April 2025 - Linux-stable-mirror

[PATCH net v1 0/2] address EEE regressions on KSZ switches since v6.9 (v6.14+)

by Oleksij Rempel

This patch series addresses a regression in Energy Efficient Ethernet (EEE) handling for KSZ switches with integrated PHYs, introduced in kernel v6.9 by commit fe0d4fd9285e ("net: phy: Keep track of EEE configuration"). The first patch updates the DSA driver to allow phylink to properly manage PHY EEE configuration. Since integrated PHYs handle LPI internally and ports without integrated PHYs do not document MAC-level LPI support, dummy MAC LPI callbacks are provided. The second patch removes outdated EEE workarounds from the micrel PHY driver, as they are no longer needed with correct phylink handling. This series addresses the regression for mainline and kernels starting from v6.14. It is not easily possible to fully fix older kernels due to missing infrastructure changes. Tested on KSZ9893 hardware. Oleksij Rempel (2): net: dsa: microchip: let phylink manage PHY EEE configuration on KSZ switches net: phy: micrel: remove KSZ9477 EEE quirks now handled by phylink drivers/net/dsa/microchip/ksz_common.c | 97 ++++++++++++++++++-------- drivers/net/phy/micrel.c | 7 -- include/linux/micrel_phy.h | 1 - 3 files changed, 69 insertions(+), 36 deletions(-) -- 2.39.5

2 months

2
5
0 0

Re: Patch "x86/Kconfig: Make CONFIG_PCI_CNB20LE_QUIRK depend on X86_32" has been added to the 6.14-stable tree

by Mateusz Jończyk

Dnia 29 kwietnia 2025 03:41:48 CEST, Sasha Levin <sashal(a)kernel.org> napisał/a: >This is a note to let you know that I've just added the patch titled > > x86/Kconfig: Make CONFIG_PCI_CNB20LE_QUIRK depend on X86_32 > >to the 6.14-stable tree which can be found at: > http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=sum… > >The filename of the patch is: > x86-kconfig-make-config_pci_cnb20le_quirk-depend-on-.patch >and it can be found in the queue-6.14 subdirectory. > >If you, or anyone else, feels it should not be added to the stable tree, >please let <stable(a)vger.kernel.org> know about it. Hello, I'd like to ask that this patch be dropped from the stable queues (for 6.14 and earlier kernels). It does not fix anything important, it is just for convenience - to hide this one option from amd64 kernel Kconfig. Greetings, Mateusz > > > >commit 8cc03c8c367ced0228a5fbaec8c02274f11b2a38 >Author: Mateusz Jończyk <mat.jonczyk(a)o2.pl> >Date: Fri Mar 21 21:48:48 2025 +0100 > > x86/Kconfig: Make CONFIG_PCI_CNB20LE_QUIRK depend on X86_32 > > [ Upstream commit d9f87802676bb23b9425aea8ad95c76ad9b50c6e ] > > I was unable to find a good description of the ServerWorks CNB20LE > chipset. However, it was probably exclusively used with the Pentium III > processor (this CPU model was used in all references to it that I > found where the CPU model was provided: dmesgs in [1] and [2]; > [3] page 2; [4]-[7]). > > As is widely known, the Pentium III processor did not support the 64-bit > mode, support for which was introduced by Intel a couple of years later. > So it is safe to assume that no systems with the CNB20LE chipset have > amd64 and the CONFIG_PCI_CNB20LE_QUIRK may now depend on X86_32. > > Additionally, I have determined that most computers with the CNB20LE > chipset did have ACPI support and this driver was inactive on them. > I have submitted a patch to remove this driver, but it was met with > resistance [8]. > > [1] Jim Studt, Re: Problem with ServerWorks CNB20LE and lost interrupts > Linux Kernel Mailing List, https://lkml.org/lkml/2002/1/11/111 > > [2] RedHat Bug 665109 - e100 problems on old Compaq Proliant DL320 > https://bugzilla.redhat.com/show_bug.cgi?id=665109 > > [3] R. Hughes-Jones, S. Dallison, G. Fairey, Performance Measurements on > Gigabit Ethernet NICs and Server Quality Motherboards, > http://datatag.web.cern.ch/papers/pfldnet2003-rhj.doc > > [4] "Hardware for Linux", > Probe #d6b5151873 of Intel STL2-bd A28808-302 Desktop Computer (STL2) > https://linux-hardware.org/?probe=d6b5151873 > > [5] "Hardware for Linux", Probe #0b5d843f10 of Compaq ProLiant DL380 > https://linux-hardware.org/?probe=0b5d843f10 > > [6] Ubuntu Forums, Dell Poweredge 2400 - Adaptec SCSI Bus AIC-7880 > https://ubuntuforums.org/showthread.php?t=1689552 > > [7] Ira W. Snyder, "BISECTED: 2.6.35 (and -git) fail to boot: APIC problems" > https://lkml.org/lkml/2010/8/13/220 > > [8] Bjorn Helgaas, "Re: [PATCH] x86/pci: drop ServerWorks / Broadcom > CNB20LE PCI host bridge driver" > https://lore.kernel.org/lkml/20220318165535.GA840063@bhelgaas/T/ > > Signed-off-by: Mateusz Jończyk <mat.jonczyk(a)o2.pl> > Signed-off-by: David Heideberg <david(a)ixit.cz> > Signed-off-by: Ingo Molnar <mingo(a)kernel.org> > Cc: "H. Peter Anvin" <hpa(a)zytor.com> > Cc: Linus Torvalds <torvalds(a)linux-foundation.org> > Link: https://lore.kernel.org/r/20250321-x86_x2apic-v3-6-b0cbaa6fa338@ixit.cz > Signed-off-by: Sasha Levin <sashal(a)kernel.org> > >diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig >index aeb95b6e55369..de47e7c435679 100644 >--- a/arch/x86/Kconfig >+++ b/arch/x86/Kconfig >@@ -2981,13 +2981,21 @@ config MMCONF_FAM10H > depends on X86_64 && PCI_MMCONFIG && ACPI > > config PCI_CNB20LE_QUIRK >- bool "Read CNB20LE Host Bridge Windows" if EXPERT >- depends on PCI >+ bool "Read PCI host bridge windows from the CNB20LE chipset" if EXPERT >+ depends on X86_32 && PCI > help > Read the PCI windows out of the CNB20LE host bridge. This allows > PCI hotplug to work on systems with the CNB20LE chipset which do > not have ACPI. > >+ The ServerWorks (later Broadcom) CNB20LE was a chipset designed >+ most probably only for Pentium III. >+ >+ To find out if you have such a chipset, search for a PCI device with >+ 1166:0009 PCI IDs, for example by executing >+ lspci -nn | grep '1166:0009' >+ The code is inactive if there is none. >+ > There's no public spec for this chipset, and this functionality > is known to be incomplete. >

