linux-stable.git/arch/sparc, branch v4.2 Linux kernel stable tree sparc64: Fix userspace FPU register corruptions. 2015-08-07T02:13:25+00:00 David S. Miller davem@davemloft.net 2015-08-07T02:13:25+00:00 44922150d87cef616fd183220d43d8fde4d41390 If we have a series of events from userpsace, with %fprs=FPRS_FEF, like follows: ETRAP ETRAP VIS_ENTRY(fprs=0x4) VIS_EXIT RTRAP (kernel FPU restore with fpu_saved=0x4) RTRAP We will not restore the user registers that were clobbered by the FPU using kernel code in the inner-most trap. Traps allocate FPU save slots in the thread struct, and FPU using sequences save the "dirty" FPU registers only. This works at the initial trap level because all of the registers get recorded into the top-level FPU save area, and we'll return to userspace with the FPU disabled so that any FPU use by the user will take an FPU disabled trap wherein we'll load the registers back up properly. But this is not how trap returns from kernel to kernel operate. The simplest fix for this bug is to always save all FPU register state for anything other than the top-most FPU save area. Getting rid of the optimized inner-slot FPU saving code ends up making VISEntryHalf degenerate into plain VISEntry. Longer term we need to do something smarter to reinstate the partial save optimizations. Perhaps the fundament error is having trap entry and exit allocate FPU save slots and restore register state. Instead, the VISEntry et al. calls should be doing that work. This bug is about two decades old. Reported-by: James Y Knight <jyknight@google.com> Signed-off-by: David S. Miller <davem@davemloft.net> 
If we have a series of events from userpsace, with %fprs=FPRS_FEF,
like follows:

ETRAP
	ETRAP
		VIS_ENTRY(fprs=0x4)
		VIS_EXIT
		RTRAP (kernel FPU restore with fpu_saved=0x4)
	RTRAP

We will not restore the user registers that were clobbered by the FPU
using kernel code in the inner-most trap.

Traps allocate FPU save slots in the thread struct, and FPU using
sequences save the "dirty" FPU registers only.

This works at the initial trap level because all of the registers
get recorded into the top-level FPU save area, and we'll return
to userspace with the FPU disabled so that any FPU use by the user
will take an FPU disabled trap wherein we'll load the registers
back up properly.

But this is not how trap returns from kernel to kernel operate.

The simplest fix for this bug is to always save all FPU register state
for anything other than the top-most FPU save area.

Getting rid of the optimized inner-slot FPU saving code ends up
making VISEntryHalf degenerate into plain VISEntry.

Longer term we need to do something smarter to reinstate the partial
save optimizations.  Perhaps the fundament error is having trap entry
and exit allocate FPU save slots and restore register state.  Instead,
the VISEntry et al. calls should be doing that work.

This bug is about two decades old.

Reported-by: James Y Knight <jyknight@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
mm: clean up per architecture MM hook header files 2015-07-17T23:39:53+00:00 Laurent Dufour ldufour@linux.vnet.ibm.com 2015-07-17T23:23:58+00:00 f2abeef9fd6f03ebf417539ed099828a56733098 Commit 2ae416b142b6 ("mm: new mm hook framework") introduced an empty header file (mm-arch-hooks.h) for every architecture, even those which doesn't need to define mm hooks. As suggested by Geert Uytterhoeven, this could be cleaned through the use of a generic header file included via each per architecture asm/include/Kbuild file. The PowerPC architecture is not impacted here since this architecture has to defined the arch_remap MM hook. Signed-off-by: Laurent Dufour <ldufour@linux.vnet.ibm.com> Suggested-by: Geert Uytterhoeven <geert@linux-m68k.org> Acked-by: Geert Uytterhoeven <geert@linux-m68k.org> Acked-by: Vineet Gupta <vgupta@synopsys.com> Cc: Oleg Nesterov <oleg@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> 
Commit 2ae416b142b6 ("mm: new mm hook framework") introduced an empty
header file (mm-arch-hooks.h) for every architecture, even those which
doesn't need to define mm hooks.

