diff --git a/tokio/src/macros/join.rs b/tokio/src/macros/join.rs
index f91b5f19145..ab85cdec26f 100644
--- a/tokio/src/macros/join.rs
+++ b/tokio/src/macros/join.rs
@@ -60,6 +60,9 @@ macro_rules! join {
         // normalization is complete.
         ( $($count:tt)* )
 
+        // The expression `0+1+1+ ... +1` equal to the number of branches.
+        ( $($total:tt)* )
+
         // Normalized join! branches
         $( ( $($skip:tt)* ) $e:expr, )*
 
@@ -71,22 +74,54 @@ macro_rules! join {
         // the requirement of `Pin::new_unchecked` called below.
         let mut futures = ( $( maybe_done($e), )* );
 
+        // Each time the future created by poll_fn is polled, a different future will be polled first
+        // to ensure every future passed to join! gets a chance to make progress even if
+        // one of the futures consumes the whole budget.
+        //
+        // This is number of futures that will be skipped in the first loop
+        // iteration the next time.
+        let mut skip_next_time: u32 = 0;
+
         poll_fn(move |cx| {
+            const COUNT: u32 = $($total)*;
+
             let mut is_pending = false;
 
+            let mut to_run = COUNT;
+
+            // The number of futures that will be skipped in the first loop iteration.
+            let mut skip = skip_next_time;
+
+            skip_next_time = if skip + 1 == COUNT { 0 } else { skip + 1 };
+
+            // This loop runs twice and the first `skip` futures
+            // are not polled in the first iteration.
+            loop {
             $(
-                // Extract the future for this branch from the tuple.
-                let ( $($skip,)* fut, .. ) = &mut futures;
+                if skip == 0 {
+                    if to_run == 0 {
+                        // Every future has been polled
+                        break;
+                    }
+                    to_run -= 1;
 
-                // Safety: future is stored on the stack above
-                // and never moved.
-                let mut fut = unsafe { Pin::new_unchecked(fut) };
+                    // Extract the future for this branch from the tuple.
+                    let ( $($skip,)* fut, .. ) = &mut futures;
+
+                    // Safety: future is stored on the stack above
+                    // and never moved.
+                    let mut fut = unsafe { Pin::new_unchecked(fut) };
 
-                // Try polling
-                if fut.poll(cx).is_pending() {
-                    is_pending = true;
+                    // Try polling
+                    if fut.poll(cx).is_pending() {
+                        is_pending = true;
+                    }
+                } else {
+                    // Future skipped, one less future to skip in the next iteration
+                    skip -= 1;
                 }
             )*
+            }
 
             if is_pending {
                 Pending
@@ -107,13 +142,13 @@ macro_rules! join {
 
     // ===== Normalize =====
 
-    (@ { ( $($s:tt)* ) $($t:tt)* } $e:expr, $($r:tt)* ) => {
-        $crate::join!(@{ ($($s)* _) $($t)* ($($s)*) $e, } $($r)*)
+    (@ { ( $($s:tt)* ) ( $($n:tt)* ) $($t:tt)* } $e:expr, $($r:tt)* ) => {
+        $crate::join!(@{ ($($s)* _) ($($n)* + 1) $($t)* ($($s)*) $e, } $($r)*)
     };
 
     // ===== Entry point =====
 
     ( $($e:expr),* $(,)?) => {
-        $crate::join!(@{ () } $($e,)*)
+        $crate::join!(@{ () (0) } $($e,)*)
     };
 }
diff --git a/tokio/src/macros/try_join.rs b/tokio/src/macros/try_join.rs
index 6d3a893b7e1..c80395c9097 100644
--- a/tokio/src/macros/try_join.rs
+++ b/tokio/src/macros/try_join.rs
@@ -106,6 +106,9 @@ macro_rules! try_join {
         // normalization is complete.
         ( $($count:tt)* )
 
+        // The expression `0+1+1+ ... +1` equal to the number of branches.
+        ( $($total:tt)* )
+
         // Normalized try_join! branches
         $( ( $($skip:tt)* ) $e:expr, )*
 
@@ -117,24 +120,56 @@ macro_rules! try_join {
         // the requirement of `Pin::new_unchecked` called below.
         let mut futures = ( $( maybe_done($e), )* );
 
+        // Each time the future created by poll_fn is polled, a different future will be polled first
+        // to ensure every future passed to join! gets a chance to make progress even if
+        // one of the futures consumes the whole budget.
+        //
+        // This is number of futures that will be skipped in the first loop
+        // iteration the next time.
+        let mut skip_next_time: u32 = 0;
+
         poll_fn(move |cx| {
+            const COUNT: u32 = $($total)*;
+
             let mut is_pending = false;
 
