You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@httpd.apache.org by Ryan Bloom <rb...@covalent.net> on 2002/04/02 06:44:35 UTC
Bump a few more tags. :-)
In httpd-2.0/server/mpm/perchild
mpm.h --> version 1.16
perchild.c --> version 1.120
Ryan
----------------------------------------------
Ryan Bloom rbb@covalent.net
645 Howard St. rbb@apache.org
San Francisco, CA
Re: Bump a few more tags. :-)
Posted by Cliff Woolley <jw...@virginia.edu>.
On Mon, 1 Apr 2002, Doug MacEachern wrote:
> nope. sounds specific to perchild, cliff does the proxy test pass for
> you with prefork and/or worker?
It does appear to be perchild-specific. It works on both prefork and
worker. Just the usual from those two: all tests passed except
ssl/http.t, which is a broken test.
--Cliff
--------------------------------------------------------------
Cliff Woolley
cliffwoolley@yahoo.com
Charlottesville, VA
RE: Bump a few more tags. :-)
Posted by Ryan Bloom <rb...@covalent.net>.
> On Mon, 1 Apr 2002, William A. Rowe, Jr. wrote:
>
> > Sounds like that could be Doug's latest changes he asked to
incorporate.
> >
> > Doug, was there an additional file to bump, beyond the three you
cited?
>
> nope. sounds specific to perchild, cliff does the proxy test pass for
> you with prefork and/or worker? i can take a look in the morning.
Perchild does some hokey stuff with sockets. The MPM is still
experimental. I am hoping to have more time with it next weekend again.
Ryan
Re: Bump a few more tags. :-)
Posted by Doug MacEachern <do...@covalent.net>.
On Mon, 1 Apr 2002, William A. Rowe, Jr. wrote:
> Sounds like that could be Doug's latest changes he asked to incorporate.
>
> Doug, was there an additional file to bump, beyond the three you cited?
nope. sounds specific to perchild, cliff does the proxy test pass for
you with prefork and/or worker? i can take a look in the morning.
Re: Bump a few more tags. :-)
Posted by "William A. Rowe, Jr." <wr...@rowe-clan.net>.
Sounds like that could be Doug's latest changes he asked to incorporate.
Doug, was there an additional file to bump, beyond the three you cited?
Bill
At 11:09 PM 4/1/2002, you wrote:
>On Mon, 1 Apr 2002, William A. Rowe, Jr. wrote:
>
> > Both are bumped... anyone else care to cvs up -r APACHE_2_0_34
> > and verify that perchild builds?
>
>perchild.c: In function `worker_thread':
>perchild.c:693: warning: variable `lr' might be clobbered by `longjmp' or
>`vfork'
>perchild.c:693: warning: variable `last_lr' might be clobbered by
>`longjmp' or `vfork'
>
>All tests pass except ssl/proxy.t. That one gets stuck here:
>
>(gdb) bt
>#0 0x403a89be in select () from /lib/libc.so.6
>#1 0x4004223c in __DTOR_END__ () from /root/apache/test/lib/libapr.so.0
>#2 0x400362f2 in apr_recv (sock=0x81cc978,
> buf=0x8274c70
>"pL?@pL?@V]nzw{6u\016[42C\205*\021\177\026\222B\223ks\236x\001\225NLM\025\023~A\r\031p\002wS\f/_V\225f\232\025\034`\004A\210i\016\005\027\207\002XtUW%\027a\\Irm3VGlC\e{9\215\217gQa-\222BV\223\fpIV\036?\ak^kj\0061hE#4\236\201\231|\231\234\216\025\e\216W\031U\232|!\216j\213mV\222}%N5"...,
>len=0xbf7fd5a0)
> at sendrecv.c:142
>#3 0x4001b14f in socket_read (a=0x827a000, str=0xbf7fd59c,
>len=0xbf7fd5a0,
> block=1127800088) at apr_buckets_socket.c:74
>#4 0x80d033b in core_input_filter (f=0x823a950, b=0x823a910,
> mode=AP_MODE_READBYTES, block=1127800088, readbytes=7) at core.c:3430
>#5 0x80c9c47 in ap_get_brigade (next=0x823a950, bb=0x823a910,
> mode=AP_MODE_READBYTES, block=1127800088, readbytes=7) at
>util_filter.c:507
>#6 0x8087cb0 in bio_bucket_in_read (bio=0x8226e00,
> in=0x826c460 "H?@H?@\005];\\~xo\a6$\031D\b \223", inl=7)
> at ssl_engine_io.c:395
>#7 0x400b2661 in BIO_read () from /usr/lib/libssl.so.0
>Cannot access memory at address 0x7
>
>
>But still, this is major progress. +1 from me!
>
>--Cliff
>
>--------------------------------------------------------------
> Cliff Woolley
> cliffwoolley@yahoo.com
> Charlottesville, VA
Re: bio_bucket_in_read bug [was Re: Bump a few more tags. :-)]
Posted by Cliff Woolley <jw...@virginia.edu>.
