qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Paolo Bonzini <pbonzini@redhat.com>
To: Peter Crosthwaite <crosthwaitepeter@gmail.com>, qemu-devel@nongnu.org
Cc: peter.maydell@linaro.org, rth@twiddle.net,
	edgar.iglesias@gmail.com, afaerber@suse.de,
	Peter Crosthwaite <crosthwaite.peter@gmail.com>
Subject: Re: [Qemu-devel] [RFC v2 22/34] microblaze: enable multi-arch
Date: Mon, 01 Jun 2015 10:16:46 +0200	[thread overview]
Message-ID: <556C14EE.2080608@redhat.com> (raw)
In-Reply-To: <b6fa26ef3fa40b5c1d841c67e00217b9dec74b2c.1433052533.git.crosthwaite.peter@gmail.com>



On 31/05/2015 08:11, Peter Crosthwaite wrote:
> Multi-arch conversion consisting of:
>  * configury
>     - defining CONFIG_ARCH_MULTI
>     - adding to MULTI_TARGETS
>     - enabling disas for MULTI_ARCH
>  * Converting target-microblaze to arch-obj-y
>  * cpu.h
>    - Compiling out all target-microblaze private contents
>      when doing multi-arch build
>    - Redefining target_ulong and cpu-def typenames with arch prefix
>    - Undeffing possibly previously defined macros
>  * Defining the QOM cpu hooks
> 
> Signed-off-by: Peter Crosthwaite <crosthwaite.peter@gmail.com>
> ---
> I guess I could split to multi patches but it will bloat this series!
> ---
>  configure                       |  6 ++++++
>  target-microblaze/Makefile.objs |  6 +++---
>  target-microblaze/cpu-qom.h     |  2 ++
>  target-microblaze/cpu.c         |  1 +
>  target-microblaze/cpu.h         | 40 ++++++++++++++++++++++++++++++++++++----
>  5 files changed, 48 insertions(+), 7 deletions(-)
> 
> diff --git a/configure b/configure
> index 1acafcd..3dbfd3e 100755
> --- a/configure
> +++ b/configure
> @@ -5440,6 +5440,9 @@ for i in $ARCH $TARGET_BASE_ARCH ; do
>    microblaze*)
>      echo "CONFIG_MICROBLAZE_DIS=y"  >> $config_target_mak
>      echo "CONFIG_MICROBLAZE_DIS=y"  >> config-all-disas.mak
> +    if [ -e $config_target_multi_mak ]; then
> +        echo "CONFIG_MICROBLAZE_DIS=y"  >> $config_target_multi_mak
> +    fi

-e looks dangerous if you have a previous build around.  Use a variable
that you can also print at the end of configure?

>    ;;
>    mips*)
>      echo "CONFIG_MIPS_DIS=y"  >> $config_target_mak
> @@ -5481,6 +5484,9 @@ if test "$tcg_interpreter" = "yes" ; then
>  fi
>  
>  case "$TARGET_BASE_ARCH" in
> +microblaze)
> +  echo "CONFIG_ARCH_MULTI=y" >> $config_target_mak
> +;;
>  *)
>    echo "CONFIG_ARCH_SINGLE=y"  >> $config_target_mak
>  ;;

How is CONFIG_ARCH_SINGLE different from $(call lnot,
$(CONFIG_ARCH_MULTI))?  And can CONFIG_ARCH_MULTI be defined in
default-configs/ rather than here?

