[Qemu-devel] [PATCH v3 0/4] Initial Windows on ARM (AArch64 64-Bit) host support

[Qemu-devel] [PATCH v3 0/4] Initial Windows on ARM (AArch64 64-Bit) host support

[Cao Jiaxi]

Initial Windows on ARM (AArch64 64-Bit) host support

This series of patches is for initial support of Windows 10 on ARM as a QEMU host.
Currently only TCG intepreter is working correctly, it crashes when TCG JIT is enabled.

For now we assume it is built using the clang aarch64-w64-mingw32 toolchain,
you can get a prebuilt toolchain at https://github.com/mstorsjo/llvm-mingw.

Cao Jiaxi (4):
  QEMU_PACKED: Remove gcc_struct attribute in Windows non x86 targets
  qga: Fix mingw compilation warnings on enum conversion
  util/cacheinfo: Use uint64_t on LLP64 model to satisfy Windows ARM64
  osdep: Fix mingw compilation regarding stdio formats

 contrib/libvhost-user/libvhost-user.h |  2 +-
 include/qemu/compiler.h               |  2 +-
 include/qemu/osdep.h                  | 10 +++++-----
 qga/commands-win32.c                  |  2 +-
 scripts/cocci-macro-file.h            |  7 ++++++-
 slirp/src/util.h                      |  2 +-
 util/cacheinfo.c                      |  2 +-
 7 files changed, 16 insertions(+), 11 deletions(-)

-- 
2.17.1

[Qemu-devel] [PATCH v3 1/4] QEMU_PACKED: Remove gcc_struct attribute in Windows non x86 targets

[Cao Jiaxi]

gcc_struct is for x86 only, and it generates an warning on ARM64 Clang/MinGW targets.

Signed-off-by: Cao Jiaxi <driver1998@foxmail.com>
---
 contrib/libvhost-user/libvhost-user.h | 2 +-
 include/qemu/compiler.h               | 2 +-
 scripts/cocci-macro-file.h            | 7 ++++++-
 slirp/src/util.h                      | 2 +-
 4 files changed, 9 insertions(+), 4 deletions(-)

diff --git a/contrib/libvhost-user/libvhost-user.h b/contrib/libvhost-user/libvhost-user.h
index 414ceb0a2f..78b33306e8 100644
--- a/contrib/libvhost-user/libvhost-user.h
+++ b/contrib/libvhost-user/libvhost-user.h
@@ -148,7 +148,7 @@ typedef struct VhostUserInflight {
     uint16_t queue_size;
 } VhostUserInflight;
 
-#if defined(_WIN32)
+#if defined(_WIN32) && (defined(__x86_64__) || defined(__i386__))
 # define VU_PACKED __attribute__((gcc_struct, packed))
 #else
 # define VU_PACKED __attribute__((packed))
diff --git a/include/qemu/compiler.h b/include/qemu/compiler.h
index 296b2fd572..09fc44cca4 100644
--- a/include/qemu/compiler.h
+++ b/include/qemu/compiler.h
@@ -28,7 +28,7 @@
 
 #define QEMU_SENTINEL __attribute__((sentinel))
 
-#if defined(_WIN32)
+#if defined(_WIN32) && (defined(__x86_64__) || defined(__i386__))
 # define QEMU_PACKED __attribute__((gcc_struct, packed))
 #else
 # define QEMU_PACKED __attribute__((packed))
diff --git a/scripts/cocci-macro-file.h b/scripts/cocci-macro-file.h
index e485cdccae..c6bbc05ba3 100644
--- a/scripts/cocci-macro-file.h
+++ b/scripts/cocci-macro-file.h
@@ -23,7 +23,12 @@
 #define QEMU_NORETURN __attribute__ ((__noreturn__))
 #define QEMU_WARN_UNUSED_RESULT __attribute__((warn_unused_result))
 #define QEMU_SENTINEL __attribute__((sentinel))
-#define QEMU_PACKED __attribute__((gcc_struct, packed))
+
+#if defined(_WIN32) && (defined(__x86_64__) || defined(__i386__))
+# define QEMU_PACKED __attribute__((gcc_struct, packed))
+#else
+# define QEMU_PACKED __attribute__((packed))
+#endif
 
 #define cat(x,y) x ## y
 #define cat2(x,y) cat(x,y)
diff --git a/slirp/src/util.h b/slirp/src/util.h
index 01f1e0e068..278828fe3f 100644
--- a/slirp/src/util.h
+++ b/slirp/src/util.h
@@ -43,7 +43,7 @@
 #include <netinet/in.h>
 #endif
 
-#if defined(_WIN32)
+#if defined(_WIN32) && (defined(__x86_64__) || defined(__i386__))
 # define SLIRP_PACKED __attribute__((gcc_struct, packed))
 #else
 # define SLIRP_PACKED __attribute__((packed))
-- 
2.17.1

[Qemu-devel] [PATCH v3 2/4] qga: Fix mingw compilation warnings on enum conversion

[Cao Jiaxi]

The win2qemu[] is supposed to be the conversion table to convert between
STORAGE_BUS_TYPE in Windows SDK and GuestDiskBusType in qga.

