From: Mark Rutland <mark.rutland@arm.com>
To: linux-kernel@vger.kernel.org, maz@kernel.org
Cc: aou@eecs.berkeley.edu, catalin.marinas@arm.com,
deanbo422@gmail.com, green.hu@gmail.com, guoren@kernel.org,
jonas@southpole.se, kernelfans@gmail.com,
linux-arm-kernel@lists.infradead.org, linux@armlinux.org.uk,
mark.rutland@arm.com, nickhu@andestech.com, palmer@dabbelt.com,
paul.walmsley@sifive.com, shorne@gmail.com,
stefan.kristiansson@saunalahti.fi, tglx@linutronix.de,
tsbogend@alpha.franken.de, vgupta@kernel.org,
vladimir.murzin@arm.com, will@kernel.org, paulmck@kernel.org,
peterz@infradead.org, torvalds@linux-foundation.org
Subject: [PATCH v2 00/17] irq: remove handle_domain_{irq,nmi}()
Date: Tue, 26 Oct 2021 10:24:47 +0100 [thread overview]
Message-ID: <20211026092504.27071-1-mark.rutland@arm.com> (raw)
The handle_domain_{irq,nmi}() functions were oringally intended as a
convenience, but recent rework to entry code across the kernel tree has
demonstrated that they cause more pain than they're worth and prevent
architectures from being able to write robust entry code.
This series reworks the irq code to remove them, handling the necessary
entry work consistently in entry code (be it architectural or generic).
Marc, on the assumption you'll take this into the irqchip tree, I've pushed
this out to my kernel.org repo, tagged as remove-handle-domain-irq-20211026,
which should be commit:
0953fb263714e1c8 ("irq: remove handle_domain_{irq,nmi}() (2021-10-26 10:13:31 +0100)")
Since v1 [1]:
* Accumulate tags
* Simplify bcm6345_l1_irq_handle()
* Don't export generic_handle_domain_nmi()
* Unexport handle_irq_desc()
[1] https://lore.kernel.org/r/20211021180236.37428-1-mark.rutland@arm.com
The gory details are in the thread starting at:
https://lore.kernel.org/r/20211011104729.GB1421@C02TD0UTHF1T.local
... which can be summarized as:
* At entry time, entry code needs to poke lockdep, RCU, and irqflag
tracing in a specific order due to interdependencies, before most
other C code can run. Part of this involves calling rcu_irq_enter().
Currently, arm64 is the only architecture which uses
handle_domain_irq() and performs the nominally correct sequence
(which is aligned with the generic entry code we aim to move to in
future).
* RCU relies on rcu_irq_enter() being called precisely once per IRQ
exception, or rcu_is_cpu_rrupt_from_idle() may fail to identify
wakeups from idle, and RCU may not trigger a reschedule when
necessary, leading to RCU stalls.
* The handle_domain_irq() helper, which is called from irqchip code
between the entry code and interrupt handlers, calls rcu_irq_enter(),
consequently causing problems for RCU on arm64.
In the thread Linus decreed:
I really think that if the rule is "we can't do accounting in
handle_domain_irq(), because it's too late for arm64", then the fix
really should be to just not do that.
Move the irq_enter()/irq_exit() to the callers - quite possibly far up
the call chain to the root of it all, and just say "architecture code
needs to do this in the low-level code before calling
handle_arch_irq".
This series does so, making entry code responsible for the IRQ entry
work, and removing the handle_domain_{irq,nmi}() functions entirely so
we're not tempted to reintroduce similar problems in future.
The rework fixes a couple of latent bugs for MIPS, fixes the problem
originally diagnosed for arm64, and should make it easier for the other
architectures using handle_domain_irq() to move to nominally correct
entry sequencing (e.g. by moving to the generic entry code).
I've added a couple of checks to generic_handle_irq() and
generic_handle_domain_irq() to verify that architectures are correctly
performing the required entry work (which fingers crossed, shouldn't
fire). Other than the above, there should be no functional change as
result of these changes (except that previously erroneous usage of RCU
in irqchip code will have become correct by virtue of RCU starting to
watch earlier).
I've given the series some light build and boot testing so far.
Thanks,
Mark.