> diff --git a/target-microblaze/Makefile.objs b/target-microblaze/Makefile.objs
> index f3d7b44..f70163d 100644
> --- a/target-microblaze/Makefile.objs
> +++ b/target-microblaze/Makefile.objs
> @@ -1,3 +1,3 @@
> -obj-y += translate.o op_helper.o helper.o cpu.o
> -obj-y += gdbstub.o
> -obj-$(CONFIG_SOFTMMU) += mmu.o
> +arch-obj-y += translate.o op_helper.o helper.o cpu.o
> +arch-obj-y += gdbstub.o
> +arch-obj-$(CONFIG_SOFTMMU) += mmu.o
> diff --git a/target-microblaze/cpu-qom.h b/target-microblaze/cpu-qom.h
> index e3e0701..88526fa 100644
> --- a/target-microblaze/cpu-qom.h
> +++ b/target-microblaze/cpu-qom.h
> @@ -67,9 +67,11 @@ static inline MicroBlazeCPU *mb_env_get_cpu(CPUMBState *env)
>      return container_of(env, MicroBlazeCPU, env);
>  }
>  
> +#ifndef TARGET_MULTI
>  #define ENV_GET_CPU(e) CPU(mb_env_get_cpu(e))
>  
>  #define ENV_OFFSET offsetof(MicroBlazeCPU, env)
> +#endif /* !TARGET_MULTI */
>  
>  void mb_cpu_do_interrupt(CPUState *cs);
>  bool mb_cpu_exec_interrupt(CPUState *cs, int int_req);
> diff --git a/target-microblaze/cpu.c b/target-microblaze/cpu.c
> index 563ad46..135233a 100644
> --- a/target-microblaze/cpu.c
> +++ b/target-microblaze/cpu.c
> @@ -135,6 +135,7 @@ static void mb_cpu_initfn(Object *obj)
>      CPUMBState *env = &cpu->env;
>      static bool tcg_initialized;
>  
> +    CPU_SET_QOM_HOOKS(cs);

Why are the hooks in the instance rather in the class?  Performance?

Also, why a macro and not a function?

>      cs->env_ptr = env;
>      cpu_exec_init(cs);
>  
> diff --git a/target-microblaze/cpu.h b/target-microblaze/cpu.h
> index 9068272..4ccbac5 100644
> --- a/target-microblaze/cpu.h
> +++ b/target-microblaze/cpu.h
> @@ -22,10 +22,21 @@
>  #include "config.h"
>  #include "qemu-common.h"
>  
> +#undef TARGET_LONG_BITS
>  #define TARGET_LONG_BITS 32
>  
> +#undef CPUArchState
>  #define CPUArchState struct CPUMBState
>  
> +#undef CPUTLBEntry
> +#undef CPUIOTLBEntry
> +#undef target_long
> +#undef target_ulong
> +#define CPUTLBEntry MBCPUTLBEntry
> +#define CPUIOTLBEntry MBCPUIOTLBEntry
> +#define target_long mb_target_long
> +#define target_ulong mb_target_ulong

Oh, this answers my previous question.  Please document it in the header.

Could the #undefs be moved to exec/cpu-defs.h and exec/target_long.h
(see softmmu_template.h for a precedent)?

