[Qemu-devel] [PULL v3 00/33] Misc patches for 2016-02-08

[Qemu-devel] [PULL v3 00/33] Misc patches for 2016-02-08

[Paolo Bonzini]

The following changes since commit ac1be2ae6b2995b99430c48329eb971b0281acf1:

  Merge remote-tracking branch 'remotes/armbru/tags/pull-qapi-2016-02-09' into staging (2016-02-09 11:42:43 +0000)

are available in the git repository at:

  git://github.com/bonzini/qemu.git tags/for-upstream

for you to fetch changes up to 150dcd1aed6f9ebcf370dbb9b666e7d7c6d908e2:

  qemu-char, io: fix ordering of arguments for UDP socket creation (2016-02-09 17:09:15 +0100)

----------------------------------------------------------------
* switch to C11 atomics (Alex)
* Coverity fixes for IPMI (Corey), i386 (Paolo), qemu-char (Paolo)
* at long last, fail on wrong .pc files if -m32 is in use (Daniel)
* qemu-char regression fix (Daniel)
* SAS1068 device (Paolo)
* memory region docs improvements (Peter)
* target-i386 cleanups (Richard)
* qemu-nbd docs improvements (Sitsofe)
* thread-safe memory hotplug (Stefan)

----------------------------------------------------------------

v2->v3: rebased due to semantic conflict, split MAINTAINERS patch

Alex Bennée (1):
      include/qemu/atomic.h: default to __atomic functions

Andrew Jones (1):
      kvm-all: trace: strerror fixup

Corey Minyard (2):
      ipmi_bmc_sim: Fix off by one in check.
      ipmi_bmc_sim: Add break to correct watchdog NMI check

Daniel P. Berrange (2):
      configure: sanity check the glib library that pkg-config finds
      char: fix repeated registration of tcp chardev I/O handlers

Janosch Frank (1):
      scripts/kvm/kvm_stat: Fix tracefs access checking

John Snow (1):
      nbd: avoid unaligned uint64_t store

Paolo Bonzini (9):
      memory: add early bail out from cpu_physical_memory_set_dirty_range
      qemu-char: Keep pty slave file descriptor open until the master is closed
      scsi: push WWN fields up to SCSIDevice
      scsi-generic: grab device and port SAS addresses from backend
      hw: Add support for LSI SAS1068 (mptsas) device
      ipmi: do not take/drop iothread lock
      target-i386: fix PSE36 mode
      get_maintainer.pl: fall back to git if only lists are found
      qemu-char, io: fix ordering of arguments for UDP socket creation

Peter Maydell (1):
      docs/memory.txt: Improve list of different memory regions

Richard Henderson (10):
      target-i386: Create gen_lea_v_seg
      target-i386: Introduce mo_stacksize
      target-i386: Use gen_lea_v_seg in gen_lea_modrm
      target-i386: Use gen_lea_v_seg in stack subroutines
      target-i386: Access segs via TCG registers
      target-i386: Use gen_lea_v_seg in pusha/popa
      target-i386: Rewrite gen_enter inline
      target-i386: Rewrite leave
      target-i386: Tidy gen_add_A0_im
      target-i386: Deconstruct the cpu_T array

Sitsofe Wheeler (3):
      qemu-nbd: Fix unintended texi verbatim formatting
      qemu-nbd: Minor texi updates
      qemu-nbd: Fix texi sentence capitalisation

Stefan Hajnoczi (1):
      memory: RCU ram_list.dirty_memory[] for safe RAM hotplug

