NetBSD Problem Report #59180
From www@netbsd.org Sat Mar 15 20:46:16 2025
Return-Path: <www@netbsd.org>
Received: from mail.netbsd.org (mail.netbsd.org [199.233.217.200])
(using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits)
key-exchange X25519 server-signature RSA-PSS (2048 bits)
client-signature RSA-PSS (2048 bits))
(Client CN "mail.NetBSD.org", Issuer "mail.NetBSD.org CA" (not verified))
by mollari.NetBSD.org (Postfix) with ESMTPS id C355B1A923C
for <gnats-bugs@gnats.NetBSD.org>; Sat, 15 Mar 2025 20:46:15 +0000 (UTC)
Message-Id: <20250315204614.782D81A923E@mollari.NetBSD.org>
Date: Sat, 15 Mar 2025 20:46:14 +0000 (UTC)
From: swaggboi@gangstalking.agency
Reply-To: swaggboi@gangstalking.agency
To: gnats-bugs@NetBSD.org
Subject: System reboots instead of shutting down
X-Send-Pr-Version: www-1.0
>Number: 59180
>Category: port-amd64
>Synopsis: System reboots instead of shutting down
>Confidential: no
>Severity: non-critical
>Priority: low
>Responsible: port-amd64-maintainer
>State: open
>Class: sw-bug
>Submitter-Id: net
>Arrival-Date: Sat Mar 15 20:50:00 +0000 2025
>Last-Modified: Mon Mar 31 14:50:07 +0000 2025
>Originator: Daniel Bowling
>Release: 10.1
>Organization:
>Environment:
NetBSD ubuntu 10.1 NetBSD 10.1 (GENERIC) #0: Mon Dec 16 13:08:11 UTC 2024 mkrepro@mkrepro.NetBSD.org:/usr/src/sys/arch/amd64/compile/GENERIC amd64
>Description:
When I enter shutdown -h now, my system performs a reboot instead of shutting down. Riastradh on IRC suggested I create a PR for this:
https://mirror.swagg.net/misc/irc_convo.txt
Here is my dmesg output:
https://mirror.swagg.net/misc/dmesg_output.txt
They noted this picture would be relevant:
https://mirror.swagg.net/Pictures/NetBSD/IMG_0846.jpeg
Here's some other photos I took in case I linked the wrong one:
https://mirror.swagg.net/Pictures/NetBSD/
>How-To-Repeat:
Enter shutdown -h now as root user
>Fix:
>Audit-Trail:
From: Taylor R Campbell <riastradh@NetBSD.org>
To: Daniel Bowling <swaggboi@gangstalking.agency>
Cc: gnats-bugs@NetBSD.org, netbsd-bugs@NetBSD.org
Subject: Re: port-amd64/59180: System reboots instead of shutting down
Date: Sun, 16 Mar 2025 14:38:40 +0000
This is a multi-part message in MIME format.
--=_QJyEfqlAfncoM8fB56hLDQdX5vxg3caH
Can you please try a kernel with the attached patch and see if it
solves the problem?
(If you need, I can provide a prebuilt kernel with this patch.)
--=_QJyEfqlAfncoM8fB56hLDQdX5vxg3caH
Content-Type: text/plain; charset="ISO-8859-1"; name="pr59180-failedohcidetachpanic-part1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: attachment; filename="pr59180-failedohcidetachpanic-part1.patch"
# HG changeset patch
# User Taylor R Campbell <riastradh@NetBSD.org>
# Date 1742134550 0
# Sun Mar 16 14:15:50 2025 +0000
# Branch trunk
# Node ID 3f8a61dfed92d463627e7bbf859c24f0255dd5aa
# Parent 4c220187a0cad6a0dbb2b1ceeda3067ff430b505
# EXP-Topic riastradh-pr59180-failedohcidetachpanic
ohci(4): Handle failed attach when we detach.
