- 23 Sep, 2005 40 commits
-
-
Nick Wilson authored
nfs_readpage_release() causes an oops while accessing a file with NFS debugging turned on (echo 32767 > /proc/sys/sunrpc/nfs_debug) and a kernel built with CONFIG_DEBUG_SLAB. This patch moves the debugging statement above nfs_release_request() to avoid accessing freed memory. Signed-off-by:
Nick Wilson <njw@osdl.org> Cc: Trond Myklebust <trond.myklebust@fys.uio.no> Signed-off-by:
Andrew Morton <akpm@osdl.org> Signed-off-by:
Linus Torvalds <torvalds@osdl.org>
-
Rob Landley authored
Problem: In some circumstances, bd_claim() is returning the wrong error code. If we try to swapon an unused block device that isn't swap formatted, we get -EINVAL. But if that same block device is already mounted, we instead get -EBUSY, even though it still isn't a valid swap device. This issue came up on the busybox list trying to get the error message from "swapon -a" right. If a swap device is already enabled, we get -EBUSY, and we shouldn't report this as an error. But we can't distinguish the two -EBUSY conditions, which are very different errors. In the code, bd_claim() returns either 0 or -EBUSY, but in this case busy means "somebody other than sys_swapon has already claimed this", and _that_ means this block device can't be a valid swap device. So return -EINVAL there. Signed-off-by:
Rob Landley <rob@landley.net> Signed-off-by:
-
Glauber de Oliveira Costa authored
Fix some warnings and a build error when EXT3_DEBUG is enabled. Signed-off-by:
-
Clemens Buchacher authored
It's deprecated. Use "%s", __FUNCTION__ instead. Signed-off-by:
Clemens Buchacher <drizzd@aon.at> Signed-off-by:
Maximilian Attems <janitor@sternwelten.at> Signed-off-by:
Domen Puncer <domen@coderock.org> Signed-off-by:
Alexey Dobriyan <adobriyan@gmail.com> Signed-off-by:
-
Chris Zankel authored
Remove io_remap_page_range() from all of Linux 2.6.x (as requested and suggested by Randy Dunlap) and minor clean-ups. Signed-off-by:
Chris Zankel <chris@zankel.net> Signed-off-by:
-
Paolo 'Blaisorblade' Giarrusso authored
User get *a lot* confused when consoles don't work but we don't report anything. And, as reported in the comment, using printk to report "your console doesn't work" isn't likely to go that far. Fix the problem on the base of this: stack consumption by host printf(). Use kernel sprintf() and os_write_file, using a wild guess that one page will be enough for the message, to preallocate the buffer with kmalloc(). Signed-off-by:
Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it> Cc: Jeff Dike <jdike@addtoit.com> Signed-off-by:
-
Paolo 'Blaisorblade' Giarrusso authored
setup_initial_poll is only called with sigio_lock() held, so use appropriate allocation. Also, parse_chan() can also be called when holding a spinlock (see line_open() -> parse_chan_pair()). I have sporadic problems (spinlock taken twice, with spinlock debugging on UP) which could be caused by a sequence like "take spinlock, alloc and go to sleep, take again the spinlock in the other thread". Signed-off-by:
-
Paolo 'Blaisorblade' Giarrusso authored
GFP_ATOMIC | GFP_KERNEL is meaningless and won't work. Actually it never worked, even in 2.4. Signed-off-by:
-
Paolo 'Blaisorblade' Giarrusso authored
Following i386, we should maybe refuse trying to fault in pages when we're doing atomic operations, because to handle the fault we could need to take already taken spinlocks. Also, if we're doing an atomic operation (in the sense of in_atomic()) we're surely in kernel mode and we're surely going to handle adequately the failed fault, so it's safe to behave this way. Currently, on UML SMP is rarely used, and we don't support PREEMPT, so this is unlikely to create problems right now, but it might in the future. Signed-off-by:
-
Paolo 'Blaisorblade' Giarrusso authored
Things are breaking horribly with sysrq called in interrupt context. I want to try to fix it, but probably this is simpler. To tell the truth, sysrq is normally run in interrupt context, so there shouldn't be any problem. There's also a warning from the fault handler because it's run in atomic context (I have a patch for that, only I deferred it). This is why I'm doing this. Signed-off-by:
-
Paolo 'Blaisorblade' Giarrusso authored
Avoid setting w = 0 twice. Spotted this (trivial) thing which is needed for another patch. Signed-off-by:
-
Paolo 'Blaisorblade' Giarrusso authored