But it was incorrectly written that it forces to set a GuestDiskBusType
value to STORAGE_BUS_TYPE, which generates an enum conversion warning in clang.

Signed-off-by: Cao Jiaxi <driver1998@foxmail.com>
---
 qga/commands-win32.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/qga/commands-win32.c b/qga/commands-win32.c
index d40d61f605..6b67f16faf 100644
--- a/qga/commands-win32.c
+++ b/qga/commands-win32.c
@@ -457,7 +457,7 @@ void qmp_guest_file_flush(int64_t handle, Error **errp)
 
 #ifdef CONFIG_QGA_NTDDSCSI
 
-static STORAGE_BUS_TYPE win2qemu[] = {
+static GuestDiskBusType win2qemu[] = {
     [BusTypeUnknown] = GUEST_DISK_BUS_TYPE_UNKNOWN,
     [BusTypeScsi] = GUEST_DISK_BUS_TYPE_SCSI,
     [BusTypeAtapi] = GUEST_DISK_BUS_TYPE_IDE,
-- 
2.17.1

[Qemu-devel] [PATCH v3 3/4] util/cacheinfo: Use uint64_t on LLP64 model to satisfy Windows ARM64

[Cao Jiaxi]

Windows ARM64 uses LLP64 model, which breaks current assumptions.

Signed-off-by: Cao Jiaxi <driver1998@foxmail.com>
---
 util/cacheinfo.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/util/cacheinfo.c b/util/cacheinfo.c
index 3cd080b83d..eebe1ce9c5 100644
--- a/util/cacheinfo.c
+++ b/util/cacheinfo.c
@@ -107,7 +107,7 @@ static void sys_cache_info(int *isize, int *dsize)
 static void arch_cache_info(int *isize, int *dsize)
 {
     if (*isize == 0 || *dsize == 0) {
-        unsigned long ctr;
+        uint64_t ctr;
 
         /* The real cache geometry is in CCSIDR_EL1/CLIDR_EL1/CSSELR_EL1,
            but (at least under Linux) these are marked protected by the
-- 
2.17.1

[Qemu-devel] [PATCH v3 4/4] osdep: Fix mingw compilation regarding stdio formats

[Cao Jiaxi]

I encountered the following compilation error on mingw:

/mnt/d/qemu/include/qemu/osdep.h:97:9: error: '__USE_MINGW_ANSI_STDIO' macro redefined [-Werror,-Wmacro-redefined]
 #define __USE_MINGW_ANSI_STDIO 1
        ^
/mnt/d/llvm-mingw/aarch64-w64-mingw32/include/_mingw.h:433:9: note: previous definition is here
 #define __USE_MINGW_ANSI_STDIO 0      /* was not defined so it should be 0 */

It turns out that __USE_MINGW_ANSI_STDIO must be set before any
system headers are included, not just before stdio.h.

Signed-off-by: Cao Jiaxi <driver1998@foxmail.com>
---
 include/qemu/osdep.h | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/include/qemu/osdep.h b/include/qemu/osdep.h
index 303d315c5d..af2b91f0b8 100644
--- a/include/qemu/osdep.h
+++ b/include/qemu/osdep.h
@@ -85,17 +85,17 @@ extern int daemon(int, int);
 #endif
 #endif
 
+/* enable C99/POSIX format strings (needs mingw32-runtime 3.15 or later) */
+#ifdef __MINGW32__
+#define __USE_MINGW_ANSI_STDIO 1
+#endif
+
 #include <stdarg.h>
 #include <stddef.h>
 #include <stdbool.h>
 #include <stdint.h>
 #include <sys/types.h>
 #include <stdlib.h>
-
-/* enable C99/POSIX format strings (needs mingw32-runtime 3.15 or later) */
-#ifdef __MINGW32__
-#define __USE_MINGW_ANSI_STDIO 1
-#endif
 #include <stdio.h>
 
 #include <string.h>
-- 
2.17.1

Re: [Qemu-devel] [PATCH v3 3/4] util/cacheinfo: Use uint64_t on LLP64 model to satisfy Windows ARM64

[Richard Henderson]

On 5/2/19 5:37 PM, Cao Jiaxi wrote:
> Windows ARM64 uses LLP64 model, which breaks current assumptions.
> 
> Signed-off-by: Cao Jiaxi <driver1998@foxmail.com>
> ---
>  util/cacheinfo.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/util/cacheinfo.c b/util/cacheinfo.c
> index 3cd080b83d..eebe1ce9c5 100644
> --- a/util/cacheinfo.c
> +++ b/util/cacheinfo.c
> @@ -107,7 +107,7 @@ static void sys_cache_info(int *isize, int *dsize)
>  static void arch_cache_info(int *isize, int *dsize)
>  {
>      if (*isize == 0 || *dsize == 0) {
> -        unsigned long ctr;
> +        uint64_t ctr;
>  
>          /* The real cache geometry is in CCSIDR_EL1/CLIDR_EL1/CSSELR_EL1,
>             but (at least under Linux) these are marked protected by the
> 

