100000000-a000000101867ce8, bias 0]
pattern 'context_switch' matches function 'context_switch'
checking instances of inline context_switch
examining inline instance of context_switch
selected inline instance of context_switch
pattern 'kernel' matches module 'kernel'
semantic error: no match for probe point
         while: resolving probe point kernel.inline("context_switch")
Pass 2: analyzed script: 0 probe(s), 0 function(s), 0 global(s) in 
177usr/18sys/196real ms.
Pass 2: analysis failed.

-- 
           Summary: inline function context_switch in IA64 platfrom
           Product: systemtap
           Version: unspecified
            Status: NEW
          Severity: normal
          Priority: P2
         Component: translator
        AssignedTo: systemtap at sources dot redhat dot com
        ReportedBy: bibo dot mao at intel dot com


http://sourceware.org/bugzilla/show_bug.cgi?id=2436

------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

[Bug translator/2436] inline function context_switch in IA64 platfrom

[bibo dot mao at intel dot com]

------- Additional Comments From bibo dot mao at intel dot com  2006-03-09 02:26 -------
I run test case in elfutils-0.119 package, elfutils-0.119/tests/dwflmodtest, by 
the command "./dwflmodtest -i -f -k", about inline function  context_switch, it 
shows:
kernel/sched.c:1511: context_switch: inline function
    inlined from ??? (lowpc => no error) (highpc => no error)
It seems that inline instance has been found, but attributes(file, 
lineno,lowpc, highpc) about the inline instance is null. It means that it can 
not find instance's position is kernel image. 
So I think the dwarf information is incomplete.
And in IA32 platform, dwarf information includes lowpc highpc attributes, it 
shows like this:
 kernel/sched.c:1518: context_switch: inline function
    inlined from ??? 0xc02ac91e..0xc02aca7c
 

-- 


http://sourceware.org/bugzilla/show_bug.cgi?id=2436

------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

[Bug translator/2436] inline function context_switch in IA64 platfrom

[aoliva at sourceware dot org]

------- Additional Comments From aoliva at sourceware dot org  2006-04-15 05:22 -------
Could it be that the problem is just that GCC uses a DW_AT_ranges tag to
represent the non-contiguous address ranges corresponding to the inlined
function code?  The info looks right to me.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |aoliva at sourceware dot org


http://sourceware.org/bugzilla/show_bug.cgi?id=2436

------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

[Bug translator/2436] inline function context_switch in IA64 platfrom

[srikar at linux dot vnet dot ibm dot com]

------- Additional Comments From srikar at linux dot vnet dot ibm dot com  2007-04-19 12:25 -------
The problem was also hit while testing of 20070414 snapshot on PPC64

-- 


http://sourceware.org/bugzilla/show_bug.cgi?id=2436

------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

[Bug translator/2436] inline function context_switch in IA64 platfrom

[fche at redhat dot com]

------- Additional Comments From fche at redhat dot com  2007-05-05 15:19 -------
Some new code in tapsets.cxx now looks at the inlined functions' "aranges" data.
Does this problem still exist on ia64?  (Separately, we know that *arguments* to
inlined functions are sometimes invisible, but that's a different bug #1155).

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |WAITING


http://sourceware.org/bugzilla/show_bug.cgi?id=2436

------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

[Bug translator/2436] inline function context_switch in IA64 platfrom

[wcohen at redhat dot com]

------- Additional Comments From wcohen at redhat dot com  2007-08-16 18:34 -------
On the RHEL4 ia64 2.6.9-55.0.2.EL kernel the "context_switch" is a real function:

$ nm /usr/lib/debug/lib/modules/2.6.9-55.0.2.EL/vmlinux |grep context_switch
a00000010006a920 t context_switch
a00000010006b4e0 T nr_context_switches

On the 2.6.23-rc3 kernel built with the RHEL4 gcc 3.4 "context_switch" is
inlined. output from 

./build-elfutils/tests/dwflmodtest -if -K2.6.23-rc3

  kernel/sched.c:1834: context_switch: inline function
    inlined at kernel/sched.c:3487 (lowpc => no error) (highpc => no error)




-- 


http://sourceware.org/bugzilla/show_bug.cgi?id=2436

------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

[Bug translator/2436] inline function context_switch in IA64 platfrom

[fche at redhat dot com]

------- Additional Comments From fche at redhat dot com  2010-05-21 19:33 -------
Long ago the translator learned heuristics about lowpc vs aranges vs. inlining;
current stap code appears to work ok on older rhels.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|WAITING                     |RESOLVED
         Resolution|                            |WORKSFORME


http://sourceware.org/bugzilla/show_bug.cgi?id=2436

------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.