2 months

2
1
0 0

Re: Patch "hardening: Disable GCC randstruct for COMPILE_TEST" has been added to the 6.14-stable tree

by Kees Cook

On April 28, 2025 6:48:46 PM PDT, Sasha Levin <sashal(a)kernel.org> wrote: >This is a note to let you know that I've just added the patch titled > > hardening: Disable GCC randstruct for COMPILE_TEST Please don't backport this to any stable kernels. There is already a fix in -next and the problem only exists due to a 6.15 landlock change. -- Kees Cook

2 months

2
1
0 0

FAILED: patch "[PATCH] serial: msm: Configure correct working mode before starting" failed to apply to 5.15-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 5.15-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. To reproduce the conflict and resubmit, you may use the following commands: git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-5.15.y git checkout FETCH_HEAD git cherry-pick -x 7094832b5ac861b0bd7ed8866c93cb15ef619996 # <resolve conflicts, build, test, etc.> git commit -s git send-email --to '<stable(a)vger.kernel.org>' --in-reply-to '2025042947-baffling-scrawny-24d8@gregkh' --subject-prefix 'PATCH 5.15.y' HEAD^.. Possible dependencies: thanks, greg k-h ------------------ original commit in Linus's tree ------------------ From 7094832b5ac861b0bd7ed8866c93cb15ef619996 Mon Sep 17 00:00:00 2001 From: Stephan Gerhold <stephan.gerhold(a)linaro.org> Date: Tue, 8 Apr 2025 19:22:47 +0200 Subject: [PATCH] serial: msm: Configure correct working mode before starting earlycon The MSM UART DM controller supports different working modes, e.g. DMA or the "single-character mode", where all reads/writes operate on a single character rather than 4 chars (32-bit) at once. When using earlycon, __msm_console_write() always writes 4 characters at a time, but we don't know which mode the bootloader was using and we don't set the mode either. This causes garbled output if the bootloader was using the single-character mode, because only every 4th character appears in the serial console, e.g. "[ 00oni pi 000xf0[ 00i s 5rm9(l)l s 1 1 SPMTA 7:C 5[ 00A ade k d[ 00ano:ameoi .Q1B[ 00ac _idaM00080oo'" If the bootloader was using the DMA ("DM") mode, output would likely fail entirely. Later, when the full serial driver probes, the port is re-initialized and output works as expected. Fix this also for earlycon by clearing the DMEN register and reset+re-enable the transmitter to apply the change. This ensures the transmitter is in the expected state before writing any output. Cc: stable <stable(a)kernel.org> Fixes: 0efe72963409 ("tty: serial: msm: Add earlycon support") Signed-off-by: Stephan Gerhold <stephan.gerhold(a)linaro.org> Reviewed-by: Neil Armstrong <neil.armstrong(a)linaro.org> Link: https://lore.kernel.org/r/20250408-msm-serial-earlycon-v1-1-429080127530@li… Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> diff --git a/drivers/tty/serial/msm_serial.c b/drivers/tty/serial/msm_serial.c index 1b137e068444..3449945493ce 100644 --- a/drivers/tty/serial/msm_serial.c +++ b/drivers/tty/serial/msm_serial.c @@ -1746,6 +1746,12 @@ msm_serial_early_console_setup_dm(struct earlycon_device *device, if (!device->port.membase) return -ENODEV; + /* Disable DM / single-character modes */ + msm_write(&device->port, 0, UARTDM_DMEN); + msm_write(&device->port, MSM_UART_CR_CMD_RESET_RX, MSM_UART_CR); + msm_write(&device->port, MSM_UART_CR_CMD_RESET_TX, MSM_UART_CR); + msm_write(&device->port, MSM_UART_CR_TX_ENABLE, MSM_UART_CR); + device->con->write = msm_serial_early_write_dm; return 0; }

