[PATCH v3 0/3] Fixes for building on Solaris 11.4.42 CBE

[PATCH v3 0/3] Fixes for building on Solaris 11.4.42 CBE

[Andrew Deason]

With these minor fixes, I can build qemu on Solaris 11.4.42 CBE
(Oracle's new rolling release thing), using '--disable-rdma
--enable-modules --disable-dbus-display --target-list=x86_64-softmmu'.
I'm just interested in the guest agent right now, so that's all I've
tested (briefly), but the rest of the build wasn't hard to get working.
With this, the guest agent runs fine using isa-serial.

Changes since v2:
- Rename new symbol to HAVE_MADVISE_WITHOUT_PROTOTYPE
- Move madvise prototype to include/qemu/osdep.h
- More comments in meson.build

Changes since v1:
- Change the CONFIG_MADVISE checks to not be platform-specific
- Add the last commit removing util/osdep.c cruft

Andrew Deason (3):
  util/osdep: Avoid madvise proto on modern Solaris
  hw/i386/acpi-build: Avoid 'sun' identifier
  util/osdep: Remove some early cruft

 hw/i386/acpi-build.c |  4 ++--
 include/qemu/osdep.h |  8 ++++++++
 meson.build          | 23 +++++++++++++++++++++--
 util/osdep.c         | 10 ----------
 4 files changed, 31 insertions(+), 14 deletions(-)

-- 
2.11.0

[PATCH v3 1/3] util/osdep: Avoid madvise proto on modern Solaris

[Andrew Deason]

On older Solaris releases (before Solaris 11), we didn't get a
prototype for madvise, and so util/osdep.c provides its own prototype.
Some time between the public Solaris 11.4 release and Solaris 11.4.42
CBE, we started getting an madvise prototype that looks like this:

    extern int madvise(void *, size_t, int);

which conflicts with the prototype in util/osdeps.c. Instead of always
declaring this prototype, check if we're missing the madvise()
prototype, and only declare it ourselves if the prototype is missing.
Move the prototype to include/qemu/osdep.h, the normal place to handle
platform-specific header quirks.

The 'missing_madvise_proto' meson check contains an obviously wrong
prototype for madvise. So if that code compiles and links, we must be
missing the actual prototype for madvise.

Signed-off-by: Andrew Deason <adeason@sinenomine.net>
---
Changes since v2:
- Rename new symbol to HAVE_MADVISE_WITHOUT_PROTOTYPE
- Move madvise prototype to include/qemu/osdep.h
- More comments in meson.build

Changes since v1:
- madvise prototype check changed to not be platforms-specific, and turned into
  CONFIG_MADVISE_MISSING_PROTOTYPE.

 include/qemu/osdep.h |  8 ++++++++
 meson.build          | 23 +++++++++++++++++++++--
 util/osdep.c         |  3 ---
 3 files changed, 29 insertions(+), 5 deletions(-)

diff --git a/include/qemu/osdep.h b/include/qemu/osdep.h
index 322103aadb..f2274b24cb 100644
--- a/include/qemu/osdep.h
+++ b/include/qemu/osdep.h
@@ -393,20 +393,28 @@ void qemu_anon_ram_free(void *ptr, size_t size);
 
 #if defined(__linux__) || defined(__FreeBSD__) ||               \
     defined(__FreeBSD_kernel__) || defined(__DragonFly__)
 #define HAVE_CHARDEV_PARPORT 1
 #endif
 
 #if defined(__HAIKU__)
 #define SIGIO SIGPOLL
 #endif
 
+#ifdef HAVE_MADVISE_WITHOUT_PROTOTYPE
+/*
+ * See MySQL bug #7156 (http://bugs.mysql.com/bug.php?id=7156) for discussion
+ * about Solaris missing the madvise() prototype.
+ */
+extern int madvise(char *, size_t, int);
+#endif
+
 #if defined(CONFIG_LINUX)
 #ifndef BUS_MCEERR_AR
 #define BUS_MCEERR_AR 4
 #endif
 #ifndef BUS_MCEERR_AO
 #define BUS_MCEERR_AO 5
 #endif
 #endif
 
