qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Nicholas Piggin <npiggin@gmail.com>
To: qemu-ppc@nongnu.org
Cc: "Nicholas Piggin" <npiggin@gmail.com>,
	"Glenn Miles" <milesg@linux.vnet.ibm.com>,
	"Cédric Le Goater" <clg@kaod.org>,
	"Frédéric Barrat" <fbarrat@linux.ibm.com>,
	qemu-devel@nongnu.org, "Glenn Miles" <milesg@linux.ibm.com>
Subject: [PATCH 1/2] ppc/pnv: Fix loss of LPC SERIRQ interrupts
Date: Tue, 28 May 2024 16:20:44 +1000	[thread overview]
Message-ID: <20240528062045.624906-2-npiggin@gmail.com> (raw)
In-Reply-To: <20240528062045.624906-1-npiggin@gmail.com>

From: Glenn Miles <milesg@linux.vnet.ibm.com>

The LPC HC irq status register bits are set when an LPC IRQSER input is
asserted. These irq status bits drive the PSI irq to the CPU interrupt
controller. The LPC HC irq status bits are cleared by software writing
to the register with 1's for the bits to clear.

Existing register write was clearing the irq status bits even when the
input was asserted, this results in interrupts being lost.

This fix changes the behavior to keep track of the device IRQ status
in internal state that is separate from the irq status register, and
only allowing the irq status bits to be cleared if the associated
input is not asserted.

[np: rebased before P9 PSI SERIRQ patch, adjust changelog/comments]
Signed-off-by: Glenn Miles <milesg@linux.ibm.com>
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
 include/hw/ppc/pnv_lpc.h |  3 +++
 hw/ppc/pnv_lpc.c         | 22 +++++++++++++++++++---
 2 files changed, 22 insertions(+), 3 deletions(-)

diff --git a/include/hw/ppc/pnv_lpc.h b/include/hw/ppc/pnv_lpc.h
index 5d22c45570..97c6872c3f 100644
--- a/include/hw/ppc/pnv_lpc.h
+++ b/include/hw/ppc/pnv_lpc.h
@@ -73,6 +73,9 @@ struct PnvLpcController {
     uint32_t opb_irq_pol;
     uint32_t opb_irq_input;
 
+    /* LPC device IRQ state */
+    uint32_t lpc_hc_irq_inputs;
+
     /* LPC HC registers */
     uint32_t lpc_hc_fw_seg_idsel;
     uint32_t lpc_hc_fw_rd_acc_size;
diff --git a/hw/ppc/pnv_lpc.c b/hw/ppc/pnv_lpc.c
index d692858bee..252690dcaa 100644
--- a/hw/ppc/pnv_lpc.c
+++ b/hw/ppc/pnv_lpc.c
@@ -505,7 +505,14 @@ static void lpc_hc_write(void *opaque, hwaddr addr, uint64_t val,
         pnv_lpc_eval_irqs(lpc);
         break;
     case LPC_HC_IRQSTAT:
-        lpc->lpc_hc_irqstat &= ~val;
+        /*
+         * This register is write-to-clear for the IRQSER (LPC device IRQ)
+         * status. However if the device has not de-asserted its interrupt
+         * that will just raise this IRQ status bit again. Model this by
+         * keeping track of the inputs and only clearing if the inputs are
+         * deasserted.
+         */
+        lpc->lpc_hc_irqstat &= ~(val & ~lpc->lpc_hc_irq_inputs);
         pnv_lpc_eval_irqs(lpc);
         break;
     case LPC_HC_ERROR_ADDRESS:
@@ -803,11 +810,20 @@ static void pnv_lpc_isa_irq_handler_cpld(void *opaque, int n, int level)
 static void pnv_lpc_isa_irq_handler(void *opaque, int n, int level)
 {
     PnvLpcController *lpc = PNV_LPC(opaque);
+    uint32_t irq_bit = LPC_HC_IRQ_SERIRQ0 >> n;
 
-    /* The Naples HW latches the 1 levels, clearing is done by SW */
     if (level) {
-        lpc->lpc_hc_irqstat |= LPC_HC_IRQ_SERIRQ0 >> n;
+        lpc->lpc_hc_irq_inputs |= irq_bit;
+
+        /*
+	 * The LPC HC in Naples and later latches LPC IRQ into a bit field in
+	 * the IRQSTAT register, and that drives the PSI IRQ to the IC.
+	 * Software clears this bit manually (see LPC_HC_IRQSTAT handler).
+         */
+        lpc->lpc_hc_irqstat |= irq_bit;
         pnv_lpc_eval_irqs(lpc);
+    } else {
+        lpc->lpc_hc_irq_inputs &= ~irq_bit;
     }
 }
 
-- 
2.43.0



  reply	other threads:[~2024-05-28  6:22 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-05-28  6:20 [PATCH 0/2] ppc/pnv: LPC interrupt fixes Nicholas Piggin
2024-05-28  6:20 ` Nicholas Piggin [this message]
2024-05-28 14:23   ` [PATCH 1/2] ppc/pnv: Fix loss of LPC SERIRQ interrupts Miles Glenn
2024-05-28  6:20 ` [PATCH 2/2] ppc/pnv: Implement POWER9 LPC PSI serirq outputs and auto-clear function Nicholas Piggin
2024-05-29 14:31   ` Miles Glenn

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=20240528062045.624906-2-npiggin@gmail.com \
    --to=npiggin@gmail.com \
    --cc=clg@kaod.org \
    --cc=fbarrat@linux.ibm.com \
    --cc=milesg@linux.ibm.com \
    --cc=milesg@linux.vnet.ibm.com \
    --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).