PR port-amd64/59180: System reboots instead of shutting down
diff -r 4c220187a0ca -r 3f8a61dfed92 sys/dev/usb/ohci.c
--- a/sys/dev/usb/ohci.c Thu Mar 13 18:41:34 2025 +0000
+++ b/sys/dev/usb/ohci.c Sun Mar 16 14:15:50 2025 +0000
@@ -373,6 +373,9 @@ ohci_detach(struct ohci_softc *sc, int f
{
int rv =3D 0;
=20
+ if (!sc->sc_attached)
+ return 0;
+
if (sc->sc_child !=3D NULL)
rv =3D config_detach(sc->sc_child, flags);
=20
@@ -1092,6 +1095,7 @@ ohci_init(ohci_softc_t *sc)
DPRINTF("enabling %#jx", sc->sc_eintrs | OHCI_MIE, 0, 0, 0);
OWRITE4(sc, OHCI_INTERRUPT_ENABLE, sc->sc_eintrs | OHCI_MIE);
=20
+ sc->sc_attached =3D true;
return 0;
=20
bad5:
diff -r 4c220187a0ca -r 3f8a61dfed92 sys/dev/usb/ohcivar.h
--- a/sys/dev/usb/ohcivar.h Thu Mar 13 18:41:34 2025 +0000
+++ b/sys/dev/usb/ohcivar.h Sun Mar 16 14:15:50 2025 +0000
@@ -90,6 +90,7 @@ typedef struct ohci_softc {
bus_space_tag_t iot;
bus_space_handle_t ioh;
bus_size_t sc_size;
+ bool sc_attached;
=20
kmutex_t sc_lock;
kmutex_t sc_intr_lock;
--=_QJyEfqlAfncoM8fB56hLDQdX5vxg3caH--
From: "swaggboi@gangstalking.agency" <swaggboi@gangstalking.agency>
To: Taylor R Campbell <riastradh@NetBSD.org>
Cc: "gnats-bugs@netbsd.org" <gnats-bugs@NetBSD.org>,
"netbsd-bugs@netbsd.org" <netbsd-bugs@NetBSD.org>
Subject: Re: port-amd64/59180: System reboots instead of shutting down
Date: Sun, 16 Mar 2025 11:36:08 -0400
Let me first read the docs and try to compile a kernel; if I cannot =
figure it out we can try the prebuilt (don=E2=80=99t want to waste your =
time if it=E2=80=99s something I can DIY.) I=E2=80=99ll report back to =
you.
- Dan
> On Mar 16, 2025, at 10:38, Taylor R Campbell <riastradh@NetBSD.org> =
wrote:
>=20
> Can you please try a kernel with the attached patch and see if it
> solves the problem?
>=20
> (If you need, I can provide a prebuilt kernel with this patch.)
> <pr59180-failedohcidetachpanic-part1.patch>
From: swagg boi <swaggboi@gangstalking.agency>
To: Taylor R Campbell <riastradh@NetBSD.org>
Cc: gnats-bugs@NetBSD.org, netbsd-bugs@NetBSD.org
Subject: Re: port-amd64/59180: System reboots instead of shutting down
Date: Mon, 17 Mar 2025 18:13:31 -0400
Hello, sorry for the delay but I have my little 'build environment'
setup with the handy build.sh so I can build kernels nice and easy now.
I gave the patch a try and still get the reboot but when I set
ddb.onpanic=1 the output looks different:
http://mirror.swagg.net/NetBSD/Pictures/03172025/IMG_0863.jpeg
http://mirror.swagg.net/NetBSD/Pictures/03172025/IMG_0864.jpeg
show panic:
http://mirror.swagg.net/NetBSD/Pictures/03172025/IMG_0865.jpeg
bt:
http://mirror.swagg.net/NetBSD/Pictures/03172025/IMG_0866.jpeg
Here's my dmesg output again just in case you'd like to see that:
http://mirror.swagg.net/NetBSD/misc/dmesg_output_03172025.txt
On 3/16/25 10:38, Taylor R Campbell wrote:
> Can you please try a kernel with the attached patch and see if it
> solves the problem?
>
> (If you need, I can provide a prebuilt kernel with this patch.)
From: Taylor R Campbell <riastradh@NetBSD.org>
To: swagg boi <swaggboi@gangstalking.agency>
Cc: gnats-bugs@NetBSD.org, netbsd-bugs@NetBSD.org
Subject: Re: port-amd64/59180: System reboots instead of shutting down
Date: Sun, 30 Mar 2025 20:34:57 +0000
This is a multi-part message in MIME format.
--=_tarl6EbRoME+6V+rh/yI3C5H/YL5MT1L
> Date: Mon, 17 Mar 2025 18:13:31 -0400
> From: swagg boi <swaggboi@gangstalking.agency>
>
> Hello, sorry for the delay but I have my little 'build environment'
> setup with the handy build.sh so I can build kernels nice and easy now.
>
> I gave the patch a try and still get the reboot but when I set
> ddb.onpanic=1 the output looks different: [...]
Thanks, can you revert the patch I gave you before and try this one
instead?
--=_tarl6EbRoME+6V+rh/yI3C5H/YL5MT1L
Content-Type: text/plain; charset="ISO-8859-1"; name="pr59180-ohcidetach-v2"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: attachment; filename="pr59180-ohcidetach-v2.patch"
diff -r 63a5958e843d -r e419e0cd0295 sys/arch/arm/xscale/pxa2x0_ohci.c
--- a/sys/arch/arm/xscale/pxa2x0_ohci.c Sun Mar 30 16:22:37 2025 +0000
+++ b/sys/arch/arm/xscale/pxa2x0_ohci.c Sun Mar 30 20:18:14 2025 +0000
@@ -123,11 +123,7 @@ pxaohci_attach(device_t parent, device_t
}
=20
#if 0
- sc->sc.sc_powerhook =3D powerhook_establish(device_xname(sc->sc.sc_bus.bd=
ev),
- pxaohci_power, sc);
- if (sc->sc.sc_powerhook =3D=3D NULL) {
- aprint_error_dev(sc->sc.sc_dev->sc_bus.bdev, "cannot establish powerhook=
\n");
- }
+ pmf_device_register1(self, ohci_suspend, ohci_resume, ohci_shutdown);
#endif
=20
sc->sc.sc_child =3D config_found(self, &sc->sc.sc_bus, usbctlprint,
@@ -151,27 +147,57 @@ pxaohci_detach(device_t self, int flags)
struct pxaohci_softc *sc =3D device_private(self);
int error;
=20
- error =3D ohci_detach(&sc->sc, flags);
+ /*
+ * Detach the USB child first. Disconnects all USB devices and
+ * prevents connecting new ones.
