- 11 Sep, 2009 25 commits
-
-
Sebastian Ott authored
Don't use kfree directly after device registration started. Signed-off-by:
Sebastian Ott <sebott@linux.vnet.ibm.com> Signed-off-by:
Martin Schwidefsky <schwidefsky@de.ibm.com>
-
Heiko Carstens authored
Merge the nearly empty C files and move everything from power/ to kernel/. That way the files are easier to handle. Signed-off-by:
Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by:
-
Heiko Carstens authored
There is no caller of do_after_copyback() anywhere. Remove it. Signed-off-by:
-
Heiko Carstens authored
Couple of coding style fixes, replace __inline__ with inline and remove #ifdef __KERNEL_- since the header file isn't exported. Signed-off-by:
-
Heiko Carstens authored
Use compare double and swap to implement efficient atomic64 ops for 31 bit. Signed-off-by:
-
Martin Schwidefsky authored
Move the 64 bit mount code from mcount.S into mcount64.S and avoid code duplication. Signed-off-by: -
Heiko Carstens authored
In the meantime gcc generates better code than the old inline assemblies do. Original inline assembly results in: lr %r1,%r2 sr %r3,%r3 lr %r2,%r1 srdl %r2,16 alr %r2,%r3 alr %r1,%r2 srl %r1,16 xilf %r1,65535 llghr %r2,%r1 br %r14 Out of the C code gcc generates this: rll %r1,%r2,16 ar %r1,%r2 srl %r1,16 xilf %r1,65535 llghr %r2,%r1 br %r14 In addition we don't have any static register allocations anymore and gcc is free to shuffle instructions around for better pipeline usage. Signed-off-by:
-
Heiko Carstens authored
Introduce get_clock_monotonic() function which can be used to get a (fast) timestamp. Resolution is the same as for get_clock(). The only difference is that the timestamps are monotonic and don't jump backward or forward. Signed-off-by:
-
Stefan Haberland authored
This patch fixes message naming so that generic dasd messages do not contain the device discipline. For this purpose the dev_ makros are replaced by pr_ makros for generic dasd messages. Signed-off-by:
Stefan Haberland <stefan.haberland@de.ibm.com> Signed-off-by:
-
Stefan Haberland authored
remove unnecessary dbf call, remove string operations for magic Signed-off-by:
-
Stefan Weinhuber authored
A DASD device that is not ready or online has no defined disk layout, so all requests that arrive in such a state need to be returned as failed. Signed-off-by:
Stefan Weinhuber <wein@de.ibm.com> Signed-off-by:
-
Sebastian Ott authored
We used the init_name to set the console ccw_device's name early at the boot stage. This patch moves the name setting (for all ccw devices) to the point where we actually register the device. At this time we can do dynamic allocations and therefore use dev_set_name. Signed-off-by:
-
Sebastian Ott authored
We use a test_and_clear_bit to prevent a device from being unregistered twice. Unfortunately in this cases the "final" put_device (from device_initialize) was issued more than once, resulting in an use after free error. Fix this by moving this put_device to ccw_device_unregister. Signed-off-by:
-
Sebastian Ott authored
We used the init_name to set the console subchannels name early at the boot stage. With the patch cio: fix memleak in subchannel validation we moved the name setting to the point where we actually register the console subchannel. At this time we can do dynamic allocations and therefore use dev_set_name. Signed-off-by:
-
Sebastian Ott authored
When scanning for new subchannels we have a code path where we allocate memory for a struct subchannel, set the device name (which is dynamically allocated now) and do a check if the underlying device is blacklisted - if so we free the subchannel structure. Since we have not set up refcounting at this stage, the device name's memory is lost. Fix this by moving the dev_set_name after the blacklist test. Note: With this patch the init_name for the console subchannel becomes virtually obsolete. Signed-off-by:
-
Sebastian Ott authored
When using s390dbf with "%s" in sprintf format strings the string itself is not copied to the dbf buffer. Since in this case only pointers are stored in the s390dbf, we should not use dev_name - which is bound to the lifetime of the device. Reading this entry from s390dbf after the device was released will cause an use after free error. Signed-off-by:
-
Jan Glauber authored
The number of qdio debugfs entries was limited. Remove this limit and group the queue files in a per device directory. Signed-off-by:
Jan Glauber <jang@linux.vnet.ibm.com> Signed-off-by:
-
Michael Ernst authored
When unit checks trigger sensing the device state is set to W4SENSE until sense completion; then the device state is set back to ONLINE. If a unit check occurs while set online or set offline requests are processed then it might happen that the device's temporary W4SENSE state causes these functions to terminate, leaving the device in an inconsistent state when the state is set back to ONLINE later on so that the device cannot be set online or offline any longer. To solve this, set online/offline and related rollback or error routines are processed only if the device is in a final or DISCONNECTED state. Signed-off-by:
Michael Ernst <mernst@de.ibm.com> Signed-off-by:
-
Sebastian Ott authored
Ensure to always hold an extra device reference for scheduling a subchannel deregistration, by moving the get_device to ccw_device_schedule_sch_unregister. This fixes an use after free error in ccw_device_call_sch_unregister where put_device was called on an already freed device structure. Signed-off-by:
-
Jan Glauber authored
With commit c38f9608 polling was stopped for the queue even if new data is available. Return immediately after scheduling the queue tasklet if the queue is not done. Signed-off-by:
-
Sebastian Ott authored
Move debug traces for start I/O and interrupt events to exclusive trace levels. Also change tracing in hot-path from sprintf (costly) to hex. Signed-off-by:
-
Sebastian Ott authored
If online/offline processing of a ccw device fails, resulting in not operational state, notify the driver and unregister the device in case the driver dosn't want to keep it. Signed-off-by:
-
Peter Oberparleiter authored
Ensure that the hardware interruption parameter for a subchannel is reset when the associated subchannel data structure is freed. Signed-off-by:
Peter Oberparleiter <peter.oberparleiter@de.ibm.com> Signed-off-by:
-
Heiko Carstens authored
All scsw helper functions are very short and usage of them shouldn't result in function calls. Therefore we move them to a separate header file. Also saves a lot of EXPORT_SYMBOLs. Signed-off-by:
-
Peter Oberparleiter authored
Path verification events occurring for offline devices are currently ignored. As a result, offline devices are not removed, even though they might no longer be accessible (for example because the last path to the device was varied offline). Fix this by scheduling a status evaluation for the affected subchannel when a path verification event occurs. Signed-off-by:
-
- 10 Sep, 2009 3 commits
-
-
Geert Uytterhoeven authored
Commit b8313b6d ("dm log: remove incorrect field from userspace table output") added a call to strstr() with a single-character "needle" string parameter. Unfortunately some versions of gcc replace such calls to strstr() by calls to strchr() behind our back. This causes linking errors if strchr() is defined as an inline function in <asm/string.h> (e.g. on m68k): | WARNING: "strchr" [drivers/md/dm-log-userspace.ko] undefined! Avoid this by explicitly calling strchr() instead. Signed-off-by:
Geert Uytterhoeven <geert@linux-m68k.org> Cc: stable@kernel.org Signed-off-by:
Linus Torvalds <torvalds@linux-foundation.org>
-
Linus Torvalds authored
* lookup-permissions-cleanup: jffs2/jfs/xfs: switch over to 'check_acl' rather than 'permission()' ext[234]: move over to 'check_acl' permission model shmfs: use 'check_acl' instead of 'permission' Make 'check_acl()' a first-class filesystem op Simplify exec_permission_lite(), part 3 Simplify exec_permission_lite() further Simplify exec_permission_lite() logic Do not call 'ima_path_check()' for each path component
-
Roland McGrath authored
In fs/binfmt_elf.c, load_elf_interp() calls padzero() for .bss even if the PT_LOAD has no PROT_WRITE and no .bss. This generates EFAULT. Here is a small test case. (Yes, there are other, useful PT_INTERP which have only .text and no .data/.bss.) ----- ptinterp.S _start: .globl _start nop int3 ----- $ gcc -m32 -nostartfiles -nostdlib -o ptinterp ptinterp.S $ gcc -m32 -Wl,--dynamic-linker=ptinterp -o hello hello.c $ ./hello Segmentation fault # during execve() itself After applying the patch: $ ./hello Trace trap # user-mode execution after execve() finishes If the ELF headers are actually self-inconsistent, then dying is fine. But having no PROT_WRITE segment is perfectly normal and correct if there is no segment with p_memsz > p_filesz (i.e. bss). John Reiser suggested checking for PROT_WRITE in the bss logic. I think it makes most sense to simply apply the bss logic only when there is bss. This patch looks less trivial than it is due to some reindentation. It just moves the "if (last_bss > elf_bss) {" test up to include the partial-page bss logic as well as the more-pages bss logic. Reported-by:John Reiser <jreiser@bitwagon.com> Signed-off-by:
Roland McGrath <roland@redhat.com> Signed-off-by:
-
- 09 Sep, 2009 3 commits
-
-
Linus Torvalds authored
-
Ed Cashin authored
Andy Whitcroft reported an oops in aoe triggered by use of an incorrectly initialised request_queue object: [ 2645.959090] kobject '<NULL>' (ffff880059ca22c0): tried to add an uninitialized object, something is seriously wrong. [ 2645.959104] Pid: 6, comm: events/0 Not tainted 2.6.31-5-generic #24-Ubuntu [ 2645.959107] Call Trace: [ 2645.959139] [<ffffffff8126ca2f>] kobject_add+0x5f/0x70 [ 2645.959151] [<ffffffff8125b4ab>] blk_register_queue+0x8b/0xf0 [ 2645.959155] [<ffffffff8126043f>] add_disk+0x8f/0x160 [ 2645.959161] [<ffffffffa01673c4>] aoeblk_gdalloc+0x164/0x1c0 [aoe] The request queue of an aoe device is not used but can be allocated in code that does not sleep. Bruno bisected this regression down to cd43e26f block: Expose stacked device queues in sysfs "This seems to generate /sys/block/$device/queue and its contents for everyone who is using queues, not just for those queues that have a non-NULL queue->request_fn." Addresses http://bugs.launchpad.net/bugs/410198 Addresses http://bugzilla.kernel.org/show_bug.cgi?id=13942 Note that embedding a queue inside another object has always been an illegal construct, since the queues are reference counted and must persist until the last reference is dropped. So aoe was always buggy in this respect (Jens). Signed-off-by:
Ed Cashin <ecashin@coraid.com> Cc: Andy Whitcroft <apw@canonical.com> Cc: "Rafael J. Wysocki" <rjw@sisk.pl> Cc: Bruno Premont <bonbons@linux-vserver.org> Cc: Martin K. Petersen <martin.petersen@oracle.com> Cc: Andrew Morton <akpm@linux-foundation.org> Signed-off-by:
Jens Axboe <jens.axboe@oracle.com>
-
Linus Torvalds authored
Reinette Chatre reports a frozen system (with blinking keyboard LEDs) when switching from graphics mode to the text console, or when suspending (which does the same thing). With netconsole, the oops turned out to be BUG: unable to handle kernel NULL pointer dereference at 0000000000000084 IP: [<ffffffffa03ecaab>] i915_driver_irq_handler+0x26b/0xd20 [i915] and it's due to the i915_gem.c code doing drm_irq_uninstall() after having done i915_gem_idle(). And the i915_gem_idle() path will do i915_gem_idle() -> i915_gem_cleanup_ringbuffer() -> i915_gem_cleanup_hws() -> dev_priv->hw_status_page = NULL; but if an i915 interrupt comes in after this stage, it may want to access that hw_status_page, and gets the above NULL pointer dereference. And since the NULL pointer dereference happens from within an interrupt, and with the screen still in graphics mode, the common end result is simply a silently hung machine. Fix it by simply uninstalling the irq handler before idling rather than after. Fixes http://bugzilla.kernel.org/show_bug.cgi?id=13819Reported-and-tested-by:Reinette Chatre <reinette.chatre@intel.com> Acked-by:
Jesse Barnes <jbarnes@virtuousgeek.org> Signed-off-by:
-
- 08 Sep, 2009 9 commits
-
-
Linus Torvalds authored
This avoids an indirect call in the VFS for each path component lookup. Well, at least as long as you own the directory in question, and the ACL check is unnecessary. Reviewed-by:
James Morris <jmorris@namei.org> Acked-by:
Serge Hallyn <serue@us.ibm.com> Signed-off-by:
-
Linus Torvalds authored
Don't implement per-filesystem 'extX_permission()' functions that have to be called for every path component operation, and instead just expose the actual ACL checking so that the VFS layer can now do it for us. Reviewed-by:
-
Linus Torvalds authored
shmfs wants purely standard POSIX ACL semantics, so we can use the new generic VFS layer POSIX ACL checking rather than cooking our own 'permission()' function. Reviewed-by:
Hugh Dickins <hugh.dickins@tiscali.co.uk> Signed-off-by:
-
Linus Torvalds authored
This is stage one in flattening out the callchains for the common permission testing. Rather than have most filesystem implement their own inode->i_op->permission function that just calls back down to the VFS layers 'generic_permission()' with the per-filesystem ACL checking function, the filesystem can just expose its 'check_acl' function directly, and let the VFS layer do everything for it. This is all just preparatory - no filesystem actually enables this yet. Reviewed-by:
-
Linus Torvalds authored
Don't call down to the generic inode_permission() function just to call the inode-specific permission function - just do it directly. The generic inode_permission() code does things like checking MAY_WRITE and devcgroup_inode_permission(), neither of which are relevant for the light pathname walk permission checks (we always do just MAY_EXEC, and the inode is never a special device). Reviewed-by:
-
Linus Torvalds authored
This function is only called for path components that are already known to be directories (they have a '->lookup' method). So don't bother doing that whole S_ISDIR() testing, the whole point of the 'lite()' version is that we know that we are looking at a directory component, and that we're only checking name lookup permission. Reviewed-by:
-
Linus Torvalds authored
Instead of returning EAGAIN and having the caller do something special for that case, just do the special case directly. Reviewed-by:
-
Linus Torvalds authored
Not only is that a supremely timing-critical path, but it's hopefully some day going to be lockless for the common case, and ima can't do that. Plus the integrity code doesn't even care about non-regular files, so it was always a total waste of time and effort. Acked-by:
Mimi Zohar <zohar@us.ibm.com> Signed-off-by:
-
Zhenyu Wang authored
eDP is exclusive connector too, and add missing crtc_mask setting for TV. This fixes http://bugzilla.kernel.org/show_bug.cgi?id=14139Signed-off-by:
Zhenyu Wang <zhenyuw@linux.intel.com> Reported-and-tested-by:
Carlos R. Mafra <crmafra2@gmail.com> Signed-off-by:
-
