[PATCH hurd v2] Add partial /proc/cpuinfo implementation

[dnietoc]

From: Diego Nieto Cid <dnietoc@gmail.com>

Hello,

On Sun, Jan 26, 2025 at 10:18:03PM +0100, Samuel Thibault wrote:
>
> dnietoc@gmail.com, le ven. 10 janv. 2025 23:52:28 +0000, a ecrit:
> > --- a/procfs/rootdir.c
> > +
> > +  m = open_memstream(contents, (size_t *) contents_len);
> 
> Please mind the GNU coding style which wants a space before opening
> parentheses, here and in other places (function calls, sizeof, etc.).
> 

I hope this time I got it better :)

> > +
> > +  memcpy(vendor + 0 * sizeof(unsigned int), (char *) &ebx, sizeof(unsigned 
> > int));
> 
> There is no need to cast &ebx, memcpy takes a void*, so the cast is
> implicit in C.
>

Ok, I removed the casts.

>
> > +
> > +  fprintf(m,
> > +    "processor       : 0\n"
> > +    "vendor_id       : %s\n"
> > +    "cpu family      : %d\n"
> > +    "model           : %d\n"
> > +    "model name      : %s\n"
> > +    "stepping        : %d\n",
> > +    vendor, family, model, model_name, stepping);
> > +
> > +  fprintf(m, "flags           :");v
> 
> Linux uses tab characters before ':', better use that too for
> compatibility.
>

Ok, I switched to tabs. Though, I had to use a mix of one and two tabs.
 
> > +
> > +  m = open_memstream(contents, (size_t *) contents_len);
> > +  if (m == NULL)
> > +    return ENOMEM;
> 
> Better return errno like on x86.
>

Done so.
 
> > +
> > +  err = aarch64_get_hwcaps(mach_host_self(), &caps, &mdir, &revdir);
> > +  if (err)
> > +    goto out;
> > +
> > +  implementer = (mdir  & 0xff000000) >> 24;
> > +  variant = (mdir      & 0x00f00000) >> 20;
> > +  architecture = (mdir & 0x000f0000) >> 16;
> > +  part_num = (mdir     & 0x0000fff0) >> 4;
> > +  revision = (mdir     & 0x0000000f) >> 0;
> 
> Better align mdir too to make it look nicer.
> 

I aligned the = sign, the variable and the semi-colon.
Now it should look better.

> > +
> > +  fprintf(m, "processor       : 0\n");
> > +  fprintf(m, "BogoMIPS        : 0\n");
> 
> I'd rather not print BogoMIPS at all rather than 0.
>

Ok, I just removed the BogoMIPS entry until we figure out how
to compute it properly.
 
>
> Also use tabs as well for compatibility.
> 

Done

> > +out:
> > +  flcose(m);
> 
> You didn't try to compile it ;)
>

Oops, it didn't cross my mind I could do that :-)

--

On Mon, Jan 27, 2025 at 12:25:15AM +0300, Sergey Bugaev wrote:
> On Mon, Jan 27, 2025 at 12:18 AM Samuel Thibault
> <samuel.thibault@gnu.org> wrote:
> >
> > Better align mdir too to make it look nicer.
> 
> Speaking of which, it's midr (main identification register) and revidr
> (revision identification register), not *dir.
>

Thanks, I renamed the variables. revidr is unused though, the docs
said it was implementation defined so I didn't know what to do with it.

--

On Sun, Jan 26, 2025 at 10:27:57PM +0100, Samuel Thibault wrote:
> Sergey Bugaev, le lun. 27 janv. 2025 00:25:15 +0300, a ecrit:
> > > > +out:
> > > > +  flcose(m);
> > >
> > > You didn't try to compile it ;)
> > 
> > I suppose that's not easy for Diego to do, given the state of aarch64-gnu :)
> 
> Sure, but he can comment the very few really-arm64-specific lines and
> try to build the rest :)
> 

It didn't occur to me about testing that way. I fiddled with the defined
macros and replaced the RPC with some harcoded data, it should work now.

