public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Max Filippov <jcmvbkbc@gmail.com>
To: linux-xtensa@linux-xtensa.org
Cc: Chris Zankel <chris@zankel.net>,
	linux-kernel@vger.kernel.org, Max Filippov <jcmvbkbc@gmail.com>
Subject: [PATCH v3 3/3] xtensa: improve call0 ABI probing
Date: Tue, 17 May 2022 03:44:58 -0700	[thread overview]
Message-ID: <20220517104458.257799-4-jcmvbkbc@gmail.com> (raw)
In-Reply-To: <20220517104458.257799-1-jcmvbkbc@gmail.com>

When call0 userspace ABI support by probing is enabled instructions that
cause illegal instruction exception when PS.WOE is clear are retried
with PS.WOE set before calling c-level exception handler. Record user pc
at which PS.WOE was set in the fast exception handler and clear PS.WOE
in the c-level exception handler if we get there from the same address.

Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
---
Changes v1->v2:

- split ABI probing improvement from the artificial division by 0

 arch/xtensa/include/asm/thread_info.h |  4 ++++
 arch/xtensa/kernel/asm-offsets.c      |  3 +++
 arch/xtensa/kernel/entry.S            |  5 +++++
 arch/xtensa/kernel/traps.c            | 12 ++++++++++++
 4 files changed, 24 insertions(+)

diff --git a/arch/xtensa/include/asm/thread_info.h b/arch/xtensa/include/asm/thread_info.h
index 52974317a6b6..326db1c1d5d8 100644
--- a/arch/xtensa/include/asm/thread_info.h
+++ b/arch/xtensa/include/asm/thread_info.h
@@ -56,6 +56,10 @@ struct thread_info {
 	/* result of the most recent exclusive store */
 	unsigned long		atomctl8;
 #endif
+#ifdef CONFIG_USER_ABI_CALL0_PROBE
+	/* Address where PS.WOE was enabled by the ABI probing code */
+	unsigned long		ps_woe_fix_addr;
+#endif
 
 	/*
 	 * If i-th bit is set then coprocessor state is loaded into the
diff --git a/arch/xtensa/kernel/asm-offsets.c b/arch/xtensa/kernel/asm-offsets.c
index 9a1db6ffcbf4..da38de20ae59 100644
--- a/arch/xtensa/kernel/asm-offsets.c
+++ b/arch/xtensa/kernel/asm-offsets.c
@@ -88,6 +88,9 @@ int main(void)
 	OFFSET(TI_STSTUS, thread_info, status);
 	OFFSET(TI_CPU, thread_info, cpu);
 	OFFSET(TI_PRE_COUNT, thread_info, preempt_count);
+#ifdef CONFIG_USER_ABI_CALL0_PROBE
+	OFFSET(TI_PS_WOE_FIX_ADDR, thread_info, ps_woe_fix_addr);
+#endif
 
 	/* struct thread_info (offset from start_struct) */
 	DEFINE(THREAD_RA, offsetof (struct task_struct, thread.ra));
diff --git a/arch/xtensa/kernel/entry.S b/arch/xtensa/kernel/entry.S
index 3224b4ceca34..e3eae648ba2e 100644
--- a/arch/xtensa/kernel/entry.S
+++ b/arch/xtensa/kernel/entry.S
@@ -1056,6 +1056,11 @@ ENTRY(fast_illegal_instruction_user)
 	movi	a3, PS_WOE_MASK
 	or	a0, a0, a3
 	wsr	a0, ps
+#ifdef CONFIG_USER_ABI_CALL0_PROBE
+	GET_THREAD_INFO(a3, a2)
+	rsr	a0, epc1
+	s32i	a0, a3, TI_PS_WOE_FIX_ADDR
+#endif
 	l32i	a3, a2, PT_AREG3
 	l32i	a0, a2, PT_AREG0
 	rsr	a2, depc
diff --git a/arch/xtensa/kernel/traps.c b/arch/xtensa/kernel/traps.c
index 2b75b252b626..f97d43a8d13d 100644
--- a/arch/xtensa/kernel/traps.c
+++ b/arch/xtensa/kernel/traps.c
@@ -317,6 +317,18 @@ static bool check_div0(struct pt_regs *regs)
 
 static void do_illegal_instruction(struct pt_regs *regs)
 {
+#ifdef CONFIG_USER_ABI_CALL0_PROBE
+	/*
+	 * When call0 application encounters an illegal instruction fast
+	 * exception handler will attempt to set PS.WOE and retry failing
+	 * instruction.
+	 * If we get here we know that that instruction is also illegal
+	 * with PS.WOE set, so it's not related to the windowed option
+	 * hence PS.WOE may be cleared.
+	 */
+	if (regs->pc == current_thread_info()->ps_woe_fix_addr)
+		regs->ps &= ~PS_WOE_MASK;
+#endif
 	if (check_div0(regs)) {
 		do_div0(regs);
 		return;
-- 
2.30.2


      parent reply	other threads:[~2022-05-17 10:46 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-05-17 10:44 [PATCH v3 0/3] xtensa: add division by zero exception handler Max Filippov
2022-05-17 10:44 ` [PATCH v3 1/3] xtensa: add trap handler for division by zero Max Filippov
2022-05-17 10:44 ` [PATCH v3 2/3] xtensa: support artificial division by 0 exception Max Filippov
2022-05-17 10:44 ` Max Filippov [this message]

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=20220517104458.257799-4-jcmvbkbc@gmail.com \
    --to=jcmvbkbc@gmail.com \
    --cc=chris@zankel.net \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-xtensa@linux-xtensa.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