From: "Ilpo Järvinen" <ilpo.jarvinen@linux.intel.com>
To: Bjorn Helgaas <bhelgaas@google.com>,
linux-pci@vger.kernel.org, Guenter Roeck <groeck@juniper.net>,
Lukas Wunner <lukas@wunner.de>,
Mika Westerberg <mika.westerberg@linux.intel.com>,
"Rafael J. Wysocki" <rafael.j.wysocki@intel.com>,
Rajat Jain <rajatxjain@gmail.com>,
Joel Mathew Thomas <proxy0@tutamail.com>,
linux-kernel@vger.kernel.org
Cc: "Jonathan Cameron" <Jonathan.Cameron@huawei.com>,
"Ilpo Järvinen" <ilpo.jarvinen@linux.intel.com>,
stable@vger.kernel.org
Subject: [PATCH 1/4] PCI/hotplug: Disable HPIE over reset
Date: Thu, 13 Mar 2025 16:23:30 +0200 [thread overview]
Message-ID: <20250313142333.5792-2-ilpo.jarvinen@linux.intel.com> (raw)
In-Reply-To: <20250313142333.5792-1-ilpo.jarvinen@linux.intel.com>
pciehp_reset_slot() disables PDCE (Presence Detect Changed Enable) and
DLLSCE (Data Link Layer State Changed Enable) for the duration of reset
and clears the related status bits PDC and DLLSC from the Slot Status
register after the reset to avoid hotplug incorrectly assuming the card
was removed.
However, hotplug shares interrupt with PME and BW notifications both of
which can make pciehp_isr() to run despite PDCE and DLLSCE bits being
off. pciehp_isr() then picks PDC or DLLSC bits from the Slot Status
register due to the events that occur during reset and caches them into
->pending_events. Later, the IRQ thread in pciehp_ist() will process
the ->pending_events and will assume the Link went Down due to a card
change (in pciehp_handle_presence_or_link_change()).
Change pciehp_reset_slot() to also clear HPIE (Hot-Plug Interrupt
Enable) as pciehp_isr() will first check HPIE to see if the interrupt
is not for it. Then synchronize with the IRQ handling to ensure no
events are pending, before invoking the reset.
Similarly, if the poll mode is in use, park the poll thread over the
duration of the reset to stop handling events.
In order to not race irq_syncronize()/kthread_{,un}park() with the irq
/ poll_thread freeing from pciehp_remove(), take reset_lock in
pciehp_free_irq() and check the irq / poll_thread variable validity in
pciehp_reset_slot().
Fixes: 06a8d89af551 ("PCI: pciehp: Disable link notification across slot reset")
Fixes: 720d6a671a6e ("PCI: pciehp: Do not handle events if interrupts are masked")
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=219765
Suggested-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Cc: stable@vger.kernel.org
---
drivers/pci/hotplug/pciehp_hpc.c | 28 ++++++++++++++++++++++++----
1 file changed, 24 insertions(+), 4 deletions(-)
diff --git a/drivers/pci/hotplug/pciehp_hpc.c b/drivers/pci/hotplug/pciehp_hpc.c
index bb5a8d9f03ad..c487e274b282 100644
--- a/drivers/pci/hotplug/pciehp_hpc.c
+++ b/drivers/pci/hotplug/pciehp_hpc.c
@@ -77,10 +77,15 @@ static inline int pciehp_request_irq(struct controller *ctrl)
static inline void pciehp_free_irq(struct controller *ctrl)
{
- if (pciehp_poll_mode)
+ down_read_nested(&ctrl->reset_lock, ctrl->depth);
+ if (pciehp_poll_mode) {
kthread_stop(ctrl->poll_thread);
- else
+ ctrl->poll_thread = NULL;
+ } else {
free_irq(ctrl->pcie->irq, ctrl);
