aboutsummaryrefslogtreecommitdiff
AgeCommit message (Collapse)AuthorFilesLines
2012-12-20CRIS: locking: fix the return value of arch_read_trylock()Wei Yongjun1-1/+1
arch_write_trylock() should return 'ret' instead of always return 1. dpatch engine is used to auto generate this patch. (https://github.com/weiyj/dpatch) Signed-off-by: Wei Yongjun <[email protected]> Signed-off-by: Jesper Nilsson <[email protected]>
2012-12-20Merge tag 'disintegrate-cris-20121009' of ↵Jesper Nilsson53-639/+682
git://git.infradead.org/users/dhowells/linux-headers into for-linus2 UAPI Disintegration 2012-10-09 * tag 'disintegrate-cris-20121009' of git://git.infradead.org/users/dhowells/linux-headers: UAPI: (Scripted) Disintegrate arch/cris/include/asm UAPI: (Scripted) Disintegrate arch/cris/include/arch-v32/arch UAPI: (Scripted) Disintegrate arch/cris/include/arch-v10/arch
2012-12-20CRIS: use kbuild.h instead of defining macros in asm-offset.cJames Hogan1-5/+1
This is modelled on commits such as the one below: Commit fc1c3a003edb8a6778e64e10ef671a38c76c969e ("sh: use kbuild.h instead of defining macros in asm-offsets.c") introduced in v2.6.26. Signed-off-by: James Hogan <[email protected]> Signed-off-by: Jesper Nilsson <[email protected]>
2012-12-20UAPI: (Scripted) Disintegrate arch/score/include/asmDavid Howells35-80/+118
Signed-off-by: David Howells <[email protected]> Acked-by: Arnd Bergmann <[email protected]> Acked-by: Thomas Gleixner <[email protected]> Acked-by: Michael Kerrisk <[email protected]> Acked-by: Paul E. McKenney <[email protected]> Acked-by: Dave Jones <[email protected]> Acked-by: Lennox Wu <[email protected]> Acked-by: Liqin Chen <[email protected]>
2012-12-20ARM: 7607/1: realview: fix private peripheral memory base for EB rev. B boardsWill Deacon1-1/+1
Commit 34ae6c96a6a7 ("ARM: 7298/1: realview: fix mapping of MPCore private memory region") accidentally broke the definition for the base address of the private peripheral region on revision B Realview-EB boards. This patch uses the correct address for REALVIEW_EB11MP_PRIV_MEM_BASE. Cc: <[email protected]> Acked-by: Marc Zyngier <[email protected]> Tested-by: Florian Fainelli <[email protected]> Signed-off-by: Will Deacon <[email protected]> Signed-off-by: Russell King <[email protected]>
2012-12-20ARM: 7606/1: cache: flush to LoUU instead of LoUIS on uniprocessor CPUsWill Deacon1-2/+4
flush_cache_louis flushes the D-side caches to the point of unification inner-shareable. On uniprocessor CPUs, this is defined as zero and therefore no flushing will take place. Rather than invent a new interface for UP systems, instead use our SMP_ON_UP patching code to read the LoUU from the CLIDR instead. Cc: <[email protected]> Cc: Lorenzo Pieralisi <[email protected]> Tested-by: Guennadi Liakhovetski <[email protected]> Signed-off-by: Will Deacon <[email protected]> Signed-off-by: Russell King <[email protected]>
2012-12-20ARM: missing ->mmap_sem around find_vma() in swp_emulate.cAl Viro1-0/+2
find_vma() is *not* safe when somebody else is removing vmas. Not just the return value might get bogus just as you are getting it (this instance doesn't try to dereference the resulting vma), the search itself can get buggered in rather spectacular ways. IOW, ->mmap_sem really, really is not optional here. Signed-off-by: Al Viro <[email protected]> Cc: <[email protected]> Signed-off-by: Russell King <[email protected]>
2012-12-19Merge branches 'cxgb4', 'ipoib' and 'mlx4' into for-nextRoland Dreier9-45/+134
2012-12-19RDMA/cxgb4: Fix bug for active and passive LE hash collision pathVipul Pandya7-88/+376
Retries active opens for INUSE errors. Logs any active ofld_connect_wr error replies. Sends ofld_connect_wr on same ctrlq. It needs to go on the same control txq as regular CPL active/passive messages. Retries on active open replies with EADDRINUSE. Uses active open fw wr only if active filter region is set. Adds stat for ofld_connect_wr failures. This patch also adds debugfs file to show endpoints. Signed-off-by: Vipul Pandya <[email protected]> Signed-off-by: Roland Dreier <[email protected]>
2012-12-19RDMA/cxgb4: Fix LE hash collision bug for passive open connectionVipul Pandya4-55/+497
It establishes passive open connection through firmware work request. Passive open connection will go through this path as now instead of listening server we create a server filter which will redirect the incoming SYN packet to the offload queue. After this driver tries to establish the connection using firmware work request. Signed-off-by: Vipul Pandya <[email protected]> Signed-off-by: Roland Dreier <[email protected]>
