Re: [evlog-dev] -- plugin problem

Hi Yi,

Looks like some version of glibc on IA64 requires apps to explicitly 
declare some of those string functions. Without declaration the compiler 
  just returns an integer (32 bits) instead of a pointer (64 bits).

Adding a #include <string.h> to backendmgr.c should fix the problem you 
see with strrchr.

Regarding tcp_rmtlog_be, it should work if both client and consolidator 
nodes are IA64 :-) . We have plan to get remote log works across 
platforms in the future, but right now it does not.



Zhu, Yi wrote:
> Hi Hien,
> 
> The problem is solved after followed your steps. 
> 
> I have a quesstion here, does current evlog support IA64? I can't find dir
> ia64 in evlog/kernel/v2.4.18/arch, though I found some "defined(__ia64__)"
> in the implementation.
> 
> Another thing is I found the glibc-2.2.4-31.2 in redhat 2.1AS has a problem.
> I don't know whether encountered this, but I had to replace the code in file
> backendmgr.c line 46 with
> 
> - ext = (char *) strrchr(name, '.');
> + ext = name + 9; /* sample_be has 9 chars */
> 
> It is ugly, but I otherwise I got a segment fault. Because the returned ext
> from strrchr() is '0x3213c', but it should be '0x600000000003213c'.
> 
> BTW, can you use the tcp_rmtlog_be in IA64? I can't, it seems to halt in
> file tcp_rmtlog_be.c line 358 when opening the fifo.
> 
> Thanks,
> -Zhu Yi
> 
> -----Original Message-----
> From: Hien Nguyen [mailto:[email protected]]
> Sent: 2003?1?17? 1:46
> To: Hien Nguyen
> Cc: Zhu, Yi; 'Steve Woodruff'; evlog mailing list
> Subject: Re: [evlog-dev] -- plugin problem
> 
> 
> Hi Steve, Yi,
> 
> I got it to work on my IA64 now. It is basically a linker problem. You 
> need to modify the libevl Makefile under evlog/user/lib/Makefile to link 
>   with the libgcc_s.so
> 
> Mine was looking like this
> 
> libevl.so: $(DOTSOS) query/made template/made util/made
>          rm -f libevl.so
>          ld -share -soname libevl.so.1 -o libevl.so $(LIB_SO) -lc 
> -L/home/janis/gnu/tools/gcc-mainline/lib -lgcc_s
> 
> I am not sure why we have to dothis on the IA64 and not IA32, could it 
> be a linker bug or gcc bug?
> 
> Thanks, Hien.
> 
> Hien Nguyen wrote:
> 
>>Yes, I got the same problem as Yi and Steve on my IA64 as I tried it 
>>last night.
>>
>>./evlogd: relocation error: /usr/lib/libevl.so.1: undefined symbol: 
>>__moddi3
>>
>>I am gooing to look for solution (work around) for this problem.
>>
>>Hien.
>>
>>Zhu, Yi wrote:
>>
>>
>>>Hi,
>>>
>>>I got the same problem with Steve. My 4-way IA64 with Redhat 2.1AS 
>>>also got
>>>the relocation error when using the plugin share libraries. Instead, my
>>>error is "relocation error: /usr/lib/libevl.so.1: undefined symbol:
>>>__moddi3".
>>>Below is the gdb trace
>>>
>>>200             for(i = 0; i < MAX_BACKENDS; i++) {
>>>(gdb) n
>>>201                     if (be_table[i].func != NULL) {
>>>(gdb) print be_table
>>>$1 = {{plugin = 0x600000000001f210, func = 0x200000000003f7b8 
>>><local+8424>,
>>>state = 0}, {plugin = 0x0,     func = 0, state = 0}, {plugin = 0x0, 
>>>func = 0, state = 0}, {plugin =
>>>0x0, func = 0, state = 0}, {
>>>    plugin = 0x0, func = 0, state = 0}, {plugin = 0x0, func = 0, state 
>>>= 0},
>>>{plugin = 0x0, func = 0,     state = 0}, {plugin = 0x0, func = 0, 
>>>state = 0}, {plugin = 0x0, func =
>>>0, state = 0}, {plugin = 0x0,     func = 0, state = 0}}
>>>(gdb) n
>>>202                             func_ret = (*(be_table[i].func))(data1,
>>>data2, evl_callback);
>>>(gdb) n
>>>------------Got events.-----------------
>>>Read 95 bytes from pipe
>>>./evlogd: relocation error: /usr/lib/libevl.so.1: undefined symbol: 
>>>__moddi3
>>>203                             if (func_ret != 0) {
>>>(gdb)
>>>
>>>But I don't think it is the evlog bug. It must be some bug related to the
>>>ia64 glibc relocation, because the __moddi3 is defined in glibc and 
>>>uses by
>>>libevl.so.1. My glibc version is 2.2.4-29.2.
>>>
>>>Thanks,
>>>-Zhu Yi
>>>
>>>-----Original Message-----
>>>From: Zhu, Yi [mailto:[email protected]]
>>>Sent: 2003?1?16? 11:39
>>>To: 'Steve Woodruff'; evlog mailing list
>>>Subject: RE: [evlog-dev] -- plugin problem
>>>
>>>
>>>Hi Steve,
>>>
>>>The defined LOG_POSIX_PRINTF v2.4.18 include/linux/evl_log.h is only 
>>>in the
>>>CVS, it is not in 1.5.0 source rpm. See the list before, I submitted a 
>>>patch
>>>to fix the bug.
>>>
>>>I have an IA64 machine with me, I'll try to setup it after lunch. :)
>>>
>>>Thanks,
>>>-Zhu Yi
>>>
>>>
>>>