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>

r~

Re: [Qemu-devel] [PATCH v3 2/4] qga: Fix mingw compilation warnings on enum conversion

[Richard Henderson]

On 5/2/19 5:36 PM, Cao Jiaxi wrote:
> The win2qemu[] is supposed to be the conversion table to convert between
> STORAGE_BUS_TYPE in Windows SDK and GuestDiskBusType in qga.
> 
> But it was incorrectly written that it forces to set a GuestDiskBusType
> value to STORAGE_BUS_TYPE, which generates an enum conversion warning in clang.
> 
> Signed-off-by: Cao Jiaxi <driver1998@foxmail.com>
> ---
>  qga/commands-win32.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/qga/commands-win32.c b/qga/commands-win32.c
> index d40d61f605..6b67f16faf 100644
> --- a/qga/commands-win32.c
> +++ b/qga/commands-win32.c
> @@ -457,7 +457,7 @@ void qmp_guest_file_flush(int64_t handle, Error **errp)
>  
>  #ifdef CONFIG_QGA_NTDDSCSI
>  
> -static STORAGE_BUS_TYPE win2qemu[] = {
> +static GuestDiskBusType win2qemu[] = {
>      [BusTypeUnknown] = GUEST_DISK_BUS_TYPE_UNKNOWN,
>      [BusTypeScsi] = GUEST_DISK_BUS_TYPE_SCSI,
>      [BusTypeAtapi] = GUEST_DISK_BUS_TYPE_IDE,
> 


Reviewed-by: Richard Henderson <richard.henderson@linaro.org>

r~

Re: [Qemu-devel] [PATCH v3 1/4] QEMU_PACKED: Remove gcc_struct attribute in Windows non x86 targets

[Thomas Huth]

On 03/05/2019 02.36, Cao Jiaxi wrote:
> gcc_struct is for x86 only, and it generates an warning on ARM64 Clang/MinGW targets.
> 
> Signed-off-by: Cao Jiaxi <driver1998@foxmail.com>
> ---
>  contrib/libvhost-user/libvhost-user.h | 2 +-
>  include/qemu/compiler.h               | 2 +-
>  scripts/cocci-macro-file.h            | 7 ++++++-
>  slirp/src/util.h                      | 2 +-
>  4 files changed, 9 insertions(+), 4 deletions(-)
> 
> diff --git a/contrib/libvhost-user/libvhost-user.h b/contrib/libvhost-user/libvhost-user.h
> index 414ceb0a2f..78b33306e8 100644
> --- a/contrib/libvhost-user/libvhost-user.h
> +++ b/contrib/libvhost-user/libvhost-user.h
> @@ -148,7 +148,7 @@ typedef struct VhostUserInflight {
>      uint16_t queue_size;
>  } VhostUserInflight;
>  
> -#if defined(_WIN32)
> +#if defined(_WIN32) && (defined(__x86_64__) || defined(__i386__))
>  # define VU_PACKED __attribute__((gcc_struct, packed))
>  #else
>  # define VU_PACKED __attribute__((packed))
> diff --git a/include/qemu/compiler.h b/include/qemu/compiler.h
> index 296b2fd572..09fc44cca4 100644
> --- a/include/qemu/compiler.h
> +++ b/include/qemu/compiler.h
> @@ -28,7 +28,7 @@
>  
>  #define QEMU_SENTINEL __attribute__((sentinel))
>  
> -#if defined(_WIN32)
> +#if defined(_WIN32) && (defined(__x86_64__) || defined(__i386__))
>  # define QEMU_PACKED __attribute__((gcc_struct, packed))
>  #else
>  # define QEMU_PACKED __attribute__((packed))
> diff --git a/scripts/cocci-macro-file.h b/scripts/cocci-macro-file.h
> index e485cdccae..c6bbc05ba3 100644
> --- a/scripts/cocci-macro-file.h
> +++ b/scripts/cocci-macro-file.h
> @@ -23,7 +23,12 @@
>  #define QEMU_NORETURN __attribute__ ((__noreturn__))
>  #define QEMU_WARN_UNUSED_RESULT __attribute__((warn_unused_result))
>  #define QEMU_SENTINEL __attribute__((sentinel))
> -#define QEMU_PACKED __attribute__((gcc_struct, packed))
> +
> +#if defined(_WIN32) && (defined(__x86_64__) || defined(__i386__))
> +# define QEMU_PACKED __attribute__((gcc_struct, packed))
> +#else
> +# define QEMU_PACKED __attribute__((packed))
> +#endif
>  
>  #define cat(x,y) x ## y
>  #define cat2(x,y) cat(x,y)
> diff --git a/slirp/src/util.h b/slirp/src/util.h
> index 01f1e0e068..278828fe3f 100644
> --- a/slirp/src/util.h
> +++ b/slirp/src/util.h
> @@ -43,7 +43,7 @@
>  #include <netinet/in.h>
>  #endif
>  
> -#if defined(_WIN32)
> +#if defined(_WIN32) && (defined(__x86_64__) || defined(__i386__))
>  # define SLIRP_PACKED __attribute__((gcc_struct, packed))
>  #else
>  # define SLIRP_PACKED __attribute__((packed))
> 

The slirp code is currently on its way into a separate module, so you
might need to provide that hunk to the libslirp folks again... I'm
putting the slirp maintainers on CC:, maybe they can pick it up from here.