2 months

1
0
0 0

FAILED: patch "[PATCH] serial: msm: Configure correct working mode before starting" failed to apply to 5.10-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 5.10-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. To reproduce the conflict and resubmit, you may use the following commands: git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-5.10.y git checkout FETCH_HEAD git cherry-pick -x 7094832b5ac861b0bd7ed8866c93cb15ef619996 # <resolve conflicts, build, test, etc.> git commit -s git send-email --to '<stable(a)vger.kernel.org>' --in-reply-to '2025042923-flagman-payroll-a044@gregkh' --subject-prefix 'PATCH 5.10.y' HEAD^.. Possible dependencies: thanks, greg k-h ------------------ original commit in Linus's tree ------------------ From 7094832b5ac861b0bd7ed8866c93cb15ef619996 Mon Sep 17 00:00:00 2001 From: Stephan Gerhold <stephan.gerhold(a)linaro.org> Date: Tue, 8 Apr 2025 19:22:47 +0200 Subject: [PATCH] serial: msm: Configure correct working mode before starting earlycon The MSM UART DM controller supports different working modes, e.g. DMA or the "single-character mode", where all reads/writes operate on a single character rather than 4 chars (32-bit) at once. When using earlycon, __msm_console_write() always writes 4 characters at a time, but we don't know which mode the bootloader was using and we don't set the mode either. This causes garbled output if the bootloader was using the single-character mode, because only every 4th character appears in the serial console, e.g. "[ 00oni pi 000xf0[ 00i s 5rm9(l)l s 1 1 SPMTA 7:C 5[ 00A ade k d[ 00ano:ameoi .Q1B[ 00ac _idaM00080oo'" If the bootloader was using the DMA ("DM") mode, output would likely fail entirely. Later, when the full serial driver probes, the port is re-initialized and output works as expected. Fix this also for earlycon by clearing the DMEN register and reset+re-enable the transmitter to apply the change. This ensures the transmitter is in the expected state before writing any output. Cc: stable <stable(a)kernel.org> Fixes: 0efe72963409 ("tty: serial: msm: Add earlycon support") Signed-off-by: Stephan Gerhold <stephan.gerhold(a)linaro.org> Reviewed-by: Neil Armstrong <neil.armstrong(a)linaro.org> Link: https://lore.kernel.org/r/20250408-msm-serial-earlycon-v1-1-429080127530@li… Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> diff --git a/drivers/tty/serial/msm_serial.c b/drivers/tty/serial/msm_serial.c index 1b137e068444..3449945493ce 100644 --- a/drivers/tty/serial/msm_serial.c +++ b/drivers/tty/serial/msm_serial.c @@ -1746,6 +1746,12 @@ msm_serial_early_console_setup_dm(struct earlycon_device *device, if (!device->port.membase) return -ENODEV; + /* Disable DM / single-character modes */ + msm_write(&device->port, 0, UARTDM_DMEN); + msm_write(&device->port, MSM_UART_CR_CMD_RESET_RX, MSM_UART_CR); + msm_write(&device->port, MSM_UART_CR_CMD_RESET_TX, MSM_UART_CR); + msm_write(&device->port, MSM_UART_CR_TX_ENABLE, MSM_UART_CR); + device->con->write = msm_serial_early_write_dm; return 0; }