Paolo

>  #include "exec/cpu-defs.h"
>  #include "fpu/softfloat.h"
>  struct CPUMBState;
> @@ -34,6 +45,7 @@ typedef struct CPUMBState CPUMBState;
>  #include "mmu.h"
>  #endif
>  
> +#ifndef TARGET_MULTI
>  #define ELF_MACHINE	EM_MICROBLAZE
>  
>  #define EXCP_MMU        1
> @@ -45,13 +57,19 @@ typedef struct CPUMBState CPUMBState;
>  /* MicroBlaze-specific interrupt pending bits.  */
>  #define CPU_INTERRUPT_NMI       CPU_INTERRUPT_TGT_EXT_3
>  
> +#endif /* TARGET_MULTI */
> +
>  /* Meanings of the MBCPU object's two inbound GPIO lines */
>  #define MB_CPU_IRQ 0
>  #define MB_CPU_FIR 1
>  
> -/* Register aliases. R0 - R15 */
> -#define R_SP     1
> +/* FIXME: Implement cpu::set_pc fn for microblaze and privatise this */
>  #define SR_PC    0
> +
> +#ifndef TARGET_MULTI
> +
> +/* Register aliases. R1 - R15 */
> +#define R_SP     1
>  #define SR_MSR   1
>  #define SR_EAR   3
>  #define SR_ESR   5
> @@ -112,6 +130,13 @@ typedef struct CPUMBState CPUMBState;
>  #define FSR_UF          (1<<1) /* Underflow */
>  #define FSR_DO          (1<<0) /* Denormalized operand error */
>  
> +#endif /* TARGET_MULTI */
> +
> +/* The Microblaze bootloader configures some of the PVRs in a board specific
> + * way as a reset process. This should go away with PVR property QOMification
> + * and then the PVRs can be made private to CPUs.
> + */
> +
>  /* Version reg.  */
>  /* Basic PVR mask */
>  #define PVR0_PVR_FULL_MASK              0x80000000
> @@ -212,6 +237,7 @@ typedef struct CPUMBState CPUMBState;
>  #define PVR11_MSR_RESET_VALUE_MASK      0x000007FF
>  
>  
> +#ifndef TARGET_MULTI
>  
>  /* CPU flags.  */
>  
> @@ -223,14 +249,17 @@ typedef struct CPUMBState CPUMBState;
>  #define CC_NE  1
>  #define CC_EQ  0
>  
> -#define NB_MMU_MODES    3
> -
>  #define STREAM_EXCEPTION (1 << 0)
>  #define STREAM_ATOMIC    (1 << 1)
>  #define STREAM_TEST      (1 << 2)
>  #define STREAM_CONTROL   (1 << 3)
>  #define STREAM_NONBLOCK  (1 << 4)
>  
> +#endif /* TARGET_MULTI */
> +
> +#undef NB_MMU_MODES
> +#define NB_MMU_MODES    3
> +
>  struct CPUMBState {
>      uint32_t debug;
>      uint32_t btaken;
> @@ -274,6 +303,8 @@ struct CPUMBState {
>  
>  #include "cpu-qom.h"
>  
> +#ifndef TARGET_MULTI
> +
>  void mb_tcg_init(void);
>  MicroBlazeCPU *cpu_mb_init(const char *cpu_model);
>  int cpu_mb_exec(CPUState *cpu);
> @@ -337,4 +368,5 @@ void mb_cpu_unassigned_access(CPUState *cpu, hwaddr addr,
>  
>  #include "exec/exec-all.h"
>  
> +#endif /* !TARGET_MULTI */
>  #endif
> 

  reply	other threads:[~2015-06-01  8:17 UTC|newest]

