qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: "Nicholas Piggin" <npiggin@gmail.com>
To: "BALATON Zoltan" <balaton@eik.bme.hu>
Cc: <qemu-devel@nongnu.org>, <qemu-ppc@nongnu.org>, <clg@kaod.org>,
	"Greg Kurz" <groug@kaod.org>,
	"Daniel Henrique Barboza" <danielhb413@gmail.com>
Subject: Re: [PATCH 08/10] target/ppc: Fix gen_sc to use correct nip
Date: Thu, 15 Jun 2023 11:43:18 +1000	[thread overview]
Message-ID: <CTCUCTKOWUT6.111CRYUYBN9HO@wheely> (raw)
In-Reply-To: <da6086be-ccd8-7cbf-8fbe-e4924d9c230f@eik.bme.hu>

On Thu Jun 15, 2023 at 7:27 AM AEST, BALATON Zoltan wrote:
> On Wed, 14 Jun 2023, Nicholas Piggin wrote:
> > On Mon Jun 12, 2023 at 8:42 AM AEST, BALATON Zoltan wrote:
> >> Most exceptions are raised with nip pointing to the faulting
> >> instruction but the sc instruction generating a syscall exception
> >> leaves nip pointing to next instruction. Fix gen_sc to not use
> >> gen_exception_err() which sets nip back but correctly set nip to
> >> pc_next so we don't have to patch this in the exception handlers.
> >>
> >> This changes the nip logged in dump_syscall and dump_hcall debug
> >> functions but now this matches how nip would be on a real CPU.
> >>
> >> Signed-off-by: BALATON Zoltan <balaton@eik.bme.hu>
> >> ---
> >>  target/ppc/excp_helper.c | 39 ---------------------------------------
> >>  target/ppc/translate.c   |  8 +++++---
> >>  2 files changed, 5 insertions(+), 42 deletions(-)
> >>
> >> diff --git a/target/ppc/excp_helper.c b/target/ppc/excp_helper.c
> >> index 885e479301..4f6a6dfb19 100644
> >> --- a/target/ppc/excp_helper.c
> >> +++ b/target/ppc/excp_helper.c
> >> @@ -493,12 +493,6 @@ static void powerpc_excp_40x(PowerPCCPU *cpu, int excp)
> >>          break;
> >>      case POWERPC_EXCP_SYSCALL:   /* System call exception                    */
> >>          dump_syscall(env);
> >> -
> >> -        /*
> >> -         * We need to correct the NIP which in this case is supposed
> >> -         * to point to the next instruction
> >> -         */
> >> -        env->nip += 4;
> >>          break;
> >>      case POWERPC_EXCP_FIT:       /* Fixed-interval timer interrupt           */
> >>          trace_ppc_excp_print("FIT");
> >> @@ -609,12 +603,6 @@ static void powerpc_excp_6xx(PowerPCCPU *cpu, int excp)
> >>          break;
> >>      case POWERPC_EXCP_SYSCALL:   /* System call exception                    */
> >>          dump_syscall(env);
> >> -
> >> -        /*
> >> -         * We need to correct the NIP which in this case is supposed
> >> -         * to point to the next instruction
> >> -         */
> >> -        env->nip += 4;
> >>          break;
> >>      case POWERPC_EXCP_FPU:       /* Floating-point unavailable exception     */
> >>      case POWERPC_EXCP_DECR:      /* Decrementer exception                    */
> >> @@ -757,13 +745,6 @@ static void powerpc_excp_7xx(PowerPCCPU *cpu, int excp)
> >>          } else {
> >>              dump_syscall(env);
> >>          }
> >> -
> >> -        /*
> >> -         * We need to correct the NIP which in this case is supposed
> >> -         * to point to the next instruction
> >> -         */
> >> -        env->nip += 4;
> >> -
> >>          /*
> >>           * The Virtual Open Firmware (VOF) relies on the 'sc 1'