 #if defined(__linux__) && \
diff --git a/meson.build b/meson.build
index bae62efc9c..282e7c4650 100644
--- a/meson.build
+++ b/meson.build
@@ -1708,25 +1708,44 @@ config_host_data.set('CONFIG_EVENTFD', cc.links('''
   int main(void) { return eventfd(0, EFD_NONBLOCK | EFD_CLOEXEC); }'''))
 config_host_data.set('CONFIG_FDATASYNC', cc.links(gnu_source_prefix + '''
   #include <unistd.h>
   int main(void) {
   #if defined(_POSIX_SYNCHRONIZED_IO) && _POSIX_SYNCHRONIZED_IO > 0
   return fdatasync(0);
   #else
   #error Not supported
   #endif
   }'''))
-config_host_data.set('CONFIG_MADVISE', cc.links(gnu_source_prefix + '''
+
+has_madvise = cc.links(gnu_source_prefix + '''
   #include <sys/types.h>
   #include <sys/mman.h>
   #include <stddef.h>
-  int main(void) { return madvise(NULL, 0, MADV_DONTNEED); }'''))
+  int main(void) { return madvise(NULL, 0, MADV_DONTNEED); }''')
+missing_madvise_proto = false
+if has_madvise
+  # Some platforms (illumos and Solaris before Solaris 11) provide madvise()
+  # but forget to prototype it. In this case, has_madvise will be true (the
+  # test program links despite a compile warning). To detect the
+  # missing-prototype case, we try again with a definitely-bogus prototype.
+  # This will only compile if the system headers don't provide the prototype;
+  # otherwise the conflicting prototypes will cause a compiler error.
+  missing_madvise_proto = cc.links(gnu_source_prefix + '''
+    #include <sys/types.h>
+    #include <sys/mman.h>
+    #include <stddef.h>
+    extern int madvise(int);
+    int main(void) { return madvise(0); }''')
+endif
+config_host_data.set('CONFIG_MADVISE', has_madvise)
+config_host_data.set('HAVE_MADVISE_WITHOUT_PROTOTYPE', missing_madvise_proto)
+
 config_host_data.set('CONFIG_MEMFD', cc.links(gnu_source_prefix + '''
   #include <sys/mman.h>
   int main(void) { return memfd_create("foo", MFD_ALLOW_SEALING); }'''))
 config_host_data.set('CONFIG_OPEN_BY_HANDLE', cc.links(gnu_source_prefix + '''
   #include <fcntl.h>
   #if !defined(AT_EMPTY_PATH)
   # error missing definition
   #else
   int main(void) { struct file_handle fh; return open_by_handle_at(0, &fh, 0); }
   #endif'''))
diff --git a/util/osdep.c b/util/osdep.c
index 7c4deda6fe..1825399bcf 100644
--- a/util/osdep.c
+++ b/util/osdep.c
@@ -21,23 +21,20 @@
  * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
  * THE SOFTWARE.
  */
 #include "qemu/osdep.h"
 #include "qapi/error.h"
 
 /* Needed early for CONFIG_BSD etc. */
 
 #ifdef CONFIG_SOLARIS
 #include <sys/statvfs.h>
-/* See MySQL bug #7156 (http://bugs.mysql.com/bug.php?id=7156) for
-   discussion about Solaris header problems */
-extern int madvise(char *, size_t, int);
 #endif
 
 #include "qemu-common.h"
 #include "qemu/cutils.h"
 #include "qemu/sockets.h"
 #include "qemu/error-report.h"
 #include "qemu/madvise.h"
 #include "qemu/mprotect.h"
 #include "qemu/hw-version.h"
 #include "monitor/monitor.h"
-- 
2.11.0

Re: [PATCH v3 1/3] util/osdep: Avoid madvise proto on modern Solaris

[Peter Maydell]

On Wed, 16 Mar 2022 at 03:53, Andrew Deason <adeason@sinenomine.net> wrote:
>
> On older Solaris releases (before Solaris 11), we didn't get a
> prototype for madvise, and so util/osdep.c provides its own prototype.
> Some time between the public Solaris 11.4 release and Solaris 11.4.42
> CBE, we started getting an madvise prototype that looks like this:
>
>     extern int madvise(void *, size_t, int);
>
> which conflicts with the prototype in util/osdeps.c. Instead of always
> declaring this prototype, check if we're missing the madvise()
> prototype, and only declare it ourselves if the prototype is missing.
> Move the prototype to include/qemu/osdep.h, the normal place to handle
> platform-specific header quirks.
>
> The 'missing_madvise_proto' meson check contains an obviously wrong
> prototype for madvise. So if that code compiles and links, we must be
> missing the actual prototype for madvise.
>
> Signed-off-by: Andrew Deason <adeason@sinenomine.net>

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>

thanks
-- PMM

[PATCH v3 2/3] hw/i386/acpi-build: Avoid 'sun' identifier

[Andrew Deason]

On Solaris, 'sun' is #define'd to 1, which causes errors if a variable
is named 'sun'. Slightly change the name of the var for the Slot User
Number so we can build on Solaris.