Regards,
Diego

-- >8 -- >8 -- >8 --

  * procfs/rootdir.c: (rootdir_gc_cpuinfo) new function
    (rootdir_entries) add entry for cpuinfo file
    (cpuinfo_x86, cpuinfo_aarch64) implementations for x86 and
    aarch64 respectively.
---
 procfs/rootdir.c | 219 +++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 219 insertions(+)

diff --git a/procfs/rootdir.c b/procfs/rootdir.c
index b8a8c4d1..05304015 100644
--- a/procfs/rootdir.c
+++ b/procfs/rootdir.c
@@ -38,6 +38,12 @@
 #include "procfs_dir.h"
 #include "main.h"
 #include <net/route.h>
+#if defined (__x86_64__) || defined (__i486__) || defined (__i586__) || 
defined (__i686__)
+#include <cpuid.h>
+#elif defined (__aarch64__)
+#warning Aarch64 port of cpuinfo is untested
+#include <mach/machine/mach_aarch64.h>
+#endif
 
 #include "mach_debug_U.h"
 #include "pfinet_U.h"
@@ -696,6 +702,212 @@ out_fclose:
   fclose (m);
   return err;
 }
+
+#if defined (__x86_64__) || defined (__i486__) || defined (__i586__) || 
defined (__i686__)
+static char *cpu_features_edx[] =
+  {
+    "fpu", "vme", "de", "pse", "tsc", "msr", "pae", "mce", "cx8", "apic",
+    NULL, "sep", "mtrr", "pge", "mca", "cmov", "pat", "pse36", "pn", "clfush",
+    NULL, "dts", "acpi", "mmx", "fxsr", "sse", "sse2", "ss", "ht", "tm",
+    "ia64", "pbe"
+  };
+
+static char *cpu_features_ecx[] =
+  {
+    "sse3", "pclmulqdq", "dtes64", "monitor", "ds_cpl", "vmx", "smx", "est",
+    "tm2", "ssse3", "cid", "sdbg", "fma", "cx16", "xtpr", "pdcm",
+    NULL, "pcid", "dca", "sse4_1", "sse4_2", "x2apic", "movbe", "popcnt",
+    "tsc_deadline_timer", "aes", "xsave", "osxsave", "avx", "f16c", "rdrand", 
"hypervisor"
+  };
+
+#define VENDOR_ID_LEN  12
+#define MODEL_NAME_LEN 48
+
+static error_t
+cpuinfo_x86 (void* hook, char **contents, ssize_t *contents_len)
+{
+  error_t err = 0;
+  FILE* m;
+  int ret, index, stepping, model, family, extended_model, extended_family;
+  unsigned int eax, ebx, ecx, edx;
+  unsigned int feature_edx, feature_ecx;
+  char vendor[VENDOR_ID_LEN + 1] = { 0 };
+  char model_name[MODEL_NAME_LEN + 1] = { 0 };
+
+  m = open_memstream (contents, (size_t *) contents_len);
+  if (m == NULL)
+    return errno;
+
+  ret = __get_cpuid (0, &eax, &ebx, &ecx, &edx);
+  if (ret != 1)
+    {
+      err = EIO;
+      goto out;
+    }
+
+  memcpy (vendor + 0 * sizeof (unsigned int), &ebx, sizeof (unsigned int));
+  memcpy (vendor + 1 * sizeof (unsigned int), &edx, sizeof (unsigned int));
+  memcpy (vendor + 2 * sizeof (unsigned int), &ecx, sizeof (unsigned int));
+
+  ret = __get_cpuid (1, &eax, &ebx, &ecx, &edx);
+  if (ret != 1)
+    {
+      err = EIO;
+      goto out;
+    }
+
+  feature_edx = edx;
+  feature_ecx = ecx;
+  stepping = eax & 0x0F;
+  model = (eax & 0xF0) >> 4;
+  family = (eax & 0xF00) >> 8;
+  extended_model = (eax & 0xF0000) >> 16;