Stephen Warren (1):
      MAINTAINERS: add all-match entry for qemu-devel@

 MAINTAINERS               |    5 +
 configure                 |   24 +
 default-configs/pci.mak   |    1 +
 docs/memory.txt           |   26 +-
 exec.c                    |   75 +-
 hw/ipmi/ipmi.c            |    2 -
 hw/ipmi/ipmi_bmc_sim.c    |    4 +-
 hw/scsi/Makefile.objs     |    1 +
 hw/scsi/mpi.h             | 1153 ++++++++++++++++++++++++++++++
 hw/scsi/mptconfig.c       |  904 ++++++++++++++++++++++++
 hw/scsi/mptendian.c       |  204 ++++++
 hw/scsi/mptsas.c          | 1441 +++++++++++++++++++++++++++++++++++++
 hw/scsi/mptsas.h          |  100 +++
 hw/scsi/scsi-disk.c       |   23 +-
 hw/scsi/scsi-generic.c    |   92 +++
 include/exec/ram_addr.h   |  193 ++++-
 include/hw/pci/pci_ids.h  |    1 +
 include/hw/scsi/scsi.h    |    3 +
 include/qemu/atomic.h     |  192 +++--
 io/channel-socket.c       |    2 +-
 kvm-all.c                 |    4 +-
 migration/ram.c           |    4 -
 nbd/server.c              |   20 +-
 qemu-char.c               |   10 +-
 qemu-nbd.texi             |   80 ++-
 scripts/get_maintainer.pl |    2 +-
 scripts/kvm/kvm_stat      |   23 +-
 target-i386/helper.c      |    4 +-
 target-i386/helper.h      |    4 -
 target-i386/seg_helper.c  |   74 --
 target-i386/translate.c   | 1725 +++++++++++++++++++++------------------------
 trace-events              |   22 +
 32 files changed, 5223 insertions(+), 1195 deletions(-)
 create mode 100644 hw/scsi/mpi.h
 create mode 100644 hw/scsi/mptconfig.c
 create mode 100644 hw/scsi/mptendian.c
 create mode 100644 hw/scsi/mptsas.c
 create mode 100644 hw/scsi/mptsas.h
-- 
2.5.0

[Qemu-devel] [PULL 20/33] target-i386: Access segs via TCG registers

[Paolo Bonzini]

From: Richard Henderson <rth@twiddle.net>

Having segs[].base as a register significantly improves code
generation for real and protected modes, particularly for TBs
that have multiple memory references where the segment base
can be held in a hard register through the TB.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <rth@twiddle.net>
Message-Id: <1450379966-28198-6-git-send-email-rth@twiddle.net>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
 target-i386/translate.c | 52 ++++++++++++++++++++++++++-----------------------
 1 file changed, 28 insertions(+), 24 deletions(-)

diff --git a/target-i386/translate.c b/target-i386/translate.c
index 83e2819..038eef2 100644
--- a/target-i386/translate.c
+++ b/target-i386/translate.c
@@ -64,6 +64,7 @@ static TCGv cpu_A0;
 static TCGv cpu_cc_dst, cpu_cc_src, cpu_cc_src2, cpu_cc_srcT;
 static TCGv_i32 cpu_cc_op;
 static TCGv cpu_regs[CPU_NB_REGS];
+static TCGv cpu_seg_base[6];
 /* local temps */
 static TCGv cpu_T[2];
 /* local register indexes (only used inside old micro ops) */
@@ -421,12 +422,11 @@ static inline void gen_op_add_reg_T0(TCGMemOp size, int reg)
 
 static inline void gen_op_addl_A0_seg(DisasContext *s, int reg)
 {
-    tcg_gen_ld_tl(cpu_tmp0, cpu_env, offsetof(CPUX86State, segs[reg].base));
     if (CODE64(s)) {
         tcg_gen_ext32u_tl(cpu_A0, cpu_A0);
-        tcg_gen_add_tl(cpu_A0, cpu_A0, cpu_tmp0);
+        tcg_gen_add_tl(cpu_A0, cpu_A0, cpu_seg_base[reg]);
     } else {
-        tcg_gen_add_tl(cpu_A0, cpu_A0, cpu_tmp0);
+        tcg_gen_add_tl(cpu_A0, cpu_A0, cpu_seg_base[reg]);
         tcg_gen_ext32u_tl(cpu_A0, cpu_A0);
     }
 }
@@ -499,9 +499,7 @@ static void gen_lea_v_seg(DisasContext *s, TCGMemOp aflag, TCGv a0,
     }
 
     if (ovr_seg >= 0) {
-        TCGv seg = tcg_temp_new();
-
-        tcg_gen_ld_tl(seg, cpu_env, offsetof(CPUX86State, segs[ovr_seg].base));
+        TCGv seg = cpu_seg_base[ovr_seg];
 
         if (aflag == MO_64) {
             tcg_gen_add_tl(cpu_A0, a0, seg);
@@ -512,8 +510,6 @@ static void gen_lea_v_seg(DisasContext *s, TCGMemOp aflag, TCGv a0,
             tcg_gen_add_tl(cpu_A0, a0, seg);
             tcg_gen_ext32u_tl(cpu_A0, cpu_A0);
         }
-
-        tcg_temp_free(seg);
     }
 }
 