2012-12-19RDMA/cxgb4: Fix LE hash collision bug for active open connectionVipul Pandya6-47/+315
It enables establishing active open connection using fw_ofld_connection work request when cpl_act_open_rpl says TCAM full error which may be because of LE hash collision. Current support is only for IPv4 active open connections. Sets ntuple bits in active open requests. For T4 firmware greater than 1.4.10.0 ntuple bits are required to be set. Adds nocong and enable_ecn module parameter options. Signed-off-by: Vipul Pandya <[email protected]> [ Move all FW return values to t4fw_api.h. - Roland ] Signed-off-by: Roland Dreier <[email protected]>
2012-12-20dma-buf: remove fallback for !CONFIG_DMA_SHARED_BUFFERMaarten Lankhorst1-99/+0
Documentation says that code requiring dma-buf should add it to select, so inline fallbacks are not going to be used. A link error will make it obvious what went wrong, instead of silently doing nothing at runtime. Signed-off-by: Maarten Lankhorst <[email protected]> Reviewed-by: Daniel Vetter <[email protected]> Reviewed-by: Rob Clark <[email protected]> Signed-off-by: Sumit Semwal <[email protected]>
2012-12-20dma-buf: might_sleep() in dma_buf_unmap_attachment()Rob Clark1-0/+2
We never really clarified if unmap could be done in atomic context. But since mapping might require sleeping, this implies mutex in use to synchronize mapping/unmapping, so unmap could sleep as well. Add a might_sleep() to clarify this. Signed-off-by: Rob Clark <[email protected]> Acked-by: Daniel Vetter <[email protected]> Reviewed-by: Maarten Lankhorst <[email protected]> Signed-off-by: Sumit Semwal <[email protected]>
2012-12-19Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparcLinus Torvalds7-12/+67
Pull sparc fixes from David Miller: "Please pull to get these sparc AES/DES/CAMELLIA crypto bug fixes as well as an addition of a pte_accessible() define for sparc64 and a hugetlb fix from Dave Kleikamp." * git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc: sparc64: Set CRYPTO_TFM_REQ_MAY_SLEEP consistently in CAMELLIA code. sparc64: Set CRYPTO_TFM_REQ_MAY_SLEEP consistently in DES code. sparc64: Fix ECB looping constructs in AES code. sparc64: Set CRYPTO_TFM_REQ_MAY_SLEEP consistently in AES code. sparc64: Fix AES ctr mode block size. sparc64: Fix unrolled AES 256-bit key loops. sparc64: Define pte_accessible() sparc: huge_ptep_set_* functions need to call set_huge_pte_at()
2012-12-19Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/netLinus Torvalds52-202/+545
Pull networking fixes from David Miller: 1) Really fix tuntap SKB use after free bug, from Eric Dumazet. 2) Adjust SKB data pointer to point past the transport header before calling icmpv6_notify() so that the headers are in the state which that function expects. From Duan Jiong. 3) Fix ambiguities in the new tuntap multi-queue APIs. From Jason Wang. 4) mISDN needs to use del_timer_sync(), from Konstantin Khlebnikov. 5) Don't destroy mutex after freeing up device private in mac802154, fix also from Konstantin Khlebnikov. 6) Fix INET request socket leak in TCP and DCCP, from Christoph Paasch. 7) SCTP HMAC kconfig rework, from Neil Horman. 8) Fix SCTP jprobes function signature, otherwise things explode, from Daniel Borkmann. 9) Fix typo in ipv6-offload Makefile variable reference, from Simon Arlott. 10) Don't fail USBNET open just because remote wakeup isn't supported, from Oliver Neukum. 11) be2net driver bug fixes from Sathya Perla. 12) SOLOS PCI ATM driver bug fixes from Nathan Williams and David Woodhouse. 13) Fix MTU changing regression in 8139cp driver, from John Greene. * git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (45 commits) solos-pci: ensure all TX packets are aligned to 4 bytes solos-pci: add firmware upgrade support for new models solos-pci: remove superfluous debug output solos-pci: add GPIO support for newer versions on Geos board 8139cp: Prevent dev_close/cp_interrupt race on MTU change net: qmi_wwan: add ZTE MF880 drivers/net: Use of_match_ptr() macro in smsc911x.c drivers/net: Use of_match_ptr() macro in smc91x.c ipv6: addrconf.c: remove unnecessary "if" bridge: Correctly encode addresses when dumping mdb entries bridge: Do not unregister all PF_BRIDGE rtnl operations use generic usbnet_manage_power() usbnet: generic manage_power() usbnet: handle PM failure gracefully ksz884x: fix receive polling race condition qlcnic: update driver version qlcnic: fix unused variable warnings net: fec: forbid FEC_PTP on SoCs that do not support be2net: fix wrong frag_idx reported by RX CQ be2net: fix be_close() to ensure all events are ack'ed ...