2 months

1
0
0 0

FAILED: patch "[PATCH] serial: msm: Configure correct working mode before starting" failed to apply to 5.4-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 5.4-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. To reproduce the conflict and resubmit, you may use the following commands: git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-5.4.y git checkout FETCH_HEAD git cherry-pick -x 7094832b5ac861b0bd7ed8866c93cb15ef619996 # <resolve conflicts, build, test, etc.> git commit -s git send-email --to '<stable(a)vger.kernel.org>' --in-reply-to '2025042956-parkway-commend-a62f@gregkh' --subject-prefix 'PATCH 5.4.y' HEAD^.. Possible dependencies: thanks, greg k-h ------------------ original commit in Linus's tree ------------------ From 7094832b5ac861b0bd7ed8866c93cb15ef619996 Mon Sep 17 00:00:00 2001 From: Stephan Gerhold <stephan.gerhold(a)linaro.org> Date: Tue, 8 Apr 2025 19:22:47 +0200 Subject: [PATCH] serial: msm: Configure correct working mode before starting earlycon The MSM UART DM controller supports different working modes, e.g. DMA or the "single-character mode", where all reads/writes operate on a single character rather than 4 chars (32-bit) at once. When using earlycon, __msm_console_write() always writes 4 characters at a time, but we don't know which mode the bootloader was using and we don't set the mode either. This causes garbled output if the bootloader was using the single-character mode, because only every 4th character appears in the serial console, e.g. "[ 00oni pi 000xf0[ 00i s 5rm9(l)l s 1 1 SPMTA 7:C 5[ 00A ade k d[ 00ano:ameoi .Q1B[ 00ac _idaM00080oo'" If the bootloader was using the DMA ("DM") mode, output would likely fail entirely. Later, when the full serial driver probes, the port is re-initialized and output works as expected. Fix this also for earlycon by clearing the DMEN register and reset+re-enable the transmitter to apply the change. This ensures the transmitter is in the expected state before writing any output. Cc: stable <stable(a)kernel.org> Fixes: 0efe72963409 ("tty: serial: msm: Add earlycon support") Signed-off-by: Stephan Gerhold <stephan.gerhold(a)linaro.org> Reviewed-by: Neil Armstrong <neil.armstrong(a)linaro.org> Link: https://lore.kernel.org/r/20250408-msm-serial-earlycon-v1-1-429080127530@li… Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> diff --git a/drivers/tty/serial/msm_serial.c b/drivers/tty/serial/msm_serial.c index 1b137e068444..3449945493ce 100644 --- a/drivers/tty/serial/msm_serial.c +++ b/drivers/tty/serial/msm_serial.c @@ -1746,6 +1746,12 @@ msm_serial_early_console_setup_dm(struct earlycon_device *device, if (!device->port.membase) return -ENODEV; + /* Disable DM / single-character modes */ + msm_write(&device->port, 0, UARTDM_DMEN); + msm_write(&device->port, MSM_UART_CR_CMD_RESET_RX, MSM_UART_CR); + msm_write(&device->port, MSM_UART_CR_CMD_RESET_TX, MSM_UART_CR); + msm_write(&device->port, MSM_UART_CR_TX_ENABLE, MSM_UART_CR); + device->con->write = msm_serial_early_write_dm; return 0; }

2 months

1
0
0 0

Build failure due to backport of "bpf: Add namespace to BPF internal symbols"

