Skip to content

Commit 6953a90

Browse files
squidadmsquidadm
committed
2026-04-03
1 parent a94b156 commit 6953a90

9 files changed

Lines changed: 529 additions & 19 deletions

File tree

‎squid-dev/2026-April.txt‎

Lines changed: 170 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -163,3 +163,173 @@ Best regards,
163163
Anthony
164164

165165

166+
From anthony.pankov at yahoo.com Fri Apr 3 14:21:05 2026
167+
From: anthony.pankov at yahoo.com (Anthony Pankov)
168+
Date: Fri, 3 Apr 2026 17:21:05 +0300
169+
Subject: [squid-dev] forward bumped traffic to parent in plain form
170+
In-Reply-To: <1805759671.20260403144632@yahoo.com>
171+
References: <1985119311.20260304190328.ref@yahoo.com>
172+
<1985119311.20260304190328@yahoo.com>
173+
<70faedf9-2e5d-4934-b837-874940f81a4b@measurement-factory.com>
174+
<1478043664.20260305122659@yahoo.com>
175+
<30a55c20-f80b-4a5f-be94-352a6a567a9c@measurement-factory.com>
176+
<38965228.20260306181359@yahoo.com>
177+
<a87212b4-41bb-4d03-95cf-cb4704d45897@measurement-factory.com>
178+
<1736177688.20260308131738@yahoo.com>
179+
<878e2ea7-2dda-410a-8bdf-1fabd429315e@measurement-factory.com>
180+
<991408460.20260310135102@yahoo.com>
181+
<f867539b-0550-43c6-9aa0-d74e8758cea6@measurement-factory.com>
182+
<1493984953.20260310175054@yahoo.com>
183+
<a728e3a6-ccc6-401d-bef4-355526f16f9a@measurement-factory.com>
184+
<803480691.20260326151406@yahoo.com>
185+
<50390f1b-105f-4384-ba35-b6a9e37a7684@measurement-factory.com>
186+
<1332811842.20260331193229@yahoo.com>
187+
<63b3703d-c87a-47ef-9e26-98b0ed4e4d18@measurement-factory.com>
188+
<1832116129.20260403142017@yahoo.com> <1805759671.20260403144632@yahoo.com>
189+
Message-ID: <562184261.20260403172105@yahoo.com>
190+
191+
192+
Friday, April 3, 2026, 2:46:32 PM, you wrote:
193+
194+
195+
> Friday, April 3, 2026, 2:20:17 PM, you wrote:
196+
197+
198+
>> I've tried to call getSslContextStart() in startPeakAndSplice().
199+
200+
>> But the problem is:
201+
>> getSslContextStart() return synchronously in most cases after calling getSslContextDone().
202+
203+
>> But in the case of #if USE_SSL_CRTD (my case) it return after sending request to CRTD without calling getSslContextDone().
204+
>> getSslContextDone() will be called further in sslCrtdHandleReplyWrapper().
205+
206+
>> So we are back in startPeakAndSplice() without getSslContextDone() and connections are closed.
207+
208+
>> Is there a preferred way to solve this problem?
209+
210+
> Sorry. It's a normal. The problem is out here.
211+
212+
213+
May be not.
214+
215+
I notice
216+
217+
26/04/03 16:54:33.312 kid1| 83,5| /client_side.cc(2993) startPeekAndSplice: Bumping on step2 done.
218+
2026/04/03 16:54:33.312 kid1| 33,5| src/base/AsyncJob.cc(177) callEnd: Http1::Server status out: [ job4]
219+
-->2026/04/03 16:54:33.312 kid1| 33,5| src/base/AsyncCallQueue.cc(28) fire: leaving Server::doClientRead(conn12
220+
local=178.218.213.200:12345 remote=217.119.19.66:63055 FD 9 flags=1, data=0x2b2273397f18)
221+
2026/04/03 16:54:33.341 kid1| 83,5| src/ssl/helper.cc(161) HandleGeneratorReply: to 0x2b2273397f18 in crtGenRq6569/1
222+
223+
and suggest that stack rewind out of Server::doClientRead() while waiting for HandleGeneratorReply. ?And connection somehow lost because next read was not triggered?
224+
I'd like to wait inside startPeekAndSplice() for ?clientNegotiateSSL()? to finish.
225+
But I don't want to introduce own "wait objects".
226+
227+
May be somebody can advice me.
228+
229+
230+
231+
--
232+
Best regards,
233+
Anthony
234+
235+
236+
From rousskov at measurement-factory.com Fri Apr 3 15:47:26 2026
237+
From: rousskov at measurement-factory.com (Alex Rousskov)
238+
Date: Fri, 3 Apr 2026 11:47:26 -0400
239+
Subject: [squid-dev] forward bumped traffic to parent in plain form
240+
In-Reply-To: <1832116129.20260403142017@yahoo.com>
241+
References: <1985119311.20260304190328.ref@yahoo.com>
242+
<1985119311.20260304190328@yahoo.com>
243+
<70faedf9-2e5d-4934-b837-874940f81a4b@measurement-factory.com>
244+
<1478043664.20260305122659@yahoo.com>
245+
<30a55c20-f80b-4a5f-be94-352a6a567a9c@measurement-factory.com>
