From: Richard Henderson <richard.henderson@linaro.org>
To: qemu-devel@nongnu.org
Cc: edgar.iglesias@xilinx.com
Subject: [PATCH 4/7] target/microblaze: Handle DISAS_EXIT_NEXT in delay slot
Date: Wed, 2 Sep 2020 09:35:56 -0700 [thread overview]
Message-ID: <20200902163559.1077728-5-richard.henderson@linaro.org> (raw)
In-Reply-To: <20200902163559.1077728-1-richard.henderson@linaro.org>
It is legal to put an mts instruction into a delay slot.
We should continue to return to the main loop in that
case so that we recognize any pending interrupts.
Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Tested-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
target/microblaze/translate.c | 34 +++++++++++++++++++++++++++++++++-
1 file changed, 33 insertions(+), 1 deletion(-)
diff --git a/target/microblaze/translate.c b/target/microblaze/translate.c
index 9c52448c06..459b25f8b9 100644
--- a/target/microblaze/translate.c
+++ b/target/microblaze/translate.c
@@ -1696,6 +1696,10 @@ static void mb_tr_translate_insn(DisasContextBase *dcb, CPUState *cs)
dc->base.pc_next += 4;
if (dc->jmp_cond != TCG_COND_NEVER && !(dc->tb_flags & D_FLAG)) {
+ /*
+ * Finish any return-from branch.
+ * TODO: Diagnose rtXd in delay slot of rtYd earlier.
+ */
if (dc->tb_flags & DRTI_FLAG) {
do_rti(dc);
} else if (dc->tb_flags & DRTB_FLAG) {
@@ -1703,7 +1707,35 @@ static void mb_tr_translate_insn(DisasContextBase *dcb, CPUState *cs)
} else if (dc->tb_flags & DRTE_FLAG) {
do_rte(dc);
}
- dc->base.is_jmp = DISAS_JUMP;
+
+ /* Complete the branch, ending the TB. */
+ switch (dc->base.is_jmp) {
+ case DISAS_NORETURN:
+ /*
+ * E.g. illegal insn in a delay slot. We've already exited
+ * and will handle D_FLAG in mb_cpu_do_interrupt.
+ */
+ break;
+ case DISAS_EXIT:
+ /*
+ * TODO: diagnose brk/brki in delay slot earlier.
+ * This would then fold into the illegal insn case above.
+ */
+ break;
+ case DISAS_NEXT:
+ /* Normal insn a delay slot. */
+ dc->base.is_jmp = DISAS_JUMP;
+ break;
+ case DISAS_EXIT_NEXT:
+ /*
+ * E.g. mts insn in a delay slot. Continue with btarget,
+ * but still return to the main loop.
+ */
+ dc->base.is_jmp = DISAS_EXIT_JUMP;
+ break;
+ default:
+ g_assert_not_reached();
+ }
}
}
--
2.25.1
next prev parent reply other threads:[~2020-09-02 16:38 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-09-02 16:35 [PATCH 0/7] target/microblaze improvements Richard Henderson
2020-09-02 16:35 ` [PATCH 1/7] target/microblaze: Rename DISAS_UPDATE to DISAS_EXIT Richard Henderson
2020-09-02 16:35 ` [PATCH 2/7] target/microblaze: Introduce DISAS_EXIT_NEXT, DISAS_EXIT_JUMP Richard Henderson
2020-09-02 16:35 ` [PATCH 3/7] target/microblaze: Replace cpustate_changed with DISAS_EXIT_NEXT Richard Henderson
2020-09-02 16:35 ` Richard Henderson [this message]
2020-09-02 16:35 ` [PATCH 5/7] target/microblaze: Force rtid, rted, rtbd to exit Richard Henderson
2020-09-02 16:35 ` [PATCH 6/7] target/microblaze: Use tcg_gen_lookup_and_goto_ptr Richard Henderson
2020-09-02 16:35 ` [PATCH 7/7] target/microblaze: Diagnose invalid insns in delay slots Richard Henderson
2020-09-02 17:13 ` Philippe Mathieu-Daudé
2020-09-02 18:03 ` Richard Henderson
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=20200902163559.1077728-5-richard.henderson@linaro.org \
--to=richard.henderson@linaro.org \
--cc=edgar.iglesias@xilinx.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).