Anyway:

Reviewed-by: Thomas Huth <thuth@redhat.com>

Re: [Qemu-devel] [PATCH v3 1/4] QEMU_PACKED: Remove gcc_struct attribute in Windows non x86 targets

[Thomas Huth]

On 03/05/2019 02.36, Cao Jiaxi wrote:
> gcc_struct is for x86 only, and it generates an warning on ARM64 Clang/MinGW targets.
> 
> Signed-off-by: Cao Jiaxi <driver1998@foxmail.com>
> ---
>  contrib/libvhost-user/libvhost-user.h | 2 +-
>  include/qemu/compiler.h               | 2 +-
>  scripts/cocci-macro-file.h            | 7 ++++++-
>  slirp/src/util.h                      | 2 +-
>  4 files changed, 9 insertions(+), 4 deletions(-)
> 
> diff --git a/contrib/libvhost-user/libvhost-user.h b/contrib/libvhost-user/libvhost-user.h
> index 414ceb0a2f..78b33306e8 100644
> --- a/contrib/libvhost-user/libvhost-user.h
> +++ b/contrib/libvhost-user/libvhost-user.h
> @@ -148,7 +148,7 @@ typedef struct VhostUserInflight {
>      uint16_t queue_size;
>  } VhostUserInflight;
>  
> -#if defined(_WIN32)
> +#if defined(_WIN32) && (defined(__x86_64__) || defined(__i386__))
>  # define VU_PACKED __attribute__((gcc_struct, packed))
>  #else
>  # define VU_PACKED __attribute__((packed))
> diff --git a/include/qemu/compiler.h b/include/qemu/compiler.h
> index 296b2fd572..09fc44cca4 100644
> --- a/include/qemu/compiler.h
> +++ b/include/qemu/compiler.h
> @@ -28,7 +28,7 @@
>  
>  #define QEMU_SENTINEL __attribute__((sentinel))
>  
> -#if defined(_WIN32)
> +#if defined(_WIN32) && (defined(__x86_64__) || defined(__i386__))
>  # define QEMU_PACKED __attribute__((gcc_struct, packed))
>  #else
>  # define QEMU_PACKED __attribute__((packed))
> diff --git a/scripts/cocci-macro-file.h b/scripts/cocci-macro-file.h
> index e485cdccae..c6bbc05ba3 100644
> --- a/scripts/cocci-macro-file.h
> +++ b/scripts/cocci-macro-file.h
> @@ -23,7 +23,12 @@
>  #define QEMU_NORETURN __attribute__ ((__noreturn__))
>  #define QEMU_WARN_UNUSED_RESULT __attribute__((warn_unused_result))
>  #define QEMU_SENTINEL __attribute__((sentinel))
> -#define QEMU_PACKED __attribute__((gcc_struct, packed))
> +
> +#if defined(_WIN32) && (defined(__x86_64__) || defined(__i386__))
> +# define QEMU_PACKED __attribute__((gcc_struct, packed))
> +#else
> +# define QEMU_PACKED __attribute__((packed))
> +#endif
>  
>  #define cat(x,y) x ## y
>  #define cat2(x,y) cat(x,y)
> diff --git a/slirp/src/util.h b/slirp/src/util.h
> index 01f1e0e068..278828fe3f 100644
> --- a/slirp/src/util.h
> +++ b/slirp/src/util.h
> @@ -43,7 +43,7 @@
>  #include <netinet/in.h>
>  #endif
>  
> -#if defined(_WIN32)
> +#if defined(_WIN32) && (defined(__x86_64__) || defined(__i386__))
>  # define SLIRP_PACKED __attribute__((gcc_struct, packed))
>  #else
>  # define SLIRP_PACKED __attribute__((packed))
> 

The slirp code is currently on its way into a separate module, so you
might need to provide that hunk to the libslirp folks again... I'm
putting the slirp maintainers on CC:, maybe they can pick it up from here.

Anyway:

Reviewed-by: Thomas Huth <thuth@redhat.com>

Re: [Qemu-devel] [PATCH v3 2/4] qga: Fix mingw compilation warnings on enum conversion

[Thomas Huth]