> >>           * instruction to communicate with QEMU. The pegasos2 machine
> >> @@ -908,13 +889,6 @@ static void powerpc_excp_74xx(PowerPCCPU *cpu, int excp)
> >>          } else {
> >>              dump_syscall(env);
> >>          }
> >> -
> >> -        /*
> >> -         * We need to correct the NIP which in this case is supposed
> >> -         * to point to the next instruction
> >> -         */
> >> -        env->nip += 4;
> >> -
> >>          /*
> >>           * The Virtual Open Firmware (VOF) relies on the 'sc 1'
> >>           * instruction to communicate with QEMU. The pegasos2 machine
> >> @@ -1073,12 +1047,6 @@ static void powerpc_excp_booke(PowerPCCPU *cpu, int excp)
> >>          break;
> >>      case POWERPC_EXCP_SYSCALL:   /* System call exception                    */
> >>          dump_syscall(env);
> >> -
> >> -        /*
> >> -         * We need to correct the NIP which in this case is supposed
> >> -         * to point to the next instruction
> >> -         */
> >> -        env->nip += 4;
> >>          break;
> >>      case POWERPC_EXCP_FPU:       /* Floating-point unavailable exception     */
> >>      case POWERPC_EXCP_APU:       /* Auxiliary processor unavailable          */
> >> @@ -1320,13 +1288,6 @@ static void powerpc_excp_books(PowerPCCPU *cpu, int excp)
> >>          } else {
> >>              dump_syscall(env);
> >>          }
> >> -
> >> -        /*
> >> -         * We need to correct the NIP which in this case is supposed
> >> -         * to point to the next instruction
> >> -         */
> >> -        env->nip += 4;
> >> -
> >>          /* "PAPR mode" built-in hypercall emulation */
> >>          if (lev == 1 && books_vhyp_handles_hcall(cpu)) {
> >>              PPCVirtualHypervisorClass *vhc =
> >> diff --git a/target/ppc/translate.c b/target/ppc/translate.c
> >> index a32a9b8a5f..4260d3d66f 100644
> >> --- a/target/ppc/translate.c
> >> +++ b/target/ppc/translate.c
> >> @@ -4419,10 +4419,12 @@ static void gen_hrfid(DisasContext *ctx)
> >>  #endif
> >>  static void gen_sc(DisasContext *ctx)
> >>  {
> >> -    uint32_t lev;
> >> +    uint32_t lev = (ctx->opcode >> 5) & 0x7F;
> >>
> >> -    lev = (ctx->opcode >> 5) & 0x7F;
> >> -    gen_exception_err(ctx, POWERPC_SYSCALL, lev);
> >> +    gen_update_nip(ctx, ctx->base.pc_next);
> >> +    gen_helper_raise_exception_err(cpu_env, tcg_constant_i32(POWERPC_SYSCALL),
> >> +                                   tcg_constant_i32(lev));
> >> +    ctx->base.is_jmp = DISAS_NORETURN;
> >
> > Generally for blame and bisect I don't like to mix cleanup with real
> > change, I guess this is pretty minor though.
> >
> > Great cleanup though, sc is certainly defined to set SRR0 to the
> > instruction past the sc unlike other interrupts so it is more natural
> > and less hacky feeling do it like this.
> >
> > Could you do scv while you are here? It has the same semantics as
> > sc.
>
> I've tried but scv seems to be a bit different as it can sometimes raise 
> POWERPC_EXCP_FU instead of POWERPC_EXCP_SYSCALL_VECTORED according to 
> excp_helper.c::helper_scv() which may need the nip to be different so I'd 
> rather leave this to somebody who knows what they are doing.