As suggested by Geert Uytterhoeven, this could be cleaned through the use
of a generic header file included via each per architecture
asm/include/Kbuild file.

The PowerPC architecture is not impacted here since this architecture has
to defined the arch_remap MM hook.

Signed-off-by: Laurent Dufour <ldufour@linux.vnet.ibm.com>
Suggested-by: Geert Uytterhoeven <geert@linux-m68k.org>
Acked-by: Geert Uytterhoeven <geert@linux-m68k.org>
Acked-by: Vineet Gupta <vgupta@synopsys.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Merge branch 'akpm' (patches from Andrew) 2015-06-26T16:52:05+00:00 Linus Torvalds torvalds@linux-foundation.org 2015-06-26T16:52:05+00:00 47a469421d792dcb91a1e73319d26134241953d2 Merge second patchbomb from Andrew Morton: - most of the rest of MM - lots of misc things - procfs updates - printk feature work - updates to get_maintainer, MAINTAINERS, checkpatch - lib/ updates * emailed patches from Andrew Morton <akpm@linux-foundation.org>: (96 commits) exit,stats: /* obey this comment */ coredump: add __printf attribute to cn_*printf functions coredump: use from_kuid/kgid when formatting corename fs/reiserfs: remove unneeded cast NILFS2: support NFSv2 export fs/befs/btree.c: remove unneeded initializations fs/minix: remove unneeded cast init/do_mounts.c: add create_dev() failure log kasan: remove duplicate definition of the macro KASAN_FREE_PAGE fs/efs: femove unneeded cast checkpatch: emit "NOTE: <types>" message only once after multiple files checkpatch: emit an error when there's a diff in a changelog checkpatch: validate MODULE_LICENSE content checkpatch: add multi-line handling for PREFER_ETHER_ADDR_COPY checkpatch: suggest using eth_zero_addr() and eth_broadcast_addr() checkpatch: fix processing of MEMSET issues checkpatch: suggest using ether_addr_equal*() checkpatch: avoid NOT_UNIFIED_DIFF errors on cover-letter.patch files checkpatch: remove local from codespell path checkpatch: add --showfile to allow input via pipe to show filenames ... 
Merge second patchbomb from Andrew Morton:

 - most of the rest of MM

 - lots of misc things

 - procfs updates

 - printk feature work

 - updates to get_maintainer, MAINTAINERS, checkpatch

 - lib/ updates

* emailed patches from Andrew Morton <akpm@linux-foundation.org>: (96 commits)
  exit,stats: /* obey this comment */
  coredump: add __printf attribute to cn_*printf functions
  coredump: use from_kuid/kgid when formatting corename
  fs/reiserfs: remove unneeded cast
  NILFS2: support NFSv2 export
  fs/befs/btree.c: remove unneeded initializations
  fs/minix: remove unneeded cast
  init/do_mounts.c: add create_dev() failure log
  kasan: remove duplicate definition of the macro KASAN_FREE_PAGE
  fs/efs: femove unneeded cast
  checkpatch: emit "NOTE: <types>" message only once after multiple files
  checkpatch: emit an error when there's a diff in a changelog
  checkpatch: validate MODULE_LICENSE content
  checkpatch: add multi-line handling for PREFER_ETHER_ADDR_COPY
  checkpatch: suggest using eth_zero_addr() and eth_broadcast_addr()
  checkpatch: fix processing of MEMSET issues
  checkpatch: suggest using ether_addr_equal*()
  checkpatch: avoid NOT_UNIFIED_DIFF errors on cover-letter.patch files
  checkpatch: remove local from codespell path
  checkpatch: add --showfile to allow input via pipe to show filenames
  ...
Merge tag 'rtc-v4.2-1' of git://git.kernel.org/pub/scm/linux/kernel/git/abelloni/linux 2015-06-26T01:55:33+00:00 Linus Torvalds torvalds@linux-foundation.org 2015-06-26T01:55:33+00:00 c13c81006314ad76c2b31824960a900385601b8b Pull RTC updates from Alexandre Belloni: "Core: - Coding style and whitespace fixes (interface, Makefile and Kconfig) - New rtc_tm_sub() helper - New CONFIG_RTC_SYSTOHC_DEVICE option - Removed rtc_set_mmss() New drivers: - Mediatek MT6397 - Cortina Gemini Drivers: - Year 2106 fixes for isl1208, pcf8563 and sunxi - update author email for at32ap700x and efi - ds1307: alarm fix - efi: use correct EFI 'epoch' - hym8563: make irq optional - imxdi: cleanups and better handling of the security/tamper monitoring - snvs: fix wakealarm - Compilation fixes or warning removal for gemini, mt6397, palmas, pfc8563 - Trivial cleanups for ab8500, ds1216, ds1286, ds1672, ep93xx, hid-sensor-time, max6900, max8998, max77686, max77802, mc13xxx, mv, mxc, s3c, spear, v3020 - Kconfig fixes for stmp3xxx and xgene" * tag 'rtc-v4.2-1' of git://git.kernel.org/pub/scm/linux/kernel/git/abelloni/linux: (48 commits) rtc: remove useless I2C dependencies rtc: whitespace fixes rtc: Properly sort Makefile MAINTAINERS: Add RTC subsystem repository rtc: pfc8563: fix uninitialized variable warning rtc: ds1307: Enable the mcp794xx alarm after programming time rtc: hym8563: make the irq optional rtc: gemini: fix cocci warnings rtc: mv: correct 24 hour error message rtc: mv: use BIT() rtc: efi: use correct EFI 'epoch' rtc: interface: Remove rtc_set_mmss() sparc: time: Replace update_persistent_clock() with CONFIG_RTC_SYSTOHC rtc: NTP: Add CONFIG_RTC_SYSTOHC_DEVICE for NTP synchronization rtc: sunxi: Replace deprecated rtc_tm_to_time() rtc: isl1208: Replace deprecated rtc_tm_to_time() rtc: Introduce rtc_tm_sub() helper function rtc: pcf8563: Replace deprecated rtc_time_to_tm() and rtc_tm_to_time() rtc: palmas: Initialise bb_charging flag before using it rtc: simplify use of devm_ioremap_resource ... 
Pull RTC updates from Alexandre Belloni:
 "Core:
   - Coding style and whitespace fixes (interface, Makefile and Kconfig)
   - New rtc_tm_sub() helper
   - New CONFIG_RTC_SYSTOHC_DEVICE option
   - Removed rtc_set_mmss()

  New drivers:
   - Mediatek MT6397
   - Cortina Gemini

  Drivers:
   - Year 2106 fixes for isl1208, pcf8563 and sunxi
   - update author email for at32ap700x and efi
   - ds1307: alarm fix
   - efi: use correct EFI 'epoch'
   - hym8563: make irq optional
   - imxdi: cleanups and better handling of the security/tamper monitoring
   - snvs: fix wakealarm
   - Compilation fixes or warning removal for gemini, mt6397, palmas, pfc8563
   - Trivial cleanups for ab8500, ds1216, ds1286, ds1672, ep93xx,
     hid-sensor-time, max6900, max8998, max77686, max77802, mc13xxx, mv,
     mxc, s3c, spear, v3020
   - Kconfig fixes for stmp3xxx and xgene"

