Alpine linux (musl libc) fixes

[ncopa]

I'm upstreaming a set of patches we have carried for alpine linux for a while.

I believe that buildroot, openwrt, gentoo, void linux and other musl libc based distros will benefit from those.

[johnwvmw]

Thanks for your contribution. Your contribution will go through a review process at VMware.

The changes will probably be reviewed individually or in small groups. An update will be provided as the reviews are completed.

[nuclearfall]

What's the hold up on this?

[neheb]

alternate patch for loff_t:

--- a/vmhgfs-fuse/fsutil.h
+++ b/vmhgfs-fuse/fsutil.h
@@ -29,6 +29,9 @@
 
 #include "request.h"
 #include "vm_basic_types.h"
+#ifndef _GNU_SOURCE
+#define _GNU_SOURCE
+#endif
 #include "hgfsProto.h"
 #include <fuse.h>

[ncopa]

I don't think enabling _GNU_SOURCE for just some of the headers is a good idea. Either it needs to be enabled globally for everything or its not enabled at all.

[ncopa]

rebased the paches to solve the conflict with almalinux introduction (63abb4f)

[neheb]

for the latest version, I have some extra fixes:

--- a/lib/err/errPosix.c
+++ b/lib/err/errPosix.c
@@ -29,6 +29,7 @@
 #endif
 
 #include <errno.h>
+#include <stdio.h>
 #include <string.h>
 #include <locale.h>
 
--- a/lib/file/fileIOPosix.c
+++ b/lib/file/fileIOPosix.c
@@ -1741,7 +1741,7 @@ FileIOPreadvInternal(
        * the library horizon this can go away.
        */
       /* coverity[func_conv] */
-      if (preadv64 == NULL) {
+      if (0) {
          fret = FileIOPreadvCoalesced(fd, entries, numEntries, offset,
                                       totalSize, &bytesRead);
          break;
@@ -1882,7 +1882,7 @@ FileIOPwritevInternal(
        * the library horizon this can go away.
        */
       /* coverity[func_conv] */
-      if (pwritev64 == NULL) {
+      if (0) {
          fret = FileIOPwritevCoalesced(fd, entries, numEntries, offset,
                                        totalSize, &bytesWritten);
          break;
--- a/lib/hgfsServer/hgfsServerLinux.c
+++ b/lib/hgfsServer/hgfsServerLinux.c
@@ -32,6 +32,7 @@
 #define _DARWIN_USE_64_BIT_INODE
 #endif
 
+#include <inttypes.h>
 #include <stdlib.h>
 #include <stdio.h>
 #include <unistd.h>
@@ -2573,7 +2574,7 @@ HgfsStatToFileAttr(struct stat *stats,
 #      define FMTTIMET "l"
 #   endif
 #else
-#   define FMTTIMET "l"
+#   define FMTTIMET PRId64
 #endif
    LOG(4, "access: %"FMTTIMET"d/%"FMT64"u \nwrite: %"FMTTIMET"d/%"FMT64"u \n"
        "attr: %"FMTTIMET"d/%"FMT64"u\n",
--- a/services/plugins/gdp/gdpPlugin.c
+++ b/services/plugins/gdp/gdpPlugin.c
@@ -32,7 +32,7 @@
 #include <sys/types.h>
 #include <sys/socket.h>
 #include <sys/eventfd.h>
-#include <sys/poll.h>
+#include <poll.h>
 #include <unistd.h>
 #endif
 

[ncopa]

for the latest version, I have some extra fixes:

--- a/lib/file/fileIOPosix.c
+++ b/lib/file/fileIOPosix.c
@@ -1741,7 +1741,7 @@ FileIOPreadvInternal(
        * the library horizon this can go away.
        */
       /* coverity[func_conv] */
-      if (preadv64 == NULL) {
+      if (0) {
          fret = FileIOPreadvCoalesced(fd, entries, numEntries, offset,
                                       totalSize, &bytesRead);
          break;
@@ -1882,7 +1882,7 @@ FileIOPwritevInternal(
        * the library horizon this can go away.
        */
       /* coverity[func_conv] */
-      if (pwritev64 == NULL) {
+      if (0) {
          fret = FileIOPwritevCoalesced(fd, entries, numEntries, offset,
                                        totalSize, &bytesWritten);
          break;
...

I don't think those changes are suitable for upstream.

[ncopa]

Would it make it easier to review this if I created a PR for each individual patch?