+ ctrl->pcie->irq = IRQ_NOTCONNECTED;
+ }
+ up_read(&ctrl->reset_lock);
}
static int pcie_poll_cmd(struct controller *ctrl, int timeout)
@@ -766,8 +771,9 @@ static int pciehp_poll(void *data)
while (!kthread_should_stop()) {
/* poll for interrupt events or user requests */
- while (pciehp_isr(IRQ_NOTCONNECTED, ctrl) == IRQ_WAKE_THREAD ||
- atomic_read(&ctrl->pending_events))
+ while (!kthread_should_park() &&
+ (pciehp_isr(IRQ_NOTCONNECTED, ctrl) == IRQ_WAKE_THREAD ||
+ atomic_read(&ctrl->pending_events)))
pciehp_ist(IRQ_NOTCONNECTED, ctrl);
if (pciehp_poll_time <= 0 || pciehp_poll_time > 60)
@@ -907,6 +913,8 @@ int pciehp_reset_slot(struct hotplug_slot *hotplug_slot, bool probe)
down_write_nested(&ctrl->reset_lock, ctrl->depth);
+ if (!pciehp_poll_mode)
+ ctrl_mask |= PCI_EXP_SLTCTL_HPIE;
if (!ATTN_BUTTN(ctrl)) {
ctrl_mask |= PCI_EXP_SLTCTL_PDCE;
stat_mask |= PCI_EXP_SLTSTA_PDC;
@@ -918,9 +926,21 @@ int pciehp_reset_slot(struct hotplug_slot *hotplug_slot, bool probe)
ctrl_dbg(ctrl, "%s: SLOTCTRL %x write cmd %x\n", __func__,
pci_pcie_cap(ctrl->pcie->port) + PCI_EXP_SLTCTL, 0);
+ /* Make sure HPIE is no longer seen by the interrupt handler. */
+ if (pciehp_poll_mode) {
+ if (ctrl->poll_thread)
+ kthread_park(ctrl->poll_thread);
+ } else {
+ if (ctrl->pcie->irq != IRQ_NOTCONNECTED)
+ synchronize_irq(ctrl->pcie->irq);
+ }
+
rc = pci_bridge_secondary_bus_reset(ctrl->pcie->port);
pcie_capability_write_word(pdev, PCI_EXP_SLTSTA, stat_mask);
+ if (pciehp_poll_mode && ctrl->poll_thread)
+ kthread_unpark(ctrl->poll_thread);
+
pcie_write_cmd_nowait(ctrl, ctrl_mask, ctrl_mask);
ctrl_dbg(ctrl, "%s: SLOTCTRL %x write cmd %x\n", __func__,
pci_pcie_cap(ctrl->pcie->port) + PCI_EXP_SLTCTL, ctrl_mask);
--
2.39.5
next prev parent reply other threads:[~2025-03-13 14:24 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-03-13 14:23 [PATCH 0/4] PCI/hotplug: Fix interrupt / event handling problems Ilpo Järvinen
2025-03-13 14:23 ` Ilpo Järvinen [this message]
2025-03-15 15:58 ` [PATCH 1/4] PCI/hotplug: Disable HPIE over reset Lukas Wunner
[not found] ` <OLQ9qyD--F-9@tutamail.com>
2025-03-15 21:12 ` Lukas Wunner
2025-03-17 18:08 ` Ilpo Järvinen
2025-03-13 14:23 ` [PATCH 2/4] PCI/hotplug: Clearing HPIE for the duration of reset is enough Ilpo Järvinen
2025-03-13 14:23 ` [PATCH 3/4] PCI/hotplug: reset_lock is not required synchronizing with irq thread Ilpo Järvinen
2025-03-14 8:32 ` Lukas Wunner
2025-03-14 11:18 ` Ilpo Järvinen
2025-03-13 14:23 ` [PATCH 4/4] PCI/hotplug: Don't enabled HPIE in poll mode Ilpo Järvinen
2025-03-15 11:57 ` Lukas Wunner
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=20250313142333.5792-2-ilpo.jarvinen@linux.intel.com \
--to=ilpo.jarvinen@linux.intel.com \
--cc=Jonathan.Cameron@huawei.com \
--cc=bhelgaas@google.com \
--cc=groeck@juniper.net \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-pci@vger.kernel.org \
--cc=lukas@wunner.de \
--cc=mika.westerberg@linux.intel.com \
--cc=proxy0@tutamail.com \
--cc=rafael.j.wysocki@intel.com \
--cc=rajatxjain@gmail.com \
--cc=stable@vger.kernel.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