* tag 'rtc-v4.2-1' of git://git.kernel.org/pub/scm/linux/kernel/git/abelloni/linux: (48 commits)
  rtc: remove useless I2C dependencies
  rtc: whitespace fixes
  rtc: Properly sort Makefile
  MAINTAINERS: Add RTC subsystem repository
  rtc: pfc8563: fix uninitialized variable warning
  rtc: ds1307: Enable the mcp794xx alarm after programming time
  rtc: hym8563: make the irq optional
  rtc: gemini: fix cocci warnings
  rtc: mv: correct 24 hour error message
  rtc: mv: use BIT()
  rtc: efi: use correct EFI 'epoch'
  rtc: interface: Remove rtc_set_mmss()
  sparc: time: Replace update_persistent_clock() with CONFIG_RTC_SYSTOHC
  rtc: NTP: Add CONFIG_RTC_SYSTOHC_DEVICE for NTP synchronization
  rtc: sunxi: Replace deprecated rtc_tm_to_time()
  rtc: isl1208: Replace deprecated rtc_tm_to_time()
  rtc: Introduce rtc_tm_sub() helper function
  rtc: pcf8563: Replace deprecated rtc_time_to_tm() and rtc_tm_to_time()
  rtc: palmas: Initialise bb_charging flag before using it
  rtc: simplify use of devm_ioremap_resource
  ...
mm/hugetlb: remove arch_prepare/release_hugepage from arch headers 2015-06-26T00:00:35+00:00 Dominik Dingel dingel@linux.vnet.ibm.com 2015-06-25T21:59:44+00:00 08bd4fc15683b9a26b9be7048d151f4ddadad96f Nobody used these hooks so they were removed from common code, and can now be removed from the architectures. Signed-off-by: Dominik Dingel <dingel@linux.vnet.ibm.com> Acked-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Acked-by: Ralf Baechle <ralf@linux-mips.org> Cc: Heiko Carstens <heiko.carstens@de.ibm.com> Cc: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> 
Nobody used these hooks so they were removed from common code, and can now
be removed from the architectures.

Signed-off-by: Dominik Dingel <dingel@linux.vnet.ibm.com>
Acked-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Acked-by: Ralf Baechle <ralf@linux-mips.org>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Merge branch 'for-4.2/sg' of git://git.kernel.dk/linux-block 2015-06-25T22:22:36+00:00 Linus Torvalds torvalds@linux-foundation.org 2015-06-25T22:22:36+00:00 ad90fb97515b732bc27a0109baa10af636c3c8cd Pull asm/scatterlist.h removal from Jens Axboe: "We don't have any specific arch scatterlist anymore, since parisc finally switched over. Kill the include" * 'for-4.2/sg' of git://git.kernel.dk/linux-block: remove scatterlist.h generation from arch Kbuild files remove <asm/scatterlist.h> 
Pull asm/scatterlist.h removal from Jens Axboe:
 "We don't have any specific arch scatterlist anymore, since parisc
  finally switched over.  Kill the include"

* 'for-4.2/sg' of git://git.kernel.dk/linux-block:
  remove scatterlist.h generation from arch Kbuild files
  remove <asm/scatterlist.h>
sparc64: perf: Use UREG_FP rather than UREG_I6 2015-06-25T13:01:03+00:00 David Ahern david.ahern@oracle.com 2015-06-15T20:15:46+00:00 2d89cd8625c4af01a2683b18c3c8194cc3b3067c perf walks userspace callchains by following frame pointers. Use the UREG_FP macro to make it clearer that the %fp is being used. Signed-off-by: David Ahern <david.ahern@oracle.com> Signed-off-by: David S. Miller <davem@davemloft.net> 
perf walks userspace callchains by following frame pointers. Use the
UREG_FP macro to make it clearer that the %fp is being used.

Signed-off-by: David Ahern <david.ahern@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
sparc64: perf: Add sanity checking on addresses in user stack 2015-06-25T13:01:02+00:00 David Ahern david.ahern@oracle.com 2015-06-15T20:15:45+00:00 b69fb7699c92f85991672fc144b0adb7c717fbc8 Processes are getting killed (sigbus or segv) while walking userspace callchains when using perf. In some instances I have seen ufp = 0x7ff which does not seem like a proper stack address. This patch adds a function to run validity checks against the address before attempting the copy_from_user. The checks are copied from the x86 version as a start point with the addition of a 4-byte alignment check. Signed-off-by: David Ahern <david.ahern@oracle.com> Signed-off-by: David S. Miller <davem@davemloft.net> 
Processes are getting killed (sigbus or segv) while walking userspace
callchains when using perf. In some instances I have seen ufp = 0x7ff
which does not seem like a proper stack address.

This patch adds a function to run validity checks against the address
before attempting the copy_from_user. The checks are copied from the
x86 version as a start point with the addition of a 4-byte alignment
check.