Mark Rutland (17):
irq: mips: avoid nested irq_enter()
irq: mips: simplify bcm6345_l1_irq_handle()
irq: mips: stop (ab)using handle_domain_irq()
irq: mips: simplify do_domain_IRQ()
irq: simplify handle_domain_{irq,nmi}()
irq: unexport handle_irq_desc()
irq: add generic_handle_arch_irq()
irq: arc: avoid CONFIG_HANDLE_DOMAIN_IRQ
irq: nds32: avoid CONFIG_HANDLE_DOMAIN_IRQ
irq: add a (temporary) CONFIG_HANDLE_DOMAIN_IRQ_IRQENTRY
irq: arm: perform irqentry in entry code
irq: arm64: perform irqentry in entry code
irq: csky: perform irqentry in entry code
irq: openrisc: perform irqentry in entry code
irq: riscv: perform irqentry in entry code
irq: remove CONFIG_HANDLE_DOMAIN_IRQ_IRQENTRY
irq: remove handle_domain_{irq,nmi}()
Documentation/core-api/irq/irq-domain.rst | 3 --
arch/arc/Kconfig | 1 -
arch/arc/kernel/irq.c | 10 +++-
arch/arm/Kconfig | 1 -
arch/arm/kernel/entry-armv.S | 5 +-
arch/arm/kernel/irq.c | 14 +++---
arch/arm/mach-imx/avic.c | 2 +-
arch/arm/mach-imx/tzic.c | 2 +-
arch/arm/mach-omap1/irq.c | 2 +-
arch/arm/mach-s3c/irq-s3c24xx.c | 2 +-
arch/arm64/Kconfig | 1 -
arch/arm64/kernel/entry-common.c | 52 ++++++++++++--------
arch/csky/Kconfig | 1 -
arch/csky/kernel/entry.S | 2 +-
arch/csky/kernel/irq.c | 5 --
arch/mips/Kconfig | 1 -
arch/mips/cavium-octeon/octeon-irq.c | 5 +-
arch/mips/kernel/irq.c | 6 +--
arch/nds32/Kconfig | 1 -
arch/openrisc/Kconfig | 1 -
arch/openrisc/kernel/entry.S | 4 +-
arch/openrisc/kernel/irq.c | 5 --
arch/riscv/Kconfig | 1 -
arch/riscv/kernel/entry.S | 3 +-
arch/riscv/kernel/smp.c | 9 +---
drivers/irqchip/irq-apple-aic.c | 20 ++++----
drivers/irqchip/irq-armada-370-xp.c | 13 ++---
drivers/irqchip/irq-aspeed-vic.c | 2 +-
drivers/irqchip/irq-ativic32.c | 22 ++++++++-
drivers/irqchip/irq-atmel-aic.c | 2 +-
drivers/irqchip/irq-atmel-aic5.c | 2 +-
drivers/irqchip/irq-bcm2835.c | 2 +-
drivers/irqchip/irq-bcm2836.c | 2 +-
drivers/irqchip/irq-bcm6345-l1.c | 6 +--
drivers/irqchip/irq-clps711x.c | 8 +--
drivers/irqchip/irq-csky-apb-intc.c | 2 +-
drivers/irqchip/irq-csky-mpintc.c | 4 +-
drivers/irqchip/irq-davinci-aintc.c | 2 +-
drivers/irqchip/irq-davinci-cp-intc.c | 2 +-
drivers/irqchip/irq-digicolor.c | 2 +-
drivers/irqchip/irq-dw-apb-ictl.c | 2 +-
drivers/irqchip/irq-ftintc010.c | 2 +-
drivers/irqchip/irq-gic-v3.c | 4 +-
drivers/irqchip/irq-gic.c | 2 +-
drivers/irqchip/irq-hip04.c | 2 +-
drivers/irqchip/irq-ixp4xx.c | 4 +-
drivers/irqchip/irq-lpc32xx.c | 2 +-
drivers/irqchip/irq-mmp.c | 4 +-
drivers/irqchip/irq-mxs.c | 2 +-
drivers/irqchip/irq-nvic.c | 17 ++++++-
drivers/irqchip/irq-omap-intc.c | 2 +-
drivers/irqchip/irq-or1k-pic.c | 2 +-
drivers/irqchip/irq-orion.c | 4 +-
drivers/irqchip/irq-rda-intc.c | 2 +-
drivers/irqchip/irq-riscv-intc.c | 2 +-
drivers/irqchip/irq-sa11x0.c | 4 +-
drivers/irqchip/irq-sun4i.c | 2 +-
drivers/irqchip/irq-versatile-fpga.c | 2 +-
drivers/irqchip/irq-vic.c | 2 +-
drivers/irqchip/irq-vt8500.c | 2 +-
drivers/irqchip/irq-wpcm450-aic.c | 2 +-
drivers/irqchip/irq-zevio.c | 2 +-
include/linux/irq.h | 1 +
include/linux/irqdesc.h | 9 +---
kernel/irq/Kconfig | 3 --
