Skip to content

GitLab

Switch branch/tag
  1. 23 Nov, 2010 4 commits
  3. 22 Nov, 2010 1 commit
    • Konrad Rzeszutek Wilk's avatar
      xen: set IO permission early (before early_cpu_init()) · ec35a69c
      Konrad Rzeszutek Wilk authored 
      This patch is based off "xen dom0: Set up basic IO permissions for dom0."
by Juan Quintela <quintela@redhat.com>.

On AMD machines when we boot the kernel as Domain 0 we get this nasty:

mapping kernel into physical memory
Xen: setup ISA identity maps
about to get started...
(XEN) traps.c:475:d0 Unhandled general protection fault fault/trap [#13] on VCPU 0 [ec=0000]
(XEN) domain_crash_sync called from entry.S
(XEN) Domain 0 (vcpu#0) crashed on cpu#0:
(XEN) ----[ Xen-4.1-101116  x86_64  debug=y  Not tainted ]----
(XEN) CPU:    0
(XEN) RIP:    e033:[<ffffffff8130271b>]
(XEN) RFLAGS: 0000000000000282   EM: 1   CONTEXT: pv guest
(XEN) rax: 000000008000c068   rbx: ffffffff8186c680   rcx: 0000000000000068
(XEN) rdx: 0000000000000cf8   rsi: 000000000000c000   rdi: 0000000000000000
(XEN) rbp: ffffffff81801e98   rsp: ffffffff81801e50   r8:  ffffffff81801eac
(XEN) r9:  ffffffff81801ea8   r10: ffffffff81801eb4   r11: 00000000ffffffff
(XEN) r12: ffffffff8186c694   r13: ffffffff81801f90   r14: ffffffffffffffff
(XEN) r15: 0000000000000000   cr0: 000000008005003b   cr4: 00000000000006f0
(XEN) cr3: 0000000221803000   cr2: 0000000000000000
(XEN) ds: 0000   es: 0000   fs: 0000   gs: 0000   ss: e02b   cs: e033
(XEN) Guest stack trace from rsp=ffffffff81801e50:

RIP points to read_pci_config() function.

The issue is that we don't set IO permissions for the Linux kernel early enough.

The call sequence used to be:

    xen_start_kernel()
	x86_init.oem.arch_setup = xen_setup_arch;
        setup_arch:
           - early_cpu_init
               - early_init_amd
                  - read_pci_config
           - x86_init.oem.arch_setup [ xen_arch_setup ]
               - set IO permissions.

We need to set the IO permissions earlier on, which this patch does.
Acked-by: default avatarJeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: default avatarKonrad Rzeszutek Wilk <konrad.wilk@oracle.com>
      ec35a69c
  5. 20 Nov, 2010 1 commit
  7. 17 Nov, 2010 13 commits
    • Alessio Igor Bogani's avatar
      powerpc: Update a BKL related comment · 0f6b77ca
      Alessio Igor Bogani authored 
      The commit 5e3d20a6 remove bkl from startup code so setup_arch() it isn't called
with bkl held anymore. Update the comment on top of that function.
Fix also a typo.

This work was supported by a hardware donation from the CE Linux Forum.
Signed-off-by: default avatarAlessio Igor Bogani <abogani@texware.it>
Signed-off-by: default avatarBenjamin Herrenschmidt <benh@kernel.crashing.org>
      0f6b77ca
    • Kumar Gala's avatar
      powerpc/mm: Fix module instruction tlb fault handling on Book-E 64 · 82ae5eaf
      Kumar Gala authored 
      We were seeing oops like the following when we did an rmmod on a module:

Unable to handle kernel paging request for instruction fetch
Faulting instruction address: 0x8000000000008010
Oops: Kernel access of bad area, sig: 11 [#1]
SMP NR_CPUS=2 P5020 DS
last sysfs file: /sys/devices/qman-portals.2/qman-pool.9/uevent
Modules linked in: qman_tester(-)
NIP: 8000000000008010 LR: c000000000074858 CTR: 8000000000008010
REGS: c00000002e29bab0 TRAP: 0400   Not tainted
(2.6.34.6-00744-g2d21f14)
MSR: 0000000080029000 <EE,ME,CE>  CR: 24000448  XER: 00000000
TASK = c00000007a8be600[4987] 'rmmod' THREAD: c00000002e298000 CPU: 1
GPR00: 8000000000008010 c00000002e29bd30 8000000000012798 c00000000035fb28
GPR04: 0000000000000002 0000000000000002 0000000024022428 c000000000009108
GPR08: fffffffffffffffe 800000000000a618 c0000000003c13c8 0000000000000000
GPR12: 0000000022000444 c00000000fffed00 0000000000000000 0000000000000000
GPR16: 00000000100c0000 0000000000000000 00000000100dabc8 0000000010099688
GPR20: 0000000000000000 00000000100cfc28 0000000000000000 0000000010011a44
GPR24: 00000000100017b2 0000000000000000 0000000000000000 0000000000000880
GPR28: c00000000035fb28 800000000000a7b8 c000000000376d80 c0000000003cce50
NIP [8000000000008010] .test_exit+0x0/0x10 [qman_tester]
LR [c000000000074858] .SyS_delete_module+0x1f8/0x2f0
Call Trace:
[c00000002e29bd30] [c0000000000748b4] .SyS_delete_module+0x254/0x2f0 (unreliable)
[c00000002e29be30] [c000000000000580] syscall_exit+0x0/0x2c
Instruction dump:
XXXXXXXX XXXXXXXX XXXXXXXX XXXXXXXX XXXXXXXX XXXXXXXX XXXXXXXX XXXXXXXX
38600000 4e800020 60000000 60000000 <4e800020> 60000000 60000000 60000000
---[ end trace 4f57124939a84dc8 ]---

This appears to be due to checking the wrong permission bits in the
instruction_tlb_miss handling if the address that faulted was in vmalloc
space.  We need to look at the supervisor execute (_PAGE_BAP_SX) bit and
not the user bit (_PAGE_BAP_UX/_PAGE_EXEC).

Also removed a branch level since it did not appear to be used.
Reported-by: default avatarJeffrey Ladouceur <Jeffrey.Ladouceur@freescale.com>
Signed-off-by: default avatarKumar Gala <galak@kernel.crashing.org>
Signed-off-by: default avatarBenjamin Herrenschmidt <benh@kernel.crashing.org>
      82ae5eaf
    • Michael Neuling's avatar
      powerpc: Fix call to subpage_protection() · 1c2c25c7
      Michael Neuling authored 
      In:
  powerpc/mm: Fix pgtable cache cleanup with CONFIG_PPC_SUBPAGE_PROT
  commit d28513bc
  Author: David Gibson <david@gibson.dropbear.id.au>

subpage_protection() was changed to to take an mm rather a pgdir but it
didn't change calling site in hashpage_preload().  The change wasn't
noticed at compile time since hashpage_preload() used a void* as the
parameter to subpage_protection().

This is obviously wrong and can trigger the following crash when
CONFIG_SLAB, CONFIG_DEBUG_SLAB, CONFIG_PPC_64K_PAGES
CONFIG_PPC_SUBPAGE_PROT are enabled.