>>>>-----Original Message-----
>>>>From: Steve Woodruff [mailto:[email protected]]
>>>>Sent: Thursday, January 16, 2003 11:32 AM
>>>>To: evlog mailing list
>>>>Subject: RE: [evlog-dev] -- plugin problem
>>>>
>>>>
>>>>Hi,
>>>>
>>>>Yes - this is a HP RX2600 (IA64).  The __umodsi3
>>>>symbol is in libgcc_s.so -- and it shows up in
>>>>/proc/ksysms.
>>>>
>>>>I didn't grab source from cvs, but rather the
>>>>1.5.0 source rpm from the project site.  You
>>>>say that the v2.4.18 include/linux/evl_log.h
>>>>has LOG_POSIX_PRINTF defined?  Hein, did u
>>>>just add that today?
>>>>
>>>>Thanks!
>>>>
>>>>/steve
>>>>
>>>>
>>>>
>>>>--- "Zhu, Yi" <[email protected]> wrote:
>>>>
>>>>
>>>>>Hi Steve,
>>>>>
>>>>>I use Redhat 8.0 on a SMP P3 machine, I have no
>>>>>problem with running the
>>>>>sample_be code. I can also run the tcp_rmtlog_be and
>>>>>udp_rmtlog_be
>>>>>successfully. I suggest you check out the latest
>>>>>evlog code from cvs, it has
>>>>>all the POSIX_LOG_PRINTF defined in all the kernels'
>>>>>evl_log.h.
>>>>>
>>>>>I did a search for umodsi3 in /usr/lib/libevl.so.1
>>>>>and /lib/i686/libc.so.6
>>>>>using "nm /lib/i686/libc.so.6 |grep umodsi3", but
>>>>>cannot find any.
>>>>>
>>>>>Are you using an IA64?
>>>>>
>>>>>Thanks,
>>>>>-Zhu Yi
>>>>>
>>>>
>>>>
>>>>
>>>>=====
>>>>----------------------------------------------------------
>>>>Steve Woodruff, N9OH @ 42.18N 88.34W, [email protected]
>>>>
>>>>__________________________________________________
>>>>Do you Yahoo!?
>>>>Yahoo! Mail Plus - Powerful. Affordable. Sign up now.
>>>>http://mailplus.yahoo.com
>>>>
>>>>
>>>>-------------------------------------------------------
>>>>This SF.NET email is sponsored by: A Thawte Code Signing Certificate 
>>>>is essential in establishing user confidence by providing assurance 
>>>>of authenticity and code integrity. Download our Free Code Signing 
>>>>guide:
>>>>http://ads.sourceforge.net/cgi-bin/redirect.pl?thaw0028en
>>>>_______________________________________________
>>>>evlog-developers mailing list
>>>>[email protected]
>>>>https://lists.sourceforge.net/lists/listinfo/evlog-developers
>>>>
>>>
>>>
>>>
>>>-------------------------------------------------------
>>>This SF.NET email is sponsored by: A Thawte Code Signing Certificate 
>>>is essential in establishing user confidence by providing assurance of 
>>>authenticity and code integrity. Download our Free Code Signing guide:
>>>http://ads.sourceforge.net/cgi-bin/redirect.pl?thaw0028en
>>>_______________________________________________
>>>evlog-developers mailing list
>>>[email protected]
>>>https://lists.sourceforge.net/lists/listinfo/evlog-developers
>>>
>>>
>>>-------------------------------------------------------
>>>This SF.NET email is sponsored by: Thawte.com
>>>Understand how to protect your customers personal information by 
>>>implementing
>>>SSL on your Apache Web Server. Click here to get our FREE Thawte 
>>>Apache Guide: http://ads.sourceforge.net/cgi-bin/redirect.pl?thaw0029en
>>>_______________________________________________
>>>evlog-developers mailing list
>>>[email protected]
>>>https://lists.sourceforge.net/lists/listinfo/evlog-developers
>>>
>>
>>
>>
>>
>>-------------------------------------------------------
>>This SF.NET email is sponsored by: Thawte.com
>>Understand how to protect your customers personal information by 
>>implementing
>>SSL on your Apache Web Server. Click here to get our FREE Thawte Apache 
>>Guide: http://ads.sourceforge.net/cgi-bin/redirect.pl?thaw0029en
>>_______________________________________________
>>evlog-developers mailing list
>>[email protected]
>>https://lists.sourceforge.net/lists/listinfo/evlog-developers
>>
> 
> 
> 
> -------------------------------------------------------
> This SF.NET email is sponsored by: Thawte.com
> Understand how to protect your customers personal information by implementing
> SSL on your Apache Web Server. Click here to get our FREE Thawte Apache 
> Guide: http://ads.sourceforge.net/cgi-bin/redirect.pl?thaw0029en
> _______________________________________________
> evlog-developers mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/evlog-developers
> 




-------------------------------------------------------
This SF.NET email is sponsored by: Thawte.com - A 128-bit supercerts will
allow you to extend the highest allowed 128 bit encryption to all your 
clients even if they use browsers that are limited to 40 bit encryption. 
Get a guide here:http://ads.sourceforge.net/cgi-bin/redirect.pl?thaw0030en