+            let mut to_run = COUNT;
+
+            // The number of futures that will be skipped in the first loop iteration
+            let mut skip = skip_next_time;
+
+            skip_next_time = if skip + 1 == COUNT { 0 } else { skip + 1 };
+
+            // This loop runs twice and the first `skip` futures
+            // are not polled in the first iteration.
+            loop {
             $(
-                // Extract the future for this branch from the tuple.
-                let ( $($skip,)* fut, .. ) = &mut futures;
-
-                // Safety: future is stored on the stack above
-                // and never moved.
-                let mut fut = unsafe { Pin::new_unchecked(fut) };
-
-                // Try polling
-                if fut.as_mut().poll(cx).is_pending() {
-                    is_pending = true;
-                } else if fut.as_mut().output_mut().expect("expected completed future").is_err() {
-                    return Ready(Err(fut.take_output().expect("expected completed future").err().unwrap()))
+                if skip == 0 {
+                    if to_run == 0 {
+                        // Every future has been polled
+                        break;
+                    }
+                    to_run -= 1;
+
+                    // Extract the future for this branch from the tuple.
+                    let ( $($skip,)* fut, .. ) = &mut futures;
+
+                    // Safety: future is stored on the stack above
+                    // and never moved.
+                    let mut fut = unsafe { Pin::new_unchecked(fut) };
+
+                    // Try polling
+                    if fut.as_mut().poll(cx).is_pending() {
+                        is_pending = true;
+                    } else if fut.as_mut().output_mut().expect("expected completed future").is_err() {
+                        return Ready(Err(fut.take_output().expect("expected completed future").err().unwrap()))
+                    }
+                } else {
+                    // Future skipped, one less future to skip in the next iteration
+                    skip -= 1;
                 }
             )*
+            }
 
             if is_pending {
                 Pending
@@ -159,13 +194,13 @@ macro_rules! try_join {
 
     // ===== Normalize =====
 
-    (@ { ( $($s:tt)* ) $($t:tt)* } $e:expr, $($r:tt)* ) => {
-        $crate::try_join!(@{ ($($s)* _) $($t)* ($($s)*) $e, } $($r)*)
+    (@ { ( $($s:tt)* ) ( $($n:tt)* ) $($t:tt)* } $e:expr, $($r:tt)* ) => {
+      $crate::try_join!(@{ ($($s)* _) ($($n)* + 1) $($t)* ($($s)*) $e, } $($r)*)
     };
 
     // ===== Entry point =====
 
     ( $($e:expr),* $(,)?) => {
-        $crate::try_join!(@{ () } $($e,)*)
+        $crate::try_join!(@{ () (0) } $($e,)*)
     };
 }
diff --git a/tokio/tests/macros_join.rs b/tokio/tests/macros_join.rs
index d4f20b3862c..56bd9c58f3c 100644
--- a/tokio/tests/macros_join.rs
+++ b/tokio/tests/macros_join.rs
@@ -1,5 +1,6 @@
 #![cfg(feature = "macros")]
 #![allow(clippy::blacklisted_name)]
+use std::sync::Arc;
 
 #[cfg(target_arch = "wasm32")]
 use wasm_bindgen_test::wasm_bindgen_test as test;
@@ -9,7 +10,7 @@ use wasm_bindgen_test::wasm_bindgen_test as maybe_tokio_test;
 #[cfg(not(target_arch = "wasm32"))]
 use tokio::test as maybe_tokio_test;
 
-use tokio::sync::oneshot;
+use tokio::sync::{oneshot, Semaphore};
 use tokio_test::{assert_pending, assert_ready, task};
 
 #[maybe_tokio_test]
@@ -71,12 +72,82 @@ fn join_size() {
         let ready = future::ready(0i32);
         tokio::join!(ready)
     };
-    assert_eq!(mem::size_of_val(&fut), 16);
+    assert_eq!(mem::size_of_val(&fut), 20);
 