@@ -2204,12 +2200,10 @@ static inline void gen_op_movl_T0_seg(int seg_reg)
 
 static inline void gen_op_movl_seg_T0_vm(int seg_reg)
 {
-    tcg_gen_andi_tl(cpu_T[0], cpu_T[0], 0xffff);
+    tcg_gen_ext16u_tl(cpu_T[0], cpu_T[0]);
     tcg_gen_st32_tl(cpu_T[0], cpu_env, 
                     offsetof(CPUX86State,segs[seg_reg].selector));
-    tcg_gen_shli_tl(cpu_T[0], cpu_T[0], 4);
-    tcg_gen_st_tl(cpu_T[0], cpu_env, 
-                  offsetof(CPUX86State,segs[seg_reg].base));
+    tcg_gen_shli_tl(cpu_seg_base[seg_reg], cpu_T[0], 4);
 }
 
 /* move T0 to seg_reg and compute if the CPU state may change. Never
@@ -7274,21 +7268,16 @@ static target_ulong disas_insn(CPUX86State *env, DisasContext *s,
                         if (s->cpl != 0) {
                             gen_exception(s, EXCP0D_GPF, pc_start - s->cs_base);
                         } else {
-                            tcg_gen_ld_tl(cpu_T[0], cpu_env,
-                                offsetof(CPUX86State,segs[R_GS].base));
-                            tcg_gen_ld_tl(cpu_T[1], cpu_env,
-                                offsetof(CPUX86State,kernelgsbase));
-                            tcg_gen_st_tl(cpu_T[1], cpu_env,
-                                offsetof(CPUX86State,segs[R_GS].base));
+                            tcg_gen_mov_tl(cpu_T[0], cpu_seg_base[R_GS]);
+                            tcg_gen_ld_tl(cpu_seg_base[R_GS], cpu_env,
+                                          offsetof(CPUX86State, kernelgsbase));
                             tcg_gen_st_tl(cpu_T[0], cpu_env,
-                                offsetof(CPUX86State,kernelgsbase));
+                                          offsetof(CPUX86State, kernelgsbase));
                         }
-                    } else
-#endif
-                    {
-                        goto illegal_op;
+                        break;
                     }
-                    break;
+#endif
+                    goto illegal_op;
                 case 1: /* rdtscp */
                     if (!(s->cpuid_ext2_features & CPUID_EXT2_RDTSCP))
                         goto illegal_op;
@@ -7737,6 +7726,14 @@ void tcg_x86_init(void)
         [R_ESP] = "esp",
 #endif
     };
+    static const char seg_base_names[6][8] = {
+        [R_CS] = "cs_base",
+        [R_DS] = "ds_base",
+        [R_ES] = "es_base",
+        [R_FS] = "fs_base",
+        [R_GS] = "gs_base",
+        [R_SS] = "ss_base",
+    };
     int i;
 
     cpu_env = tcg_global_reg_new_ptr(TCG_AREG0, "env");
@@ -7755,6 +7752,13 @@ void tcg_x86_init(void)
                                          reg_names[i]);
     }
 
+    for (i = 0; i < 6; ++i) {
+        cpu_seg_base[i]
+            = tcg_global_mem_new(cpu_env,
+                                 offsetof(CPUX86State, segs[i].base),
+                                 seg_base_names[i]);
+    }
+
     helper_lock_init();
 }
 
-- 
2.5.0

[Qemu-devel] [PULL 31/33] get_maintainer.pl: fall back to git if only lists are found

[Paolo Bonzini]

It's not 100% obvious to project newcomers that all patches should be sent
there; checkpatch doesn't say so, and since it mentions other lists to CC,
the wording "the list" from the SubmitAPatch wiki page can be taken
to mean only those lists, not the main list too.  We would like therefore
to add a catch-all entry for qemu-devel@nongnu.org.