+ */
+ error =3D config_detach_children(self, flags);
if (error)
return error;
=20
+ /*
+ * Shut down the controller and block interrupts at the device
+ * level. Once we have shut down the controller, the shutdown
+ * handler no longer needed -- deregister it from PMF.
+ * (Harmless to call ohci_shutdown more than once, so no
+ * synchronization needed.)
+ */
+ ohci_shutdown(self, 0);
#if 0
- if (sc->sc.sc_powerhook) {
- powerhook_disestablish(sc->sc.sc_powerhook);
- sc->sc.sc_powerhook =3D NULL;
- }
+ pmf_device_deregister(self);
#endif
=20
+ /*
+ * Interrupts are blocked at the device level by ohci_shutdown.
+ * Disestablish the interrupt handler. This waits for it to
+ * complete on all CPUs.
+ */
if (sc->sc_ih) {
pxa2x0_intr_disestablish(sc->sc_ih);
sc->sc_ih =3D NULL;
}
=20
+ /*
+ * Free the bus-independent ohci(4) state now that the
+ * interrupt handler has ceased to run on all CPUs.
+ */
+ ohci_detach(&sc->sc);
+
+ /*
+ * Issue a Full Host Reset to disable the host controller
+ * interface.
+ *
+ * XXX Is this necessary or is it redundant with ohci_shutdown?
+ * Should it be done in ohci_shutdown as well?
+ */
pxaohci_disable(sc);
=20
/* stop clock */
pxa2x0_clkman_config(CKEN_USBHC, 0);
=20
+ /*
+ * Unmap the registers now that we're all done with them.
+ */
if (sc->sc.sc_size) {
bus_space_unmap(sc->sc.iot, sc->sc.ioh, sc->sc.sc_size);
sc->sc.sc_size =3D 0;
diff -r 63a5958e843d -r e419e0cd0295 sys/arch/mips/ralink/ralink_ehci.c
--- a/sys/arch/mips/ralink/ralink_ehci.c Sun Mar 30 16:22:37 2025 +0000
+++ b/sys/arch/mips/ralink/ralink_ehci.c Sun Mar 30 20:18:14 2025 +0000
@@ -57,15 +57,11 @@ static void ralink_ehci_attach(device_t,
static int ralink_ehci_detach(device_t, int);
=20
CFATTACH_DECL2_NEW(ralink_ehci, sizeof(struct ralink_ehci_softc),
- ralink_ehci_match, ralink_ehci_attach, ralink_ehci_detach,
- ehci_activate, NULL, ehci_childdet);
+ ralink_ehci_match, ralink_ehci_attach, ralink_ehci_detach,
+ ehci_activate, NULL, ehci_childdet);
=20
static TAILQ_HEAD(, ralink_usb_hc) ralink_usb_alldevs =3D
- TAILQ_HEAD_INITIALIZER(ralink_usb_alldevs);
-
-#if 0
-struct usb_hc_alldevs ralink_usb_alldevs =3D TAILQ_HEAD_INITIALIZER(ralink=
_usb_alldevs);
-#endif
+ TAILQ_HEAD_INITIALIZER(ralink_usb_alldevs);
=20
/*
* ralink_ehci_match
@@ -144,6 +140,7 @@ ralink_ehci_attach(device_t parent, devi
* have lower function numbers so they should be enumerated already.
*/
int ncomp =3D 0;
+ KASSERT(KERNEL_LOCKED_P()); /* XXXSMP ralink_usb_alldevs */
TAILQ_FOREACH(ruh, &ralink_usb_alldevs, next) {
aprint_normal_dev(self, "companion %s\n",
device_xname(ruh->usb));
@@ -207,6 +204,9 @@ ralink_ehci_detach(device_t self, int fl
void
ralink_usb_hc_add(struct ralink_usb_hc *ruh, device_t usbp)
{
+
+ KASSERT(KERNEL_LOCKED_P()); /* XXXSMP ralink_usb_alldevs */
+
TAILQ_INSERT_TAIL(&ralink_usb_alldevs, ruh, next);
ruh->usb =3D usbp;
}
@@ -214,5 +214,9 @@ ralink_usb_hc_add(struct ralink_usb_hc *
void
ralink_usb_hc_rem(struct ralink_usb_hc *ruh)
{
- TAILQ_REMOVE(&ralink_usb_alldevs, ruh, next);
+
+ KASSERT(KERNEL_LOCKED_P()); /* XXXSMP ralink_usb_alldevs */
+
+ if (ruh->usb)
+ TAILQ_REMOVE(&ralink_usb_alldevs, ruh, next);
}
diff -r 63a5958e843d -r e419e0cd0295 sys/arch/mips/ralink/ralink_ohci.c
--- a/sys/arch/mips/ralink/ralink_ohci.c Sun Mar 30 16:22:37 2025 +0000
+++ b/sys/arch/mips/ralink/ralink_ohci.c Sun Mar 30 20:18:14 2025 +0000
@@ -162,28 +162,58 @@ static int
ralink_ohci_detach(device_t self, int flags)
{
struct ralink_ohci_softc *sc =3D device_private(self);
- int rv;
+ int error;
+
+ /*
+ * Detach the USB child first. Disconnects all USB devices and
+ * prevents connecting new ones.
