linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Benjamin Herrenschmidt <benh@kernel.crashing.org>
To: linux-kernel@vger.kernel.org
Cc: tglx@linutronix.de,
	Linus Torvalds <torvalds@linux-foundation.org>,
	Frederic Weisbecker <fweisbec@gmail.com>,
	davem@davemloft.net, Paul Mackerras <paulus@au1.ibm.com>
Subject: do_softirq() vs __do_softirq() in irq_exit() and stack overflow
Date: Thu, 05 Sep 2013 07:39:56 +1000	[thread overview]
Message-ID: <1378330796.4321.50.camel@pasglop> (raw)

Hi Folks !

It appears that the current version of irq_exit() calls __do_softirq()
directly rather than do_softirq().

That means we are going to call the softirq's in the current interrupt
frame rather than on the separate softirq stack.

The current frame is also still the normal kernel stack, because
do_IRQ() itself only switches to the interrupt stack for processing
the handlers (it's back to the original stack by the time it calls
irq_exit).

That means that we end up stacking the normal stack, the actually HW
interrupt stack frame (which can be pretty big on ppc) + do_IRQ's own,
then the softirq (networks stack can create HUGE stack frames) and ...
we are in softirq, so HW irqs are enable, we can thus can another irq
stack frame piled up on top of that (or a perf stack).

We are observing actual overflows, here's an example blowing up our 16k
stack on ppc64, you notice that it's all on the normal kernel stack:

[ 1002.364577] do_IRQ: stack overflow: 1920
[ 1002.364653] CPU: 0 PID: 1602 Comm: qemu-system-ppc Not tainted 3.10.4-300.1.fc19.ppc64p7 #1
[ 1002.364734] Call Trace:
[ 1002.364770] [c0000000050a8740] [c0000000000157c0] .show_stack+0x130/0x200 (unreliable)
[ 1002.364872] [c0000000050a8810] [c00000000082f6d0] .dump_stack+0x28/0x3c
[ 1002.364955] [c0000000050a8880] [c000000000010b98] .do_IRQ+0x2b8/0x2c0
[ 1002.365039] [c0000000050a8930] [c0000000000026d4] hardware_interrupt_common+0x154/0x180
[ 1002.365148] --- Exception: 501 at .cp_start_xmit+0x3a4/0x820 [8139cp]
[ 1002.365148]     LR = .cp_start_xmit+0x390/0x820 [8139cp]
[ 1002.365359] [c0000000050a8d40] [c0000000006d7f14] .dev_hard_start_xmit+0x394/0x640
[ 1002.365433] [c0000000050a8e00] [c0000000007028c0] .sch_direct_xmit+0x110/0x260
[ 1002.365499] [c0000000050a8ea0] [c0000000006d8420] .dev_queue_xmit+0x260/0x630
[ 1002.365571] [c0000000050a8f40] [d0000000027d30d4] .br_dev_queue_push_xmit+0xc4/0x130 [bridge]
[ 1002.365641] [c0000000050a8fc0] [d0000000027d01f8] .br_dev_xmit+0x198/0x270 [bridge]
[ 1002.365707] [c0000000050a9070] [c0000000006d7f14] .dev_hard_start_xmit+0x394/0x640
[ 1002.365774] [c0000000050a9130] [c0000000006d85e8] .dev_queue_xmit+0x428/0x630
[ 1002.365834] [c0000000050a91d0] [c000000000729764] .ip_finish_output+0x2a4/0x550
[ 1002.365902] [c0000000050a9290] [c00000000072aaf0] .ip_local_out+0x50/0x70
[ 1002.365960] [c0000000050a9310] [c00000000072aed8] .ip_queue_xmit+0x148/0x420
[ 1002.366018] [c0000000050a93b0] [c000000000749524] .tcp_transmit_skb+0x4e4/0xaf0
[ 1002.366085] [c0000000050a94a0] [c00000000073de9c] .__tcp_ack_snd_check+0x7c/0xf0
[ 1002.366152] [c0000000050a9520] [c0000000007451d8] .tcp_rcv_established+0x1e8/0x930
[ 1002.366217] [c0000000050a95f0] [c00000000075326c] .tcp_v4_do_rcv+0x21c/0x570
[ 1002.366274] [c0000000050a96c0] [c000000000754a44] .tcp_v4_rcv+0x734/0x930
[ 1002.366332] [c0000000050a97a0] [c000000000724144] .ip_local_deliver_finish+0x184/0x360
[ 1002.366398] [c0000000050a9840] [c000000000724468] .ip_rcv_finish+0x148/0x400
[ 1002.366457] [c0000000050a98d0] [c0000000006d3248] .__netif_receive_skb_core+0x4f8/0xb00
[ 1002.366523] [c0000000050a99d0] [c0000000006d5414] .netif_receive_skb+0x44/0x110
[ 1002.366594] [c0000000050a9a70] [d0000000027d4e2c] .br_handle_frame_finish+0x2bc/0x3f0 [bridge]
[ 1002.366674] [c0000000050a9b20] [d0000000027de5ac] .br_nf_pre_routing_finish+0x2ac/0x420 [bridge]
[ 1002.366754] [c0000000050a9bd0] [d0000000027df5ec] .br_nf_pre_routing+0x4dc/0x7d0 [bridge]
[ 1002.366820] [c0000000050a9c70] [c000000000717aa4] .nf_iterate+0x114/0x130
[ 1002.366877] [c0000000050a9d30] [c000000000717b74] .nf_hook_slow+0xb4/0x1e0
[ 1002.366938] [c0000000050a9e00] [d0000000027d51f0] .br_handle_frame+0x290/0x330 [bridge]
[ 1002.367005] [c0000000050a9ea0] [c0000000006d309c] .__netif_receive_skb_core+0x34c/0xb00
[ 1002.367072] [c0000000050a9fa0] [c0000000006d5414] .netif_receive_skb+0x44/0x110
[ 1002.367138] [c0000000050aa040] [c0000000006d6218] .napi_gro_receive+0xe8/0x120
[ 1002.367210] [c0000000050aa0c0] [d00000000208536c] .cp_rx_poll+0x31c/0x590 [8139cp]
[ 1002.367276] [c0000000050aa1d0] [c0000000006d59cc] .net_rx_action+0x1dc/0x310
[ 1002.367335] [c0000000050aa2b0] [c0000000000a0568] .__do_softirq+0x158/0x330
[ 1002.367394] [c0000000050aa3b0] [c0000000000a0978] .irq_exit+0xc8/0x110
[ 1002.367452] [c0000000050aa430] [c0000000000109bc] .do_IRQ+0xdc/0x2c0
[ 1002.367510] [c0000000050aa4e0] [c0000000000026d4] hardware_interrupt_common+0x154/0x180
[ 1002.367580] --- Exception: 501 at .bad_range+0x1c/0x110
[ 1002.367580]     LR = .get_page_from_freelist+0x908/0xbb0
[ 1002.367658] [c0000000050aa7d0] [c00000000041d758] .list_del+0x18/0x50 (unreliable)
[ 1002.367725] [c0000000050aa850] [c0000000001bfa98] .get_page_from_freelist+0x908/0xbb0
[ 1002.367792] [c0000000050aa9e0] [c0000000001bff5c] .__alloc_pages_nodemask+0x21c/0xae0
[ 1002.367860] [c0000000050aaba0] [c0000000002126d0] .alloc_pages_vma+0xd0/0x210
[ 1002.367918] [c0000000050aac60] [c0000000001e93f4] .handle_pte_fault+0x814/0xb70
[ 1002.367985] [c0000000050aad50] [c0000000001eade4] .__get_user_pages+0x1a4/0x640
[ 1002.368052] [c0000000050aae60] [c00000000004606c] .get_user_pages_fast+0xec/0x160
[ 1002.368130] [c0000000050aaf10] [d000000001f73930] .__gfn_to_pfn_memslot+0x3b0/0x430 [kvm]
[ 1002.368205] [c0000000050aafd0] [d000000001f7e214] .kvmppc_gfn_to_pfn+0x64/0x130 [kvm]
[ 1002.368280] [c0000000050ab070] [d000000001f8a824] .kvmppc_mmu_map_page+0x94/0x530 [kvm]
[ 1002.368354] [c0000000050ab190] [d000000001f85064] .kvmppc_handle_pagefault+0x174/0x610 [kvm]
[ 1002.368429] [c0000000050ab270] [d000000001f85b74] .kvmppc_handle_exit_pr+0x464/0x9b0 [kvm]
[ 1002.368504] [c0000000050ab320] [d000000001f88ec4] kvm_start_lightweight+0x1ec/0x1fc [kvm]
[ 1002.368578] [c0000000050ab4f0] [d000000001f86a58] .kvmppc_vcpu_run_pr+0x168/0x3b0 [kvm]
[ 1002.368652] [c0000000050ab9c0] [d000000001f7f218] .kvmppc_vcpu_run+0xc8/0xf0 [kvm]
[ 1002.368725] [c0000000050aba50] [d000000001f7bdac] .kvm_arch_vcpu_ioctl_run+0x5c/0x1a0 [kvm]
[ 1002.368797] [c0000000050abae0] [d000000001f74618] .kvm_vcpu_ioctl+0x478/0x730 [kvm]
[ 1002.368865] [c0000000050abc90] [c00000000025302c] .do_vfs_ioctl+0x4ec/0x7c0
[ 1002.368923] [c0000000050abd80] [c0000000002533d4] .SyS_ioctl+0xd4/0xf0
[ 1002.368981] [c0000000050abe30] [c000000000009ed4] syscall_exit+0x0/0x98
[ 1002.369117] ------------[ cut here ]------------