On its own, this would break fallback to git, because now every file
has a maintainer of sorts.  Modify get_maintainer.pl so that mailing
lists (L: lines) no longer prevent the fallback, only humans (M:
entries).

Several pre-existing entries have a list but no human.  These now
fall back to git.  That's a feature.

Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Markus Armbruster <armbru@redhat.com>
Cc: John Snow <jsnow@redhat.com>
Signed-off-by: Stephen Warren <swarren@wwwdotorg.org>
Message-Id: <1454987065-12961-1-git-send-email-swarren@wwwdotorg.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
 scripts/get_maintainer.pl | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/scripts/get_maintainer.pl b/scripts/get_maintainer.pl
index 7dacf32..8261bcb 100755
--- a/scripts/get_maintainer.pl
+++ b/scripts/get_maintainer.pl
@@ -636,7 +636,7 @@ sub get_maintainers {
 
     if ($email) {
 	if (! $interactive) {
-	    $email_git_fallback = 0 if @email_to > 0 || @list_to > 0 || $email_git || $email_git_blame;
+	    $email_git_fallback = 0 if @email_to > 0 || $email_git || $email_git_blame;
 	    if ($email_git_fallback) {
 	        print STDERR "get_maintainer.pl: No maintainers found, printing recent contributors.\n";
 	        print STDERR "get_maintainer.pl: Do not blindly cc: them on patches!  Use common sense.\n";
-- 
2.5.0

[Qemu-devel] [PULL 32/33] MAINTAINERS: add all-match entry for qemu-devel@

[Paolo Bonzini]

From: Stephen Warren <swarren@wwwdotorg.org>

Add an entry to MAINTAINERS that matches every patch, and requests the
user send patches to qemu-devel@nongnu.org.

It's not 100% obvious to project newcomers that all patches should be sent
there; checkpatch doesn't say so, and since it mentions other lists to CC,
the wording "the list" from the SubmitAPatch wiki page can be taken
to mean only those lists, not the main list too.

The F: entries were taken from a similar entry in the Linux kernel.

Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Markus Armbruster <armbru@redhat.com>
Cc: John Snow <jsnow@redhat.com>
Signed-off-by: Stephen Warren <swarren@wwwdotorg.org>
Message-Id: <1454987065-12961-1-git-send-email-swarren@wwwdotorg.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
 MAINTAINERS | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/MAINTAINERS b/MAINTAINERS
index b6ed87a..2d78eea 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -52,6 +52,11 @@ General Project Administration
 ------------------------------
 M: Peter Maydell <peter.maydell@linaro.org>
 
+All patches CC here
+L: qemu-devel@nongnu.org
+F: *
+F: */
+
 Responsible Disclosure, Reporting Security Issues
 ------------------------------
 W: http://wiki.qemu.org/SecurityProcess
-- 
2.5.0

Re: [Qemu-devel] [PULL v3 00/33] Misc patches for 2016-02-08

[Paolo Bonzini]

