CLEANUP: fd: remove leftovers of the fdcache

The "cache" entry was still present in the fdtab struct and it was
reported in "show sess". Removing it broke the cache-line alignment
on 64-bit machines which is important for threads, so it was fixed
by adding an attribute(aligned()) when threads are in use. Doing it
only in this case allows 32-bit thread-less platforms to see the
struct fit into 32 bytes.
diff --git a/include/types/fd.h b/include/types/fd.h
index 5336461..35d3bc2 100644
--- a/include/types/fd.h
+++ b/include/types/fd.h
@@ -120,7 +120,6 @@
 	__decl_hathreads(HA_SPINLOCK_T lock);
 	unsigned long thread_mask;           /* mask of thread IDs authorized to process the task */
 	unsigned long update_mask;           /* mask of thread IDs having an update for fd */
-	struct fdlist_entry cache;           /* Entry in the fdcache */
 	struct fdlist_entry update;          /* Entry in the global update list */
 	void (*iocb)(int fd);                /* I/O handler */
 	void *owner;                         /* the connection or listener associated with this fd, NULL if closed */
@@ -128,7 +127,14 @@
 	unsigned char ev;                    /* event seen in return of poll() : FD_POLL_* */
 	unsigned char linger_risk:1;         /* 1 if we must kill lingering before closing */
 	unsigned char cloned:1;              /* 1 if a cloned socket, requires EPOLL_CTL_DEL on close */
-};
+}
+#ifdef USE_THREAD
+/* only align on cache lines when using threads; 32-bit small archs
+ * can put everything in 32-bytes when threads are disabled.
+ */
+__attribute__((aligned(64)))
+#endif
+;
 
 /* less often used information */
 struct fdinfo {
diff --git a/src/fd.c b/src/fd.c
index 620de59..17cf52b 100644
--- a/src/fd.c
+++ b/src/fd.c
@@ -596,7 +596,6 @@
 	for (p = 0; p < global.maxsock; p++) {
 		HA_SPIN_INIT(&fdtab[p].lock);
 		/* Mark the fd as out of the fd cache */
-		fdtab[p].cache.next = -3;
 		fdtab[p].update.next = -3;
 	}
 
diff --git a/src/stream.c b/src/stream.c
index 095565a..75cacc1 100644
--- a/src/stream.c
+++ b/src/stream.c
@@ -3214,11 +3214,10 @@
 			              obj_base_ptr(conn->target));
 
 			chunk_appendf(&trash,
-			              "      flags=0x%08x fd=%d fd.state=%02x fd.cache=%d updt=%d fd.tmask=0x%lx\n",
+			              "      flags=0x%08x fd=%d fd.state=%02x updt=%d fd.tmask=0x%lx\n",
 			              conn->flags,
 			              conn->handle.fd,
 			              conn->handle.fd >= 0 ? fdtab[conn->handle.fd].state : 0,
-			              conn->handle.fd >= 0 ? fdtab[conn->handle.fd].cache.next >= -2 : 0,
 			              conn->handle.fd >= 0 ? !!(fdtab[conn->handle.fd].update_mask & tid_bit) : 0,
 				      conn->handle.fd >= 0 ? fdtab[conn->handle.fd].thread_mask: 0);
 
@@ -3251,11 +3250,10 @@
 			              obj_base_ptr(conn->target));
 
 			chunk_appendf(&trash,
-			              "      flags=0x%08x fd=%d fd.state=%02x fd.cache=%d updt=%d fd.tmask=0x%lx\n",
+			              "      flags=0x%08x fd=%d fd.state=%02x updt=%d fd.tmask=0x%lx\n",
 			              conn->flags,
 			              conn->handle.fd,
 			              conn->handle.fd >= 0 ? fdtab[conn->handle.fd].state : 0,
-			              conn->handle.fd >= 0 ? fdtab[conn->handle.fd].cache.next >= -2 : 0,
 			              conn->handle.fd >= 0 ? !!(fdtab[conn->handle.fd].update_mask & tid_bit) : 0,
 				      conn->handle.fd >= 0 ? fdtab[conn->handle.fd].thread_mask: 0);