Thread overview: 99+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-05-31  6:11 [Qemu-devel] [RFC v2 00/34] Multi Architecture System Emulation Peter Crosthwaite
2015-05-31  6:11 ` [Qemu-devel] [RFC v2 01/34] cpu-defs: Move CPU_TEMP_BUF_NLONGS to tcg Peter Crosthwaite
2015-06-01 18:57   ` Richard Henderson
2015-05-31  6:11 ` [Qemu-devel] [RFC v2 02/34] cpu-exec: Purge all uses of CPU_GET_ENV Peter Crosthwaite
2015-06-01 19:03   ` Richard Henderson
2015-06-07 22:57     ` Peter Crosthwaite
2015-05-31  6:11 ` [Qemu-devel] [RFC v2 03/34] Makefile.target: Introduce arch-obj Peter Crosthwaite
2015-06-01 19:10   ` Richard Henderson
2015-06-02  9:40     ` Peter Crosthwaite
2015-06-02 10:34       ` Paolo Bonzini
2015-05-31  6:11 ` [Qemu-devel] [RFC v2 04/34] cpu-exec: Migrate some generic fns to cpus.c Peter Crosthwaite
2015-06-01  7:45   ` Paolo Bonzini
2015-07-10  6:37     ` Peter Crosthwaite
2015-06-01 19:12   ` Richard Henderson
2015-05-31  6:11 ` [Qemu-devel] [RFC v2 05/34] translate: Listify tcg_exec_init Peter Crosthwaite
2015-06-01 19:17   ` Richard Henderson
2015-07-10 10:15   ` Peter Crosthwaite
2015-07-11  9:16     ` Peter Crosthwaite
2015-05-31  6:11 ` [Qemu-devel] [RFC v2 06/34] cpu-common: Define tb_page_addr_t for everyone Peter Crosthwaite
2015-06-01  7:51   ` Paolo Bonzini
2015-06-07 23:06     ` Peter Crosthwaite
2015-06-24  9:48       ` Paolo Bonzini
2015-05-31  6:11 ` [Qemu-devel] [RFC v2 07/34] exec-all: Move cpu_can_do_io to qom/cpu.h Peter Crosthwaite
2015-06-01 19:18   ` Richard Henderson
2015-06-24  9:11   ` Paolo Bonzini
2015-06-24 11:40     ` Pavel Dovgaluk
2015-06-24 11:41       ` Paolo Bonzini
2015-06-24 11:59         ` Pavel Dovgaluk
2015-05-31  6:11 ` [Qemu-devel] [RFC v2 08/34] translate-all: Move tcg_handle_interrupt to -common Peter Crosthwaite
2015-06-01 19:20   ` Richard Henderson
2015-05-31  6:11 ` [Qemu-devel] [RFC v2 09/34] include/exec: Move standard exceptions to cpu-all.h Peter Crosthwaite
2015-06-01 19:20   ` Richard Henderson
2015-05-31  6:11 ` [Qemu-devel] [RFC v2 10/34] include/exec: Split target_long def to new header Peter Crosthwaite
2015-06-01 19:24   ` Richard Henderson
2015-06-01 19:51     ` Paolo Bonzini
2015-06-01 20:25       ` Peter Maydell
2015-06-01 20:27         ` Paolo Bonzini
2015-06-01 20:32       ` Richard Henderson
2015-06-01 20:39         ` Paolo Bonzini
2015-06-02 10:14         ` Peter Crosthwaite
2015-06-03 10:01           ` Paolo Bonzini
2015-05-31  6:11 ` [Qemu-devel] [RFC v2 11/34] include/exec: Move cputlb exec.c defs out Peter Crosthwaite
2015-06-01  7:56   ` Paolo Bonzini
2015-06-02  9:44     ` Peter Crosthwaite
2015-06-02 10:38       ` Paolo Bonzini
2015-06-01 19:25   ` Richard Henderson
2015-06-24  9:25   ` Paolo Bonzini
2015-05-31  6:11 ` [Qemu-devel] [RFC v2 12/34] include/exec: Move tb hash functions out Peter Crosthwaite
2015-06-01  7:56   ` Paolo Bonzini
2015-06-01 19:25   ` Richard Henderson
2015-06-24  9:05   ` Paolo Bonzini
2015-05-31  6:11 ` [Qemu-devel] [RFC v2 13/34] cpu-defs: Move out TB_JMP defines Peter Crosthwaite
2015-06-01 19:25   ` Richard Henderson
2015-05-31  6:11 ` [Qemu-devel] [RFC v2 14/34] cpu-defs: Allow multiple inclusions Peter Crosthwaite
2015-06-01  7:59   ` Paolo Bonzini
2015-06-01 19:29     ` Richard Henderson
2015-05-31  6:11 ` [Qemu-devel] [RFC v2 15/34] HACK: monitor: Comment out TCG profile ops Peter Crosthwaite