by Shung-Hsi Yu

Hi Sasha, bpf-add-namespace-to-bpf-internal-symbols.patch is stable-queue.git seems like the culprit responsible for build failure in stable 6.12 and earlier (log below). The reason is likely due to the lack of commit cdd30ebb1b9f ("module: Convert symbol namespace to string literal") before v6.13. Getting rid of the quotes s/"BPF_INTERNAL"/BPF_INTERNAL/ probably would be enough to fix it. In file included from .vmlinux.export.c:1: .vmlinux.export.c:1697:33: error: expected ‘)’ before ‘BPF_INTERNAL’ 1697 | KSYMTAB_FUNC(bpf_map_get, "", ""BPF_INTERNAL""); | ^~~~~~~~~~~~ ./include/linux/export-internal.h:45:28: note: in definition of macro ‘__KSYMTAB’ 45 | " .asciz \"" ns "\"" "\n" \ | ^~ .vmlinux.export.c:1697:1: note: in expansion of macro ‘KSYMTAB_FUNC’ 1697 | KSYMTAB_FUNC(bpf_map_get, "", ""BPF_INTERNAL""); | ^~~~~~~~~~~~ ./include/linux/export-internal.h:41:12: note: to match this ‘(’ 41 | asm(" .section \"__ksymtab_strings\",\"aMS\",%progbits,1" "\n" \ | ^ ./include/linux/export-internal.h:62:41: note: in expansion of macro ‘__KSYMTAB’ 62 | #define KSYMTAB_FUNC(name, sec, ns) __KSYMTAB(name, KSYM_FUNC(name), sec, ns) | ^~~~~~~~~ .vmlinux.export.c:1697:1: note: in expansion of macro ‘KSYMTAB_FUNC’ 1697 | KSYMTAB_FUNC(bpf_map_get, "", ""BPF_INTERNAL""); | ^~~~~~~~~~~~ .vmlinux.export.c:1706:42: error: expected ‘)’ before ‘BPF_INTERNAL’ 1706 | KSYMTAB_FUNC(bpf_link_get_from_fd, "", ""BPF_INTERNAL""); | ^~~~~~~~~~~~ ./include/linux/export-internal.h:45:28: note: in definition of macro ‘__KSYMTAB’ 45 | " .asciz \"" ns "\"" "\n" \ | ^~ .vmlinux.export.c:1706:1: note: in expansion of macro ‘KSYMTAB_FUNC’ 1706 | KSYMTAB_FUNC(bpf_link_get_from_fd, "", ""BPF_INTERNAL""); | ^~~~~~~~~~~~ ./include/linux/export-internal.h:41:12: note: to match this ‘(’ 41 | asm(" .section \"__ksymtab_strings\",\"aMS\",%progbits,1" "\n" \ | ^ ./include/linux/export-internal.h:62:41: note: in expansion of macro ‘__KSYMTAB’ 62 | #define KSYMTAB_FUNC(name, sec, ns) __KSYMTAB(name, KSYM_FUNC(name), sec, ns) | ^~~~~~~~~ .vmlinux.export.c:1706:1: note: in expansion of macro ‘KSYMTAB_FUNC’ 1706 | KSYMTAB_FUNC(bpf_link_get_from_fd, "", ""BPF_INTERNAL""); | ^~~~~~~~~~~~ .vmlinux.export.c:1708:34: error: expected ‘)’ before ‘BPF_INTERNAL’ 1708 | KSYMTAB_FUNC(kern_sys_bpf, "", ""BPF_INTERNAL""); | ^~~~~~~~~~~~ ./include/linux/export-internal.h:45:28: note: in definition of macro ‘__KSYMTAB’ 45 | " .asciz \"" ns "\"" "\n" \ | ^~ .vmlinux.export.c:1708:1: note: in expansion of macro ‘KSYMTAB_FUNC’ 1708 | KSYMTAB_FUNC(kern_sys_bpf, "", ""BPF_INTERNAL""); | ^~~~~~~~~~~~ ./include/linux/export-internal.h:41:12: note: to match this ‘(’ 41 | asm(" .section \"__ksymtab_strings\",\"aMS\",%progbits,1" "\n" \ | ^ ./include/linux/export-internal.h:62:41: note: in expansion of macro ‘__KSYMTAB’ 62 | #define KSYMTAB_FUNC(name, sec, ns) __KSYMTAB(name, KSYM_FUNC(name), sec, ns) | ^~~~~~~~~ .vmlinux.export.c:1708:1: note: in expansion of macro ‘KSYMTAB_FUNC’ 1708 | KSYMTAB_FUNC(kern_sys_bpf, "", ""BPF_INTERNAL""); | ^~~~~~~~~~~~ make[2]: *** [scripts/Makefile.vmlinux:18: .vmlinux.export.o] Error 1 make[1]: *** [/home/runner/work/libbpf/libbpf/.kernel/Makefile:1184: vmlinux] Error 2 make: *** [Makefile:224: __sub-make] Error 2 Shung-Hsi Yu

