From 0792ff02ff863106cf22076862909abc1f611c31 Mon Sep 17 00:00:00 2001
From: Angelos Katharopoulos <a_katharopoulos@apple.com>
Date: Wed, 5 Mar 2025 13:16:19 -0800
Subject: [PATCH 1/3] Only fail when 10 consecutive socket errors occur (#1928)

---
 mlx/distributed/ring/ring.cpp | 12 ++++++++++--
 python/mlx/distributed_run.py |  7 ++++++-
 2 files changed, 16 insertions(+), 3 deletions(-)

diff --git a/mlx/distributed/ring/ring.cpp b/mlx/distributed/ring/ring.cpp
index 1f1c1b0b6..5bf08200e 100644
--- a/mlx/distributed/ring/ring.cpp
+++ b/mlx/distributed/ring/ring.cpp
@@ -199,6 +199,7 @@ class SocketThread {
   }
 
   void worker() {
+    int error_count = 0;
     bool delete_recv = false;
     bool delete_send = false;
     while (true) {
@@ -235,10 +236,11 @@ class SocketThread {
           task.buffer = static_cast<char*>(task.buffer) + r;
           task.size -= r;
           delete_recv = task.size == 0;
+          error_count = 0;
         } else if (errno != EAGAIN) {
+          error_count++;
           log_info(
               true, "Receiving from socket", fd_, "failed with errno", errno);
-          return;
         }
       }
       if (!sends_.empty()) {
@@ -248,11 +250,17 @@ class SocketThread {
           task.buffer = static_cast<char*>(task.buffer) + r;
           task.size -= r;
           delete_send = task.size == 0;
+          error_count = 0;
         } else if (errno != EAGAIN) {
+          error_count++;
           log_info(true, "Sending to socket", fd_, "failed with errno", errno);
-          return;
         }
       }
+
+      if (error_count >= 10) {
+        log_info(true, "Too many send/recv errors. Aborting...");
+        return;
+      }
     }
   }
 
diff --git a/python/mlx/distributed_run.py b/python/mlx/distributed_run.py
index 1a749beed..5d6bc4383 100644
--- a/python/mlx/distributed_run.py
+++ b/python/mlx/distributed_run.py
@@ -112,7 +112,12 @@ def extract_rings(hosts, index):
                     break
             if not ring:
                 break
-            rings.append(normalize(concretize(ring, used_ports)))
+            try:
+                rings.append(normalize(concretize(ring, used_ports)))
+            except RuntimeError:
+                if len(rings) > 0:
+                    return rings
+                raise
 
     return rings
 

From f599c11bc874a81b15a5c030257802e6e25f28d3 Mon Sep 17 00:00:00 2001
From: Awni Hannun <awni@apple.com>
Date: Wed, 5 Mar 2025 13:16:53 -0800
Subject: [PATCH 2/3] bump (#1931)

---
 mlx/version.h | 2 +-
 setup.py      | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/mlx/version.h b/mlx/version.h
index 35b026149..f244dcb16 100644
--- a/mlx/version.h
+++ b/mlx/version.h
@@ -3,7 +3,7 @@
 #pragma once
 
 #define MLX_VERSION_MAJOR 0
-#define MLX_VERSION_MINOR 24
+#define MLX_VERSION_MINOR 23
 #define MLX_VERSION_PATCH 2
 #define MLX_VERSION_NUMERIC \
   (100000 * MLX_VERSION_MAJOR + 1000 * MLX_VERSION_MINOR + MLX_VERSION_PATCH)
diff --git a/setup.py b/setup.py
index d4b5e15dd..787e0545b 100644
--- a/setup.py
+++ b/setup.py
@@ -173,7 +173,7 @@ if __name__ == "__main__":
 
     setup(
         name="mlx",
-        version=get_version("0.23.1"),
+        version=get_version("0.23.2"),
         author="MLX Contributors",
         author_email="mlx@group.apple.com",
         description="A framework for machine learning on Apple silicon.",

From 85b34d59bcea059603cd15db7ceaceed885113a0 Mon Sep 17 00:00:00 2001
From: Chunyang Wen <chunyang.wen@gmail.com>
Date: Thu, 6 Mar 2025 05:48:03 +0800
Subject: [PATCH 3/3] Clean unused sys (#1929)

---
 setup.py | 1 -
 1 file changed, 1 deletion(-)

diff --git a/setup.py b/setup.py
index 787e0545b..72bc2dba3 100644
--- a/setup.py
+++ b/setup.py
@@ -5,7 +5,6 @@ import os
 import platform
 import re
 import subprocess
-import sys
 from pathlib import Path
 from subprocess import run