+ */
+ error =3D config_detach_children(self, flags);
+ if (error)
+ return error;
=20
+ /*
+ * Stop listing this as a possible companion controller for
+ * ehci(4).
+ */
+#if NEHCI > 0
+ ralink_usb_hc_rem(&sc->sc_hc);
+#endif
+
+ /*
+ * Shut down the controller and block interrupts at the device
+ * level. Once we have shut down the controller, the shutdown
+ * handler no longer needed -- deregister it from PMF.
+ * (Harmless to call ohci_shutdown more than once, so no
+ * synchronization needed.)
+ */
+ ohci_shutdown(self, 0);
pmf_device_deregister(self);
=20
- rv =3D ohci_detach(&sc->sc_ohci, flags);
- if (rv !=3D 0)
- return rv;
-
+ /*
+ * Interrupts are blocked at the device level by ohci_shutdown.
+ * Disestablish the interrupt handler. This waits for it to
+ * complete on all CPUs.
+ */
if (sc->sc_ih !=3D NULL) {
ra_intr_disestablish(sc->sc_ih);
sc->sc_ih =3D NULL;
}
=20
- if (sc->sc_ohci.sc_size =3D=3D 0) {
+ /*
+ * Free the bus-independent ohci(4) state now that the
+ * interrupt handler has ceased to run on all CPUs.
+ */
+ ohci_detach(&sc->sc_ohci);
+
+ /*
+ * Unmap the registers now that we're all done with them.
+ */
+ if (sc->sc_ohci.sc_size) {
bus_space_unmap(sc->sc_ohci.iot, sc->sc_ohci.ioh,
- sc->sc_ohci.sc_size);
+ sc->sc_ohci.sc_size);
sc->sc_ohci.sc_size =3D 0;
}
=20
-#if NEHCI > 0
- ralink_usb_hc_rem(&sc->sc_hc);
-#endif
-
return 0;
}
diff -r 63a5958e843d -r e419e0cd0295 sys/dev/cardbus/ehci_cardbus.c
--- a/sys/dev/cardbus/ehci_cardbus.c Sun Mar 30 16:22:37 2025 +0000
+++ b/sys/dev/cardbus/ehci_cardbus.c Sun Mar 30 20:18:14 2025 +0000
@@ -82,7 +82,7 @@ CFATTACH_DECL_NEW(ehci_cardbus, sizeof(s
ehci_activate);
=20
static TAILQ_HEAD(, usb_cardbus) ehci_cardbus_alldevs =3D
- TAILQ_HEAD_INITIALIZER(ehci_cardbus_alldevs);
+ TAILQ_HEAD_INITIALIZER(ehci_cardbus_alldevs);
=20
int
ehci_cardbus_match(device_t parent, cfdata_t match, void *aux)
@@ -175,6 +175,7 @@ ehci_cardbus_attach(device_t parent, dev
* have lower function numbers so they should be enumerated already.
*/
ncomp =3D 0;
+ KASSERT(KERNEL_LOCKED_P()); /* XXXSMP ehci_cardbus_alldevs */
TAILQ_FOREACH(up, &ehci_cardbus_alldevs, next) {
if (up->bus =3D=3D ca->ca_bus) {
DPRINTF(("ehci_cardbus_attach: companion %s\n",
@@ -232,6 +233,9 @@ void
usb_cardbus_add(struct usb_cardbus *up, struct cardbus_attach_args *ca,
device_t bu)
{
+
+ KASSERT(KERNEL_LOCKED_P()); /* XXXSMP ehci_cardbus_alldevs */
+
TAILQ_INSERT_TAIL(&ehci_cardbus_alldevs, up, next);
up->bus =3D ca->ca_bus;
up->function =3D ca->ca_function;
@@ -241,5 +245,9 @@ usb_cardbus_add(struct usb_cardbus *up,=20
void
usb_cardbus_rem(struct usb_cardbus *up)
{
- TAILQ_REMOVE(&ehci_cardbus_alldevs, up, next);
+
+ KASSERT(KERNEL_LOCKED_P()); /* XXXSMP ehci_cardbus_alldevs */
+
+ if (up->usb)
+ TAILQ_REMOVE(&ehci_cardbus_alldevs, up, next);
}
diff -r 63a5958e843d -r e419e0cd0295 sys/dev/cardbus/ohci_cardbus.c
--- a/sys/dev/cardbus/ohci_cardbus.c Sun Mar 30 16:22:37 2025 +0000
+++ b/sys/dev/cardbus/ohci_cardbus.c Sun Mar 30 20:18:14 2025 +0000
@@ -175,22 +175,58 @@ ohci_cardbus_detach(device_t self, int f
{
struct ohci_cardbus_softc *sc =3D device_private(self);
struct cardbus_devfunc *ct =3D sc->sc_ct;
- int rv;
+ int error;
+
+ /*
+ * Detach the USB child first. Disconnects all USB devices and
+ * prevents connecting new ones.
