Date: Mon, 14 Apr 2025 16:10:20 GMT From: Colin Percival <cperciva@FreeBSD.org> To: src-committers@FreeBSD.org, dev-commits-src-all@FreeBSD.org, dev-commits-src-branches@FreeBSD.org Subject: git: 43933d89a325 - stable/14 - acpi_pci: Add quirk for DELAY-after-EJ0 Message-ID: <202504141610.53EGAKuI066273@gitrepo.freebsd.org>
next in thread | raw e-mail | index | archive | help
The branch stable/14 has been updated by cperciva: URL: https://cgit.FreeBSD.org/src/commit/?id=43933d89a3256f33d0eb29473ac56b3e0e7b9318 commit 43933d89a3256f33d0eb29473ac56b3e0e7b9318 Author: Colin Percival <cperciva@FreeBSD.org> AuthorDate: 2025-03-06 05:22:33 +0000 Commit: Colin Percival <cperciva@FreeBSD.org> CommitDate: 2025-04-14 16:09:49 +0000 acpi_pci: Add quirk for DELAY-after-EJ0 On some EC2 instances, there is a race between removing a device from the system and making the PCI bus stop reporting the presence of the device. As a result, a PCI BUS_RESCAN performed immediately after the _EJ0 method returns "sees" the device which is being ejected, which then causes problems later (e.g. we won't recognize a new device being plugged into that slot because we never knew it was vacant). On other operating systems the bus is synchronously marked as needing to be rescanned but the rescan does not occur until O(1) seconds later. Create a new ACPI_Q_DELAY_BEFORE_EJECT_RESCAN quirk and set it in EC2 AMIs, and add a 10 ms DELAY between _EJ0 and BUS_RESCAN when tht quirk is set. Reviewed by: jhb MFC after: 1 month Sponsored by: Amazon Differential Revision: https://reviews.freebsd.org/D49252 (cherry picked from commit 55c3348ed78fb1d0891e8bb51a8948f95da3560b) --- release/tools/ec2.conf | 5 +++-- sys/dev/acpica/acpi_pci.c | 2 ++ sys/dev/acpica/acpivar.h | 3 +++ 3 files changed, 8 insertions(+), 2 deletions(-) diff --git a/release/tools/ec2.conf b/release/tools/ec2.conf index 21ea56d109ca..3675736ab986 100644 --- a/release/tools/ec2.conf +++ b/release/tools/ec2.conf @@ -68,8 +68,9 @@ ec2_common() { # (in fact the PL061 has no pullup/pulldown resistors). Graviton 1 # through Graviton 3 have non-functional PCI _EJ0 and need a value # written to the PCI power status register in order to eject a - # device. - echo 'debug.acpi.quirks="24"' >> ${DESTDIR}/boot/loader.conf + # device. EC2 instances with PCI (not PCIe) buses need a short + # delay before rescanning upon device detach. + echo 'debug.acpi.quirks="56"' >> ${DESTDIR}/boot/loader.conf # Load the kernel module for the Amazon "Elastic Network Adapter" echo 'if_ena_load="YES"' >> ${DESTDIR}/boot/loader.conf diff --git a/sys/dev/acpica/acpi_pci.c b/sys/dev/acpica/acpi_pci.c index 97704111839b..b7a2bf70b4e0 100644 --- a/sys/dev/acpica/acpi_pci.c +++ b/sys/dev/acpica/acpi_pci.c @@ -432,6 +432,8 @@ acpi_pci_device_notify_handler(ACPI_HANDLE h, UINT32 notify, void *context) acpi_name(h), AcpiFormatException(status)); return; } + if (acpi_quirks & ACPI_Q_DELAY_BEFORE_EJECT_RESCAN) + DELAY(10 * 1000); BUS_RESCAN(dev); bus_topo_unlock(); break; diff --git a/sys/dev/acpica/acpivar.h b/sys/dev/acpica/acpivar.h index 830764434f48..d35504127c9c 100644 --- a/sys/dev/acpica/acpivar.h +++ b/sys/dev/acpica/acpivar.h @@ -232,6 +232,8 @@ extern struct mtx acpi_mutex; * as "PullUp" and they should be treated as "NoPull" instead. * ACPI_Q_CLEAR_PME_ON_DETACH: Specifies that PCIM_PSTAT_(PME & ~PMEENABLE) * should be written to the power status register as part of ACPI Eject. + * ACPI_Q_DELAY_BEFORE_EJECT_RESCAN: Specifies that we need a short (10ms) + * delay after _EJ0 returns before rescanning the PCI bus. */ extern int acpi_quirks; #define ACPI_Q_OK 0 @@ -240,6 +242,7 @@ extern int acpi_quirks; #define ACPI_Q_MADT_IRQ0 (1 << 2) #define ACPI_Q_AEI_NOPULL (1 << 3) #define ACPI_Q_CLEAR_PME_ON_DETACH (1 << 4) +#define ACPI_Q_DELAY_BEFORE_EJECT_RESCAN (1 << 5) #if defined(__amd64__) || defined(__i386__) /*
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?202504141610.53EGAKuI066273>