Signed-off-by: David Ahern <david.ahern@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
sparc64: Convert BUG_ON to warning 2015-06-25T13:01:02+00:00 David Ahern david.ahern@oracle.com 2015-06-15T20:15:44+00:00 2bf7c3efc393937d1e5f92681501a914dbfbae07 Pagefault handling has a BUG_ON path that panics the system. Convert it to a warning instead. There is no need to bring down the system for this kind of failure. The following was hit while running: perf sched record -g -- make -j 16 [3609412.782801] kernel BUG at /opt/dahern/linux.git/arch/sparc/mm/fault_64.c:416! [3609412.782833] \|/ ____ \|/ [3609412.782833] "@'/ .. \`@" [3609412.782833] /_| \__/ |_\ [3609412.782833] \__U_/ [3609412.782870] cat(4516): Kernel bad sw trap 5 [#1] [3609412.782889] CPU: 0 PID: 4516 Comm: cat Tainted: G E 4.1.0-rc8+ #6 [3609412.782909] task: fff8000126e31f80 ti: fff8000110d90000 task.ti: fff8000110d90000 [3609412.782931] TSTATE: 0000004411001603 TPC: 000000000096b164 TNPC: 000000000096b168 Y: 0000004e Tainted: G E [3609412.782964] TPC: <do_sparc64_fault+0x5e4/0x6a0> [3609412.782979] g0: 000000000096abe0 g1: 0000000000d314c4 g2: 0000000000000000 g3: 0000000000000001 [3609412.783009] g4: fff8000126e31f80 g5: fff80001302d2000 g6: fff8000110d90000 g7: 00000000000000ff [3609412.783045] o0: 0000000000aff6a8 o1: 00000000000001a0 o2: 0000000000000001 o3: 0000000000000054 [3609412.783080] o4: fff8000100026820 o5: 0000000000000001 sp: fff8000110d935f1 ret_pc: 000000000096b15c [3609412.783117] RPC: <do_sparc64_fault+0x5dc/0x6a0> [3609412.783137] l0: 000007feff996000 l1: 0000000000030001 l2: 0000000000000004 l3: fff8000127bd0120 [3609412.783174] l4: 0000000000000054 l5: fff8000127bd0188 l6: 0000000000000000 l7: fff8000110d9dba8 [3609412.783210] i0: fff8000110d93f60 i1: fff8000110ca5530 i2: 000000000000003f i3: 0000000000000054 [3609412.783244] i4: fff800010000081a i5: fff8000100000398 i6: fff8000110d936a1 i7: 0000000000407c6c [3609412.783286] I7: <sparc64_realfault_common+0x10/0x20> [3609412.783308] Call Trace: [3609412.783329] [0000000000407c6c] sparc64_realfault_common+0x10/0x20 [3609412.783353] Disabling lock debugging due to kernel taint [3609412.783379] Caller[0000000000407c6c]: sparc64_realfault_common+0x10/0x20 [3609412.783449] Caller[fff80001002283e4]: 0xfff80001002283e4 [3609412.783471] Instruction DUMP: 921021a0 7feaff91 901222a8 <91d02005> 82086100 02f87f7b 808a2873 81cfe008 01000000 [3609412.783542] Kernel panic - not syncing: Fatal exception [3609412.784605] Press Stop-A (L1-A) to return to the boot prom [3609412.784615] ---[ end Kernel panic - not syncing: Fatal exception With this patch rather than a panic I occasionally get something like this: perf sched record -g -m 1024 -- make -j N where N is based on number of cpus (128 to 1024 for a T7-4 and 8 for an 8 cpu VM on a T5-2). WARNING: CPU: 211 PID: 52565 at /opt/dahern/linux.git/arch/sparc/mm/fault_64.c:417 do_sparc64_fault+0x340/0x70c() address (7feffcd6000) != regs->tpc (fff80001004873c0) Modules linked in: ipt_REJECT nf_reject_ipv4 nf_conntrack_ipv4 nf_defrag_ipv4 iptable_filter ip_tables ip6t_REJECT nf_reject_ipv6 xt_tcpudp nf_conntrack_ipv6 nf_defrag_ipv6 xt_state nf_conntrack ip6table_filter ip6_tables x_tables ipv6 cdc_ether usbnet mii ixgbe mdio igb i2c_algo_bit i2c_core ptp crc32c_sparc64 camellia_sparc64 des_sparc64 des_generic md5_sparc64 sha512_sparc64 sha1_sparc64 uio_pdrv_genirq uio usb_storage mpt3sas scsi_transport_sas raid_class aes_sparc64 sunvnet sunvdc sha256_sparc64(E) sha256_generic(E) CPU: 211 PID: 52565 Comm: ld Tainted: G W E 4.1.0-rc8+ #19 Call Trace: [000000000045ce30] warn_slowpath_common+0x7c/0xa0 [000000000045ceec] warn_slowpath_fmt+0x30/0x40 [000000000098ad64] do_sparc64_fault+0x340/0x70c [0000000000407c2c] sparc64_realfault_common+0x10/0x20 ---[ end trace 62ee02065a01a049 ]--- ld[52565]: segfault at fff80001004873c0 ip fff80001004873c0 (rpc fff8000100158868) sp 000007feffcd70e1 error 30002 in libc-2.12.so[fff8000100410000+184000] The segfault is horrible, but better than a system panic. An 8-cpu VM on a T5-2 also showed the above traces from time to time, so it is a general problem and not specific to the T7 or baremetal. Signed-off-by: David Ahern <david.ahern@oracle.com> Signed-off-by: David S. Miller <davem@davemloft.net> 
Pagefault handling has a BUG_ON path that panics the system. Convert it to
a warning instead. There is no need to bring down the system for this kind
of failure.