Reviewed-by: Ani Sinha <ani@anisinha.ca>
Signed-off-by: Andrew Deason <adeason@sinenomine.net>
---
 hw/i386/acpi-build.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/hw/i386/acpi-build.c b/hw/i386/acpi-build.c
index 4ad4d7286c..dcf6ece3d0 100644
--- a/hw/i386/acpi-build.c
+++ b/hw/i386/acpi-build.c
@@ -576,32 +576,32 @@ static void build_append_pci_bus_devices(Aml *parent_scope, PCIBus *bus,
 }
 
 Aml *aml_pci_device_dsm(void)
 {
     Aml *method, *UUID, *ifctx, *ifctx1, *ifctx2, *ifctx3, *elsectx;
     Aml *acpi_index = aml_local(0);
     Aml *zero = aml_int(0);
     Aml *bnum = aml_arg(4);
     Aml *func = aml_arg(2);
     Aml *rev = aml_arg(1);
-    Aml *sun = aml_arg(5);
+    Aml *sunum = aml_arg(5);
 
     method = aml_method("PDSM", 6, AML_SERIALIZED);
 
     /*
      * PCI Firmware Specification 3.1
      * 4.6.  _DSM Definitions for PCI
      */
     UUID = aml_touuid("E5C937D0-3553-4D7A-9117-EA4D19C3434D");
     ifctx = aml_if(aml_equal(aml_arg(0), UUID));
     {
-        aml_append(ifctx, aml_store(aml_call2("AIDX", bnum, sun), acpi_index));
+        aml_append(ifctx, aml_store(aml_call2("AIDX", bnum, sunum), acpi_index));
         ifctx1 = aml_if(aml_equal(func, zero));
         {
             uint8_t byte_list[1];
 
             ifctx2 = aml_if(aml_equal(rev, aml_int(2)));
             {
                 /*
                  * advertise function 7 if device has acpi-index
                  * acpi_index values:
                  *            0: not present (default value)
-- 
2.11.0

Re: [PATCH v3 2/3] hw/i386/acpi-build: Avoid 'sun' identifier

[Michael S. Tsirkin]

On Tue, Mar 15, 2022 at 10:52:26PM -0500, Andrew Deason wrote:
> On Solaris, 'sun' is #define'd to 1, which causes errors if a variable
> is named 'sun'. Slightly change the name of the var for the Slot User
> Number so we can build on Solaris.
> 
> Reviewed-by: Ani Sinha <ani@anisinha.ca>
> Signed-off-by: Andrew Deason <adeason@sinenomine.net>


Reviewed-by: Michael S. Tsirkin <mst@redhat.com>


> ---
>  hw/i386/acpi-build.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/hw/i386/acpi-build.c b/hw/i386/acpi-build.c
> index 4ad4d7286c..dcf6ece3d0 100644
> --- a/hw/i386/acpi-build.c
> +++ b/hw/i386/acpi-build.c
> @@ -576,32 +576,32 @@ static void build_append_pci_bus_devices(Aml *parent_scope, PCIBus *bus,
>  }
>  
>  Aml *aml_pci_device_dsm(void)
>  {
>      Aml *method, *UUID, *ifctx, *ifctx1, *ifctx2, *ifctx3, *elsectx;
>      Aml *acpi_index = aml_local(0);
>      Aml *zero = aml_int(0);
>      Aml *bnum = aml_arg(4);
>      Aml *func = aml_arg(2);
>      Aml *rev = aml_arg(1);
> -    Aml *sun = aml_arg(5);
> +    Aml *sunum = aml_arg(5);
>  
>      method = aml_method("PDSM", 6, AML_SERIALIZED);
>  
>      /*
>       * PCI Firmware Specification 3.1
>       * 4.6.  _DSM Definitions for PCI
>       */
>      UUID = aml_touuid("E5C937D0-3553-4D7A-9117-EA4D19C3434D");
>      ifctx = aml_if(aml_equal(aml_arg(0), UUID));
>      {
> -        aml_append(ifctx, aml_store(aml_call2("AIDX", bnum, sun), acpi_index));
> +        aml_append(ifctx, aml_store(aml_call2("AIDX", bnum, sunum), acpi_index));
>          ifctx1 = aml_if(aml_equal(func, zero));
>          {
>              uint8_t byte_list[1];
>  
>              ifctx2 = aml_if(aml_equal(rev, aml_int(2)));
>              {
>                  /*
>                   * advertise function 7 if device has acpi-index
>                   * acpi_index values:
>                   *            0: not present (default value)
> -- 
> 2.11.0