The current code doesn't handle well general protection faults on the host - it thinks that cr2 is always the address of a page fault. While actually, on general protection faults, that address is not accessible, so we'd better assume we couldn't satisfy the fault. Currently instead we think we've fixed it, so we go back, retry the instruction and fault again endlessly. This leads to the kernel hanging when doing copy_from_user(dest, -1, ...) in TT mode, since reading *(-1) causes a GFP, and we don't support kernel preemption. Thanks to Luo Xin for testing UML with LTP and reporting the failures he got. Cc: Luo Xin <luothing@sina.com> Signed-off-by:
-
Paolo 'Blaisorblade' Giarrusso authored
pte_modify marks a page as needing flush, which is redundant because the resulting PTE is still set with set_pte, which already handles that. Signed-off-by:
-
Paolo 'Blaisorblade' Giarrusso authored
Simplify the code by using strlcat() instead of strncat() and manual appending. Signed-off-by:
-
Paolo 'Blaisorblade' Giarrusso authored
Only remove the UML pidfile and management socket if we created them. Currently in case two UMLs are started with the same umid, the second will remove the first's ones. Probably we should also panic() at that point, not sure however. Signed-off-by:
-
Pavel Machek authored
Fix comments in swsusp. Signed-off-by:
Pavel Machek <pavel@suse.cz> Signed-off-by:
-
Christoph Lameter authored
I had an issue on ia64 where I got a bug in kernel/workqueue because kzalloc returned a NULL pointer due to the task structure getting too big for the slab allocator. Usually these cases are caught by the kmalloc macro in include/linux/slab.h. Compilation will fail if a too big value is passed to kmalloc. However, kzalloc uses __kmalloc which has no check for that. This patch makes __kmalloc bug if a too large entity is requested. Signed-off-by:
Christoph Lameter <clameter@sgi.com> Signed-off-by:
-
Benjamin Herrenschmidt authored
The PowerMac mesh SCSI driver had some missing error handling which would trigger warnings due to lack of handling of return value from scsi_add_host. This fixes it. Signed-off-by:
Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: James Bottomley <James.Bottomley@SteelEye.com> Signed-off-by:
-
Chuck Ebbert authored
Correctly identify atiixp_modem in its error messages. Signed-off-by:
Chuck Ebbert <76306.1226@compuserve.com> Cc: Vojtech Pavlik <vojtech@suse.cz> Signed-off-by:
-
Randy Dunlap authored
Corrections to the recent top-level README changes. Signed-off-by:
Randy Dunlap <rdunlap@xenotime.net> Signed-off-by:
-
Benjamin Herrenschmidt authored
The SMU is the "system controller" chip used by Apple recent G5 machines including the iMac G5. It drives things like fans, i2c busses, real time clock, etc... The current kernel contains a very crude driver that doesn't do much more than reading the real time clock synchronously. This is a completely rewritten driver that provides interrupt based command queuing, a userland interface, and an i2c/smbus driver for accessing the devices hanging off the SMU i2c busses like temperature sensors. This driver is a basic block for upcoming work on thermal control for those machines, among others. Signed-off-by:
-
Anton Blanchard authored
Fix my stupid bug in the 64bit version of PTRACE_SET_DEBUGREG. Signed-off-by:
Anton Blanchard <anton@samba.org> Signed-off-by:
-
Anton Blanchard authored
Fix build when iommu debug is enabled. Signed-off-by:
-
Anton Blanchard authored
The recent iommu fix broke booting on some POWER4 and POWER5 LPAR boxes. It looks like we have been calling the non LPAR iommu_dev_setup on LPAR machines for a while. The recent iommu fix caused that code path to fail. It looks like we just need to hook up the devices iommu_table to the parents one, so do that instead of calling iommu_dev_setup_pSeries and crossing the streams. Signed-off-by:
-
OGAWA Hirofumi authored
EXT3_MOUNT_DATA_FLAGS is not a boolean. This fixes it. Signed-off-by:
OGAWA Hirofumi <hirofumi@mail.parknet.co.jp> Signed-off-by:
-
Christoph Lameter authored
The numa slab allocator may allocate pages from foreign nodes onto the lists for a particular node if a node runs out of memory. Inspecting the slab->nodeid field will not reflect that the page is now in use for the slabs of another node. This patch fixes that issue by adding a node field to free_block so that the caller can indicate which node currently uses a slab. Also removes the check for the current node from kmalloc_cache_node since the process may shift later to another node which may lead to an allocation on another node than intended. Signed-off-by:
-
Rafael J. Wysocki authored
The following patch makes swsusp avoid triggering the BUG_ON() in swsusp_suspend() if there is not enough memory for suspend. Signed-off-by:
Rafael J. Wysocki <rjw@sisk.pl> Cc: Pavel Machek <pavel@ucw.cz> Signed-off-by:
-
Robert Love authored
- Handle dmi_system_check() elegantly, now that my bugfix is upstream. - Add support for the X41 and R52. - Cleanup some comments do I do not have to keep updating them with each new whitelisted laptop. Signed-off-by:
Robert Love <rml@novell.com> Signed-off-by:
-
Keir Fraser authored
The address passed to io_remap_pfn_range() in hpet_mmap() does not need to be converted using __pa(): it is already a physical address. This bug was found and the patch suggested by Clay Harris. I introduced this particular bug when making io_remap_pfn_range changes a few months ago. In fact mmap()ing /dev/hpet has *never* previously worked: before my changes __pa() was being executed on an ioremap()ed virtual address, which is also invalid. Signed-off-by:
Keir Fraser <keir@xensource.com> Cc: Robert Picco <Robert.Picco@hp.com> Signed-off-by:
-
Ivan Kokshaysky authored
It is essential that index_of() be inlined. But alpha undoes the gcc inlining hackery and index_of() ends up out-of-line. So fiddle with things to make that function inline again. Cc: Richard Henderson <rth@twiddle.net> Signed-off-by:
-
Ivan Kokshaysky authored
acquire_console_sem() does BUG() in interrupt context now, as in the case of SysRq-b. Cc: Richard Henderson <rth@twiddle.net> Signed-off-by:
-
Randy Dunlap authored
Signed-off-by:
Pavel Machek <pavel@ucw.cz> Signed-off-by:
-
Benjamin Herrenschmidt authored
Current -git tree doesn't build when enabling oprofile on a non-bookE CPU (like on a PowerMac for example). While there is no performance counter support for these CPUs implemented yet, it's still nice to be able to use the timer based sampling, and that got broken. Signed-off-by:
-
Grant Coady authored
Fix namespace clash: drivers/mtd/devices/docecc.c:43:1: warning: "DEBUG" redefined In file included from drivers/mtd/devices/docecc.c:40: include/linux/mtd/mtd.h:219:1: warning: this is the location of the previous definition Signed-off-by:
Grant Coady <gcoady@gmail.com> Cc: David Woodhouse <dwmw2@infradead.org> Signed-off-by:
-
Bernd Petrovitsch authored
Rename the (apparently) bttv intern #define vprintk to verbprintk to resolve a name clash. Reason: vprintk() is defined in include/linux/kernel.h similar to printk but with a va_list argument. (akpm: I changed it to bttv_printk) Cc: Mauro Carvalho Chehab <mchehab@brturbo.com.br> Signed-off-by:
-
Latchesar Ionkov authored
If error occurs while in v9fs_get_sb after it calles sget, the dentry object of the root and its inode may be freed twice -- once while handling the error in v9fs_get_sb, and second time when v9fs_get_sb calles deactivate_super (which in turn calls v9fs_kill_super) The patch removes the unnecessary code that frees the root dentry and its inode. Signed-off-by:
Latchesar Ionkov <lucho@ionkov.net> Cc: Eric Van Hensbergen <ericvh@gmail.com> Signed-off-by:
-
Latchesar Ionkov authored
v9fs_vfs_readlink allocates space for the link using __getname and errorneously uses strlen on the newly allocated buffer to check if the buffer passed by the user is bigger than the one returned by __getname. The patch replaces the strlen usage to PATH_MAX, which is the actual size of the buffers returned by __getname. Signed-off-by:
-
Latchesar Ionkov authored
When a new session is created it uses a template object of the specified transport type to instantiate its own copy. The code for the making a copy of the template object was lost, and the object itself is attached to the v9fs session. This leads to many sessions using the same transport instead of having their own copy. The patch puts back the code that makes a copy of the template object. Signed-off-by:
-
Latchesar Ionkov authored
When v9fs_deserealize_fcall deserializes a Rwalk message, it incorrectly allocates space for the qid array in the source instead of the destination buffer. Signed-off-by:
-
Latchesar Ionkov authored
buf_check_size function checks if the conv buffer has enough space for the performed operation, but it doesn't return the result back to the calling function, only logs an error in the log. The report-back-error functionality was lost when buf_check_size was converted from macro to inline function. The return in the macro used to exit from the functions that include it, after the conversion it just exits from the inline function itself. The patch makes buf_check_size to return flag and all functions that use it check if they should perform the operation, or exit. Signed-off-by:
-