The following was hit while running:
    perf sched record -g -- make -j 16

[3609412.782801] kernel BUG at /opt/dahern/linux.git/arch/sparc/mm/fault_64.c:416!
[3609412.782833]               \|/ ____ \|/
[3609412.782833]               "@'/ .. \`@"
[3609412.782833]               /_| \__/ |_\
[3609412.782833]                  \__U_/
[3609412.782870] cat(4516): Kernel bad sw trap 5 [#1]
[3609412.782889] CPU: 0 PID: 4516 Comm: cat Tainted: G            E   4.1.0-rc8+ #6
[3609412.782909] task: fff8000126e31f80 ti: fff8000110d90000 task.ti: fff8000110d90000
[3609412.782931] TSTATE: 0000004411001603 TPC: 000000000096b164 TNPC: 000000000096b168 Y: 0000004e    Tainted: G            E
[3609412.782964] TPC: <do_sparc64_fault+0x5e4/0x6a0>
[3609412.782979] g0: 000000000096abe0 g1: 0000000000d314c4 g2: 0000000000000000 g3: 0000000000000001
[3609412.783009] g4: fff8000126e31f80 g5: fff80001302d2000 g6: fff8000110d90000 g7: 00000000000000ff
[3609412.783045] o0: 0000000000aff6a8 o1: 00000000000001a0 o2: 0000000000000001 o3: 0000000000000054
[3609412.783080] o4: fff8000100026820 o5: 0000000000000001 sp: fff8000110d935f1 ret_pc: 000000000096b15c
[3609412.783117] RPC: <do_sparc64_fault+0x5dc/0x6a0>
[3609412.783137] l0: 000007feff996000 l1: 0000000000030001 l2: 0000000000000004 l3: fff8000127bd0120
[3609412.783174] l4: 0000000000000054 l5: fff8000127bd0188 l6: 0000000000000000 l7: fff8000110d9dba8
[3609412.783210] i0: fff8000110d93f60 i1: fff8000110ca5530 i2: 000000000000003f i3: 0000000000000054
[3609412.783244] i4: fff800010000081a i5: fff8000100000398 i6: fff8000110d936a1 i7: 0000000000407c6c
[3609412.783286] I7: <sparc64_realfault_common+0x10/0x20>
[3609412.783308] Call Trace:
[3609412.783329]  [0000000000407c6c] sparc64_realfault_common+0x10/0x20
[3609412.783353] Disabling lock debugging due to kernel taint
[3609412.783379] Caller[0000000000407c6c]: sparc64_realfault_common+0x10/0x20
[3609412.783449] Caller[fff80001002283e4]: 0xfff80001002283e4
[3609412.783471] Instruction DUMP: 921021a0  7feaff91  901222a8 <91d02005> 82086100  02f87f7b  808a2873  81cfe008  01000000
[3609412.783542] Kernel panic - not syncing: Fatal exception
[3609412.784605] Press Stop-A (L1-A) to return to the boot prom
[3609412.784615] ---[ end Kernel panic - not syncing: Fatal exception