+ */
+ error =3D config_detach_children(self, flags);
+ if (error)
+ return error;
=20
- rv =3D ohci_detach(&sc->sc, flags);
- if (rv)
- return rv;
+ /*
+ * Stop listing this as a possible companion controller for
+ * ehci(4).
+ */
+#if NEHCI_CARDBUS > 0
+ usb_cardbus_rem(&sc->sc_cardbus);
+#endif
+
+ /*
+ * Shut down the controller and block interrupts at the device
+ * level. Once we have shut down the controller, the shutdown
+ * handler no longer needed -- deregister it from PMF.
+ * (Harmless to call ohci_shutdown more than once, so no
+ * synchronization needed.)
+ */
+ ohci_shutdown(self, 0);
+ pmf_device_deregister(self);
+
+ /*
+ * Interrupts are blocked at the device level by ohci_shutdown.
+ * Disestablish the interrupt handler. This waits for it to
+ * complete on all CPUs.
+ */
if (sc->sc_ih !=3D NULL) {
Cardbus_intr_disestablish(ct, sc->sc_ih);
sc->sc_ih =3D NULL;
}
+
+ /*
+ * Free the bus-independent ohci(4) state now that the
+ * interrupt handler has ceased to run on all CPUs.
+ */
+ ohci_detach(&sc->sc);
+
+ /*
+ * Unmap the registers now that we're all done with them.
+ */
if (sc->sc.sc_size) {
Cardbus_mapreg_unmap(ct, PCI_CBMEM, sc->sc.iot,
sc->sc.ioh, sc->sc.sc_size);
sc->sc.sc_size =3D 0;
}
-#if NEHCI_CARDBUS > 0
- usb_cardbus_rem(&sc->sc_cardbus);
-#endif
+
return 0;
}
diff -r 63a5958e843d -r e419e0cd0295 sys/dev/pci/ehci_pci.c
--- a/sys/dev/pci/ehci_pci.c Sun Mar 30 16:22:37 2025 +0000
+++ b/sys/dev/pci/ehci_pci.c Sun Mar 30 20:18:14 2025 +0000
@@ -243,6 +243,7 @@ ehci_pci_attach(device_t parent, device_
const u_int maxncomp =3D EHCI_HCS_N_CC(EREAD4(&sc->sc, EHCI_HCSPARAMS));
KASSERT(maxncomp <=3D EHCI_COMPANION_MAX);
ncomp =3D 0;
+ KASSERT(KERNEL_LOCKED_P()); /* XXXSMP ehci_pci_alldevs */
TAILQ_FOREACH(up, &ehci_pci_alldevs, next) {
if (up->bus =3D=3D pa->pa_bus && up->device =3D=3D pa->pa_device &&
!up->claimed) {
diff -r 63a5958e843d -r e419e0cd0295 sys/dev/pci/ohci_pci.c
--- a/sys/dev/pci/ohci_pci.c Sun Mar 30 16:22:37 2025 +0000
+++ b/sys/dev/pci/ohci_pci.c Sun Mar 30 20:18:14 2025 +0000
@@ -187,33 +187,58 @@ static int
ohci_pci_detach(device_t self, int flags)
{
struct ohci_pci_softc *sc =3D device_private(self);
- int rv;
+ int error;
+
+ /*
+ * Detach the USB child first. Disconnects all USB devices and
+ * prevents connecting new ones.
+ */
+ error =3D config_detach_children(self, flags);
+ if (error)
+ return error;
=20
- rv =3D ohci_detach(&sc->sc, flags);
- if (rv)
- return rv;
+ /*
+ * Stop listing this as a possible companion controller for
+ * ehci(4).
+ */
+#if NEHCI > 0
+ usb_pci_rem(&sc->sc_pci);
+#endif
=20
+ /*
+ * Shut down the controller and block interrupts at the device
+ * level. Once we have shut down the controller, the shutdown
+ * handler no longer needed -- deregister it from PMF.
+ * (Harmless to call ohci_shutdown more than once, so no
+ * synchronization needed.)
+ */
+ ohci_shutdown(self, 0);
pmf_device_deregister(self);
=20
- ohci_shutdown(self, flags);
-
- if (sc->sc.sc_size) {
- /* Disable interrupts, so we don't get any spurious ones. */
- bus_space_write_4(sc->sc.iot, sc->sc.ioh,
- OHCI_INTERRUPT_DISABLE, OHCI_ALL_INTRS);
- }
-
+ /*
+ * Interrupts are blocked at the device level by ohci_shutdown.