246+
<38965228.20260306181359@yahoo.com>
247+
<a87212b4-41bb-4d03-95cf-cb4704d45897@measurement-factory.com>
248+
<1736177688.20260308131738@yahoo.com>
249+
<878e2ea7-2dda-410a-8bdf-1fabd429315e@measurement-factory.com>
250+
<991408460.20260310135102@yahoo.com>
251+
<f867539b-0550-43c6-9aa0-d74e8758cea6@measurement-factory.com>
252+
<1493984953.20260310175054@yahoo.com>
253+
<a728e3a6-ccc6-401d-bef4-355526f16f9a@measurement-factory.com>
254+
<803480691.20260326151406@yahoo.com>
255+
<50390f1b-105f-4384-ba35-b6a9e37a7684@measurement-factory.com>
256+
<1332811842.20260331193229@yahoo.com>
257+
<63b3703d-c87a-47ef-9e26-98b0ed4e4d18@measurement-factory.com>
258+
<1832116129.20260403142017@yahoo.com>
259+
Message-ID: <7785cda9-25eb-4eb7-bac7-0d5d6b82a4ad@measurement-factory.com>
260+
261+
On 2026-04-03 07:20, Anthony Pankov wrote:
262+
>
263+
> I've tried to call getSslContextStart() in startPeakAndSplice().
264+
>
265+
> But the problem is:
266+
> getSslContextStart() return synchronously in most cases after calling getSslContextDone().
267+
>
268+
> But in the case of #if USE_SSL_CRTD (my case) it return after sending request to CRTD without calling getSslContextDone().
269+
> getSslContextDone() will be called further in sslCrtdHandleReplyWrapper().
270+
>
271+
> So we are back in startPeakAndSplice() without getSslContextDone() and connections are closed.
272+
>
273+
> Is there a preferred way to solve this problem?
274+
275+
AFAICT, in your case, startPeekAndSplice() should do nothing when the
276+
added getSslContextStart() call returns. From
277+
https://ml-archives.squid-cache.org/squid-dev/2026-March/009992.html :
278+
getSslContextStart() should eventually trigger a call to
279+
getSslContextDone() with a non-nil Security::ContextPointer that calls
280+
httpsCreate() with the transaction-specific security context (SNI, etc.).
281+
282+
That eventual httpsCreate() call should advance the transaction further.
283+
284+
I am sure more changes will be needed.
285+
286+
287+
HTH,
288+
289+
Alex.
290+
291+
292+
293+
From rousskov at measurement-factory.com Fri Apr 3 15:51:08 2026
294+
From: rousskov at measurement-factory.com (Alex Rousskov)
295+
Date: Fri, 3 Apr 2026 11:51:08 -0400
296+
Subject: [squid-dev] forward bumped traffic to parent in plain form
297+
In-Reply-To: <562184261.20260403172105@yahoo.com>
298+
References: <1985119311.20260304190328.ref@yahoo.com>
299+
<1985119311.20260304190328@yahoo.com>
300+
<70faedf9-2e5d-4934-b837-874940f81a4b@measurement-factory.com>
301+
<1478043664.20260305122659@yahoo.com>
302+
<30a55c20-f80b-4a5f-be94-352a6a567a9c@measurement-factory.com>
303+
<38965228.20260306181359@yahoo.com>
304+
<a87212b4-41bb-4d03-95cf-cb4704d45897@measurement-factory.com>
305+
<1736177688.20260308131738@yahoo.com>
306+
<878e2ea7-2dda-410a-8bdf-1fabd429315e@measurement-factory.com>
307+
<991408460.20260310135102@yahoo.com>
308+
<f867539b-0550-43c6-9aa0-d74e8758cea6@measurement-factory.com>
309+
<1493984953.20260310175054@yahoo.com>
310+
<a728e3a6-ccc6-401d-bef4-355526f16f9a@measurement-factory.com>
311+
<803480691.20260326151406@yahoo.com>
312+
<50390f1b-105f-4384-ba35-b6a9e37a7684@measurement-factory.com>
313+
<1332811842.20260331193229@yahoo.com>
314+
<63b3703d-c87a-47ef-9e26-98b0ed4e4d18@measurement-factory.com>
315+
<1832116129.20260403142017@yahoo.com> <1805759671.20260403144632@yahoo.com>
316+
<562184261.20260403172105@yahoo.com>
317+
Message-ID: <b32244e9-752b-47d9-a6f7-5b14b5767b6b@measurement-factory.com>
318+
319+
On 2026-04-03 10:21, Anthony Pankov wrote:
320+
321+
> I'd like to wait inside startPeekAndSplice() for ?clientNegotiateSSL()? to finish.
322+
> But I don't want to introduce own "wait objects".
323+
324+
Squid code does not normally block to wait for asynchronous operations
325+
(e.g., communication with the helper) to finish. I assume that existing
326+
code already handles asynchronous helper communication correctly, so I
327+
am not sure why you want Squid to block and wait inside
328+
startPeekAndSplice().
329+
330+
Alex.
331+
332+
333+
334+
335+