2015-06-01  8:03   ` Paolo Bonzini
2015-06-01 18:49     ` Richard Henderson
2015-07-11 20:26     ` Peter Crosthwaite
2015-05-31  6:11 ` [Qemu-devel] [RFC v2 16/34] HACK: Disable list_cpus Peter Crosthwaite
2015-06-01  8:04   ` Paolo Bonzini
2015-05-31  6:11 ` [Qemu-devel] [RFC v2 17/34] HACK: globalise TCG page size variables Peter Crosthwaite
2015-06-01 18:52   ` Richard Henderson
2015-06-01 18:58     ` Peter Crosthwaite
2015-05-31  6:11 ` [Qemu-devel] [RFC v2 18/34] HACK: monitor: uninclude cpu_ldst Peter Crosthwaite
2015-07-09  3:08   ` Peter Crosthwaite
2015-05-31  6:11 ` [Qemu-devel] [RFC v2 19/34] HACK: disas: Defeature print_target_address Peter Crosthwaite
2015-06-01  8:08   ` Paolo Bonzini
2015-06-01 18:55   ` Richard Henderson
2015-06-08  0:21     ` Peter Crosthwaite
2015-05-31  6:11 ` [Qemu-devel] [RFC v2 20/34] HACK: exec: comment out use of cpu_get_tb_cpu_from_state Peter Crosthwaite
2015-06-01  8:09   ` Paolo Bonzini
2015-06-08  0:22     ` Peter Crosthwaite
2015-05-31  6:11 ` [Qemu-devel] [RFC v2 21/34] core: virtualise CPU interfaces completely Peter Crosthwaite
2015-06-01  8:11   ` Paolo Bonzini
2015-06-02  9:52     ` Peter Crosthwaite
2015-06-02 10:38       ` Paolo Bonzini
2015-06-01 19:53   ` Richard Henderson
2015-06-02  9:56     ` Peter Crosthwaite
2015-05-31  6:11 ` [Qemu-devel] [RFC v2 22/34] microblaze: enable multi-arch Peter Crosthwaite
2015-06-01  8:16   ` Paolo Bonzini [this message]
2015-05-31  6:11 ` [Qemu-devel] [RFC v2 23/34] arm: cpu: static inline cpu_arm_init Peter Crosthwaite
2015-06-01  8:18   ` Paolo Bonzini
2015-05-31  6:11 ` [Qemu-devel] [RFC v2 24/34] target-arm: Split cp helper API to new C file Peter Crosthwaite
2015-05-31  6:11 ` [Qemu-devel] [RFC v2 25/34] arm: enable multi-arch Peter Crosthwaite
2015-06-01  8:22   ` Paolo Bonzini
2015-05-31  6:11 ` [Qemu-devel] [RFC v2 26/34] core: Introduce multi-arch build Peter Crosthwaite
2015-06-01  8:26   ` Paolo Bonzini
2015-05-31  6:12 ` [Qemu-devel] [RFC v2 27/34] hw: arm: Explicitly include cpu.h for consumers Peter Crosthwaite
2015-05-31  6:12 ` [Qemu-devel] [RFC v2 28/34] arm: Remove ELF_MACHINE from cpu.h Peter Crosthwaite
2015-06-01  8:39   ` Paolo Bonzini
2015-05-31  6:12 ` [Qemu-devel] [RFC v2 29/34] hw: mb: Explicitly include cpu.h for consumers Peter Crosthwaite
2015-05-31  6:12 ` [Qemu-devel] [RFC v2 30/34] mb: Remove ELF_MACHINE from cpu.h Peter Crosthwaite
2015-05-31  6:12 ` [Qemu-devel] [RFC v2 31/34] arm: boot: Don't assume all CPUs are ARM Peter Crosthwaite
2015-05-31  6:12 ` [Qemu-devel] [RFC v2 32/34] arm: xilinx_zynq: Add a Microblaze Peter Crosthwaite
2015-05-31  6:12 ` [Qemu-devel] [RFC v2 33/34] HACK: mb: boot: Assume using -firmware for mb software Peter Crosthwaite
2015-05-31  6:12 ` [Qemu-devel] [RFC v2 34/34] HACK: mb: boot: Disable dtb load in multi-arch Peter Crosthwaite
2015-06-01  8:44 ` [Qemu-devel] [RFC v2 00/34] Multi Architecture System Emulation Paolo Bonzini

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=556C14EE.2080608@redhat.com \
    --to=pbonzini@redhat.com \
    --cc=afaerber@suse.de \
    --cc=crosthwaite.peter@gmail.com \
    --cc=crosthwaitepeter@gmail.com \
    --cc=edgar.iglesias@gmail.com \
    --cc=peter.maydell@linaro.org \
    --cc=qemu-devel@nongnu.org \
    --cc=rth@twiddle.net \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).