On 03/05/2019 02.36, Cao Jiaxi wrote:
> The win2qemu[] is supposed to be the conversion table to convert between
> STORAGE_BUS_TYPE in Windows SDK and GuestDiskBusType in qga.
> 
> But it was incorrectly written that it forces to set a GuestDiskBusType
> value to STORAGE_BUS_TYPE, which generates an enum conversion warning in clang.
> 
> Signed-off-by: Cao Jiaxi <driver1998@foxmail.com>
> ---
>  qga/commands-win32.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/qga/commands-win32.c b/qga/commands-win32.c
> index d40d61f605..6b67f16faf 100644
> --- a/qga/commands-win32.c
> +++ b/qga/commands-win32.c
> @@ -457,7 +457,7 @@ void qmp_guest_file_flush(int64_t handle, Error **errp)
>  
>  #ifdef CONFIG_QGA_NTDDSCSI
>  
> -static STORAGE_BUS_TYPE win2qemu[] = {
> +static GuestDiskBusType win2qemu[] = {
>      [BusTypeUnknown] = GUEST_DISK_BUS_TYPE_UNKNOWN,
>      [BusTypeScsi] = GUEST_DISK_BUS_TYPE_SCSI,
>      [BusTypeAtapi] = GUEST_DISK_BUS_TYPE_IDE,

Reviewed-by: Thomas Huth <thuth@redhat.com>

Re: [Qemu-devel] [PATCH v3 2/4] qga: Fix mingw compilation warnings on enum conversion

[Thomas Huth]

On 03/05/2019 02.36, Cao Jiaxi wrote:
> The win2qemu[] is supposed to be the conversion table to convert between
> STORAGE_BUS_TYPE in Windows SDK and GuestDiskBusType in qga.
> 
> But it was incorrectly written that it forces to set a GuestDiskBusType
> value to STORAGE_BUS_TYPE, which generates an enum conversion warning in clang.
> 
> Signed-off-by: Cao Jiaxi <driver1998@foxmail.com>
> ---
>  qga/commands-win32.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/qga/commands-win32.c b/qga/commands-win32.c
> index d40d61f605..6b67f16faf 100644
> --- a/qga/commands-win32.c
> +++ b/qga/commands-win32.c
> @@ -457,7 +457,7 @@ void qmp_guest_file_flush(int64_t handle, Error **errp)
>  
>  #ifdef CONFIG_QGA_NTDDSCSI
>  
> -static STORAGE_BUS_TYPE win2qemu[] = {
> +static GuestDiskBusType win2qemu[] = {
>      [BusTypeUnknown] = GUEST_DISK_BUS_TYPE_UNKNOWN,
>      [BusTypeScsi] = GUEST_DISK_BUS_TYPE_SCSI,
>      [BusTypeAtapi] = GUEST_DISK_BUS_TYPE_IDE,

Reviewed-by: Thomas Huth <thuth@redhat.com>

Re: [Qemu-devel] [PATCH v3 3/4] util/cacheinfo: Use uint64_t on LLP64 model to satisfy Windows ARM64

[Thomas Huth]

On 03/05/2019 02.37, Cao Jiaxi wrote:
> Windows ARM64 uses LLP64 model, which breaks current assumptions.
> 
> Signed-off-by: Cao Jiaxi <driver1998@foxmail.com>
> ---
>  util/cacheinfo.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/util/cacheinfo.c b/util/cacheinfo.c
> index 3cd080b83d..eebe1ce9c5 100644
> --- a/util/cacheinfo.c
> +++ b/util/cacheinfo.c
> @@ -107,7 +107,7 @@ static void sys_cache_info(int *isize, int *dsize)
>  static void arch_cache_info(int *isize, int *dsize)
>  {
>      if (*isize == 0 || *dsize == 0) {
> -        unsigned long ctr;
> +        uint64_t ctr;
>  
>          /* The real cache geometry is in CCSIDR_EL1/CLIDR_EL1/CSSELR_EL1,
>             but (at least under Linux) these are marked protected by the
> 

Reviewed-by: Thomas Huth <thuth@redhat.com>

Re: [Qemu-devel] [PATCH v3 4/4] osdep: Fix mingw compilation regarding stdio formats

[Thomas Huth]

On 03/05/2019 02.37, Cao Jiaxi wrote:
> I encountered the following compilation error on mingw:
> 
> /mnt/d/qemu/include/qemu/osdep.h:97:9: error: '__USE_MINGW_ANSI_STDIO' macro redefined [-Werror,-Wmacro-redefined]
>  #define __USE_MINGW_ANSI_STDIO 1
>         ^
> /mnt/d/llvm-mingw/aarch64-w64-mingw32/include/_mingw.h:433:9: note: previous definition is here
>  #define __USE_MINGW_ANSI_STDIO 0      /* was not defined so it should be 0 */
> 
> It turns out that __USE_MINGW_ANSI_STDIO must be set before any
> system headers are included, not just before stdio.h.
> 
> Signed-off-by: Cao Jiaxi <driver1998@foxmail.com>
> ---
>  include/qemu/osdep.h | 10 +++++-----
>  1 file changed, 5 insertions(+), 5 deletions(-)
> 
> diff --git a/include/qemu/osdep.h b/include/qemu/osdep.h
> index 303d315c5d..af2b91f0b8 100644
> --- a/include/qemu/osdep.h
> +++ b/include/qemu/osdep.h
> @@ -85,17 +85,17 @@ extern int daemon(int, int);
>  #endif
>  #endif
>  
> +/* enable C99/POSIX format strings (needs mingw32-runtime 3.15 or later) */
> +#ifdef __MINGW32__
> +#define __USE_MINGW_ANSI_STDIO 1
> +#endif
> +
>  #include <stdarg.h>
>  #include <stddef.h>
>  #include <stdbool.h>
>  #include <stdint.h>
>  #include <sys/types.h>
>  #include <stdlib.h>
> -
> -/* enable C99/POSIX format strings (needs mingw32-runtime 3.15 or later) */
> -#ifdef __MINGW32__
> -#define __USE_MINGW_ANSI_STDIO 1
> -#endif
>  #include <stdio.h>
>  
>  #include <string.h>
> 