Yes and the fscr check requires a helper so no point doing that
twice, but I think it would still better match the new sc code to
have something like this (patch is tested, you can squash it in
yours if you like):

Thanks,
Nick
---

diff --git a/target/ppc/excp_helper.c b/target/ppc/excp_helper.c
index 12d8a7257b..55c6edfacb 100644
--- a/target/ppc/excp_helper.c
+++ b/target/ppc/excp_helper.c
@@ -1507,7 +1507,6 @@ static void powerpc_excp_books(PowerPCCPU *cpu, int excp)
     case POWERPC_EXCP_SYSCALL_VECTORED: /* scv exception                     */
         lev = env->error_code;
         dump_syscall(env);
-        env->nip += 4;
         new_msr |= env->msr & ((target_ulong)1 << MSR_EE);
         new_msr |= env->msr & ((target_ulong)1 << MSR_RI);
 
@@ -2623,6 +2622,7 @@ void helper_ppc_maybe_interrupt(CPUPPCState *env)
 void helper_scv(CPUPPCState *env, uint32_t lev)
 {
     if (env->spr[SPR_FSCR] & (1ull << FSCR_SCV)) {
+        env->nip += 4;
         raise_exception_err(env, POWERPC_EXCP_SYSCALL_VECTORED, lev);
     } else {
         raise_exception_err(env, POWERPC_EXCP_FU, FSCR_IC_SCV);
diff --git a/target/ppc/translate.c b/target/ppc/translate.c
index b591f2e496..90d07a3e56 100644
--- a/target/ppc/translate.c
+++ b/target/ppc/translate.c
@@ -4432,7 +4432,11 @@ static void gen_scv(DisasContext *ctx)
 {
     uint32_t lev = (ctx->opcode >> 5) & 0x7F;
 
-    /* Set the PC back to the faulting instruction. */
+    /*
+     * Set the PC back to the scv instruction (unlike sc), because a facility
+     * unavailable interrupt must be generated if FSCR[SCV]=0. The helper
+     * advances nip if the FSCR check passes.
+     */
     gen_update_nip(ctx, ctx->cia);
     gen_helper_scv(cpu_env, tcg_constant_i32(lev));
 


  reply	other threads:[~2023-06-15  1:44 UTC|newest]

Thread overview: 34+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-06-11 22:42 [PATCH 00/10] Misc clean ups to target/ppc exception handling BALATON Zoltan
2023-06-11 22:42 ` [PATCH 01/10] target/ppc: Remove some superfluous parentheses BALATON Zoltan
2023-06-14  3:25   ` Nicholas Piggin
2023-06-11 22:42 ` [PATCH 02/10] target/ppc: Remove unneeded parameter from powerpc_reset_wakeup() BALATON Zoltan
2023-06-14  3:26   ` Nicholas Piggin
2023-06-11 22:42 ` [PATCH 03/10] target/ppc: Move common check in exception handlers to a function BALATON Zoltan
2023-06-12  9:28   ` Philippe Mathieu-Daudé
2023-06-12 10:07     ` BALATON Zoltan
2023-06-14  3:35   ` Nicholas Piggin
2023-06-14  6:25     ` Cédric Le Goater
2023-06-15  1:50       ` Nicholas Piggin
2023-06-11 22:42 ` [PATCH 04/10] target/ppc: Use env_cpu for cpu_abort in excp_helper BALATON Zoltan
2023-06-14  3:36   ` Nicholas Piggin
2023-06-14 10:13     ` BALATON Zoltan
2023-06-15  1:01       ` Nicholas Piggin
2023-06-11 22:42 ` [PATCH 05/10] target/ppc: Change parameter of cpu_interrupt_exittb() to an env pointer BALATON Zoltan
2023-06-11 22:42 ` [PATCH 06/10] target/ppc: Readability improvements in exception handlers BALATON Zoltan
2023-06-14  3:42   ` Nicholas Piggin
2023-06-14 10:07     ` BALATON Zoltan
2023-06-15  0:59       ` Nicholas Piggin
2023-06-11 22:42 ` [PATCH 07/10] target/ppd: Remove unused define BALATON Zoltan
2023-06-12  9:29   ` Philippe Mathieu-Daudé
2023-06-14  3:42   ` Nicholas Piggin
2023-06-11 22:42 ` [PATCH 08/10] target/ppc: Fix gen_sc to use correct nip BALATON Zoltan
2023-06-14  3:47   ` Nicholas Piggin
2023-06-14 10:05     ` BALATON Zoltan
2023-06-14 21:27     ` BALATON Zoltan
2023-06-15  1:43       ` Nicholas Piggin [this message]
2023-06-15 23:02         ` BALATON Zoltan
2023-06-11 22:42 ` [PATCH 09/10] target/ppc: Simplify syscall exception handlers BALATON Zoltan
2023-06-14  4:18   ` Nicholas Piggin
2023-06-14 21:33     ` BALATON Zoltan
2023-06-15  1:47       ` Nicholas Piggin
2023-06-11 22:42 ` [PATCH 10/10] target/ppc: Get CPUState in one step BALATON Zoltan

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=CTCUCTKOWUT6.111CRYUYBN9HO@wheely \
    --to=npiggin@gmail.com \
    --cc=balaton@eik.bme.hu \
    --cc=clg@kaod.org \
    --cc=danielhb413@gmail.com \
    --cc=groug@kaod.org \
    --cc=qemu-devel@nongnu.org \
    --cc=qemu-ppc@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).