[core] add comment for FastCGI mem use in hctx->rb (#3033)
x-ref: "Memory Growth with PUT and full buffered streams" https://redmine.lighttpd.net/issues/3033personal/stbuehler/tests-path
parent
7ba521ffb4
commit
ceab1e700d
|
@ -2052,6 +2052,11 @@ handler_t gw_handle_subrequest(request_st * const r, void *p_d) {
|
|||
if ((r->conf.stream_response_body & FDEVENT_STREAM_RESPONSE_BUFMIN)
|
||||
&& r->resp_body_started) {
|
||||
if (chunkqueue_length(&r->write_queue) > 65536 - 4096) {
|
||||
/* Note: if apps inheriting gw_handle use hctx->rb, then those apps
|
||||
* are responsible for limiting amount of data buffered in memory
|
||||
* in hctx->rb. Currently, mod_fastcgi is the only core app doing
|
||||
* so, and the maximum FCGI_Record size is 8 + 65535 + 255 = 65798
|
||||
* (FCGI_HEADER_LEN(8)+contentLength(65535)+paddingLength(255)) */
|
||||
fdevent_fdnode_event_clr(hctx->ev, hctx->fdn, FDEVENT_IN);
|
||||
}
|
||||
else if (!(fdevent_fdnode_interest(hctx->fdn) & FDEVENT_IN)) {
|
||||
|
|
Loading…
Reference in New Issue