MINOR: cache: Process the If-Modified-Since header in conditional requests

If a client sends a conditional request containing an If-Modified-Since
header (and no If-None-Match header), we try to compare the date with
the one stored in the cache entry (coming either from a Last-Modified
head, or a Date header, or corresponding to the first response's
reception time). If the request's date is earlier than the stored one,
we send a "304 Not Modified" response back. Otherwise, the stored is sent
(through a 200 OK response).

This resolves GitHub issue #821.

reg-tests/cache/if-modified-since.vtc

@@ -0,0 +1,145 @@
+varnishtest "If-Modified-Since support"
+
+#REQUIRE_VERSION=2.3
+
+feature ignore_unknown_macro
+
+server s1 {
+       # Response containing a "Last-Modified" field
+       rxreq
+       expect req.url == "/last_modified"
+       txresp -nolen -hdr "Transfer-Encoding: chunked" \
+               -hdr "Last-Modified: Thu, 15 Oct 2020 22:23:24 GMT"
+       chunkedlen 15
+       chunkedlen 15
+       chunkedlen 15
+       chunkedlen 0
+
+       # Response containing a "Date" field
+       rxreq
+       expect req.url == "/date"
+       txresp -nolen -hdr "Transfer-Encoding: chunked" \
+               -hdr "Date: Thu, 22 Oct 2020 16:51:12 GMT"
+       chunkedlen 16
+       chunkedlen 16
+       chunkedlen 16
+       chunkedlen 0
+
+       # Response containing both a "Last-Modified" and a "Date" fields
+       # Should behave the same way as if the "Date" field was not here.
+       rxreq
+       expect req.url == "/last_modified_and_date"
+       txresp -nolen -hdr "Transfer-Encoding: chunked" \
+               -hdr "Last-Modified: Thu, 15 Oct 2020 14:24:38 GMT" \
+               -hdr "Date: Thu, 22 Oct 2020 16:51:12 GMT"
+       chunkedlen 17
+       chunkedlen 17
+       chunkedlen 17
+       chunkedlen 0
+} -start
+
+haproxy h1 -conf {
+       defaults
+               mode http
+               ${no-htx} option http-use-htx
+               timeout connect 1s
+               timeout client  1s
+               timeout server  1s
+
+       frontend fe
+               bind "fd@${fe}"
+               default_backend test
+
+       backend test
+               http-request cache-use my_cache
+               server www ${s1_addr}:${s1_port}
+               http-response cache-store my_cache
+
+       cache my_cache
+               total-max-size 3
+               max-age 20
+               max-object-size 3072
+} -start
+
+
+client c1 -connect ${h1_fe_sock} {
+       txreq -url "/last_modified"
+       rxresp
+       expect resp.status == 200
+       expect resp.bodylen == 45
+
+       txreq -url "/date"
+       rxresp
+       expect resp.status == 200
+       expect resp.bodylen == 48
+
+       txreq -url "/last_modified_and_date"
+       rxresp
+       expect resp.status == 200
+       expect resp.bodylen == 51
+
+
+       # Earlier date
+       # "Last-Modified" version
+       txreq -url "/last_modified" \
+              -hdr "If-Modified-Since: Thu, 15 Oct 2020 00:00:01 GMT"
+       rxresp
+       expect resp.status == 200
+       expect resp.bodylen == 45
+       # "Date" version
+       txreq -url "/date" \
+              -hdr "If-Modified-Since: Thu, 01 Oct 2020 00:00:01 GMT"
+       rxresp
+       expect resp.status == 200
+       expect resp.bodylen == 48
+       # "Last-Modified" and "Date" version
+       txreq -url "/last_modified_and_date" \
+              -hdr "If-Modified-Since: Thu, 15 Oct 2020 00:00:01 GMT"
+       rxresp
+       expect resp.status == 200
+       expect resp.bodylen == 51
+
+
+       # Same date
+       # "Last-Modified" version
+       txreq -url "/last_modified" \
+              -hdr "If-Modified-Since: Thu, 15 Oct 2020 22:23:24 GMT"
+       rxresp
+       expect resp.status == 304
+       expect resp.bodylen == 0
+       # "Date" version
+       txreq -url "/date" \
+              -hdr "If-Modified-Since: Thu, 22 Oct 2020 16:51:12 GMT"
+       rxresp
+       expect resp.status == 304
+       expect resp.bodylen == 0
+       # "Last-Modified" and "Date" version
+       txreq -url "/last_modified_and_date" \
+              -hdr "If-Modified-Since: Thu, 15 Oct 2020 16:51:12 GMT"
+       rxresp
+       expect resp.status == 304
+       expect resp.bodylen == 0
+
+
+       # Later date
+       # "Last-Modified" version
+       txreq -url "/last_modified" \
+              -hdr "If-Modified-Since: Thu, 22 Oct 2020 23:00:00 GMT"
+       rxresp
+       expect resp.status == 304
+       expect resp.bodylen == 0
+       # "Date" version
+       txreq -url "/date" \
+              -hdr "If-Modified-Since: Thu, 22 Oct 2020 23:00:00 GMT"
+       rxresp
+       expect resp.status == 304
+       expect resp.bodylen == 0
+       # "Last-Modified" and "Date" version
+       txreq -url "/last_modified_and_date" \
+              -hdr "If-Modified-Since: Thu, 22 Oct 2020 23:00:00 GMT"
+       rxresp
+       expect resp.status == 304
+       expect resp.bodylen == 0
+
+} -run
+