+  extended_family = (eax &0xFF00000) >> 20;
+
+  if (family == 6 || family == 15)
+    model += (extended_model << 4);
+  if (family == 15)
+    family += extended_family;
+
+  __get_cpuid (0x80000000, &eax, &ebx, &ecx, &edx);
+  if (eax >= 0x80000004)
+    {
+      __get_cpuid (0x80000002, &eax, &ebx, &ecx, &edx);
+      memcpy (model_name + 0 * sizeof (unsigned int), &eax, sizeof (unsigned 
int));
+      memcpy (model_name + 1 * sizeof (unsigned int), &ebx, sizeof (unsigned 
int));
+      memcpy (model_name + 2 * sizeof (unsigned int), &ecx, sizeof (unsigned 
int));
+      memcpy (model_name + 3 * sizeof (unsigned int), &edx, sizeof (unsigned 
int));
+
+      __get_cpuid (0x80000003, &eax, &ebx, &ecx, &edx);
+      memcpy (model_name + 4 * sizeof (unsigned int), &eax, sizeof (unsigned 
int));
+      memcpy (model_name + 5 * sizeof (unsigned int), &ebx, sizeof (unsigned 
int));
+      memcpy (model_name + 6 * sizeof (unsigned int), &ecx, sizeof (unsigned 
int));
+      memcpy (model_name + 7 * sizeof (unsigned int), &edx, sizeof (unsigned 
int));
+
+      __get_cpuid (0x80000004, &eax, &ebx, &ecx, &edx);
+      memcpy (model_name + 8 * sizeof (unsigned int), &eax, sizeof (unsigned 
int));
+      memcpy (model_name + 9 * sizeof (unsigned int), &ebx, sizeof (unsigned 
int));
+      memcpy (model_name + 10 * sizeof (unsigned int), &ecx, sizeof (unsigned 
int));
+      memcpy (model_name + 11 * sizeof (unsigned int), &edx, sizeof (unsigned 
int));
+    }
+
+  fprintf (m,
+    "processor\t: 0\n"
+    "vendor_id\t: %s\n"
+    "cpu family\t: %d\n"
+    "model\t\t: %d\n"
+    "model name\t: %s\n"
+    "stepping\t: %d\n",
+    vendor, family, model, model_name, stepping);
+
+  fprintf (m, "flags\t\t:");
+  for (index = 0; index < (sizeof(cpu_features_edx)/sizeof(char*)); index++)
+    {
+      if (cpu_features_edx[index] == NULL)
+        continue;
+      if (feature_edx & (1ul << index))
+        fprintf(m, " %s", cpu_features_edx[index]);
+    }
+  for (index = 0; index < (sizeof(cpu_features_ecx)/sizeof(char*)); index++)
+    {
+      if (cpu_features_ecx[index] == NULL)
+        continue;
+      if (feature_ecx & (1ul << index))
+        fprintf(m, " %s", cpu_features_ecx[index]);
+    }
+
+  fprintf(m, "\n\n");
+
+out:
+  fclose(m);
+  return err;
+}
+#endif
+
+#if defined (__aarch64__)
+
+static char *cpu_features_1[] =
+  {
+    "fp", "asimd", "evtstrm", "aes", "pmul", "sha1", "sha2", "crc32",
+    "atomics", "fphp", "asimdhp", "cpuid", "asimdrdm", "jscvt", "fcma", 
"lrcpc",
+    "dcpop", "sha3", "sm3", "sm4", "asimddp", "sha512", "sve", "asimdfhm",
+    "dit", "uscat", "ilrcpc", "flagm", "ssbs", "sb", "paca", "pacg"
+  };
+
+static char *cpu_features_2[] =
+  {
+    "dcpodp", "sve2", "sveaes", "svepmull", "svebitperm", "svesha3", "svesm4", 
"flagm2",
+    "frint", "svei8mm", "svef32mm", "svef64mm", "svebf16", "i8mm", "bf16", 
