httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "MATHIHALLI,MADHUSUDAN (HP-Cupertino,ex1)" <mad...@hp.com>
Subject RE: Help with DAV SEGV
Date Tue, 10 Jun 2003 23:33:27 GMT
I think we've identified the problem (.. and a solution too). The problem
was because of a faulty compiler.

For those interested in knowing more about the problem : the bug shows up
when we forward declare a structure, and the compiler doesn't manage the
literals properly.

The following patch helped fix the segv problem (but, the real fix is in the
compiler)

--- repos.c@@/main/5    Tue Apr  1 17:55:55 2003
+++ repos.c     Tue Jun 10 16:21:30 2003
@@ -134,8 +134,8 @@
 extern const dav_hooks_locks dav_hooks_locks_fs;
 
 /* forward-declare the hook structures */
-static const dav_hooks_repository dav_hooks_repository_fs;
-static const dav_hooks_liveprop dav_hooks_liveprop_fs;
+static dav_hooks_repository dav_hooks_repository_fs;
+static dav_hooks_liveprop dav_hooks_liveprop_fs;
 
 /*
 ** The namespace URIs that we use. This list and the enumeration must
@@ -1795,7 +1795,7 @@
     return apr_psprintf(ctx->pool, "\"%lx\"", (unsigned long)
ctx->finfo.mtime);
 }
 
-static const dav_hooks_repository dav_hooks_repository_fs =
+static dav_hooks_repository dav_hooks_repository_fs =
 {
     DEBUG_GET_HANDLER,   /* normally: special GET handling not required */
     dav_fs_get_resource,
@@ -2080,7 +2080,7 @@
 }
 
 
-static const dav_hooks_liveprop dav_hooks_liveprop_fs =
+static dav_hooks_liveprop dav_hooks_liveprop_fs =
 {
     dav_fs_insert_prop,
     dav_fs_is_writable,


Thanks
-Madhu

>-----Original Message-----
>From: MATHIHALLI,MADHUSUDAN (HP-Cupertino,ex1) [mailto:madhum@hp.com]
>Sent: Tuesday, June 10, 2003 2:42 PM
>To: 'dev@httpd.apache.org'
>Subject: Help with DAV SEGV
>
>
>Hi,
>	I'm running the litmus test for WebDAV (with Apache 2.0.46), and
>seeing a strange behaviour. The tests pass for 64-bit (PA-RISC 
>arch., OS Ver
>11.23 / 11.11), 64-bit (IA64 arch., OS Ver 11.22), but dumps 
>core on 64-bit
>(IA64 OS Ver 11.23). (The same tests were passing with 2.0.45)
>
>Any idea what might be happening ?.
>
>(gdb) where
>#0  0x0 in <unknown_procedure> ()
>warning: Attempting to unwind past bad PC 0x0 
>#1  0xc000000001e4ac90:0 in dav_get_resource (r=0x60000000001ebf50, 
>    label_allowed=0, use_checked_in=0, res_p=0x8fffffffef5f8f50)
>    at mod_dav.c:705
>#2  0xc000000001e4c6b0:0 in dav_method_delete (r=0x60000000001ebf50)
>    at mod_dav.c:1086
>#3  0xc000000001e5c7a0:0 in dav_handler (r=0x60000000001ebf50)
>    at mod_dav.c:4512
>#4  0x400000000005fe80:0 in ap_run_handler+0x120 ()
>#5  0x4000000000061560:0 in ap_invoke_handler+0x2f0 ()
>#6  0x4000000000051d80:0 in ap_process_request+0x90 ()
>#7  0x40000000000429b0:0 in ap_process_http_connection+0x100 ()
>#8  0x4000000000081230:0 in ap_run_process_connection+0x120 ()
>#9  0x4000000000081d50:0 in ap_process_connection+0xf0 ()
>#10 0x4000000000055740:0 in process_socket+0x1d0 ()
>#11 0x4000000000056d80:0 in worker_thread+0x410 ()
>#12 0xc000000001cb8480:0 in dummy_worker+0x70 ()
>   from /opt/hpws/apache/lib/libapr-0.9
>#13 0xc0000000000b4ae0:0 in __pthread_unbound_body+0x4a0 ()
>   from /usr/lib/hpux64/libpthread.so.1
>(gdb) p *conf
>$1 = {provider_name = 0xc000000001e43440 "filesystem", 
>  provider = 0x8fffffffef7213f0, 
>  dir = 0x6000000000072538 "/opt/hpws/apache/htdocs", locktimeout = 0, 
>  allow_depthinfinity = 0}
>(gdb) p *conf->provider
>$2 = {repos = 0xc000000001ea9610, propdb = 0x8fffffffef721120, 
>  locks = 0x8fffffffef721290, vsn = 0x0, binding = 0x0, search = 0x0, 
>  ctx = 0x0}
>(gdb) p *conf->provider->repos
>$3 = {handle_get = 0, get_resource = 0, get_parent_resource = 0, 
>  is_same_resource = 0, is_parent_resource = 0, open_stream = 0, 
>  close_stream = 0, write_stream = 0, seek_stream = 0, 
>set_headers = 0, 
>  deliver = 0, create_collection = 0, copy_resource = 0, 
>move_resource = 0, 
>  remove_resource = 0, walk = 0, getetag = 0, ctx = 0x0}
>

Mime
View raw message