[PATCH v3 3/3] util/osdep: Remove some early cruft

[Andrew Deason]

The include for statvfs.h has not been needed since all statvfs calls
were removed in commit 4a1418e07bdc ("Unbreak large mem support by
removing kqemu").

The comment mentioning CONFIG_BSD hasn't made sense since an include
for config-host.h was removed in commit aafd75841001 ("util: Clean up
includes").

Remove this cruft.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Andrew Deason <adeason@sinenomine.net>
---
 util/osdep.c | 7 -------
 1 file changed, 7 deletions(-)

diff --git a/util/osdep.c b/util/osdep.c
index 1825399bcf..394804d32e 100644
--- a/util/osdep.c
+++ b/util/osdep.c
@@ -16,27 +16,20 @@
  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
  * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
  * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
  * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
  * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
  * THE SOFTWARE.
  */
 #include "qemu/osdep.h"
 #include "qapi/error.h"
-
-/* Needed early for CONFIG_BSD etc. */
-
-#ifdef CONFIG_SOLARIS
-#include <sys/statvfs.h>
-#endif
-
 #include "qemu-common.h"
 #include "qemu/cutils.h"
 #include "qemu/sockets.h"
 #include "qemu/error-report.h"
 #include "qemu/madvise.h"
 #include "qemu/mprotect.h"
 #include "qemu/hw-version.h"
 #include "monitor/monitor.h"
 
 static bool fips_enabled = false;
-- 
2.11.0

Re: [PATCH v3 0/3] Fixes for building on Solaris 11.4.42 CBE

[Michael S. Tsirkin]

On Tue, Mar 15, 2022 at 10:52:24PM -0500, Andrew Deason wrote:
> With these minor fixes, I can build qemu on Solaris 11.4.42 CBE
> (Oracle's new rolling release thing), using '--disable-rdma
> --enable-modules --disable-dbus-display --target-list=x86_64-softmmu'.
> I'm just interested in the guest agent right now, so that's all I've
> tested (briefly), but the rest of the build wasn't hard to get working.
> With this, the guest agent runs fine using isa-serial.

Who's merging this? Me?

> Changes since v2:
> - Rename new symbol to HAVE_MADVISE_WITHOUT_PROTOTYPE
> - Move madvise prototype to include/qemu/osdep.h
> - More comments in meson.build
> 
> Changes since v1:
> - Change the CONFIG_MADVISE checks to not be platform-specific
> - Add the last commit removing util/osdep.c cruft
> 
> Andrew Deason (3):
>   util/osdep: Avoid madvise proto on modern Solaris
>   hw/i386/acpi-build: Avoid 'sun' identifier
>   util/osdep: Remove some early cruft
> 
>  hw/i386/acpi-build.c |  4 ++--
>  include/qemu/osdep.h |  8 ++++++++
>  meson.build          | 23 +++++++++++++++++++++--
>  util/osdep.c         | 10 ----------
>  4 files changed, 31 insertions(+), 14 deletions(-)
> 
> -- 
> 2.11.0

Re: [PATCH v3 0/3] Fixes for building on Solaris 11.4.42 CBE

[Peter Maydell]

On Wed, 16 Mar 2022 at 08:55, Michael S. Tsirkin <mst@redhat.com> wrote:
>
> On Tue, Mar 15, 2022 at 10:52:24PM -0500, Andrew Deason wrote:
> > With these minor fixes, I can build qemu on Solaris 11.4.42 CBE
> > (Oracle's new rolling release thing), using '--disable-rdma
> > --enable-modules --disable-dbus-display --target-list=x86_64-softmmu'.
> > I'm just interested in the guest agent right now, so that's all I've
> > tested (briefly), but the rest of the build wasn't hard to get working.
> > With this, the guest agent runs fine using isa-serial.
>
> Who's merging this? Me?

I'm putting together a pullreq this morning anyway, so I can
take it through target-arm.next.

thanks
-- PMM