With this patch rather than a panic I occasionally get something like this:
    perf sched record -g -m 1024  -- make -j N

where N is based on number of cpus (128 to 1024 for a T7-4 and 8 for an 8 cpu
VM on a T5-2).

WARNING: CPU: 211 PID: 52565 at /opt/dahern/linux.git/arch/sparc/mm/fault_64.c:417 do_sparc64_fault+0x340/0x70c()
address (7feffcd6000) != regs->tpc (fff80001004873c0)
Modules linked in: ipt_REJECT nf_reject_ipv4 nf_conntrack_ipv4 nf_defrag_ipv4 iptable_filter ip_tables ip6t_REJECT nf_reject_ipv6 xt_tcpudp nf_conntrack_ipv6 nf_defrag_ipv6 xt_state nf_conntrack ip6table_filter ip6_tables x_tables ipv6 cdc_ether usbnet mii ixgbe mdio igb i2c_algo_bit i2c_core ptp crc32c_sparc64 camellia_sparc64 des_sparc64 des_generic md5_sparc64 sha512_sparc64 sha1_sparc64 uio_pdrv_genirq uio usb_storage mpt3sas scsi_transport_sas raid_class aes_sparc64 sunvnet sunvdc sha256_sparc64(E) sha256_generic(E)
CPU: 211 PID: 52565 Comm: ld Tainted: G        W   E   4.1.0-rc8+ #19
Call Trace:
 [000000000045ce30] warn_slowpath_common+0x7c/0xa0
 [000000000045ceec] warn_slowpath_fmt+0x30/0x40
 [000000000098ad64] do_sparc64_fault+0x340/0x70c
 [0000000000407c2c] sparc64_realfault_common+0x10/0x20
---[ end trace 62ee02065a01a049 ]---
ld[52565]: segfault at fff80001004873c0 ip fff80001004873c0 (rpc fff8000100158868) sp 000007feffcd70e1 error 30002 in libc-2.12.so[fff8000100410000+184000]

The segfault is horrible, but better than a system panic.

An 8-cpu VM on a T5-2 also showed the above traces from time to time,
so it is a general problem and not specific to the T7 or baremetal.

