[newlib-cygwin/cygwin-3_4-branch] Cygwin: /proc/cpuinfo: Always print topology info

[newlib-cygwin/cygwin-3_4-branch] Cygwin: /proc/cpuinfo: Always print topology info

[Corinna Vinschen]

https://sourceware.org/git/gitweb.cgi?p=newlib-cygwin.git;h=7c870ee90f7c0f8264b58cf43616fa754993b9d1

commit 7c870ee90f7c0f8264b58cf43616fa754993b9d1
Author:     Corinna Vinschen <corinna@vinschen.de>
AuthorDate: Mon Jan 16 14:23:38 2023 +0100
Commit:     Corinna Vinschen <corinna@vinschen.de>
CommitDate: Mon Jan 16 14:23:50 2023 +0100

    Cygwin: /proc/cpuinfo: Always print topology info
    
    vmstat from proc-ps-4.0.x prints "Unable to create system stat structure"
    if the /proc/cpuinfo output fails to contain topology info.  While
    Linux always prints topology info if the kernel has been built with
    CONFIG_SMP, Cygwin only prints topology info if the CPU is known to
    be multi-core (i. e., the HT feature flag is set).
    
    Fix that by printing topology info all the time, even for single-core
    CPUs.
    
    Fixes: e0d48debedfa ("Fix /proc/cpuinfo topology and cache size info")
    Signed-off-by: Corinna Vinschen <corinna@vinschen.de>

Diff:
---
 winsup/cygwin/fhandler/proc.cc | 18 +++++++++++++++---
 winsup/cygwin/release/3.4.4    |  3 +++
 2 files changed, 18 insertions(+), 3 deletions(-)

diff --git a/winsup/cygwin/fhandler/proc.cc b/winsup/cygwin/fhandler/proc.cc
index 24067f68751d..66272125cb20 100644
--- a/winsup/cygwin/fhandler/proc.cc
+++ b/winsup/cygwin/fhandler/proc.cc
@@ -672,7 +672,8 @@ format_proc_cpuinfo (void *, char *&destbuf)
 
       if (!SetThreadGroupAffinity (GetCurrentThread (), &affinity,
 				   &orig_group_affinity))
-	system_printf ("SetThreadGroupAffinity(%x,%d (%x/%d)) failed %E", cpu_mask, cpu_group, cpu_number, cpu_number);
+	system_printf ("SetThreadGroupAffinity(%x,%d (%x/%d)) failed %E",
+		       cpu_mask, cpu_group, cpu_number, cpu_number);
       orig_affinity_mask = 1; /* Just mark success. */
       /* I'm not sure whether the thread changes processor immediately
 	 and I'm not sure whether this function will cause the thread
@@ -951,8 +952,7 @@ format_proc_cpuinfo (void *, char *&destbuf)
 	  core_id = (initial_apic_id & ((1 << logical_bits) - 1)) >> ht_bits;
 
 	  bufptr += __small_sprintf (bufptr, "physical id\t: %d\n", phys_id);
-	  if (siblings > 0)
-	    bufptr += __small_sprintf (bufptr, "siblings\t: %u\n", siblings);
+	  bufptr += __small_sprintf (bufptr, "siblings\t: %u\n", siblings);
 	  bufptr += __small_sprintf (bufptr, "core id\t\t: %d\n"
 					     "cpu cores\t: %d\n",
 				     core_id, cpu_cores);
@@ -962,6 +962,18 @@ format_proc_cpuinfo (void *, char *&destbuf)
 				       apic_id, initial_apic_id);
 
 	}
+      else
+	{
+	  /* Linux prints all this info depending on CONFIG_SMP.  There's no
+	     check if the CPU is actually a multicore CPU. */
+	  bufptr += __small_sprintf (bufptr, "physical id\t: %d\n", cpu_number);
+	  bufptr += __small_sprintf (bufptr, "siblings\t: 1\n");
+	  bufptr += __small_sprintf (bufptr, "core id\t\t: 0\n"
+					     "cpu cores\t: 1\n");
+	  bufptr += __small_sprintf (bufptr, "apicid\t\t: %d\n"
+					     "initial apicid\t: %d\n",
+				     apic_id, apic_id);
+	}
 
       /* level is number of non-zero leafs exc. sub-leafs */
       int level = maxf + 1 + (maxe & 0x7fffffff) + 1;
diff --git a/winsup/cygwin/release/3.4.4 b/winsup/cygwin/release/3.4.4
index 35d746ba08fd..2c93cd0663c2 100644
--- a/winsup/cygwin/release/3.4.4
+++ b/winsup/cygwin/release/3.4.4
@@ -13,3 +13,6 @@ Bug Fixes
 
 - Avoid hangs when reading /proc/<PID>/status.
   Addresses: https://cygwin.com/pipermail/cygwin/2022-December/252756.html
+
+- Fix vmstat(1) printing an error message on single core CPUs.
+  Addresses: https://cygwin.com/pipermail/cygwin/2023-January/252857.html