‎squid-dev/2026-April/010001.html‎

Lines changed: 6 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -13,7 +13,7 @@
1313
</style>
1414
<META http-equiv="Content-Type" content="text/html; charset=us-ascii">
1515
<LINK REL="Previous" HREF="010000.html">
16-
16+
<LINK REL="Next" HREF="010002.html">
1717
</HEAD>
1818
<BODY BGCOLOR="#ffffff">
1919
<H1>[squid-dev] forward bumped traffic to parent in plain form</H1>
@@ -25,7 +25,8 @@ <H1>[squid-dev] forward bumped traffic to parent in plain form</H1>
2525
<P><UL>
2626
<LI>Previous message (by thread): <A HREF="010000.html">[squid-dev] forward bumped traffic to parent in plain form
2727
</A></li>
28-
28+
<LI>Next message (by thread): <A HREF="010002.html">[squid-dev] forward bumped traffic to parent in plain form
29+
</A></li>
2930
<LI> <B>Messages sorted by:</B>
3031
<a href="date.html#10001">[ date ]</a>
3132
<a href="thread.html#10001">[ thread ]</a>
@@ -61,13 +62,15 @@ <H1>[squid-dev] forward bumped traffic to parent in plain form</H1>
6162

6263
</PRE>
6364

65+
6466
<!--endarticle-->
6567
<HR>
6668
<P><UL>
6769
<!--threads-->
6870
<LI>Previous message (by thread): <A HREF="010000.html">[squid-dev] forward bumped traffic to parent in plain form
6971
</A></li>
70-
72+
<LI>Next message (by thread): <A HREF="010002.html">[squid-dev] forward bumped traffic to parent in plain form
73+
</A></li>
7174
<LI> <B>Messages sorted by:</B>
7275
<a href="date.html#10001">[ date ]</a>
7376
<a href="thread.html#10001">[ thread ]</a>