2 months

2
1
0 0

[PATCH V3] arm64/boot: Enable EL2 requirements for FEAT_PMUv3p9

by Anshuman Khandual

FEAT_PMUv3p9 registers such as PMICNTR_EL0, PMICFILTR_EL0, and PMUACR_EL1 access from EL1 requires appropriate EL2 fine grained trap configuration via FEAT_FGT2 based trap control registers HDFGRTR2_EL2 and HDFGWTR2_EL2. Otherwise such register accesses will result in traps into EL2. Add a new helper __init_el2_fgt2() which initializes FEAT_FGT2 based fine grained trap control registers HDFGRTR2_EL2 and HDFGWTR2_EL2 (setting the bits nPMICNTR_EL0, nPMICFILTR_EL0 and nPMUACR_EL1) to enable access into PMICNTR_EL0, PMICFILTR_EL0, and PMUACR_EL1 registers. Also update booting.rst with SCR_EL3.FGTEn2 requirement for all FEAT_FGT2 based registers to be accessible in EL2. Cc: Catalin Marinas <catalin.marinas(a)arm.com> Cc: Will Deacon <will(a)kernel.org> Cc: Mark Rutland <mark.rutland(a)arm.com> Cc: Rob Herring <robh(a)kernel.org> Cc: Jonathan Corbet <corbet(a)lwn.net> Cc: Marc Zyngier <maz(a)kernel.org> Cc: Oliver Upton <oliver.upton(a)linux.dev> Cc: linux-arm-kernel(a)lists.infradead.org Cc: linux-doc(a)vger.kernel.org Cc: linux-kernel(a)vger.kernel.org Cc: kvmarm(a)lists.linux.dev Fixes: 0bbff9ed8165 ("perf/arm_pmuv3: Add PMUv3.9 per counter EL0 access control") Fixes: d8226d8cfbaf ("perf: arm_pmuv3: Add support for Armv9.4 PMU instruction counter") Cc: stable(a)vger.kernel.org Tested-by: Rob Herring (Arm) <robh(a)kernel.org> Reviewed-by: Rob Herring (Arm) <robh(a)kernel.org> Signed-off-by: Anshuman Khandual <anshuman.khandual(a)arm.com> --- Changes in V3: - Added 'MDCR_EL3.EnPM2 = 0b1' as a booting requirement per Mark - Added 'Fixes:' and 'CC: stable' tags per Mark Changes in V2: https://lore.kernel.org/all/20250203050828.1049370-8-anshuman.khandual@arm.… Documentation/arch/arm64/booting.rst | 22 ++++++++++++++++++++++ arch/arm64/include/asm/el2_setup.h | 25 +++++++++++++++++++++++++ 2 files changed, 47 insertions(+) diff --git a/Documentation/arch/arm64/booting.rst b/Documentation/arch/arm64/booting.rst index cad6fdc96b98..dee7b6de864f 100644 --- a/Documentation/arch/arm64/booting.rst +++ b/Documentation/arch/arm64/booting.rst @@ -288,6 +288,12 @@ Before jumping into the kernel, the following conditions must be met: - SCR_EL3.FGTEn (bit 27) must be initialised to 0b1. + For CPUs with the Fine Grained Traps 2 (FEAT_FGT2) extension present: + + - If EL3 is present and the kernel is entered at EL2: + + - SCR_EL3.FGTEn2 (bit 59) must be initialised to 0b1. + For CPUs with support for HCRX_EL2 (FEAT_HCX) present: - If EL3 is present and the kernel is entered at EL2: @@ -382,6 +388,22 @@ Before jumping into the kernel, the following conditions must be met: - SMCR_EL2.EZT0 (bit 30) must be initialised to 0b1. + For CPUs with the Performance Monitors Extension (FEAT_PMUv3p9): + + - If EL3 is present: + + - MDCR_EL3.EnPM2 (bit 7) must be initialised to 0b1. + + - If the kernel is entered at EL1 and EL2 is present: + + - HDFGRTR2_EL2.nPMICNTR_EL0 (bit 2) must be initialised to 0b1. + - HDFGRTR2_EL2.nPMICFILTR_EL0 (bit 3) must be initialised to 0b1. + - HDFGRTR2_EL2.nPMUACR_EL1 (bit 4) must be initialised to 0b1. + + - HDFGWTR2_EL2.nPMICNTR_EL0 (bit 2) must be initialised to 0b1. + - HDFGWTR2_EL2.nPMICFILTR_EL0 (bit 3) must be initialised to 0b1. + - HDFGWTR2_EL2.nPMUACR_EL1 (bit 4) must be initialised to 0b1. + For CPUs with Memory Copy and Memory Set instructions (FEAT_MOPS): - If the kernel is entered at EL1 and EL2 is present: diff --git a/arch/arm64/include/asm/el2_setup.h b/arch/arm64/include/asm/el2_setup.h index 25e162651750..1a0071faf57e 100644 --- a/arch/arm64/include/asm/el2_setup.h +++ b/arch/arm64/include/asm/el2_setup.h @@ -233,6 +233,30 @@ .Lskip_fgt_\@: .endm +.macro __init_el2_fgt2 + mrs x1, id_aa64mmfr0_el1 + ubfx x1, x1, #ID_AA64MMFR0_EL1_FGT_SHIFT, #4 + cmp x1, #ID_AA64MMFR0_EL1_FGT_FGT2 + b.lt .Lskip_fgt2_\@ + + mov x0, xzr + mrs x1, id_aa64dfr0_el1 + ubfx x1, x1, #ID_AA64DFR0_EL1_PMUVer_SHIFT, #4 + cmp x1, #ID_AA64DFR0_EL1_PMUVer_V3P9 + b.lt .Lskip_pmuv3p9_\@ + + orr x0, x0, #HDFGRTR2_EL2_nPMICNTR_EL0 + orr x0, x0, #HDFGRTR2_EL2_nPMICFILTR_EL0 + orr x0, x0, #HDFGRTR2_EL2_nPMUACR_EL1 +.Lskip_pmuv3p9_\@: + msr_s SYS_HDFGRTR2_EL2, x0 + msr_s SYS_HDFGWTR2_EL2, x0 + msr_s SYS_HFGRTR2_EL2, xzr + msr_s SYS_HFGWTR2_EL2, xzr + msr_s SYS_HFGITR2_EL2, xzr +.Lskip_fgt2_\@: +.endm + .macro __init_el2_gcs mrs_s x1, SYS_ID_AA64PFR1_EL1 ubfx x1, x1, #ID_AA64PFR1_EL1_GCS_SHIFT, #4 @@ -283,6 +307,7 @@ __init_el2_nvhe_idregs __init_el2_cptr __init_el2_fgt + __init_el2_fgt2 __init_el2_gcs .endm -- 2.25.1