Reviewed-by: Thomas Huth <thuth@redhat.com>

Re: [Qemu-devel] [PATCH v3 4/4] osdep: Fix mingw compilation regarding stdio formats

[Stefan Weil]

[-- Attachment #1: Type: text/plain, Size: 1672 bytes --]

On 03.05.19 07:11, Thomas Huth wrote:
> On 03/05/2019 02.37, Cao Jiaxi wrote:
>> I encountered the following compilation error on mingw:
>>
>> /mnt/d/qemu/include/qemu/osdep.h:97:9: error: '__USE_MINGW_ANSI_STDIO' macro redefined [-Werror,-Wmacro-redefined]
>>  #define __USE_MINGW_ANSI_STDIO 1
>>         ^
>> /mnt/d/llvm-mingw/aarch64-w64-mingw32/include/_mingw.h:433:9: note: previous definition is here
>>  #define __USE_MINGW_ANSI_STDIO 0      /* was not defined so it should be 0 */
>>
>> It turns out that __USE_MINGW_ANSI_STDIO must be set before any
>> system headers are included, not just before stdio.h.
>>
>> Signed-off-by: Cao Jiaxi <driver1998@foxmail.com>
>> ---
>>  include/qemu/osdep.h | 10 +++++-----
>>  1 file changed, 5 insertions(+), 5 deletions(-)
>>
>> diff --git a/include/qemu/osdep.h b/include/qemu/osdep.h
>> index 303d315c5d..af2b91f0b8 100644
>> --- a/include/qemu/osdep.h
>> +++ b/include/qemu/osdep.h
>> @@ -85,17 +85,17 @@ extern int daemon(int, int);
>>  #endif
>>  #endif
>>  
>> +/* enable C99/POSIX format strings (needs mingw32-runtime 3.15 or later) */
>> +#ifdef __MINGW32__
>> +#define __USE_MINGW_ANSI_STDIO 1
>> +#endif
>> +
>>  #include <stdarg.h>
>>  #include <stddef.h>
>>  #include <stdbool.h>
>>  #include <stdint.h>
>>  #include <sys/types.h>
>>  #include <stdlib.h>
>> -
>> -/* enable C99/POSIX format strings (needs mingw32-runtime 3.15 or later) */
>> -#ifdef __MINGW32__
>> -#define __USE_MINGW_ANSI_STDIO 1
>> -#endif
>>  #include <stdio.h>
>>  
>>  #include <string.h>
>>
> 
> Reviewed-by: Thomas Huth <thuth@redhat.com>

Reviewed-by: Stefan Weil <sw@weilnetz.de>


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [Qemu-devel] [PATCH v3 0/4] Initial Windows on ARM (AArch64 64-Bit) host support

[Philippe Mathieu-Daudé]

On 5/3/19 2:22 AM, Cao Jiaxi wrote:
> Initial Windows on ARM (AArch64 64-Bit) host support
> 
> This series of patches is for initial support of Windows 10 on ARM as a QEMU host.
> Currently only TCG intepreter is working correctly, it crashes when TCG JIT is enabled.
> 
> For now we assume it is built using the clang aarch64-w64-mingw32 toolchain,
> you can get a prebuilt toolchain at https://github.com/mstorsjo/llvm-mingw.
> 

I'm a bit confused since those patches appear 2 times in my mailbox.
Assuming this is a git send-email setup mistake, please consider adding
my Reviewed-by/Tested-by here too.

These are:

> Cao Jiaxi (4):
>   QEMU_PACKED: Remove gcc_struct attribute in Windows non x86 targets

No tag.

>   qga: Fix mingw compilation warnings on enum conversion

Suggested-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>

>   util/cacheinfo: Use uint64_t on LLP64 model to satisfy Windows ARM64

Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>

>   osdep: Fix mingw compilation regarding stdio formats

Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>

Thanks,

Phil.

Re: [Qemu-devel] [PATCH v3 0/4] Initial Windows on ARM (AArch64 64-Bit) host support

[Peter Maydell]

On Fri, 3 May 2019 at 11:20, Philippe Mathieu-Daudé <philmd@redhat.com> wrote:
>
> On 5/3/19 2:22 AM, Cao Jiaxi wrote:
> > Initial Windows on ARM (AArch64 64-Bit) host support
> >
> > This series of patches is for initial support of Windows 10 on ARM as a QEMU host.
> > Currently only TCG intepreter is working correctly, it crashes when TCG JIT is enabled.
> >
> > For now we assume it is built using the clang aarch64-w64-mingw32 toolchain,
> > you can get a prebuilt toolchain at https://github.com/mstorsjo/llvm-mingw.
> >
>
> I'm a bit confused since those patches appear 2 times in my mailbox.
> Assuming this is a git send-email setup mistake, please consider adding
> my Reviewed-by/Tested-by here too.