On 09/02/2016 17:13, Paolo Bonzini wrote:
> The following changes since commit ac1be2ae6b2995b99430c48329eb971b0281acf1:
> 
>   Merge remote-tracking branch 'remotes/armbru/tags/pull-qapi-2016-02-09' into staging (2016-02-09 11:42:43 +0000)
> 
> are available in the git repository at:
> 
>   git://github.com/bonzini/qemu.git tags/for-upstream
> 
> for you to fetch changes up to 150dcd1aed6f9ebcf370dbb9b666e7d7c6d908e2:
> 
>   qemu-char, io: fix ordering of arguments for UDP socket creation (2016-02-09 17:09:15 +0100)
> 
> ----------------------------------------------------------------
> * switch to C11 atomics (Alex)
> * Coverity fixes for IPMI (Corey), i386 (Paolo), qemu-char (Paolo)
> * at long last, fail on wrong .pc files if -m32 is in use (Daniel)
> * qemu-char regression fix (Daniel)
> * SAS1068 device (Paolo)
> * memory region docs improvements (Peter)
> * target-i386 cleanups (Richard)
> * qemu-nbd docs improvements (Sitsofe)
> * thread-safe memory hotplug (Stefan)
> 
> ----------------------------------------------------------------
> 
> v2->v3: rebased due to semantic conflict, split MAINTAINERS patch
> 
> Alex Bennée (1):
>       include/qemu/atomic.h: default to __atomic functions
> 
> Andrew Jones (1):
>       kvm-all: trace: strerror fixup
> 
> Corey Minyard (2):
>       ipmi_bmc_sim: Fix off by one in check.
>       ipmi_bmc_sim: Add break to correct watchdog NMI check
> 
> Daniel P. Berrange (2):
>       configure: sanity check the glib library that pkg-config finds
>       char: fix repeated registration of tcp chardev I/O handlers
> 
> Janosch Frank (1):
>       scripts/kvm/kvm_stat: Fix tracefs access checking
> 
> John Snow (1):
>       nbd: avoid unaligned uint64_t store
> 
> Paolo Bonzini (9):
>       memory: add early bail out from cpu_physical_memory_set_dirty_range
>       qemu-char: Keep pty slave file descriptor open until the master is closed
>       scsi: push WWN fields up to SCSIDevice
>       scsi-generic: grab device and port SAS addresses from backend
>       hw: Add support for LSI SAS1068 (mptsas) device
>       ipmi: do not take/drop iothread lock
>       target-i386: fix PSE36 mode
>       get_maintainer.pl: fall back to git if only lists are found
>       qemu-char, io: fix ordering of arguments for UDP socket creation
> 
> Peter Maydell (1):
>       docs/memory.txt: Improve list of different memory regions
> 
> Richard Henderson (10):
>       target-i386: Create gen_lea_v_seg
>       target-i386: Introduce mo_stacksize
>       target-i386: Use gen_lea_v_seg in gen_lea_modrm
>       target-i386: Use gen_lea_v_seg in stack subroutines
>       target-i386: Access segs via TCG registers
>       target-i386: Use gen_lea_v_seg in pusha/popa
>       target-i386: Rewrite gen_enter inline
>       target-i386: Rewrite leave
>       target-i386: Tidy gen_add_A0_im
>       target-i386: Deconstruct the cpu_T array
> 
> Sitsofe Wheeler (3):
>       qemu-nbd: Fix unintended texi verbatim formatting
>       qemu-nbd: Minor texi updates
>       qemu-nbd: Fix texi sentence capitalisation
> 
> Stefan Hajnoczi (1):
>       memory: RCU ram_list.dirty_memory[] for safe RAM hotplug
> 
> Stephen Warren (1):
>       MAINTAINERS: add all-match entry for qemu-devel@
> 
>  MAINTAINERS               |    5 +
>  configure                 |   24 +
>  default-configs/pci.mak   |    1 +
>  docs/memory.txt           |   26 +-
>  exec.c                    |   75 +-
>  hw/ipmi/ipmi.c            |    2 -
>  hw/ipmi/ipmi_bmc_sim.c    |    4 +-
>  hw/scsi/Makefile.objs     |    1 +
>  hw/scsi/mpi.h             | 1153 ++++++++++++++++++++++++++++++
>  hw/scsi/mptconfig.c       |  904 ++++++++++++++++++++++++
>  hw/scsi/mptendian.c       |  204 ++++++
>  hw/scsi/mptsas.c          | 1441 +++++++++++++++++++++++++++++++++++++
>  hw/scsi/mptsas.h          |  100 +++
>  hw/scsi/scsi-disk.c       |   23 +-
>  hw/scsi/scsi-generic.c    |   92 +++
>  include/exec/ram_addr.h   |  193 ++++-
>  include/hw/pci/pci_ids.h  |    1 +
>  include/hw/scsi/scsi.h    |    3 +
>  include/qemu/atomic.h     |  192 +++--
>  io/channel-socket.c       |    2 +-
>  kvm-all.c                 |    4 +-
>  migration/ram.c           |    4 -
>  nbd/server.c              |   20 +-
>  qemu-char.c               |   10 +-
>  qemu-nbd.texi             |   80 ++-
>  scripts/get_maintainer.pl |    2 +-
>  scripts/kvm/kvm_stat      |   23 +-
>  target-i386/helper.c      |    4 +-
>  target-i386/helper.h      |    4 -
>  target-i386/seg_helper.c  |   74 --
>  target-i386/translate.c   | 1725 +++++++++++++++++++++------------------------
>  trace-events              |   22 +
>  32 files changed, 5223 insertions(+), 1195 deletions(-)
>  create mode 100644 hw/scsi/mpi.h
>  create mode 100644 hw/scsi/mptconfig.c
>  create mode 100644 hw/scsi/mptendian.c
>  create mode 100644 hw/scsi/mptsas.c
>  create mode 100644 hw/scsi/mptsas.h
> 

Self-NACK, this breaks (at least) FreeDOS.