2012-12-19Merge tags 'dt-for-linus', 'gpio-for-linus' and 'spi-for-linus' of ↵Linus Torvalds12-29/+57
git://git.secretlab.ca/git/linux-2.6 Pull devicetree, gpio and spi bugfixes from Grant Likely: "Device tree v3.8 bug fix: - Fixes an undefined struct device build error and a missing symbol export. GPIO device driver bug fixes: - gpio/mvebu-gpio: Make mvebu-gpio depend on OF_CONFIG - gpio/ich: Add missing spinlock init SPI device driver bug fixes: - Most of this is bug fixes to the core code and the sh-hspi and s3c64xx device drivers. - There is also a patch here to add DT support to the Atmel driver. This one should have been in the first round, but I missed it. It's a low risk change contained within a single driver and the Atmel maintainer has requested it." * tag 'dt-for-linus' of git://git.secretlab.ca/git/linux-2.6: of: define struct device in of_platform.h if !OF_DEVICE and !OF_ADDRESS of: Fix export of of_find_matching_node_and_match() * tag 'gpio-for-linus' of git://git.secretlab.ca/git/linux-2.6: gpio/mvebu-gpio: Make mvebu-gpio depend on OF_CONFIG gpio/ich: Add missing spinlock init * tag 'spi-for-linus' of git://git.secretlab.ca/git/linux-2.6: spi/sh-hspi: fix return value check in hspi_probe(). spi: fix tegra SPI binding examples spi/atmel: add DT support of/spi: Fix SPI module loading by using proper "spi:" modalias prefixes. spi: Change FIFO flush operation and spi channel off spi: Keep chipselect assertion during one message
2012-12-19Merge branch 'drm-next' of git://people.freedesktop.org/~airlied/linuxLinus Torvalds1-1/+1
Pull drm bugfix from Dave Airlie: "Just a single urgent regression fix, seeing a few wierd behaviours I'd like not to persist." * 'drm-next' of git://people.freedesktop.org/~airlied/linux: drm/ttm: fix delayed ttm_bo_cleanup_refs_and_unlock delayed handling
2012-12-19Merge tag 'random_for_linus' of ↵Linus Torvalds2-13/+30
git://git.kernel.org/pub/scm/linux/kernel/git/tytso/random Pull random updates from Ted Ts'o: "A few /dev/random improvements for the v3.8 merge window." * tag 'random_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/random: random: Mix cputime from each thread that exits to the pool random: prime last_data value per fips requirements random: fix debug format strings random: make it possible to enable debugging without rebuild
2012-12-19sparc64: Set CRYPTO_TFM_REQ_MAY_SLEEP consistently in CAMELLIA code.David S. Miller1-0/+3
We use the FPU and therefore cannot sleep during the crypto loops. Signed-off-by: David S. Miller <[email protected]>
2012-12-19sparc64: Set CRYPTO_TFM_REQ_MAY_SLEEP consistently in DES code.David S. Miller1-0/+6
We use the FPU and therefore cannot sleep during the crypto loops. Signed-off-by: David S. Miller <[email protected]>
2012-12-19sparc64: Fix ECB looping constructs in AES code.David S. Miller1-0/+1
Things works better when you increment the source buffer pointer properly. Signed-off-by: David S. Miller <[email protected]>
2012-12-19sparc64: Set CRYPTO_TFM_REQ_MAY_SLEEP consistently in AES code.David S. Miller1-0/+4
We use the FPU and therefore cannot sleep during the crypto loops. Signed-off-by: David S. Miller <[email protected]>
2012-12-19sparc64: Fix AES ctr mode block size.David S. Miller1-3/+24
Like the generic versions, we need to support a block size of '1' for CTR mode AES. This was discovered thanks to all of the new test cases added by Jussi Kivilinna. Signed-off-by: David S. Miller <[email protected]>
2012-12-19sparc64: Fix unrolled AES 256-bit key loops.David S. Miller1-6/+14