Cheers,
Ben.



             reply	other threads:[~2013-09-04 21:40 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-09-04 21:39 Benjamin Herrenschmidt [this message]
2013-09-05  0:27 ` do_softirq() vs __do_softirq() in irq_exit() and stack overflow Benjamin Herrenschmidt
2013-09-05 13:29 ` Frederic Weisbecker
2013-09-05 15:33 ` [RFC PATCH 0/3] irq: Fix stack overflow due to softirq called on current stack Frederic Weisbecker
2013-09-05 15:33   ` [PATCH 1/3] irq: Consolidate do_softirq() arch overriden implementations Frederic Weisbecker
2013-09-05 15:33   ` [PATCH 2/3] irq: Execute softirq on its own stack on irq exit Frederic Weisbecker
2013-09-05 15:33   ` [PATCH 3/3] irq: Comment on the use of inline stack for ksoftirq Frederic Weisbecker
2013-09-05 22:56     ` Frederic Weisbecker
2013-09-05 22:18   ` [RFC PATCH 0/3] irq: Fix stack overflow due to softirq called on current stack Benjamin Herrenschmidt
2013-09-18  6:51   ` Paul Mackerras
2013-09-19 15:38     ` Frederic Weisbecker

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=1378330796.4321.50.camel@pasglop \
    --to=benh@kernel.crashing.org \
    --cc=davem@davemloft.net \
    --cc=fweisbec@gmail.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=paulus@au1.ibm.com \
    --cc=tglx@linutronix.de \
    --cc=torvalds@linux-foundation.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;
as well as URLs for NNTP newsgroup(s).