+ * Disestablish the interrupt handler. This waits for it to
+ * complete on all CPUs.
+ */
if (sc->sc_ih !=3D NULL) {
pci_intr_disestablish(sc->sc_pc, sc->sc_ih);
sc->sc_ih =3D NULL;
}
+
+ /*
+ * Free the bus-independent ohci(4) state now that the
+ * interrupt handler has ceased to run on all CPUs.
+ */
+ ohci_detach(&sc->sc);
+
+ /*
+ * Unmap the registers now that we're all done with them.
+ */
if (sc->sc.sc_size) {
bus_space_unmap(sc->sc.iot, sc->sc.ioh, sc->sc.sc_size);
sc->sc.sc_size =3D 0;
}
-#if NEHCI > 0
- usb_pci_rem(&sc->sc_pci);
-#endif
+
return 0;
}
=20
diff -r 63a5958e843d -r e419e0cd0295 sys/dev/pci/usb_pci.c
--- a/sys/dev/pci/usb_pci.c Sun Mar 30 16:22:37 2025 +0000
+++ b/sys/dev/pci/usb_pci.c Sun Mar 30 20:18:14 2025 +0000
@@ -33,28 +33,31 @@
__KERNEL_RCSID(0, "$NetBSD: usb_pci.c,v 1.7 2008/04/28 20:23:55 martin Exp=
$");
=20
#include <sys/param.h>
+#include <sys/kernel.h>
+#include <sys/proc.h>
+#include <sys/queue.h>
#include <sys/systm.h>
-#include <sys/kernel.h>
-#include <sys/queue.h>
-#include <sys/proc.h>
=20
#include <dev/pci/pcivar.h>
#include <dev/pci/usb_pci.h>
=20
#include <dev/usb/usb.h>
+#include <dev/usb/usb_mem.h>
#include <dev/usb/usbdi.h>
#include <dev/usb/usbdivar.h>
-#include <dev/usb/usb_mem.h>
=20
#include <dev/usb/ehcireg.h>
#include <dev/usb/ehcivar.h>
=20
struct usb_pci_alldevs ehci_pci_alldevs =3D
- TAILQ_HEAD_INITIALIZER(ehci_pci_alldevs);
+ TAILQ_HEAD_INITIALIZER(ehci_pci_alldevs);
=20
void
usb_pci_add(struct usb_pci *up, struct pci_attach_args *pa, device_t bu)
{
+
+ KASSERT(KERNEL_LOCKED_P()); /* XXXSMP ehci_pci_alldevs */
+
TAILQ_INSERT_TAIL(&ehci_pci_alldevs, up, next);
up->bus =3D pa->pa_bus;
up->device =3D pa->pa_device;
@@ -65,5 +68,9 @@ usb_pci_add(struct usb_pci *up, struct p
void
usb_pci_rem(struct usb_pci *up)
{
- TAILQ_REMOVE(&ehci_pci_alldevs, up, next);
+
+ KASSERT(KERNEL_LOCKED_P()); /* XXXSMP ehci_pci_alldevs */
+
+ if (up->usb)
+ TAILQ_REMOVE(&ehci_pci_alldevs, up, next);
}
diff -r 63a5958e843d -r e419e0cd0295 sys/dev/usb/ohci.c
--- a/sys/dev/usb/ohci.c Sun Mar 30 16:22:37 2025 +0000
+++ b/sys/dev/usb/ohci.c Sun Mar 30 20:18:14 2025 +0000
@@ -368,16 +368,14 @@ ohci_childdet(device_t self, device_t ch
sc->sc_child =3D NULL;
}
=20
-int
-ohci_detach(struct ohci_softc *sc, int flags)
+void
+ohci_detach(struct ohci_softc *sc)
{
- int rv =3D 0;
-
- if (sc->sc_child !=3D NULL)
- rv =3D config_detach(sc->sc_child, flags);
-
- if (rv !=3D 0)
- return rv;
+
+ KASSERT(sc->sc_child =3D=3D NULL);
+
+ if (!sc->sc_attached)
+ return;
=20
softint_disestablish(sc->sc_rhsc_si);
=20
@@ -391,8 +389,6 @@ ohci_detach(struct ohci_softc *sc, int f
usb_freemem(&sc->sc_hccadma);
pool_cache_destroy(sc->sc_xferpool);
cv_destroy(&sc->sc_abort_cv);
-
- return rv;
}
=20
ohci_soft_ed_t *
@@ -1092,6 +1088,7 @@ ohci_init(ohci_softc_t *sc)
DPRINTF("enabling %#jx", sc->sc_eintrs | OHCI_MIE, 0, 0, 0);
OWRITE4(sc, OHCI_INTERRUPT_ENABLE, sc->sc_eintrs | OHCI_MIE);
=20
+ sc->sc_attached =3D true;
return 0;
=20
bad5:
@@ -1166,6 +1163,9 @@ ohci_shutdown(device_t self, int flags)
=20
OHCIHIST_FUNC(); OHCIHIST_CALLED();
=20
+ if (!sc->sc_attached)
+ return true;
+
DPRINTF("stopping the HC", 0, 0, 0, 0);
OWRITE4(sc, OHCI_INTERRUPT_DISABLE, OHCI_ALL_INTRS);
OWRITE4(sc, OHCI_CONTROL, OHCI_SET_HCFS(OHCI_HCFS_RESET));