2 months

2
3
0 0

[PATCH v2] drm/amdgpu: Remove redundant return value checks for amdgpu_ras_error_data_init

by Wentao Liang

The function amdgpu_ras_error_data_init() always returns 0, making its return value checks redundant. This patch changes its return type to void and removes all unnecessary checks in the callers. This simplifies the code and avoids confusion about the function's behavior. Additionally, this change keeps the usage consistent within amdgpu_ras_do_page_retirement(), which also does not check the return value. Fixes: 5b1270beb380 ("drm/amdgpu: add ras_err_info to identify RAS error source") Cc: stable(a)vger.kernel.org # 6.7+ Signed-off-by: Wentao Liang <vulab(a)iscas.ac.cn> --- v2: Add a missing semicolon drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 19 +++++-------------- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h | 2 +- drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c | 8 ++------ drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c | 3 +-- drivers/gpu/drm/amd/amdgpu/nbio_v7_9.c | 3 +-- 5 files changed, 10 insertions(+), 25 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index 4c9fa24dd972..5be391ebeca3 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -182,9 +182,7 @@ static int amdgpu_reserve_page_direct(struct amdgpu_device *adev, uint64_t addre return 0; } - ret = amdgpu_ras_error_data_init(&err_data); - if (ret) - return ret; + amdgpu_ras_error_data_init(&err_data); memset(&err_rec, 0x0, sizeof(struct eeprom_table_record)); err_data.err_addr = &err_rec; @@ -687,8 +685,7 @@ static struct ras_manager *amdgpu_ras_create_obj(struct amdgpu_device *adev, if (alive_obj(obj)) return NULL; - if (amdgpu_ras_error_data_init(&obj->err_data)) - return NULL; + amdgpu_ras_error_data_init(&obj->err_data); obj->head = *head; obj->adev = adev; @@ -1428,9 +1425,7 @@ static int amdgpu_ras_query_error_status_with_event(struct amdgpu_device *adev, if (!obj) return -EINVAL; - ret = amdgpu_ras_error_data_init(&err_data); - if (ret) - return ret; + amdgpu_ras_error_data_init(&err_data); if (!amdgpu_ras_get_error_query_mode(adev, &error_query_mode)) return -EINVAL; @@ -2255,9 +2250,7 @@ static void amdgpu_ras_interrupt_umc_handler(struct ras_manager *obj, if (!data->cb) return; - ret = amdgpu_ras_error_data_init(&err_data); - if (ret) - return; + amdgpu_ras_error_data_init(&err_data); /* Let IP handle its data, maybe we need get the output * from the callback to update the error type/count, etc @@ -4623,13 +4616,11 @@ void amdgpu_ras_inst_reset_ras_error_count(struct amdgpu_device *adev, } } -int amdgpu_ras_error_data_init(struct ras_err_data *err_data) +void amdgpu_ras_error_data_init(struct ras_err_data *err_data) { memset(err_data, 0, sizeof(*err_data)); INIT_LIST_HEAD(&err_data->err_node_list); - - return 0; } static void amdgpu_ras_error_node_release(struct ras_err_node *err_node) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h index 6db772ecfee4..5f88e70fbf5c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h @@ -931,7 +931,7 @@ void amdgpu_ras_inst_reset_ras_error_count(struct amdgpu_device *adev, uint32_t reg_list_size, uint32_t instance); -int amdgpu_ras_error_data_init(struct ras_err_data *err_data); +void amdgpu_ras_error_data_init(struct ras_err_data *err_data); void amdgpu_ras_error_data_fini(struct ras_err_data *err_data); int amdgpu_ras_error_statistic_ce_count(struct ras_err_data *err_data, struct amdgpu_smuio_mcm_config_info *mcm_info, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c index 896f3609b0ee..5de6e332c2cd 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c @@ -52,9 +52,7 @@ int amdgpu_umc_page_retirement_mca(struct amdgpu_device *adev, struct ras_err_data err_data; int ret; - ret = amdgpu_ras_error_data_init(&err_data); - if (ret) - return ret; + amdgpu_ras_error_data_init(&err_data); err_data.err_addr = kcalloc(adev->umc.max_ras_err_cnt_per_query, @@ -230,9 +228,7 @@ int amdgpu_umc_pasid_poison_handler(struct amdgpu_device *adev, }; struct ras_manager *obj = amdgpu_ras_find_obj(adev, &head); - ret = amdgpu_ras_error_data_init(&err_data); - if (ret) - return ret; + amdgpu_ras_error_data_init(&err_data); ret = amdgpu_umc_do_page_retirement(adev, &err_data, NULL, reset); diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c b/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c index a26a9be58eac..d4bdfe280c88 100644 --- a/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c +++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c @@ -364,8 +364,7 @@ static void nbio_v7_4_handle_ras_controller_intr_no_bifring(struct amdgpu_device struct ras_err_data err_data; struct amdgpu_ras *ras = amdgpu_ras_get_context(adev); - if (amdgpu_ras_error_data_init(&err_data)) - return; + amdgpu_ras_error_data_init(&err_data); if (adev->asic_type == CHIP_ALDEBARAN) bif_doorbell_intr_cntl = RREG32_SOC15(NBIO, 0, mmBIF_DOORBELL_INT_CNTL_ALDE); diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_9.c b/drivers/gpu/drm/amd/amdgpu/nbio_v7_9.c index 8a0a63ac88d2..c79ed1adf681 100644 --- a/drivers/gpu/drm/amd/amdgpu/nbio_v7_9.c +++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_9.c @@ -537,8 +537,7 @@ static void nbio_v7_9_handle_ras_controller_intr_no_bifring(struct amdgpu_device struct ras_err_data err_data; struct amdgpu_ras *ras = amdgpu_ras_get_context(adev); - if (amdgpu_ras_error_data_init(&err_data)) - return; + amdgpu_ras_error_data_init(&err_data); bif_doorbell_intr_cntl = RREG32_SOC15(NBIO, 0, regBIF_BX0_BIF_DOORBELL_INT_CNTL); -- 2.42.0.windows.2