     let fut = async {
         let ready1 = future::ready(0i32);
         let ready2 = future::ready(0i32);
         tokio::join!(ready1, ready2)
     };
-    assert_eq!(mem::size_of_val(&fut), 28);
+    assert_eq!(mem::size_of_val(&fut), 32);
+}
+
+async fn non_cooperative_task(permits: Arc<Semaphore>) -> usize {
+    let mut exceeded_budget = 0;
+
+    for _ in 0..5 {
+        // Another task should run after after this task uses its whole budget
+        for _ in 0..128 {
+            let _permit = permits.clone().acquire_owned().await.unwrap();
+        }
+
+        exceeded_budget += 1;
+    }
+
+    exceeded_budget
+}
+
+async fn poor_little_task(permits: Arc<Semaphore>) -> usize {
+    let mut how_many_times_i_got_to_run = 0;
+
+    for _ in 0..5 {
+        let _permit = permits.clone().acquire_owned().await.unwrap();
+        how_many_times_i_got_to_run += 1;
+    }
+
+    how_many_times_i_got_to_run
+}
+
+#[tokio::test]
+async fn join_does_not_allow_tasks_to_starve() {
+    let permits = Arc::new(Semaphore::new(1));
+
+    // non_cooperative_task should yield after its budget is exceeded and then poor_little_task should run.
+    let (non_cooperative_result, little_task_result) = tokio::join!(
+        non_cooperative_task(Arc::clone(&permits)),
+        poor_little_task(permits)
+    );
+
+    assert_eq!(5, non_cooperative_result);
+    assert_eq!(5, little_task_result);
+}
+
+#[tokio::test]
+async fn a_different_future_is_polled_first_every_time_poll_fn_is_polled() {
+    let poll_order = Arc::new(std::sync::Mutex::new(vec![]));
+
+    let fut = |x, poll_order: Arc<std::sync::Mutex<Vec<i32>>>| async move {
+        for _ in 0..4 {
+            {
+                let mut guard = poll_order.lock().unwrap();
+
+                guard.push(x);
+            }
+
+            tokio::task::yield_now().await;
+        }
+    };
+
+    tokio::join!(
+        fut(1, Arc::clone(&poll_order)),
+        fut(2, Arc::clone(&poll_order)),
+        fut(3, Arc::clone(&poll_order)),
+    );
+
+    // Each time the future created by join! is polled, it should start
+    // by polling a different future first.
+    assert_eq!(
+        vec![1, 2, 3, 2, 3, 1, 3, 1, 2, 1, 2, 3],
+        *poll_order.lock().unwrap()
+    );
 }
diff --git a/tokio/tests/macros_try_join.rs b/tokio/tests/macros_try_join.rs
index 60a726b659a..556436d2207 100644
--- a/tokio/tests/macros_try_join.rs
+++ b/tokio/tests/macros_try_join.rs
@@ -1,7 +1,9 @@
 #![cfg(feature = "macros")]
 #![allow(clippy::blacklisted_name)]
 
-use tokio::sync::oneshot;
+use std::sync::Arc;
+
+use tokio::sync::{oneshot, Semaphore};
 use tokio_test::{assert_pending, assert_ready, task};
 
 #[cfg(target_arch = "wasm32")]
@@ -94,16 +96,89 @@ fn join_size() {
         let ready = future::ready(ok(0i32));
         tokio::try_join!(ready)
     };
-    assert_eq!(mem::size_of_val(&fut), 16);
+    assert_eq!(mem::size_of_val(&fut), 20);
 
     let fut = async {
         let ready1 = future::ready(ok(0i32));
         let ready2 = future::ready(ok(0i32));
         tokio::try_join!(ready1, ready2)
     };
-    assert_eq!(mem::size_of_val(&fut), 28);
+    assert_eq!(mem::size_of_val(&fut), 32);
 }
 
 fn ok<T>(val: T) -> Result<T, ()> {
     Ok(val)
 }
+
+async fn non_cooperative_task(permits: Arc<Semaphore>) -> Result<usize, String> {
+    let mut exceeded_budget = 0;
+
+    for _ in 0..5 {
+        // Another task should run after after this task uses its whole budget
+        for _ in 0..128 {
+            let _permit = permits.clone().acquire_owned().await.unwrap();
+        }
+
+        exceeded_budget += 1;
+    }
+
+    Ok(exceeded_budget)
+}
+
+async fn poor_little_task(permits: Arc<Semaphore>) -> Result<usize, String> {
+    let mut how_many_times_i_got_to_run = 0;
+
+    for _ in 0..5 {
+        let _permit = permits.clone().acquire_owned().await.unwrap();
+
+        how_many_times_i_got_to_run += 1;
+    }
+
+    Ok(how_many_times_i_got_to_run)
+}
+
+#[tokio::test]
+async fn try_join_does_not_allow_tasks_to_starve() {
+    let permits = Arc::new(Semaphore::new(10));
+
+    // non_cooperative_task should yield after its budget is exceeded and then poor_little_task should run.
+    let result = tokio::try_join!(
+        non_cooperative_task(Arc::clone(&permits)),
+        poor_little_task(permits)
+    );
+
+    let (non_cooperative_result, little_task_result) = result.unwrap();
+
+    assert_eq!(5, non_cooperative_result);
+    assert_eq!(5, little_task_result);
+}
+
+#[tokio::test]
+async fn a_different_future_is_polled_first_every_time_poll_fn_is_polled() {
+    let poll_order = Arc::new(std::sync::Mutex::new(vec![]));
+
+    let fut = |x, poll_order: Arc<std::sync::Mutex<Vec<i32>>>| async move {
+        for _ in 0..4 {
+            {
+                let mut guard = poll_order.lock().unwrap();
+
+                guard.push(x);
+            }
+
+            tokio::task::yield_now().await;
+        }
+    };
+
+    tokio::join!(
+        fut(1, Arc::clone(&poll_order)),
+        fut(2, Arc::clone(&poll_order)),
+        fut(3, Arc::clone(&poll_order)),
+    );
+
+    // Each time the future created by join! is polled, it should start
+    // by polling a different future first.
+    assert_eq!(
+        vec![1, 2, 3, 2, 3, 1, 3, 1, 2, 1, 2, 3],
+        *poll_order.lock().unwrap()
+    );
+}