Freeing unused kernel memory: 704k freed
Unable to handle kernel paging request for data at address 0x6b6b6b6b6b6c49b7
Faulting instruction address: 0xc0000000000410f4
cpu 0x2: Vector: 300 (Data Access) at [c00000004233f590]
    pc: c0000000000410f4: .hash_preload+0x258/0x338
    lr: c000000000041054: .hash_preload+0x1b8/0x338
    sp: c00000004233f810
   msr: 8000000000009032
   dar: 6b6b6b6b6b6c49b7
 dsisr: 40000000
  current = 0xc00000007e2c0070
  paca    = 0xc000000007fe0500
    pid   = 1, comm = init
enter ? for help
[c00000004233f810] c000000000041020 .hash_preload+0x184/0x338 (unreliable)
[c00000004233f8f0] c00000000003ed98 .update_mmu_cache+0xb0/0xd0
[c00000004233f990] c000000000157754 .__do_fault+0x48c/0x5dc
[c00000004233faa0] c000000000158fd0 .handle_mm_fault+0x508/0xa8c
[c00000004233fb90] c0000000006acdd4 .do_page_fault+0x428/0x6ac
[c00000004233fe30] c000000000005260 handle_page_fault+0x20/0x74
Signed-off-by: default avatarBenjamin Herrenschmidt <benh@kernel.crashing.org>
      1c2c25c7
    • kerstin jonsson's avatar
      powerpc: Set CONFIG_32BIT on ppc32 · 234a71a7
      kerstin jonsson authored 
      commit ffe8018c of the -mm tree
fixes the initramfs size calculation for e.g. s390 but breaks it
for 32bit architectures which do not define CONFIG_32BIT.

This patch fix the problem for PPC32 which will elsewise end up
with a __initramfs_size of 0.
Signed-off-by: default avatarKerstin Jonsson <kerstin.jonsson@ericsson.com>
Signed-off-by: default avatarBenjamin Herrenschmidt <benh@kernel.crashing.org>
      234a71a7
    • Kumar Gala's avatar
      powerpc/mm: Fix build error in setup_initial_memory_limit · 4a89261b
      Kumar Gala authored 
      arch/powerpc/mm/tlb_nohash.c: In function 'setup_initial_memory_limit':
arch/powerpc/mm/tlb_nohash.c:588:29: error: 'ppc64_memblock_base' undeclared (first use in this function)
arch/powerpc/mm/tlb_nohash.c:588:29: note: each undeclared identifier is reported only once for each function it appears in

Due to a copy/paste typo with the following commit:

	commit cd3db0c4
	Author: Benjamin Herrenschmidt <benh@kernel.crashing.org>
	Date:   Tue Jul 6 15:39:02 2010 -0700

	    memblock: Remove rmo_size, burry it in arch/powerpc where it belongs
Signed-off-by: default avatarKumar Gala <galak@kernel.crashing.org>
Signed-off-by: default avatarBenjamin Herrenschmidt <benh@kernel.crashing.org>
      4a89261b
    • Nishanth Aravamudan's avatar
      powerpc/pseries: Don't override CONFIG_PPC_PSERIES_DEBUG · 01cf6fe8
      Nishanth Aravamudan authored 
      EEH and pci_dlpar #undef DEBUG, but I think they were added before the
ability to control this from Kconfig. It's really annoying to only get
some of the debug messages from these files. Leave the lpar.c #undef
alone as it produces so much output as to make the kernel unusable.
Update the Kconfig text to indicate this particular quirk :)
Signed-off-by: default avatarNishanth Aravamudan <nacc@us.ibm.com>
Acked-by: default avatarMichael Ellerman <michael@ellerman.id.au>
Signed-off-by: default avatarBenjamin Herrenschmidt <benh@kernel.crashing.org>
      01cf6fe8
    • Benjamin Herrenschmidt's avatar
      powerpc: Fix div64 in bootloader · acbfd58e
      Benjamin Herrenschmidt authored 
      The code is missing a fix that went into the main kernel variant
(we should try to share that code again at some stage)
Reported-by: default avatarAlbert Cahalan <acahalan@gmail.com>
Signed-off-by: default avatarBenjamin Herrenschmidt <benh@kernel.crashing.org>
      acbfd58e
    • Avi Kivity's avatar
      KVM: VMX: Fix host userspace gsbase corruption · c8770e7b
      Avi Kivity authored 
      We now use load_gs_index() to load gs safely; unfortunately this also
changes MSR_KERNEL_GS_BASE, which we managed separately.  This resulted
in confusion and breakage running 32-bit host userspace on a 64-bit kernel.