2 months

1
0
0 0

[PATCH] fs/erofs/fileio: call erofs_onlinefolio_split() after bio_add_folio()

by Max Kellermann

If bio_add_folio() fails (because it is full), erofs_fileio_scan_folio() needs to submit the I/O request via erofs_fileio_rq_submit() and allocate a new I/O request with an empty `struct bio`. Then it retries the bio_add_folio() call. However, at this point, erofs_onlinefolio_split() has already been called which increments `folio->private`; the retry will call erofs_onlinefolio_split() again, but there will never be a matching erofs_onlinefolio_end() call. This leaves the folio locked forever and all waiters will be stuck in folio_wait_bit_common(). This bug has been added by commit ce63cb62d794 ("erofs: support unencoded inodes for fileio"), but was practically unreachable because there was room for 256 folios in the `struct bio` - until commit 9f74ae8c9ac9 ("erofs: shorten bvecs[] for file-backed mounts") which reduced the array capacity to 16 folios. It was now trivial to trigger the bug by manually invoking readahead from userspace, e.g.: posix_fadvise(fd, 0, st.st_size, POSIX_FADV_WILLNEED); This should be fixed by invoking erofs_onlinefolio_split() only after bio_add_folio() has succeeded. This is safe: asynchronous completions invoking erofs_onlinefolio_end() will not unlock the folio because erofs_fileio_scan_folio() is still holding a reference to be released by erofs_onlinefolio_end() at the end. Fixes: ce63cb62d794 ("erofs: support unencoded inodes for fileio") Fixes: 9f74ae8c9ac9 ("erofs: shorten bvecs[] for file-backed mounts") Cc: stable(a)vger.kernel.org Signed-off-by: Max Kellermann <max.kellermann(a)ionos.com> --- fs/erofs/fileio.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/fs/erofs/fileio.c b/fs/erofs/fileio.c index 4fa0a0121288..60c7cc4c105c 100644 --- a/fs/erofs/fileio.c +++ b/fs/erofs/fileio.c @@ -150,10 +150,10 @@ static int erofs_fileio_scan_folio(struct erofs_fileio *io, struct folio *folio) io->rq->bio.bi_iter.bi_sector = io->dev.m_pa >> 9; attached = 0; } - if (!attached++) - erofs_onlinefolio_split(folio); if (!bio_add_folio(&io->rq->bio, folio, len, cur)) goto io_retry; + if (!attached++) + erofs_onlinefolio_split(folio); io->dev.m_pa += len; } cur += len; -- 2.47.2

2 months

2
1
0 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

Linux-stable-mirror April 2025