kernel/irq/handle.c | 18 +++++++
kernel/irq/irqdesc.c | 81 +++++++------------------------
67 files changed, 192 insertions(+), 219 deletions(-)
--
2.11.0
next reply other threads:[~2021-10-26 9:25 UTC|newest]
Thread overview: 32+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-10-26 9:24 Mark Rutland [this message]
2021-10-26 9:24 ` [PATCH v2 01/17] irq: mips: avoid nested irq_enter() Mark Rutland
2021-10-26 9:24 ` [PATCH v2 02/17] irq: mips: simplify bcm6345_l1_irq_handle() Mark Rutland
2021-10-26 9:24 ` [PATCH v2 03/17] irq: mips: stop (ab)using handle_domain_irq() Mark Rutland
2021-10-26 9:24 ` [PATCH v2 04/17] irq: mips: simplify do_domain_IRQ() Mark Rutland
2021-10-26 9:24 ` [PATCH v2 05/17] irq: simplify handle_domain_{irq,nmi}() Mark Rutland
2021-10-26 9:24 ` [PATCH v2 06/17] irq: unexport handle_irq_desc() Mark Rutland
2021-10-26 9:24 ` [PATCH v2 07/17] irq: add generic_handle_arch_irq() Mark Rutland
2021-10-26 9:24 ` [PATCH v2 08/17] irq: arc: avoid CONFIG_HANDLE_DOMAIN_IRQ Mark Rutland
2021-10-26 9:24 ` [PATCH v2 09/17] irq: nds32: " Mark Rutland
2021-10-26 9:24 ` [PATCH v2 10/17] irq: add a (temporary) CONFIG_HANDLE_DOMAIN_IRQ_IRQENTRY Mark Rutland
2021-10-26 9:24 ` [PATCH v2 11/17] irq: arm: perform irqentry in entry code Mark Rutland
2021-10-26 9:24 ` [PATCH v2 12/17] irq: arm64: " Mark Rutland
2021-10-26 9:25 ` [PATCH v2 13/17] irq: csky: " Mark Rutland
2021-10-26 9:25 ` [PATCH v2 14/17] irq: openrisc: " Mark Rutland
2021-10-26 9:25 ` [PATCH v2 15/17] irq: riscv: " Mark Rutland
2021-10-26 9:25 ` [PATCH v2 16/17] irq: remove CONFIG_HANDLE_DOMAIN_IRQ_IRQENTRY Mark Rutland
2021-10-26 9:25 ` [PATCH v2 17/17] irq: remove handle_domain_{irq,nmi}() Mark Rutland
2022-05-06 20:32 ` Lukas Wunner
2022-05-09 8:54 ` Mark Rutland
2022-05-09 9:09 ` Marc Zyngier
2022-05-09 13:12 ` Thomas Gleixner
2022-05-10 23:36 ` Thomas Gleixner
2022-05-10 12:13 ` Lukas Wunner
2022-05-10 14:15 ` Mark Rutland
2022-05-10 22:52 ` Thomas Gleixner
2022-05-11 8:23 ` Mark Rutland
2022-05-11 8:57 ` Lukas Wunner
2022-05-11 9:27 ` Mark Rutland
2022-05-11 0:11 ` Thomas Gleixner
2022-05-11 8:11 ` Mark Rutland
2021-10-26 10:12 ` [PATCH v2 00/17] " Marc Zyngier
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=20211026092504.27071-1-mark.rutland@arm.com \
--to=mark.rutland@arm.com \
--cc=aou@eecs.berkeley.edu \
--cc=catalin.marinas@arm.com \
--cc=deanbo422@gmail.com \
--cc=green.hu@gmail.com \
--cc=guoren@kernel.org \
--cc=jonas@southpole.se \
--cc=kernelfans@gmail.com \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux@armlinux.org.uk \
--cc=maz@kernel.org \
--cc=nickhu@andestech.com \
--cc=palmer@dabbelt.com \
--cc=paul.walmsley@sifive.com \
--cc=paulmck@kernel.org \
--cc=peterz@infradead.org \
--cc=shorne@gmail.com \
--cc=stefan.kristiansson@saunalahti.fi \
--cc=tglx@linutronix.de \
--cc=torvalds@linux-foundation.org \
--cc=tsbogend@alpha.franken.de \
--cc=vgupta@kernel.org \
--cc=vladimir.murzin@arm.com \
--cc=will@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
all inboxes | Powered by JetHome®