Thanks -- I've applied these patches to target-arm.next, and have
consolidated the various tags that people have applied to the
two sets of emails (and checked that they actually are the same
patches and not different code :-))

-- PMM

Re: [Qemu-devel] [PATCH v3 1/4] QEMU_PACKED: Remove gcc_struct attribute in Windows non x86 targets

[Peter Maydell]

On Fri, 3 May 2019 at 06:07, Thomas Huth <thuth@redhat.com> wrote:
>
> On 03/05/2019 02.36, Cao Jiaxi wrote:
> > gcc_struct is for x86 only, and it generates an warning on ARM64 Clang/MinGW targets.
> >
> > Signed-off-by: Cao Jiaxi <driver1998@foxmail.com>
> > ---
> >  contrib/libvhost-user/libvhost-user.h | 2 +-
> >  include/qemu/compiler.h               | 2 +-
> >  scripts/cocci-macro-file.h            | 7 ++++++-
> >  slirp/src/util.h                      | 2 +-
> >  4 files changed, 9 insertions(+), 4 deletions(-)

> > diff --git a/slirp/src/util.h b/slirp/src/util.h
> > index 01f1e0e068..278828fe3f 100644
> > --- a/slirp/src/util.h
> > +++ b/slirp/src/util.h
> > @@ -43,7 +43,7 @@
> >  #include <netinet/in.h>
> >  #endif
> >
> > -#if defined(_WIN32)
> > +#if defined(_WIN32) && (defined(__x86_64__) || defined(__i386__))
> >  # define SLIRP_PACKED __attribute__((gcc_struct, packed))
> >  #else
> >  # define SLIRP_PACKED __attribute__((packed))
> >
>
> The slirp code is currently on its way into a separate module, so you
> might need to provide that hunk to the libslirp folks again... I'm
> putting the slirp maintainers on CC:, maybe they can pick it up from here.

Yes, the slirp module has now landed in master, so this patch
definitely needs to be split into two. I've kept in my
target-arm.next tree the parts which are applicable to
the QEMU repo itself (ie everything except the slirp/ change),
so we just need a new patch for the slirp submodule part.

Marc-André, Samuel -- what's the process for submitting and
getting reviewed changes to the slirp submodule now it's a
separate project ?

thanks
-- PMM

Re: [Qemu-devel] [PATCH v3 1/4] QEMU_PACKED: Remove gcc_struct attribute in Windows non x86 targets

[Peter Maydell]

On Fri, 3 May 2019 at 06:07, Thomas Huth <thuth@redhat.com> wrote:
>
> On 03/05/2019 02.36, Cao Jiaxi wrote:
> > gcc_struct is for x86 only, and it generates an warning on ARM64 Clang/MinGW targets.
> >
> > Signed-off-by: Cao Jiaxi <driver1998@foxmail.com>
> > ---
> >  contrib/libvhost-user/libvhost-user.h | 2 +-
> >  include/qemu/compiler.h               | 2 +-
> >  scripts/cocci-macro-file.h            | 7 ++++++-
> >  slirp/src/util.h                      | 2 +-
> >  4 files changed, 9 insertions(+), 4 deletions(-)

> > diff --git a/slirp/src/util.h b/slirp/src/util.h
> > index 01f1e0e068..278828fe3f 100644
> > --- a/slirp/src/util.h
> > +++ b/slirp/src/util.h
> > @@ -43,7 +43,7 @@
> >  #include <netinet/in.h>
> >  #endif
> >
> > -#if defined(_WIN32)
> > +#if defined(_WIN32) && (defined(__x86_64__) || defined(__i386__))
> >  # define SLIRP_PACKED __attribute__((gcc_struct, packed))
> >  #else
> >  # define SLIRP_PACKED __attribute__((packed))
> >
>
> The slirp code is currently on its way into a separate module, so you
> might need to provide that hunk to the libslirp folks again... I'm
> putting the slirp maintainers on CC:, maybe they can pick it up from here.

Yes, the slirp module has now landed in master, so this patch
definitely needs to be split into two. I've kept in my
target-arm.next tree the parts which are applicable to
the QEMU repo itself (ie everything except the slirp/ change),
so we just need a new patch for the slirp submodule part.