diff -r 63a5958e843d -r e419e0cd0295 sys/dev/usb/ohcivar.h
--- a/sys/dev/usb/ohcivar.h Sun Mar 30 16:22:37 2025 +0000
+++ b/sys/dev/usb/ohcivar.h Sun Mar 30 20:18:14 2025 +0000
@@ -90,6 +90,7 @@ typedef struct ohci_softc {
bus_space_tag_t iot;
bus_space_handle_t ioh;
bus_size_t sc_size;
+ bool sc_attached;
=20
kmutex_t sc_lock;
kmutex_t sc_intr_lock;
@@ -172,7 +173,7 @@ struct ohci_xfer {
=20
int ohci_init(ohci_softc_t *);
int ohci_intr(void *);
-int ohci_detach(ohci_softc_t *, int);
+void ohci_detach(ohci_softc_t *);
bool ohci_shutdown(device_t, int);
void ohci_childdet(device_t, device_t);
int ohci_activate(device_t, enum devact);
diff -r 63a5958e843d -r e419e0cd0295 sys/dev/usb/usbdivar.h
--- a/sys/dev/usb/usbdivar.h Sun Mar 30 16:22:37 2025 +0000
+++ b/sys/dev/usb/usbdivar.h Sun Mar 30 20:18:14 2025 +0000
@@ -74,9 +74,11 @@
*
*/
=20
+#include <sys/bus.h>
#include <sys/callout.h>
#include <sys/mutex.h>
-#include <sys/bus.h>
+
+#include <dev/usb/usbdi.h>
=20
/* From usb_mem.h */
struct usb_dma_block;
--=_tarl6EbRoME+6V+rh/yI3C5H/YL5MT1L--
From: Daniel Bowling <swaggboi@gangstalking.agency>
To: Taylor R Campbell <riastradh@NetBSD.org>
Cc: gnats-bugs@NetBSD.org, netbsd-bugs@NetBSD.org
Subject: Re: port-amd64/59180: System reboots instead of shutting down
Date: Sun, 30 Mar 2025 18:34:42 -0400
This patch appears to work; thank you. I applied the patch and did both
a reboot and a normal shutdown with ddb.onpanic=1 set without any issues.
On 3/30/25 16:34, Taylor R Campbell wrote:
>> Date: Mon, 17 Mar 2025 18:13:31 -0400
>> From: swagg boi <swaggboi@gangstalking.agency>
>>
>> Hello, sorry for the delay but I have my little 'build environment'
>> setup with the handy build.sh so I can build kernels nice and easy now.
>>
>> I gave the patch a try and still get the reboot but when I set
>> ddb.onpanic=1 the output looks different: [...]
> Thanks, can you revert the patch I gave you before and try this one
> instead?
From: "Taylor R Campbell" <riastradh@netbsd.org>
To: gnats-bugs@gnats.NetBSD.org
Cc:
Subject: PR/59180 CVS commit: src/sys/dev/usb
Date: Mon, 31 Mar 2025 14:45:14 +0000
Module Name: src
Committed By: riastradh
Date: Mon Mar 31 14:45:14 UTC 2025
Modified Files:
src/sys/dev/usb: usbdivar.h
Log Message:
usbdivar.h: Add missing usbdi.h include for usbd_status &c.
Sort includes while here.
Prompted by:
PR port-amd64/59180: System reboots instead of shutting down
To generate a diff of this commit:
cvs rdiff -u -r1.138 -r1.139 src/sys/dev/usb/usbdivar.h
Please note that diffs are not public domain; they are subject to the
copyright notices on the relevant files.
From: "Taylor R Campbell" <riastradh@netbsd.org>
To: gnats-bugs@gnats.NetBSD.org
Cc:
Subject: PR/59180 CVS commit: src/sys
Date: Mon, 31 Mar 2025 14:45:36 +0000
Module Name: src
Committed By: riastradh
Date: Mon Mar 31 14:45:35 UTC 2025
Modified Files:
src/sys/arch/mips/ralink: ralink_ehci.c
src/sys/dev/cardbus: ehci_cardbus.c
src/sys/dev/pci: ehci_pci.c usb_pci.c
Log Message:
ehci(4): Sprinkle KERNEL_LOCKED_P and KNF on usb_*_add/rem.
Mark with XXXSMP comments to be fixed later without the kernel lock.
No functional change intended: callers generally do this from
autoconf *_attach/detach routines, which run kernel-locked anyway
(for the moment).