"dgh",
+    "rng", "bti", "mie", "ecv", "afp", "rpres", "mte3", "sme",
+    "sme_i16i64", "sme_f64f64", "sme_i8i32", "sme_f16f32", "sme_b16f32", 
"sme_f32f32", "sme_fa64", "wfxt",
+    "ebf16", "sve_ebf16", "cssc", "rprfm", "sve2p1", "sme2", "sme2p1", 
"sme_i15i32",
+    "sme_bi32i32", "sme_b16b16", "sme_f16f16", "mops", "hbc", "sve_b16b16", 
"lrcpc3", "lse123"
+  };
+
+static error_t
+cpuinfo_aarch64 (void *hook, char **contents, ssize_t *contents_len)
+{
+  error_t err = 0;
+  hwcaps_t caps;
+  uint64_t midr, revidr;
+  int index;
+  unsigned int implementer, variant, architecture, part_num, revision;
+  FILE *m;
+
+  m = open_memstream (contents, (size_t *) contents_len);
+  if (m == NULL)
+    return errno;
+
+  err = aarch64_get_hwcaps (mach_host_self(), &caps, &midr, &revidr);
+  if (err)
+    goto out;
+
+  implementer  = (midr & 0xff000000) >> 24;
+  variant      = (midr & 0x00f00000) >> 20;
+  architecture = (midr & 0x000f0000) >> 16;
+  part_num     = (midr & 0x0000fff0) >>  4;
+  revision     = (midr & 0x0000000f) >>  0;
+
+  fprintf(m, "processor\t\t: 0\n");
+  fprintf(m, "Features\t\t:");
+  for (index = 0; index < (sizeof (cpu_features_1) / sizeof (char *)); index++)
+    {
+      if (cpu_features_1[index] == NULL)
+        continue;
+      if (caps[0] & (1ul << index))
+        fprintf(m, " %s", cpu_features_1[index]);
+    }
+  for (index = 0; index < (sizeof(cpu_features_2) / sizeof(char *)); index++)
+    {
+      if (cpu_features_2[index] == NULL)
+        continue;
+      if (caps[1] & (1ul << index))
+        fprintf(m, " %s", cpu_features_2[index]);
+    }
+  fprintf(m, "\n");
+  fprintf(m, "CPU implementer\t\t: 0x%x\n", implementer);
+  fprintf(m, "CPU architecture\t: %d\n", architecture);
+  fprintf(m, "CPU variant\t\t: 0x%x\n", variant);
+  fprintf(m, "CPU part\t\t: 0x%x\n", part_num);
+  fprintf(m, "CPU revision\t\t: %d\n", revision);
+  fprintf(m, "\n");
+out:
+  fclose(m);
+  return err;
+}
+#endif
+
+static error_t
+rootdir_gc_cpuinfo (void *hook, char **contents, ssize_t *contents_len)
+{
+#if defined (__x86_64__) || defined (__i486__) || defined (__i586__) || 
defined (__i686__)
+  return cpuinfo_x86(hook, contents, contents_len);
+#elif defined (__aarch64__)
+  return cpuinfo_aarch64(hook, contents, contents_len);
+#else
+  return ENOTSUP;
+#endif
+}
 
 /* Glue logic and entries table */
 
@@ -888,6 +1100,13 @@ static const struct procfs_dir_entry rootdir_entries[] = {
       .cleanup_contents = procfs_cleanup_contents_with_free,
     },
   },
+  {
+    .name = "cpuinfo",
+    .hook = & (struct procfs_node_ops) {
+      .get_contents = rootdir_gc_cpuinfo,
+      .cleanup_contents = procfs_cleanup_contents_with_free,
+    },
+  },
 #ifdef PROFILE
   /* In order to get a usable gmon.out file, we must apparently use exit(). */
   {
-- 
2.47.1