VSL fails to get hold of SHM

[bsdphk]

Sometimes the VSL thread in varnishtest simply doesn't get hold of anything.

I have tried to postpone startup of the VSL thread until I am sure there is something in the VSL, but it still happens.

To provoke the bug add this patch:

diff --git a/bin/varnishtest/vtc_varnish.c b/bin/varnishtest/vtc_varnish.c
index 544a63c40..1a466ba4e 100644
--- a/bin/varnishtest/vtc_varnish.c
+++ b/bin/varnishtest/vtc_varnish.c
@@ -86,6 +86,7 @@ struct varnish {
    unsigned                vsl_tag_count[256];
 
    volatile int            vsl_idle;
+   volatile int            vsl_rec;
 };
 
 #define NONSENSE   "%XJEIFLH|)Xspa8P"
@@ -233,6 +234,7 @@ varnishlog_thread(void *priv)
                            break;
 
                    v->vsl_idle = 0;
+                   v->vsl_rec = 1;
 
                    tag = VSL_TAG(c->rec.ptr);
                    vxid = VSL_ID(c->rec.ptr);
@@ -571,7 +573,7 @@ varnish_start(struct varnish *v)
    macro_def(v->vl, v->name, "port", "%s", p);
    macro_def(v->vl, v->name, "sock", "%s %s", h, p);
    /* Wait for vsl logging to get underway */
-   while (v->vsl_idle == 0)
+   while (v->vsl_rec == 0)
            VTIM_sleep(.1);
 }

Create this VTC:

varnishtest "foo"

server s1 {
    rxreq
    txresp
} -start

varnish v1 -vcl+backend {} -start

client c1 {
    txreq
    rxresp
} -run

varnish v1 -stop

And run varnishtest -i -n 1000 -j 10 foo.vtc

[hermunn]

Backport review: The fix acd2198 has been backported as 3e19f33.