Cleanup in preparation for:
PR port-amd64/59180: System reboots instead of shutting down
To generate a diff of this commit:
cvs rdiff -u -r1.9 -r1.10 src/sys/arch/mips/ralink/ralink_ehci.c
cvs rdiff -u -r1.37 -r1.38 src/sys/dev/cardbus/ehci_cardbus.c
cvs rdiff -u -r1.77 -r1.78 src/sys/dev/pci/ehci_pci.c
cvs rdiff -u -r1.7 -r1.8 src/sys/dev/pci/usb_pci.c
Please note that diffs are not public domain; they are subject to the
copyright notices on the relevant files.
From: "Taylor R Campbell" <riastradh@netbsd.org>
To: gnats-bugs@gnats.NetBSD.org
Cc:
Subject: PR/59180 CVS commit: src/sys
Date: Mon, 31 Mar 2025 14:45:57 +0000
Module Name: src
Committed By: riastradh
Date: Mon Mar 31 14:45:57 UTC 2025
Modified Files:
src/sys/arch/mips/ralink: ralink_ehci.c
src/sys/dev/cardbus: ehci_cardbus.c
src/sys/dev/pci: usb_pci.c
Log Message:
ehci(4): Make usb_*_rem work as noop on zero-initialized input.
This way, if some *hci_attach function fails before usb_*_add, the
caller need not maintain boolean state to decide whether to call
usb_*_rem on detach -- it'll just work without extra effort.
Cleanup in preparation for:
PR port-amd64/59180: System reboots instead of shutting down
To generate a diff of this commit:
cvs rdiff -u -r1.10 -r1.11 src/sys/arch/mips/ralink/ralink_ehci.c
cvs rdiff -u -r1.38 -r1.39 src/sys/dev/cardbus/ehci_cardbus.c
cvs rdiff -u -r1.8 -r1.9 src/sys/dev/pci/usb_pci.c
Please note that diffs are not public domain; they are subject to the
copyright notices on the relevant files.
From: "Taylor R Campbell" <riastradh@netbsd.org>
To: gnats-bugs@gnats.NetBSD.org
Cc:
Subject: PR/59180 CVS commit: src/sys
Date: Mon, 31 Mar 2025 14:46:42 +0000
Module Name: src
Committed By: riastradh
Date: Mon Mar 31 14:46:42 UTC 2025
Modified Files:
src/sys/arch/arm/xscale: pxa2x0_ohci.c
src/sys/arch/mips/ralink: ralink_ohci.c
src/sys/dev/cardbus: ohci_cardbus.c
src/sys/dev/pci: ohci_pci.c
src/sys/dev/usb: ohci.c ohcivar.h
Log Message:
ohci(4): Rework detach logic and justify the ordering.
Handle failed attach when we detach.
This changes the signature of the ohci_detach function, but it is
only ever used by statically linked ohci bus attachments, never by
modules so far, so no kernel revbump.
PR port-amd64/59180: System reboots instead of shutting down
To generate a diff of this commit:
cvs rdiff -u -r1.13 -r1.14 src/sys/arch/arm/xscale/pxa2x0_ohci.c
cvs rdiff -u -r1.7 -r1.8 src/sys/arch/mips/ralink/ralink_ohci.c
cvs rdiff -u -r1.46 -r1.47 src/sys/dev/cardbus/ohci_cardbus.c
cvs rdiff -u -r1.59 -r1.60 src/sys/dev/pci/ohci_pci.c
cvs rdiff -u -r1.329 -r1.330 src/sys/dev/usb/ohci.c
cvs rdiff -u -r1.63 -r1.64 src/sys/dev/usb/ohcivar.h
Please note that diffs are not public domain; they are subject to the
copyright notices on the relevant files.
From: "Taylor R Campbell" <riastradh@netbsd.org>
To: gnats-bugs@gnats.NetBSD.org
Cc:
Subject: PR/59180 CVS commit: src/sys/dev/pci
Date: Mon, 31 Mar 2025 14:48:50 +0000
Module Name: src
Committed By: riastradh
Date: Mon Mar 31 14:48:50 UTC 2025
Modified Files:
src/sys/dev/pci: ohci_pci.c
Log Message:
ohci@pci: Don't detach on shutdown.
If there's some USB device that requires detach on shutdown (which
would be surprising, because USB devices should generally be safe to
yank at any time, and _storage_ drivers such as ld(4) or sd(4) should
already be synced at shutdown some other way) then that driver should
have DVF_DETACH_SHUTDOWN -- not the entire HCI.
PR port-amd64/59180: System reboots instead of shutting down
PR kern/59179: DVF_DETACH_SHUTDOWN: audit drivers for unnecessary use
To generate a diff of this commit:
cvs rdiff -u -r1.60 -r1.61 src/sys/dev/pci/ohci_pci.c
Please note that diffs are not public domain; they are subject to the
copyright notices on the relevant files.
(Contact us)
$NetBSD: query-full-pr,v 1.47 2022/09/11 19:34:41 kim Exp $
$NetBSD: gnats_config.sh,v 1.9 2014/08/02 14:16:04 spz Exp $
Copyright © 1994-2025
The NetBSD Foundation, Inc. ALL RIGHTS RESERVED.
Home