Fix by
- saving guest MSR_KERNEL_GS_BASE before we we reload the host's gs
- doing the host save/load unconditionally, instead of only when in guest
  long mode

Things can be cleaned up further, but this is the minmal fix for now.
Signed-off-by: default avatarAvi Kivity <avi@redhat.com>
Signed-off-by: default avatarMarcelo Tosatti <mtosatti@redhat.com>
      c8770e7b
    • Avi Kivity's avatar
      KVM: Correct ordering of ldt reload wrt fs/gs reload · 0a77fe4c
      Avi Kivity authored 
      If fs or gs refer to the ldt, they must be reloaded after the ldt.  Reorder
the code to that effect.

Userspace code that uses the ldt with kvm is nonexistent, so this doesn't fix
a user-visible bug.
Signed-off-by: default avatarAvi Kivity <avi@redhat.com>
Signed-off-by: default avatarMarcelo Tosatti <mtosatti@redhat.com>
      0a77fe4c
    • Dongdong Deng's avatar
      kgdb,ppc: Fix regression in evr register handling · e3839ed8
      Dongdong Deng authored 
      Commit ff10b88b (kgdb,ppc: Individual
register get/set for ppc) introduced a problem where memcpy was used
incorrectly to read and write the evr registers with a kernel that
has:

CONFIG_FSL_BOOKE=y
CONFIG_SPE=y
CONFIG_KGDB=y

This patch also fixes the following compilation problems:

arch/powerpc/kernel/kgdb.c: In function 'dbg_get_reg':
arch/powerpc/kernel/kgdb.c:341: error: passing argument 2 of 'memcpy' makes pointer from integer without a cast
arch/powerpc/kernel/kgdb.c: In function 'dbg_set_reg':
arch/powerpc/kernel/kgdb.c:366: error: passing argument 1 of 'memcpy' makes pointer from integer without a cast

[jason.wessel@windriver.com: Remove void * casts and fix patch header]
Reported-by: default avatarMilton Miller <miltonm@bga.com>
Signed-off-by: default avatarDongdong Deng <dongdong.deng@windriver.com>
Acked-by: default avatarKumar Gala <galak@kernel.crashing.org>
Signed-off-by: default avatarJason Wessel <jason.wessel@windriver.com>
CC: linuxppc-dev@lists.ozlabs.org
      e3839ed8
    • Jason Wessel's avatar
      kgdb,x86: fix regression in detach handling · 10a6e676
      Jason Wessel authored 
      The fix from ba773f7c
(x86,kgdb: Fix hw breakpoint regression) was not entirely complete.

The kgdb_remove_all_hw_break() function also needs to call the
hw_break_release_slot() or else a breakpoint can get activated again
after the debugger has detached.

The kgdb test suite exposes the behavior in the form of either a hang
or repetitive failure.  The kernel config that exposes the problem
contains all of the following:

CONFIG_DEBUG_RODATA=y
CONFIG_KGDB_TESTS=y
CONFIG_KGDB_TESTS_ON_BOOT=y
CONFIG_KGDB_TESTS_BOOT_STRING="V1F100"
Reported-by: default avatarFrederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: default avatarJason Wessel <jason.wessel@windriver.com>
Tested-by: default avatarFrederic Weisbecker <fweisbec@gmail.com>
      10a6e676
    • Chris Paulson-Ellis's avatar
      ASoC: davinci: fixes for multi-component · bedad0ca
      Chris Paulson-Ellis authored 
      Multi-component commit f0fba2ad broke a few things which this patch should
fix. Tested on the DM355 EVM. I've been as careful as I can, but it would be
good if those with access to other Davinci boards could test.

--

The multi-component commit put the initialisation of
snd_soc_dai.[capture|playback]_dma_data into snd_soc_dai_ops.hw_params of the
McBSP, McASP & VCIF drivers (davinci-i2s.c, davinci-mcasp.c & davinci-vcif.c).
The initialisation had to be moved from the probe function in these drivers
because davinci_*_dai changed from snd_soc_dai to snd_soc_dai_driver.