The basic scheme of the block mode assembler is that we start by enabling the FPU, loading the key into the floating point registers, then iterate calling the encrypt/decrypt routine for each block. For the 256-bit key cases, we run short on registers in the unrolled loops. So the {ENCRYPT,DECRYPT}_256_2() macros reload the key registers that get clobbered. The unrolled macros, {ENCRYPT,DECRYPT}_256(), are not mindful of this. So if we have a mix of multi-block and single-block calls, the single-block unrolled 256-bit encrypt/decrypt can run with some of the key registers clobbered. Handle this by always explicitly loading those registers before using the non-unrolled 256-bit macro. This was discovered thanks to all of the new test cases added by Jussi Kivilinna. Signed-off-by: David S. Miller <[email protected]>
2012-12-19alpha: switch to generic sigaltstackAl Viro2-9/+2
Signed-off-by: Al Viro <[email protected]>
2012-12-19new helpers: __save_altstack/__compat_save_altstack, switch x86 and um to thoseAl Viro6-25/+25
note that they are relying on access_ok() already checked by caller. Signed-off-by: Al Viro <[email protected]>
2012-12-19generic compat_sys_sigaltstack()Al Viro10-67/+67
Again, conditional on CONFIG_GENERIC_SIGALTSTACK Signed-off-by: Al Viro <[email protected]>
2012-12-19introduce generic sys_sigaltstack(), switch x86 and um to itAl Viro14-21/+19
Conditional on CONFIG_GENERIC_SIGALTSTACK; architectures that do not select it are completely unaffected Signed-off-by: Al Viro <[email protected]>
2012-12-19new helper: compat_user_stack_pointer()Al Viro3-2/+13
Compat counterpart of current_user_stack_pointer(); for most of the biarch architectures those two are identical, but e.g. arm64 and arm use different registers for stack pointer... Note that amd64 variants of current_user_stack_pointer/compat_user_stack_pointer do *not* rely on pt_regs having been through FIXUP_TOP_OF_STACK. Signed-off-by: Al Viro <[email protected]>
2012-12-19new helper: restore_altstack()Al Viro2-0/+9
to be used by rt_sigreturn instances Signed-off-by: Al Viro <[email protected]>
2012-12-19unify SS_ONSTACK/SS_DISABLE definitionsAl Viro18-103/+2
Signed-off-by: Al Viro <[email protected]>
2012-12-19new helper: current_user_stack_pointer()Al Viro7-0/+14
Cross-architecture equivalent of rdusp(); default is user_stack_pointer(current_pt_regs()) - that works for almost all platforms that have usp saved in pt_regs. The only exception from that is ia64 - we want memory stack, not the backing store for register one. Signed-off-by: Al Viro <[email protected]>
2012-12-19missing user_stack_pointer() instancesAl Viro9-0/+10
for the architectures that have usp in pt_regs and do not have user_stack_pointer() already defined. Signed-off-by: Al Viro <[email protected]>
2012-12-19Bury the conditionals from kernel_thread/kernel_execve seriesAl Viro63-131/+6
All architectures have CONFIG_GENERIC_KERNEL_THREAD CONFIG_GENERIC_KERNEL_EXECVE __ARCH_WANT_SYS_EXECVE None of them have __ARCH_WANT_KERNEL_EXECVE and there are only two callers of kernel_execve() (which is a trivial wrapper for do_execve() now) left. Kill the conditionals and make both callers use do_execve(). Signed-off-by: Al Viro <[email protected]>
2012-12-19COMPAT_SYSCALL_DEFINE: infrastructureAl Viro2-0/+45
Signed-off-by: Al Viro <[email protected]>
2012-12-19solos-pci: ensure all TX packets are aligned to 4 bytesDavid Woodhouse1-4/+23
The FPGA can't handled unaligned DMA (yet). So copy into an aligned buffer, if skb->data isn't suitably aligned. Signed-off-by: David Woodhouse <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2012-12-19solos-pci: add firmware upgrade support for new modelsNathan Williams1-11/+42
Signed-off-by: Nathan Williams <[email protected]> Signed-off-by: David Woodhouse <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2012-12-19solos-pci: remove superfluous debug outputNathan Williams1-1/+0