Paolo

Re: [Qemu-devel] [PULL v3 00/33] Misc patches for 2016-02-08

[Peter Maydell]

On 10 February 2016 at 12:48, Paolo Bonzini <pbonzini@redhat.com> wrote:
>
>
> On 09/02/2016 17:13, Paolo Bonzini wrote:
>> The following changes since commit ac1be2ae6b2995b99430c48329eb971b0281acf1:
>>
>>   Merge remote-tracking branch 'remotes/armbru/tags/pull-qapi-2016-02-09' into staging (2016-02-09 11:42:43 +0000)
>>
>> are available in the git repository at:
>>
>>   git://github.com/bonzini/qemu.git tags/for-upstream
>>
>> for you to fetch changes up to 150dcd1aed6f9ebcf370dbb9b666e7d7c6d908e2:
>>
>>   qemu-char, io: fix ordering of arguments for UDP socket creation (2016-02-09 17:09:15 +0100)
>>
>> ----------------------------------------------------------------
>> * switch to C11 atomics (Alex)
>> * Coverity fixes for IPMI (Corey), i386 (Paolo), qemu-char (Paolo)
>> * at long last, fail on wrong .pc files if -m32 is in use (Daniel)
>> * qemu-char regression fix (Daniel)
>> * SAS1068 device (Paolo)
>> * memory region docs improvements (Peter)
>> * target-i386 cleanups (Richard)
>> * qemu-nbd docs improvements (Sitsofe)
>> * thread-safe memory hotplug (Stefan)
>>
>> ----------------------------------------------------------------

> Self-NACK, this breaks (at least) FreeDOS.

Oops, I read this email five seconds after pushing the merge to master.
Can you send out reverts for the appropriate patches and I'll apply
them direct to master?

thanks
-- PMM

Re: [Qemu-devel] [PULL v3 00/33] Misc patches for 2016-02-08

[Paolo Bonzini]

On 10/02/2016 14:49, Peter Maydell wrote:
>> > Self-NACK, this breaks (at least) FreeDOS.
> Oops, I read this email five seconds after pushing the merge to master.
> Can you send out reverts for the appropriate patches and I'll apply
> them direct to master?

I can send the fix.

Paolo

Re: [Qemu-devel] [PULL v3 00/33] Misc patches for 2016-02-08

[John Snow]

On 02/10/2016 07:48 AM, Paolo Bonzini wrote:
> 
> 
> On 09/02/2016 17:13, Paolo Bonzini wrote:
>> The following changes since commit ac1be2ae6b2995b99430c48329eb971b0281acf1:
>>
>>   Merge remote-tracking branch 'remotes/armbru/tags/pull-qapi-2016-02-09' into staging (2016-02-09 11:42:43 +0000)
>>
>> are available in the git repository at:
>>
>>   git://github.com/bonzini/qemu.git tags/for-upstream
>>
>> for you to fetch changes up to 150dcd1aed6f9ebcf370dbb9b666e7d7c6d908e2:
>>
>>   qemu-char, io: fix ordering of arguments for UDP socket creation (2016-02-09 17:09:15 +0100)
>>
>> ----------------------------------------------------------------
>> * switch to C11 atomics (Alex)
>> * Coverity fixes for IPMI (Corey), i386 (Paolo), qemu-char (Paolo)
>> * at long last, fail on wrong .pc files if -m32 is in use (Daniel)
>> * qemu-char regression fix (Daniel)
>> * SAS1068 device (Paolo)
>> * memory region docs improvements (Peter)
>> * target-i386 cleanups (Richard)
>> * qemu-nbd docs improvements (Sitsofe)
>> * thread-safe memory hotplug (Stefan)
>>
>> ----------------------------------------------------------------
>>
>> v2->v3: rebased due to semantic conflict, split MAINTAINERS patch
>>
>> Alex Bennée (1):
>>       include/qemu/atomic.h: default to __atomic functions
>>
>> Andrew Jones (1):
>>       kvm-all: trace: strerror fixup
>>
>> Corey Minyard (2):
>>       ipmi_bmc_sim: Fix off by one in check.
>>       ipmi_bmc_sim: Add break to correct watchdog NMI check
>>
>> Daniel P. Berrange (2):
>>       configure: sanity check the glib library that pkg-config finds

