qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Blue Swirl <blauwirbel@gmail.com>
To: Artyom Tarasenko <atar4qemu@googlemail.com>
Cc: qemu-devel@nongnu.org, Artyom Tarasenko <atar4qemu@gmail.com>
Subject: [Qemu-devel] Re: sparc32 do_unassigned_access overhaul
Date: Fri, 15 Jan 2010 19:45:47 +0000	[thread overview]
Message-ID: <f43fc5581001151145v77ea594dk7288373974413e8f@mail.gmail.com> (raw)
In-Reply-To: <1263581172-16129-1-git-send-email-atar4qemu@google.com>

On Fri, Jan 15, 2010 at 6:46 PM, Artyom Tarasenko
<atar4qemu@googlemail.com> wrote:
> According to pages 9-31 - 9-34 of "SuperSPARC & MultiCache Controller
> User's Manual":
>
> 1. "A lower priority fault may not overwrite the
>    MFSR status of a higher priority fault."
> 2. The MFAR is overwritten according to the policy defined for the MFSR
> 3. The overwrite bit is asserted if the fault status register (MFSR)
>   has been written more than once by faults of the same class
> 4. SuperSPARC will never place instruction fault addresses in the MFAR.
>
> Implementation of points 1-3 allows booting Solaris 2.6 and 2.5.1.

Nice work! This also passes my tests. However, there are some
CODING_STYLE issues.

>
> Signed-off-by: Artyom Tarasenko <atar4qemu@gmail.com>
> ---
> diff --git a/target-sparc/op_helper.c b/target-sparc/op_helper.c
> index 381e6c4..3a56ce9 100644
> --- a/target-sparc/op_helper.c
> +++ b/target-sparc/op_helper.c
> @@ -3714,6 +3714,7 @@ void do_unassigned_access(target_phys_addr_t addr, int is_write, int is_exec,
>                           int is_asi, int size)
>  {
>     CPUState *saved_env;
> +    int fault_type;
>
>     /* XXX: hack to restore env in all cases, even if not called from
>        generated code */
> @@ -3731,18 +3732,27 @@ void do_unassigned_access(target_phys_addr_t addr, int is_write, int is_exec,
>                is_exec ? "exec" : is_write ? "write" : "read", size,
>                size == 1 ? "" : "s", addr, env->pc);
>  #endif
> -    if (env->mmuregs[3]) /* Fault status register */
> -        env->mmuregs[3] = 1; /* overflow (not read before another fault) */
> -    if (is_asi)
> -        env->mmuregs[3] |= 1 << 16;
> -    if (env->psrs)
> -        env->mmuregs[3] |= 1 << 5;
> -    if (is_exec)
> -        env->mmuregs[3] |= 1 << 6;
> -    if (is_write)
> -        env->mmuregs[3] |= 1 << 7;
> -    env->mmuregs[3] |= (5 << 2) | 2;
> -    env->mmuregs[4] = addr; /* Fault address register */
> +    /* Don't overwrite translation and access faults */
> +    fault_type=(env->mmuregs[3]&0x1c)>>2;

Must have spaces around '=', '&' and '>>'.

> +    if ((fault_type > 4) || (fault_type==0)) {

Must have spaces around '=='.

> +        env->mmuregs[3]=0; /* Fault status register */

and '='

> +        if (is_asi)
> +            env->mmuregs[3] |= 1 << 16;
> +        if (env->psrs)
> +            env->mmuregs[3] |= 1 << 5;
> +        if (is_exec)
> +            env->mmuregs[3] |= 1 << 6;
> +        if (is_write)
> +            env->mmuregs[3] |= 1 << 7;

Here you could add the {} which the original lacked, but as this is
only code movement it's not needed.

> +        env->mmuregs[3] |= (5 << 2) | 2;
> +        /* SuperSPARC will never place instruction fault addresses in the FAR */
> +        if (!is_exec)
> +            env->mmuregs[4] = addr; /* Fault address register */

But this is new code so {} must be added.

> +    }
> +    /* overflow (same type fault was not read before another fault) */
> +    if (fault_type==((env->mmuregs[3]&0x1c))>>2)

Must have spaces around '=', '&' and '>>'.

> +        env->mmuregs[3] |= 1;
> +
>     if ((env->mmuregs[0] & MMU_E) && !(env->mmuregs[0] & MMU_NF)) {
>         if (is_exec)
>             raise_exception(TT_CODE_ACCESS);
> @@ -3750,6 +3760,10 @@ void do_unassigned_access(target_phys_addr_t addr, int is_write, int is_exec,
>             raise_exception(TT_DATA_ACCESS);
>     }
>     env = saved_env;
> +    /* flush neverland mappings created during no-fault mode,
> +       so the sequential MMU faults report proper fault types */
> +    if (env->mmuregs[0] & MMU_NF)
> +        tlb_flush(env, 1);

New code, {}.

>  }
>  #else
>  void do_unassigned_access(target_phys_addr_t addr, int is_write, int is_exec,
>

  reply	other threads:[~2010-01-15 19:46 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-01-15 18:46 [Qemu-devel] sparc32 do_unassigned_access overhaul Artyom Tarasenko
2010-01-15 19:45 ` Blue Swirl [this message]
2010-01-15 21:11   ` [Qemu-devel] " Artyom Tarasenko
2010-01-15 21:26     ` Blue Swirl
2010-01-15 21:48       ` Artyom Tarasenko
2010-01-19 17:30         ` Artyom Tarasenko
2010-01-19 19:32           ` Blue Swirl
2010-01-19 21:44             ` Artyom Tarasenko
2010-01-20 18:29               ` Blue Swirl
2010-01-22 18:00               ` Blue Swirl
2010-01-22 20:51                 ` Artyom Tarasenko
2010-01-23  7:55                   ` Blue Swirl
2010-01-23 16:46                     ` Artyom Tarasenko
2010-01-23 18:39                       ` Blue Swirl

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=f43fc5581001151145v77ea594dk7288373974413e8f@mail.gmail.com \
    --to=blauwirbel@gmail.com \
    --cc=atar4qemu@gmail.com \
    --cc=atar4qemu@googlemail.com \
    --cc=qemu-devel@nongnu.org \
    /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).