‎squid-dev/2026-April/010002.html‎

Lines changed: 106 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,106 @@
1+
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
2+
<HTML>
3+
<HEAD>
4+
<TITLE> [squid-dev] forward bumped traffic to parent in plain form
5+
</TITLE>
6+
<LINK REL="Index" HREF="index.html" >
7+
<LINK REL="made" HREF="mailto:squid-dev%40lists.squid-cache.org?Subject=Re%3A%20%5Bsquid-dev%5D%20forward%20bumped%20traffic%20to%20parent%20in%20plain%20form&In-Reply-To=%3C562184261.20260403172105%40yahoo.com%3E">
8+
<META NAME="robots" CONTENT="index,nofollow">
9+
<style type="text/css">
10+
pre {
11+
white-space: pre-wrap; /* css-2.1, curent FF, Opera, Safari */
12+
}
13+
</style>
14+
<META http-equiv="Content-Type" content="text/html; charset=us-ascii">
15+
<LINK REL="Previous" HREF="010001.html">
16+
<LINK REL="Next" HREF="010004.html">
17+
</HEAD>
18+
<BODY BGCOLOR="#ffffff">
19+
<H1>[squid-dev] forward bumped traffic to parent in plain form</H1>
20+
<B>Anthony Pankov</B>
21+
<A HREF="mailto:squid-dev%40lists.squid-cache.org?Subject=Re%3A%20%5Bsquid-dev%5D%20forward%20bumped%20traffic%20to%20parent%20in%20plain%20form&In-Reply-To=%3C562184261.20260403172105%40yahoo.com%3E"
22+
TITLE="[squid-dev] forward bumped traffic to parent in plain form">anthony.pankov at yahoo.com
23+
</A><BR>
24+
<I>Fri Apr 3 14:21:05 UTC 2026</I>
25+
<P><UL>
26+
<LI>Previous message (by thread): <A HREF="010001.html">[squid-dev] forward bumped traffic to parent in plain form
27+
</A></li>
28+
<LI>Next message (by thread): <A HREF="010004.html">[squid-dev] forward bumped traffic to parent in plain form
29+
</A></li>
30+
<LI> <B>Messages sorted by:</B>
31+
<a href="date.html#10002">[ date ]</a>
32+
<a href="thread.html#10002">[ thread ]</a>
33+
<a href="subject.html#10002">[ subject ]</a>
34+
<a href="author.html#10002">[ author ]</a>
35+
</LI>
36+
</UL>
37+
<HR>
38+
<!--beginarticle-->
39+
<PRE>
40+
Friday, April 3, 2026, 2:46:32 PM, you wrote:
41+
42+
43+
&gt;<i> Friday, April 3, 2026, 2:20:17 PM, you wrote:
44+
</I>
45+
46+
&gt;&gt;<i> I've tried to call getSslContextStart() in startPeakAndSplice().
47+
</I>
48+
&gt;&gt;<i> But the problem is:
49+
</I>&gt;&gt;<i> getSslContextStart() return synchronously in most cases after calling getSslContextDone().
50+
</I>
51+
&gt;&gt;<i> But in the case of #if USE_SSL_CRTD (my case) it return after sending request to CRTD without calling getSslContextDone().
52+
</I>&gt;&gt;<i> getSslContextDone() will be called further in sslCrtdHandleReplyWrapper().
53+
</I>
54+
&gt;&gt;<i> So we are back in startPeakAndSplice() without getSslContextDone() and connections are closed.
55+
</I>
56+
&gt;&gt;<i> Is there a preferred way to solve this problem?
57+
</I>
58+
&gt;<i> Sorry. It's a normal. The problem is out here.
59+
</I>
60+
61+
May be not.
62+
63+
I notice
64+
65+
26/04/03 16:54:33.312 kid1| 83,5| /client_side.cc(2993) startPeekAndSplice: Bumping on step2 done.
66+
2026/04/03 16:54:33.312 kid1| 33,5| src/base/AsyncJob.cc(177) callEnd: Http1::Server status out: [ job4]
67+
--&gt;2026/04/03 16:54:33.312 kid1| 33,5| src/base/AsyncCallQueue.cc(28) fire: leaving Server::doClientRead(conn12
68+
local=178.218.213.200:12345 remote=217.119.19.66:63055 FD 9 flags=1, data=0x2b2273397f18)
69+
2026/04/03 16:54:33.341 kid1| 83,5| src/ssl/helper.cc(161) HandleGeneratorReply: to 0x2b2273397f18 in crtGenRq6569/1
70+
71+
and suggest that stack rewind out of Server::doClientRead() while waiting for HandleGeneratorReply. ?And connection somehow lost because next read was not triggered?
72+
I'd like to wait inside startPeekAndSplice() for ?clientNegotiateSSL()? to finish.
73+
But I don't want to introduce own &quot;wait objects&quot;.
74+
75+
May be somebody can advice me.
76+
77+
78+
79+
--
80+
Best regards,
81+
Anthony
82+
83+
</PRE>
84+
85+
86+
87+
<!--endarticle-->
88+
<HR>
89+
<P><UL>
90+
<!--threads-->
91+
<LI>Previous message (by thread): <A HREF="010001.html">[squid-dev] forward bumped traffic to parent in plain form
92+
</A></li>
93+
<LI>Next message (by thread): <A HREF="010004.html">[squid-dev] forward bumped traffic to parent in plain form
94+
</A></li>
95+
<LI> <B>Messages sorted by:</B>
96+
<a href="date.html#10002">[ date ]</a>
97+
<a href="thread.html#10002">[ thread ]</a>
98+
<a href="subject.html#10002">[ subject ]</a>
99+
<a href="author.html#10002">[ author ]</a>
100+
</LI>
101+
</UL>
102+
103+
<hr>
104+
<a href="https://lists.squid-cache.org/listinfo/squid-dev">More information about the squid-dev
105+
mailing list</a><br>
106+
</body></html>

0 commit comments

Comments
 (0)