Appears to break clang 3.5.0 on F22;

jsnow@scv ((977a82a...)) ~/s/q/b/git> ../../configure --cxx=clang++
--cc=clang --host-cc=clang --extra-cflags=-Werror
--extra-cflags=-fsanitize=undefined
--extra-cflags=-Wno-deprecated-declarations
--extra-cflags=-fno-sanitize=float-divide-by-zero
--target-list="x86_64-softmmu"

ERROR: sizeof(size_t) doesn't match GLIB_SIZEOF_SIZE_T.
       You probably need to set PKG_CONFIG_LIBDIR
       to point to the right pkg-config files for your
       build target



Problem appears to be this:

/usr/include/glib-2.0/glib/gmem.h:76:58: error: unknown attribute
'__alloc_size__' ignored [-Werror,-Wunknown-attributes]
gpointer g_malloc         (gsize         n_bytes) G_GNUC_MALLOC
G_GNUC_ALLOC_SIZE(1);
                                                                ^
/usr/include/glib-2.0/glib/gmacros.h:67:45: note: expanded from macro
'G_GNUC_ALLOC_SIZE'
#define G_GNUC_ALLOC_SIZE(x) __attribute__((__alloc_size__(x)))


>>       char: fix repeated registration of tcp chardev I/O handlers
>>
>> Janosch Frank (1):
>>       scripts/kvm/kvm_stat: Fix tracefs access checking
>>
>> John Snow (1):
>>       nbd: avoid unaligned uint64_t store
>>
>> Paolo Bonzini (9):
>>       memory: add early bail out from cpu_physical_memory_set_dirty_range
>>       qemu-char: Keep pty slave file descriptor open until the master is closed
>>       scsi: push WWN fields up to SCSIDevice
>>       scsi-generic: grab device and port SAS addresses from backend
>>       hw: Add support for LSI SAS1068 (mptsas) device
>>       ipmi: do not take/drop iothread lock
>>       target-i386: fix PSE36 mode
>>       get_maintainer.pl: fall back to git if only lists are found
>>       qemu-char, io: fix ordering of arguments for UDP socket creation
>>
>> Peter Maydell (1):
>>       docs/memory.txt: Improve list of different memory regions
>>
>> Richard Henderson (10):
>>       target-i386: Create gen_lea_v_seg
>>       target-i386: Introduce mo_stacksize
>>       target-i386: Use gen_lea_v_seg in gen_lea_modrm
>>       target-i386: Use gen_lea_v_seg in stack subroutines
>>       target-i386: Access segs via TCG registers
>>       target-i386: Use gen_lea_v_seg in pusha/popa
>>       target-i386: Rewrite gen_enter inline
>>       target-i386: Rewrite leave
>>       target-i386: Tidy gen_add_A0_im
>>       target-i386: Deconstruct the cpu_T array
>>
>> Sitsofe Wheeler (3):
>>       qemu-nbd: Fix unintended texi verbatim formatting
>>       qemu-nbd: Minor texi updates
>>       qemu-nbd: Fix texi sentence capitalisation
>>
>> Stefan Hajnoczi (1):
>>       memory: RCU ram_list.dirty_memory[] for safe RAM hotplug
>>
>> Stephen Warren (1):
>>       MAINTAINERS: add all-match entry for qemu-devel@
>>
>>  MAINTAINERS               |    5 +
>>  configure                 |   24 +
>>  default-configs/pci.mak   |    1 +
>>  docs/memory.txt           |   26 +-
>>  exec.c                    |   75 +-
>>  hw/ipmi/ipmi.c            |    2 -
>>  hw/ipmi/ipmi_bmc_sim.c    |    4 +-
>>  hw/scsi/Makefile.objs     |    1 +
>>  hw/scsi/mpi.h             | 1153 ++++++++++++++++++++++++++++++
>>  hw/scsi/mptconfig.c       |  904 ++++++++++++++++++++++++
>>  hw/scsi/mptendian.c       |  204 ++++++
>>  hw/scsi/mptsas.c          | 1441 +++++++++++++++++++++++++++++++++++++
>>  hw/scsi/mptsas.h          |  100 +++
>>  hw/scsi/scsi-disk.c       |   23 +-
>>  hw/scsi/scsi-generic.c    |   92 +++
>>  include/exec/ram_addr.h   |  193 ++++-
>>  include/hw/pci/pci_ids.h  |    1 +
>>  include/hw/scsi/scsi.h    |    3 +
>>  include/qemu/atomic.h     |  192 +++--
>>  io/channel-socket.c       |    2 +-
>>  kvm-all.c                 |    4 +-
>>  migration/ram.c           |    4 -
>>  nbd/server.c              |   20 +-
>>  qemu-char.c               |   10 +-
>>  qemu-nbd.texi             |   80 ++-
>>  scripts/get_maintainer.pl |    2 +-
>>  scripts/kvm/kvm_stat      |   23 +-
>>  target-i386/helper.c      |    4 +-
>>  target-i386/helper.h      |    4 -
>>  target-i386/seg_helper.c  |   74 --
>>  target-i386/translate.c   | 1725 +++++++++++++++++++++------------------------
>>  trace-events              |   22 +
>>  32 files changed, 5223 insertions(+), 1195 deletions(-)
>>  create mode 100644 hw/scsi/mpi.h
>>  create mode 100644 hw/scsi/mptconfig.c
>>  create mode 100644 hw/scsi/mptendian.c
>>  create mode 100644 hw/scsi/mptsas.c
>>  create mode 100644 hw/scsi/mptsas.h
>>
> 
> Self-NACK, this breaks (at least) FreeDOS.
> 
> Paolo
> 