src/cache.c

@@ -1183,7 +1183,13 @@ int sha1_hosturi(struct stream *s)
  * matches, a "304 Not Modified" response should be sent instead of the cached
  * data.
  * Although unlikely in a GET/HEAD request, the "If-None-Match: *" syntax is
- * valid and should receive a "304 Not Modified" response (RFC 7434#4.3.2).
+ * valid and should receive a "304 Not Modified" response (RFC 7234#4.3.2).
+ *
+ * If no "If-None-Match" header was found, look for an "If-Modified-Since"
+ * header and compare its value (date) to the one stored in the cache_entry.
+ * If the request's date is later than the cached one, we also send a
+ * "304 Not Modified" response (see RFCs 7232#3.3 and 7234#4.3.2).
+ *
  * Returns 1 if "304 Not Modified" should be sent, 0 otherwise.
  */
 static int should_send_notmodified_response(struct cache *cache, struct htx *htx,
@@ -1194,14 +1200,16 @@ static int should_send_notmodified_response(struct cache *cache, struct htx *htx
 	struct http_hdr_ctx ctx = { .blk = NULL };
 	struct ist cache_entry_etag = IST_NULL;
 	struct buffer *etag_buffer = NULL;
+	int if_none_match_found = 0;
 
-	if (entry->etag_length == 0)
+	struct tm tm = {};
-		return 0;
+	time_t if_modified_since = 0;
 
 	/* If we find a "If-None-Match" header in the request, rebuild the
-	 * cache_entry's ETag in order to perform comparisons. */
+	 * cache_entry's ETag in order to perform comparisons.
-	/* There could be multiple "if-none-match" header lines. */
+	 * There could be multiple "if-none-match" header lines. */
 	while (http_find_header(htx, ist("if-none-match"), &ctx, 0)) {
+		if_none_match_found = 1;
 
 		/* A '*' matches everything. */
 		if (isteq(ctx.value, ist("*")) != 0) {
@@ -1209,6 +1217,10 @@ static int should_send_notmodified_response(struct cache *cache, struct htx *htx
 			break;
 		}
 
+		/* No need to rebuild an etag if none was stored in the cache. */
+		if (entry->etag_length == 0)
+			break;
+
 		/* Rebuild the stored ETag. */
 		if (etag_buffer == NULL) {
 			etag_buffer = get_trash_chunk();
@@ -1230,6 +1242,23 @@ static int should_send_notmodified_response(struct cache *cache, struct htx *htx
 		}
 	}
 
+	/* If the request did not contain an "If-None-Match" header, we look for
+	 * an "If-Modified-Since" header (see RFC 7232#3.3). */
+	if (retval == 0 && if_none_match_found == 0) {
+		ctx.blk = NULL;
+		if (http_find_header(htx, ist("if-modified-since"), &ctx, 1)) {
+			if (parse_http_date(istptr(ctx.value), istlen(ctx.value), &tm)) {
+				if_modified_since = my_timegm(&tm);
+
+				/* We send a "304 Not Modified" response if the
+				 * entry's last modified date is earlier than
+				 * the one found in the "If-Modified-Since"
+				 * header. */
+				retval = (entry->last_modified <= if_modified_since);
+			}
+		}
+	}
+
 	return retval;
 }