On Mon, 1 Apr 2002, Doug MacEachern wrote:
> your grep pattern missed this one:
> ctx->inbio.block = block;
Doh! :)
Regardless, the value of ->block that's getting passed in is fried. Some
of the others might be too, I'll have to go back and poke through gdb more
carefully.
Right now I'm running httpd-test against worker and prefork.
--------------------------------------------------------------
Cliff Woolley
cliffwoolley@yahoo.com
Charlottesville, VA
Re: bio_bucket_in_read bug [was Re: Bump a few more tags. :-)]
Posted by Doug MacEachern <do...@covalent.net>.
your grep pattern missed this one:
ctx->inbio.block = block;
bio_bucket_in_read bug [was Re: Bump a few more tags. :-)]
Posted by Cliff Woolley <jw...@virginia.edu>.
On Tue, 2 Apr 2002, Cliff Woolley wrote:
> #3 0x4001b14f in socket_read (a=0x827a000, str=0xbf7fd59c,
> len=0xbf7fd5a0,
> block=1127800088) at apr_buckets_socket.c:74
> #4 0x80d033b in core_input_filter (f=0x823a950, b=0x823a910,
> mode=AP_MODE_READBYTES, block=1127800088, readbytes=7) at core.c:3430
> #5 0x80c9c47 in ap_get_brigade (next=0x823a950, bb=0x823a910,
> mode=AP_MODE_READBYTES, block=1127800088, readbytes=7) at
> util_filter.c:507
> #6 0x8087cb0 in bio_bucket_in_read (bio=0x8226e00,
> in=0x826c460 "H?@H?@\005];\\~xo\a6$\031D\b \223", inl=7)
> at ssl_engine_io.c:395
> #7 0x400b2661 in BIO_read () from /usr/lib/libssl.so.0
> Cannot access memory at address 0x7
There appears to be a bug in bio_bucket_in_read: it passes inbio->block as
the block parameter to ap_get_brigade(), but inbio->block is
uninitialized:
/root/apache/httpd-2.0.34/modules/ssl# grep -- "->block" *.[ch]
ssl_engine_io.c: AP_MODE_READBYTES, inbio->block,
ssl_engine_io.c: &buf, &buf_len, inbio->block);
--Cliff
--------------------------------------------------------------
Cliff Woolley
cliffwoolley@yahoo.com
Charlottesville, VA
Re: Bump a few more tags. :-)
Posted by Cliff Woolley <jw...@virginia.edu>.
On Mon, 1 Apr 2002, William A. Rowe, Jr. wrote:
> Both are bumped... anyone else care to cvs up -r APACHE_2_0_34
> and verify that perchild builds?
perchild.c: In function `worker_thread':
perchild.c:693: warning: variable `lr' might be clobbered by `longjmp' or
`vfork'
perchild.c:693: warning: variable `last_lr' might be clobbered by
`longjmp' or `vfork'
All tests pass except ssl/proxy.t. That one gets stuck here:
(gdb) bt
#0 0x403a89be in select () from /lib/libc.so.6
#1 0x4004223c in __DTOR_END__ () from /root/apache/test/lib/libapr.so.0
#2 0x400362f2 in apr_recv (sock=0x81cc978,
buf=0x8274c70
"pL?@pL?@V]nzw{6u\016[42C\205*\021\177\026\222B\223ks\236x\001\225NLM\025\023~A\r\031p\002wS\f/_V\225f\232\025\034`\004A\210i\016\005\027\207\002XtUW%\027a\\Irm3VGlC\e{9\215\217gQa-\222BV\223\fpIV\036?\ak^kj\0061hE#4\236\201\231|\231\234\216\025\e\216W\031U\232|!\216j\213mV\222}%N5"...,
len=0xbf7fd5a0)
at sendrecv.c:142
#3 0x4001b14f in socket_read (a=0x827a000, str=0xbf7fd59c,
len=0xbf7fd5a0,
block=1127800088) at apr_buckets_socket.c:74
#4 0x80d033b in core_input_filter (f=0x823a950, b=0x823a910,
mode=AP_MODE_READBYTES, block=1127800088, readbytes=7) at core.c:3430
#5 0x80c9c47 in ap_get_brigade (next=0x823a950, bb=0x823a910,
mode=AP_MODE_READBYTES, block=1127800088, readbytes=7) at
util_filter.c:507
#6 0x8087cb0 in bio_bucket_in_read (bio=0x8226e00,
in=0x826c460 "H?@H?@\005];\\~xo\a6$\031D\b \223", inl=7)
at ssl_engine_io.c:395
#7 0x400b2661 in BIO_read () from /usr/lib/libssl.so.0
Cannot access memory at address 0x7
But still, this is major progress. +1 from me!
--Cliff
--------------------------------------------------------------
Cliff Woolley
cliffwoolley@yahoo.com
Charlottesville, VA
Re: Bump a few more tags. :-)
Posted by "William A. Rowe, Jr." <wr...@rowe-clan.net>.
At 10:44 PM 4/1/2002, you wrote:
>In httpd-2.0/server/mpm/perchild
>
>mpm.h --> version 1.16
>
>perchild.c --> version 1.120
Both are bumped... anyone else care to cvs up -r APACHE_2_0_34
and verify that perchild builds?
Bill