Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > linux.kernel > #1615942 > unrolled thread

Re: [PATCH] MIPS: IRQ Stack: Unwind IRQ stack onto task stack

Started by"Jason A. Donenfeld" <Jason@zx2c4.com>
First post2017-04-04 14:00 +0200
Last post2017-04-04 14:50 +0200
Articles 3 — 3 participants

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.


Contents

  Re: [PATCH] MIPS: IRQ Stack: Unwind IRQ stack onto task stack "Jason A. Donenfeld" <Jason@zx2c4.com> - 2017-04-04 14:00 +0200
    Re: [PATCH] MIPS: IRQ Stack: Unwind IRQ stack onto task stack Matt Redfearn <matt.redfearn@imgtec.com> - 2017-04-04 14:20 +0200
    Re: [PATCH] MIPS: IRQ Stack: Unwind IRQ stack onto task stack Ralf Baechle <ralf@linux-mips.org> - 2017-04-04 14:50 +0200

#1615942 — Re: [PATCH] MIPS: IRQ Stack: Unwind IRQ stack onto task stack

From"Jason A. Donenfeld" <Jason@zx2c4.com>
Date2017-04-04 14:00 +0200
SubjectRe: [PATCH] MIPS: IRQ Stack: Unwind IRQ stack onto task stack
Message-ID<tsvhU-4mH-15@gated-at.bofh.it>
This indeed is useful. Out of curiosity, are other archs using a
similar technique? In anycase,

Acked-by: Jason A. Donenfeld <Jason@zx2c4.com>

[toc] | [next] | [standalone]


#1615948

FromMatt Redfearn <matt.redfearn@imgtec.com>
Date2017-04-04 14:20 +0200
Message-ID<tsvBg-4KM-11@gated-at.bofh.it>
In reply to#1615942
Hi Jason,


On 04/04/17 12:58, Jason A. Donenfeld wrote:
> This indeed is useful. Out of curiosity, are other archs using a
> similar technique? In anycase,
>
> Acked-by: Jason A. Donenfeld <Jason@zx2c4.com>

Yes, at least x86 and ARM64 do the same thing (probably more)

x86 saves the previous stack pointer on the IRQ stack in 
http://lxr.free-electrons.com/source/arch/x86/kernel/irq_32.c#L70
which is then unwound in 
http://lxr.free-electrons.com/source/arch/x86/kernel/dumpstack.c#L51

ARM64 saves the task SP in 
http://lxr.free-electrons.com/source/arch/arm64/kernel/entry.S#L249
And unwinds it in 
http://lxr.free-electrons.com/source/arch/arm64/kernel/traps.c#L140

Thanks,
Matt

[toc] | [prev] | [next] | [standalone]


#1615997

FromRalf Baechle <ralf@linux-mips.org>
Date2017-04-04 14:50 +0200
Message-ID<tsw4i-4Ve-33@gated-at.bofh.it>
In reply to#1615942
On Tue, Apr 04, 2017 at 01:58:04PM +0200, Jason A. Donenfeld wrote:

> This indeed is useful. Out of curiosity, are other archs using a
> similar technique? In anycase,
> 
> Acked-by: Jason A. Donenfeld <Jason@zx2c4.com>

Unfortunately MIPS doesn't have anything like a frame pointer or similar
to make backtracing easy so proper stacktraces have always been more of
rocket science than they really ought to ...

  Ralf

[toc] | [prev] | [standalone]


Back to top | Article view | linux.kernel


csiph-web