Unfortunately, the DMA params pointer is needed by davinci_pcm_open (in
davinci-pcm.c) before hw_params is called. I have moved the initialisation to
a new snd_soc_dai_ops.startup function in each of these drivers. This fix
indicates that all platforms that use davinci-pcm must have been broken and
need to test with this fix.

--

The multi-component commit also changed the McBSP driver name from
"davinci-asp" to "davinci-i2s" in davinci-i2s.c without updating the board
level references to the driver name. This change is understandable, as there
is a similarly named "davinci-mcasp" driver in davinci-mcasp.c.

There is probably no 'correct' name for this driver. The DM6446 datasheet
calls it the "ASP" and describes it as a "specialised McBSP". The DM355
datasheet calls it the "ASP" and describes it as a "specialised ASP". The
DM365 datasheet calls it the "McBSP". Rather than fix this problem by
reverting to "davinci-asp", I've elected to avoid future confusion with the
"davinci-mcasp" driver by changing it to "davinci-mcbsp", which is also
consistent with the names of the functions in the driver. There are other
fixes required, so it was never going to be as simple as a revert anyway.

--

The DM365 only has one McBSP port (of the McBSP platforms, only the DM355 has
2 ports), so I've changed the the id of the platform_device from 0 to -1.

--

In davinci-evm.c, the DM6446 EVM can no longer share a snd_soc_dai_link
structure with the DM355 EVM as they use different cpu DAI names (the DM355
has 2 ports and the EVM uses the second port, but the DM6446 only has 1 port).
This also means that the 2 boards need different snd_soc_card structures.

--

The codec_name entries in davinci-evm.c didn't match the i2c ids in the board
files. I have only checked and fixed the details of the names used for the
McBSP based platforms. Someone with a McASP based platform (eg DA8xx) should
check the others.
Signed-off-by: default avatarChris Paulson-Ellis <chris@edesix.com>
Acked-by: default avatarLiam Girdwood <lrg@slimlogic.co.uk>
Signed-off-by: default avatarMark Brown <broonie@opensource.wolfsonmicro.com>
      bedad0ca
    • Arnd Bergmann's avatar
      BKL: remove extraneous #include <smp_lock.h> · 451a3c24
      Arnd Bergmann authored 
      The big kernel lock has been removed from all these files at some point,
leaving only the #include.

Remove this too as a cleanup.
Signed-off-by: default avatarArnd Bergmann <arnd@arndb.de>
Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
      451a3c24
  9. 16 Nov, 2010 1 commit
    • Jeff Garzik's avatar
      SCSI host lock push-down · f281233d
      Jeff Garzik authored 
      Move the mid-layer's ->queuecommand() invocation from being locked
with the host lock to being unlocked to facilitate speeding up the
critical path for drivers who don't need this lock taken anyway.

The patch below presents a simple SCSI host lock push-down as an
equivalent transformation.  No locking or other behavior should change
with this patch.  All existing bugs and locking orders are preserved.

Additionally, add one parameter to queuecommand,
	struct Scsi_Host *
and remove one parameter from queuecommand,
	void (*done)(struct scsi_cmnd *)

Scsi_Host* is a convenient pointer that most host drivers need anyway,
and 'done' is redundant to struct scsi_cmnd->scsi_done.

Minimal code disturbance was attempted with this change.  Most drivers
needed only two one-line modifications for their host lock push-down.
Signed-off-by: default avatarJeff Garzik <jgarzik@redhat.com>
Acked-by: default avatarJames Bottomley <James.Bottomley@suse.de>
Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
      f281233d
  11. 14 Nov, 2010 1 commit
  13. 13 Nov, 2010 3 commits
  15. 12 Nov, 2010 3 commits
  17. 11 Nov, 2010 2 commits
  19. 10 Nov, 2010 11 commits