DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG· RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT <http://issues.apache.org/bugzilla/show_bug.cgi?id=38403>. ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND· INSERTED IN THE BUG DATABASE.
http://issues.apache.org/bugzilla/show_bug.cgi?id=38403 ------- Additional Comments From [EMAIL PROTECTED] 2006-02-04 12:44 ------- (In reply to comment #11) Thanks for the update. > > ap_core_input_filter (f=0x195740b8, b=0x9feddc8, mode=AP_MODE_GETLINE, > block=APR_BLOCK_READ, readbytes=0) at core_filters.c:141 > 141 BRIGADE_NORMALIZE(ctx->b); > > > code says this: > > /* ### This is bad. */ > BRIGADE_NORMALIZE(ctx->b); > > > but I'm not sure whether "this is bad" refers to a possible bug or just a > performance issue. > > Is there a simple way to inspect the BRIGADE in gdb ? I guess there's a > special Yes, there is a helpful macro to inspect a brigade in gdb (dump_brigade). Please have a look at the end of the section of http://httpd.apache.org/dev/debugging.html#gdb BRIGADE_NORMALIZE is a macro that does a loop. May it would be helpful to replace BRIGADE_NORMALIZE(ctx->b); with the expanded macro to see if the loop is never left. do { apr_bucket *e = APR_BRIGADE_FIRST(ctx->b); do { if (e->length == 0 && !APR_BUCKET_IS_METADATA(e)) { apr_bucket *d; d = APR_BUCKET_NEXT(e); apr_bucket_delete(e); e = d; } else { e = APR_BUCKET_NEXT(e); } } while (!APR_BRIGADE_EMPTY(ctx->b) && (e != APR_BRIGADE_SENTINEL(ctx->b))); } while (0) Furthermore it would be better to use step / next instead of stepi to check where the code circles. stepi only executes one assembler instruction, whereas step executes one line of C code. In your initial comment to this bug you mentioned that a thread is looping with close(-1). Would it be possible to get a backtrace of this thread? So I would recommend the following next steps: 1. Expand the BRIGADE_NORMALIZE macro and recompile. 2. If the error occurs again do a dump_brigade ctx->b and attach the output 3. Check with if the BRIGADE_NORMALIZE loop is left. For this set a breakpoint after the loop and cont. If you reach it the loop is left if not we circle in the loop. 4. Try to get a backtrace of the thread that loops with close(-1) -- Configure bugmail: http://issues.apache.org/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee. --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