Marc-André, Samuel -- what's the process for submitting and
getting reviewed changes to the slirp submodule now it's a
separate project ?

thanks
-- PMM

Re: [Qemu-devel] [PATCH v3 1/4] QEMU_PACKED: Remove gcc_struct attribute in Windows non x86 targets

[Marc-André Lureau]

Hi

Le ven. 3 mai 2019 à 17:23, Peter Maydell <peter.maydell@linaro.org> a
écrit :

> On Fri, 3 May 2019 at 06:07, Thomas Huth <thuth@redhat.com> wrote:
> >
> > On 03/05/2019 02.36, Cao Jiaxi wrote:
> > > gcc_struct is for x86 only, and it generates an warning on ARM64
> Clang/MinGW targets.
> > >
> > > Signed-off-by: Cao Jiaxi <driver1998@foxmail.com>
> > > ---
> > >  contrib/libvhost-user/libvhost-user.h | 2 +-
> > >  include/qemu/compiler.h               | 2 +-
> > >  scripts/cocci-macro-file.h            | 7 ++++++-
> > >  slirp/src/util.h                      | 2 +-
> > >  4 files changed, 9 insertions(+), 4 deletions(-)
>
> > > diff --git a/slirp/src/util.h b/slirp/src/util.h
> > > index 01f1e0e068..278828fe3f 100644
> > > --- a/slirp/src/util.h
> > > +++ b/slirp/src/util.h
> > > @@ -43,7 +43,7 @@
> > >  #include <netinet/in.h>
> > >  #endif
> > >
> > > -#if defined(_WIN32)
> > > +#if defined(_WIN32) && (defined(__x86_64__) || defined(__i386__))
> > >  # define SLIRP_PACKED __attribute__((gcc_struct, packed))
> > >  #else
> > >  # define SLIRP_PACKED __attribute__((packed))
> > >
> >
> > The slirp code is currently on its way into a separate module, so you
> > might need to provide that hunk to the libslirp folks again... I'm
> > putting the slirp maintainers on CC:, maybe they can pick it up from
> here.
>
> Yes, the slirp module has now landed in master, so this patch
> definitely needs to be split into two. I've kept in my
> target-arm.next tree the parts which are applicable to
> the QEMU repo itself (ie everything except the slirp/ change),
> so we just need a new patch for the slirp submodule part.
>
> Marc-André, Samuel -- what's the process for submitting and
> getting reviewed changes to the slirp submodule now it's a
> separate project ?
>

It's hosted on gitlab.freedesktop.org, with some CI. It's fine to send
patches on qemu devel, as long as Samuel or I are in CC. But in the long
term, I think gitlab MR will be favoured (after a while using it, I think
gitlab is better than ML for patches & bug tracking tbh)

>

Re: [Qemu-devel] [PATCH v3 1/4] QEMU_PACKED: Remove gcc_struct attribute in Windows non x86 targets

[Philippe Mathieu-Daudé]

On 5/3/19 5:56 PM, Marc-André Lureau wrote:
> Hi
> 
> Le ven. 3 mai 2019 à 17:23, Peter Maydell <peter.maydell@linaro.org> a
> écrit :
> 
>> On Fri, 3 May 2019 at 06:07, Thomas Huth <thuth@redhat.com> wrote:
>>>
>>> On 03/05/2019 02.36, Cao Jiaxi wrote:
>>>> gcc_struct is for x86 only, and it generates an warning on ARM64
>> Clang/MinGW targets.
>>>>
>>>> Signed-off-by: Cao Jiaxi <driver1998@foxmail.com>
>>>> ---
>>>>  contrib/libvhost-user/libvhost-user.h | 2 +-
>>>>  include/qemu/compiler.h               | 2 +-
>>>>  scripts/cocci-macro-file.h            | 7 ++++++-
>>>>  slirp/src/util.h                      | 2 +-
>>>>  4 files changed, 9 insertions(+), 4 deletions(-)
>>
>>>> diff --git a/slirp/src/util.h b/slirp/src/util.h
>>>> index 01f1e0e068..278828fe3f 100644
>>>> --- a/slirp/src/util.h
>>>> +++ b/slirp/src/util.h
>>>> @@ -43,7 +43,7 @@
>>>>  #include <netinet/in.h>
>>>>  #endif
>>>>
>>>> -#if defined(_WIN32)
>>>> +#if defined(_WIN32) && (defined(__x86_64__) || defined(__i386__))
>>>>  # define SLIRP_PACKED __attribute__((gcc_struct, packed))
>>>>  #else
>>>>  # define SLIRP_PACKED __attribute__((packed))
>>>>
>>>
>>> The slirp code is currently on its way into a separate module, so you
>>> might need to provide that hunk to the libslirp folks again... I'm
>>> putting the slirp maintainers on CC:, maybe they can pick it up from
>> here.
>>
>> Yes, the slirp module has now landed in master, so this patch
>> definitely needs to be split into two. I've kept in my
>> target-arm.next tree the parts which are applicable to
>> the QEMU repo itself (ie everything except the slirp/ change),
>> so we just need a new patch for the slirp submodule part.
>>
>> Marc-André, Samuel -- what's the process for submitting and
>> getting reviewed changes to the slirp submodule now it's a
>> separate project ?
>>
> 
> It's hosted on gitlab.freedesktop.org, with some CI. It's fine to send
> patches on qemu devel, as long as Samuel or I are in CC. But in the long
> term, I think gitlab MR will be favoured (after a while using it, I think
> gitlab is better than ML for patches & bug tracking tbh)

Except when working offline (or with poor intermittent link).