Signed-off-by: Nathan Williams <[email protected]> Signed-off-by: David Woodhouse <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2012-12-19solos-pci: add GPIO support for newer versions on Geos boardNathan Williams1-0/+105
dwmw2: Tidy up a little, simpler matching on which GPIO is being accessed, only register on newer boards, register under PCI device instead of duplicating them under each ATM device. Signed-off-by: Nathan Williams <[email protected]> Signed-off-by: David Woodhouse <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2012-12-198139cp: Prevent dev_close/cp_interrupt race on MTU changeJohn Greene1-7/+11
commit: cb64edb6b89491edfdbae52ba7db9a8b8391d339 upstream Above commit may introduce a race between cp_interrupt and dev_close / change MTU / dev_open up state. Changes cp_interrupt to tolerate this. Change spin_locking in cp_interrupt to avoid possible but unobserved race. Reported-by: "Francois Romieu" <[email protected]> Tested on virtual hardware, Tx MTU size up to 4096, max tx payload was ping -s 4068 for MTU of 4096. No real hardware, need test assist. Signed-off-by: "John Greene" <[email protected]> CC: "David S. Miller" <[email protected]> CC: "David Woodhouse" <[email protected]> Tested-by: David Woodhouse <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2012-12-19gpio/mvebu-gpio: Make mvebu-gpio depend on OF_CONFIGGregory CLEMENT2-17/+1
When building without device tree support (for the mv78x00 SoCs for example), the build failed because of_irq_count is undeclared. However mvebu-gpio is not designed to build without device tree support. So make it depends on OF_CONFIG, remove the #ifdef OF_CONFIG line and the platform_device_id. Tested on RD-78x00-mASA, DB-78460-BP, DB-88F6710-BP-DDR3, DB-MV784MP-GP, Mirabox and OpenBlocks AX3. Signed-off-by: Gregory CLEMENT <[email protected]> Signed-off-by: Grant Likely <[email protected]>
2012-12-19gpio/ich: Add missing spinlock initJean Delvare1-0/+1
As reported by CONFIG_DEBUG_SPINLOCK=y. Signed-off-by: Jean Delvare <[email protected]> Acked-by: Peter Tyser <[email protected]> Signed-off-by: Grant Likely <[email protected]>
2012-12-20drm/ttm: fix delayed ttm_bo_cleanup_refs_and_unlock delayed handlingMaarten Lankhorst1-1/+1
Fix regression introduced by 85b144f860176 "drm/ttm: call ttm_bo_cleanup_refs with reservation and lru lock held, v3" Slowpath ttm_bo_cleanup_refs_and_unlock accidentally tried to increase refcount on &bo->sync_obj instead of bo->sync_obj. The compiler didn't complain since sync_obj_ref takes a void pointer, so it was still valid c. This could result in lockups, memory corruptions, and warnings like these when graphics card VRAM usage is high: ------------[ cut here ]------------ WARNING: at include/linux/kref.h:42 radeon_fence_ref+0x2c/0x40() Hardware name: System Product Name Pid: 157, comm: X Not tainted 3.7.0-rc7-00520-g85b144f-dirty #174 Call Trace: [<ffffffff81058c84>] ? warn_slowpath_common+0x74/0xb0 [<ffffffff8129273c>] ? radeon_fence_ref+0x2c/0x40 [<ffffffff8125e95c>] ? ttm_bo_cleanup_refs_and_unlock+0x18c/0x2d0 [<ffffffff8125f17c>] ? ttm_mem_evict_first+0x1dc/0x2a0 [<ffffffff81264452>] ? ttm_bo_man_get_node+0x62/0xb0 [<ffffffff8125f4ce>] ? ttm_bo_mem_space+0x28e/0x340 [<ffffffff8125fb0c>] ? ttm_bo_move_buffer+0xfc/0x170 [<ffffffff810de172>] ? kmem_cache_alloc+0xb2/0xc0 [<ffffffff8125fc15>] ? ttm_bo_validate+0x95/0x110 [<ffffffff8125ff7c>] ? ttm_bo_init+0x2ec/0x3b0 [<ffffffff8129419a>] ? radeon_bo_create+0x18a/0x200 [<ffffffff81293e80>] ? radeon_bo_clear_va+0x40/0x40 [<ffffffff812a5342>] ? radeon_gem_object_create+0x92/0x160 [<ffffffff812a575c>] ? radeon_gem_create_ioctl+0x6c/0x150 [<ffffffff812a529f>] ? radeon_gem_object_free+0x2f/0x40 [<ffffffff81246b60>] ? drm_ioctl+0x420/0x4f0 [<ffffffff812a56f0>] ? radeon_gem_pwrite_ioctl+0x20/0x20 [<ffffffff810f53a4>] ? do_vfs_ioctl+0x2e4/0x4e0 [<ffffffff810e5588>] ? vfs_read+0x118/0x160 [<ffffffff810f55ec>] ? sys_ioctl+0x4c/0xa0 [<ffffffff810e5851>] ? sys_read+0x51/0xa0 [<ffffffff814b0612>] ? system_call_fastpath+0x16/0x1b Signed-off-by: Maarten Lankhorst <[email protected]> Reported-by: Markus Trippelsdorf <[email protected]> Acked-by: Paul Menzel <[email protected]> Signed-off-by: Dave Airlie <[email protected]>
2012-12-19watchdog: mei: avoid oops in watchdog unregister code pathTomas Winkler1-1/+1
With commit c7d3df3 "mei: use internal watchdog device registration tracking" will crash the kernel on shutdown path on systems where ME watchdog is not present. Since the watchdog was never initialized in such case the WDOG_UNREGISTERED bit is never set and the system crashes on access to uninitialized variables down the path. To solve the issue we query for NULL on watchdog driver driver_data to check whether the device is registered. This is handled in the driver and doesn't depend on watchdog core internals. Cc: Borislav Petkov <[email protected]> Cc: Wanlong Gao <[email protected]> Signed-off-by: Jerry Snitselaar <[email protected]> Signed-off-by: Tomas Winkler <[email protected]> Signed-off-by: Wim Van Sebroeck <[email protected]>
2012-12-19watchdog: Orion: Fix possible null-deference in orion_wdt_probeJason Gunthorpe1-0/+2
If the DT does not include a regs parameter then the null res would be dereferenced. Signed-off-by: Jason Gunthorpe <[email protected]> Signed-off-by: Wim Van Sebroeck <[email protected]>
2012-12-19watchdog: sp5100_tco: Add SB8x0 chipset supportTakahisa Tanaka2-61/+306
The current sp5100_tco driver only supports SP5100/SB7x0 chipset, doesn't support SB8x0 chipset, because current sp5100_tco driver doesn't know that the offset address for watchdog timer was changed from SB8x0 chipset. The offset address of SP5100 and SB7x0 chipsets are as follows, quote from the AMD SB700/710/750 Register Reference Guide (Page 164) and the AMD SP5100 Register Reference Guide (Page 166). WatchDogTimerControl 69h WatchDogTimerBase0 6Ch WatchDogTimerBase1 6Dh WatchDogTimerBase2 6Eh WatchDogTimerBase3 6Fh In contrast, the offset address of SB8x0 chipset is as follows, quote from AMD SB800-Series Southbridges Register Reference Guide (Page 147). WatchDogTimerEn 48h WatchDogTimerConfig 4Ch So, In the case of SB8x0 chipset, sp5100_tco reads meaningless MMIO address (for example, 0xbafe00) from wrong offset address, and the following message is logged. SP5100 TCO timer: mmio address 0xbafe00 already in use With this patch, sp5100_tco driver supports SB8x0 chipset, and can avoid iomem resource conflict. The processing of this patch is as follows. Step 1) Attempt to get the watchdog base address from indirect I/O (0xCD6 and 0xCD7). - Go to the step 7 if obtained address hasn't conflicted with other resource. But, currently, the address (0xfec000f0) conflicts with the IOAPIC MMIO address, and the following message is logged. SP5100 TCO timer: mmio address 0xfec000f0 already in use 0xfec000f0 is recommended by AMD BIOS Developer's Guide. So, go to the next step. Step 2) Attempt to get the SBResource_MMIO base address from AcpiMmioEN (for SB8x0, PM_Reg:24h) or SBResource_MMIO (SP5100/SB7x0, PCI_Reg:9Ch) register. - Go to the step 7 if these register has enabled by BIOS, and obtained address hasn't conflicted with other resource. - If above condition isn't true, go to the next step. Step 3) Attempt to get the free MMIO address from