Signed-off-by: David Ahern <david.ahern@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
sparc: perf: Disable pagefaults while walking userspace stacks 2015-06-25T13:01:02+00:00 David Ahern david.ahern@oracle.com 2015-06-15T20:15:43+00:00 c17af4dd96aa99e6e58b5d715a7c66db63a15106 Page faults generated walking userspace stacks can call schedule to switch out the task. When collecting callchains for scheduler tracepoints this causes a deadlock as the tracepoints can be hit with the runqueue lock held: [ 8138.159054] WARNING: CPU: 758 PID: 12488 at /opt/dahern/linux.git/arch/sparc/kernel/nmi.c:80 perfctr_irq+0x1f8/0x2b4() [ 8138.203152] Watchdog detected hard LOCKUP on cpu 758 [ 8138.410969] CPU: 758 PID: 12488 Comm: perf Not tainted 4.0.0-rc6+ #6 [ 8138.437146] Call Trace: [ 8138.447193] [000000000045cdd4] warn_slowpath_common+0x7c/0xa0 [ 8138.471238] [000000000045ce90] warn_slowpath_fmt+0x30/0x40 [ 8138.494189] [0000000000983e38] perfctr_irq+0x1f8/0x2b4 [ 8138.515716] [00000000004209f4] tl0_irq15+0x14/0x20 [ 8138.535791] [00000000009839ec] _raw_spin_trylock_bh+0x68/0x108 [ 8138.560180] [0000000000980018] __schedule+0xcc/0x710 [ 8138.580981] [00000000009806dc] preempt_schedule_common+0x10/0x3c [ 8138.606082] [000000000098077c] _cond_resched+0x34/0x44 [ 8138.627603] [0000000000565990] kmem_cache_alloc_node+0x24/0x1a0 [ 8138.652345] [0000000000450b60] tsb_grow+0xac/0x488 [ 8138.672429] [0000000000985040] do_sparc64_fault+0x4dc/0x6e4 [ 8138.695736] [0000000000407c2c] sparc64_realfault_common+0x10/0x20 [ 8138.721202] [00000000006f2e24] NG4copy_from_user+0xa4/0x3c0 [ 8138.744510] [000000000044f900] perf_callchain_user+0x5c/0x6c [ 8138.768182] [0000000000517b5c] perf_callchain+0x16c/0x19c [ 8138.790774] [0000000000515f84] perf_prepare_sample+0x68/0x218 [ 8138.814801] ---[ end trace 42ca6294b1ff7573 ]--- As with PowerPC (b59a1bfcc240, "powerpc/perf: Disable pagefaults during callchain stack read") disable pagefaults while walking userspace stacks. Signed-off-by: David Ahern <david.ahern@oracle.com> Signed-off-by: David S. Miller <davem@davemloft.net> 
Page faults generated walking userspace stacks can call schedule to switch
out the task. When collecting callchains for scheduler tracepoints this
causes a deadlock as the tracepoints can be hit with the runqueue lock held:

[ 8138.159054] WARNING: CPU: 758 PID: 12488 at /opt/dahern/linux.git/arch/sparc/kernel/nmi.c:80 perfctr_irq+0x1f8/0x2b4()

[ 8138.203152] Watchdog detected hard LOCKUP on cpu 758

[ 8138.410969] CPU: 758 PID: 12488 Comm: perf Not tainted 4.0.0-rc6+ #6
[ 8138.437146] Call Trace:
[ 8138.447193]  [000000000045cdd4] warn_slowpath_common+0x7c/0xa0
[ 8138.471238]  [000000000045ce90] warn_slowpath_fmt+0x30/0x40
[ 8138.494189]  [0000000000983e38] perfctr_irq+0x1f8/0x2b4
[ 8138.515716]  [00000000004209f4] tl0_irq15+0x14/0x20
[ 8138.535791]  [00000000009839ec] _raw_spin_trylock_bh+0x68/0x108
[ 8138.560180]  [0000000000980018] __schedule+0xcc/0x710
[ 8138.580981]  [00000000009806dc] preempt_schedule_common+0x10/0x3c
[ 8138.606082]  [000000000098077c] _cond_resched+0x34/0x44
[ 8138.627603]  [0000000000565990] kmem_cache_alloc_node+0x24/0x1a0
[ 8138.652345]  [0000000000450b60] tsb_grow+0xac/0x488
[ 8138.672429]  [0000000000985040] do_sparc64_fault+0x4dc/0x6e4
[ 8138.695736]  [0000000000407c2c] sparc64_realfault_common+0x10/0x20
[ 8138.721202]  [00000000006f2e24] NG4copy_from_user+0xa4/0x3c0
[ 8138.744510]  [000000000044f900] perf_callchain_user+0x5c/0x6c
[ 8138.768182]  [0000000000517b5c] perf_callchain+0x16c/0x19c
[ 8138.790774]  [0000000000515f84] perf_prepare_sample+0x68/0x218
[ 8138.814801] ---[ end trace 42ca6294b1ff7573 ]---

As with PowerPC (b59a1bfcc240, "powerpc/perf: Disable pagefaults during
callchain stack read") disable pagefaults while walking userspace stacks.

Signed-off-by: David Ahern <david.ahern@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>