Re: [Qemu-devel] [PULL v3 00/33] Misc patches for 2016-02-08

[Paolo Bonzini]

On 10/02/2016 20:16, John Snow wrote:
> 
> jsnow@scv ((977a82a...)) ~/s/q/b/git> ../../configure --cxx=clang++
> --cc=clang --host-cc=clang --extra-cflags=-Werror
> --extra-cflags=-fsanitize=undefined
> --extra-cflags=-Wno-deprecated-declarations
> --extra-cflags=-fno-sanitize=float-divide-by-zero
> --target-list="x86_64-softmmu"
> 
> ERROR: sizeof(size_t) doesn't match GLIB_SIZEOF_SIZE_T.
>        You probably need to set PKG_CONFIG_LIBDIR
>        to point to the right pkg-config files for your
>        build target
> 
> 
> 
> Problem appears to be this:
> 
> /usr/include/glib-2.0/glib/gmem.h:76:58: error: unknown attribute
> '__alloc_size__' ignored [-Werror,-Wunknown-attributes]
> gpointer g_malloc         (gsize         n_bytes) G_GNUC_MALLOC
> G_GNUC_ALLOC_SIZE(1);
>                                                                 ^
> /usr/include/glib-2.0/glib/gmacros.h:67:45: note: expanded from macro
> 'G_GNUC_ALLOC_SIZE'
> #define G_GNUC_ALLOC_SIZE(x) __attribute__((__alloc_size__(x)))

F23 wraps it like this:

/* Clang feature detection: http://clang.llvm.org/docs/LanguageExtensions.html */
#ifndef __has_feature
#define __has_feature(x) 0
#endif

#if     (!defined(__clang__) && ((__GNUC__ > 4) || (__GNUC__ == 4 && __GNUC_MINO
R__ >= 3))) || \
        (defined(__clang__) && __has_feature(__alloc_size__))
#define G_GNUC_ALLOC_SIZE(x) __attribute__((__alloc_size__(x)))
#define G_GNUC_ALLOC_SIZE2(x,y) __attribute__((__alloc_size__(x,y)))
#else
#define G_GNUC_ALLOC_SIZE(x)
#define G_GNUC_ALLOC_SIZE2(x,y)
#endif

Please open a bug on the distro to backport this fix.

Paolo