allocate_resource(). - Go to the step 7 if these register has enabled by BIOS, and obtained address hasn't conflicted with other resource. - Driver initialization has failed if obtained address has conflicted with other resource, and no 'force_addr' parameter is specified. Step 4) Use the specified address If 'force_addr' parameter is specified. - allocate_resource() function may fail, when the PCI bridge device occupies iomem resource from 0xf0000000 to 0xffffffff. To handle such a case, I added 'force_addr' parameter to sp5100_tco driver. With 'force_addr' parameter, sp5100_tco driver directly can assign MMIO address for watchdog timer from free iomem region. Note that It's dangerous to specify wrong address in the 'force_addr' parameter. Example of force_addr parameter use # cat /proc/iomem ...snip... fec00000-fec003ff : IOAPIC 0 <--- free MMIO region fec10000-fec1001f : pnp 00:0b fec20000-fec203ff : IOAPIC 1 ...snip... # cat /etc/modprobe.d/sp5100_tco.conf options sp5100_tco force_addr=0xfec00800 # modprobe sp5100_tco # cat /proc/iomem ...snip... fec00000-fec003ff : IOAPIC 0 fec00800-fec00807 : SP5100 TCO <--- watchdog timer MMIO address fec10000-fec1001f : pnp 00:0b fec20000-fec203ff : IOAPIC 1 ...snip... # - Driver initialization has failed if specified address has conflicted with other resource. Step 5) Disable the watchdog timer - To rewrite the watchdog timer register of the chipset, absolutely guarantee that the watchdog timer is disabled. Step 6) Re-program the watchdog timer MMIO address to chipset. - Re-program the obtained MMIO address in Step 3 or Step 4 to chipset via indirect I/O (0xCD6 and 0xCD7). Step 7) Enable and setup the watchdog timer This patch has worked fine on my test environment (ASUS M4A89GTD-PRO/USB3 and DL165G7). therefore I believe that it's no problem to re-program the MMIO address for watchdog timer to chipset during disabled watchdog. However, I'm not sure about it, because I don't know much about chipset programming. So, any comments will be welcome. Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=43176 Tested-by: Arkadiusz Miskiewicz <[email protected]> Tested-by: Paul Menzel <[email protected]> Signed-off-by: Takahisa Tanaka <[email protected]> Signed-off-by: Wim Van Sebroeck <[email protected]>
2012-12-19watchdog: davinci_wdt: add OF supportMurali Karicheri2-0/+19
This adds OF support for davinci_wdt driver. Signed-off-by: Murali Karicheri <[email protected]> Acked-by: Grant Likely <[email protected]> Signed-off-by: Wim Van Sebroeck <[email protected]>
2012-12-19watchdog: da9052: Fix invalid free of devm_ allocated dataTushar Behera1-4/+0
It is not required to free devm_ allocated data. Since kref_put needs a valid release function, da9052_wdt_release_resources() is not deleted. Fixes following warning. drivers/watchdog/da9052_wdt.c:59:1-6: WARNING: invalid free of devm_ allocated data Signed-off-by: Tushar Behera <[email protected]> Signed-off-by: Wim Van Sebroeck <[email protected]>
2012-12-19watchdog: twl4030_wdt: Change TWL4030_MODULE_PM_RECEIVER to ↵Peter Ujfalusi1-1/+1
TWL_MODULE_PM_RECEIVER To facilitate upcoming cleanup in twl stack. No functional changes. Signed-off-by: Peter Ujfalusi <[email protected]> Signed-off-by: Wim Van Sebroeck <[email protected]>
2012-12-19watchdog: remove depends on CONFIG_EXPERIMENTALKees Cook1-2/+2
The CONFIG_EXPERIMENTAL config item has not carried much meaning for a while now and is almost always enabled by default. As agreed during the Linux kernel summit, remove it from any "depends on" lines in Kconfigs. Signed-off-by: Kees Cook <[email protected]> Signed-off-by: Wim Van Sebroeck <[email protected]>
Powered by cgit, Gruvboxed by Blaster4385.