- 06 Oct, 2004 40 commits
-
-
Andi Kleen authored
x86-64 in_gate_vma would take a read lock on the VMA when the passed address was inside the 32bit vsyscall page. This would be called by get_user_pages, which already holds the mmap_sem. Unfortunately some callers of get_user_pages hold the mmap_sem for writing, which could in theory cause a deadlock. I think it can currently not happen because the only users who hold it for write before calling gup() are coredump and AIO in the ring setup, and both should not ever access the vsyscall page. But not taking the semaphore is safer and avoid this here. Signed-off-by:
Andi Kleen <ak@suse.de> Signed-off-by:
Andrew Morton <akpm@osdl.org> Signed-off-by:
Linus Torvalds <torvalds@osdl.org>
-
Andi Kleen authored
Don't corrupt interrupt flag in time resume Signed-off-by:
-
Andi Kleen authored
Avoid a deadlock during panic. Don't take the smp_call_function lock in smp_send_stop() Also fix the documentation to conform to Linux standards. Signed-off-by:
-
Andi Kleen authored
Fix a circular dependency Signed-off-by:
-
Andi Kleen authored
CONFIG_FRAME_POINTER has never worked on x86-64 because it never passed -fno-omit-frame-pointer to the compiler, and that is the only way to get a frame pointer on x86-64. It also causes complications with profiling. Drop it. Signed-off-by:
-
Andi Kleen authored
This fixes profile_pc to work properly on x86-64 and not crash. It does now a simple backtrace to the caller of the spin lock without requiring a frame pointer for this. Frame pointer support has been dropped because it never worked. There is still a small race window, but the only way to avoid it would be to rewrite kernel/spinlock.c in assembler again. The race will account a profile tick the the parent of the spinlock caller. Signed-off-by:
-
Andi Kleen authored
Fix oops when multiple MCE entries are logged. Signed-off-by:
-
Andi Kleen authored
From: <mg@iceni.pl> The following patch fixes compilation of time.c with HPET_HACK_ENABLE_DANGEROUS turned on. Signed-off-by:
-
David Gibson authored
arch/ppc64/kernel/process.c has an #ifdef CONFIG_ALTIVEC within an #ifdef CONFIG_ALTIVEC. This patch removes the inner one. Signed-off-by:
David Gibson <david@gibson.dropbear.id.au> Signed-off-by:
-
David Gibson authored
A slightly non-ideal version of the recent patch which fixed EEH being a no-op went in. The srcsave variable in eeh_memcpy_to_io() is now never referenced on non-pSeries machines, and so spews hundreds of warnings. The variable doesn't actually accomplish anything, so this patch gets rid of it. Signed-off-by:
David Gibson <dwg@au1.ibm.com> Signed-off-by:
-
Matt Porter authored
Fixes some annoying warnings due to unitialized variables. Signed-off-by:
Matt Porter <mporter@kernel.crashing.org> Signed-off-by:
-
Matt Porter authored
Adds support for booting the same Ocotea kernel from either the default PIBS f/w or U-Boot. Signed-off-by:
-
Matt Porter authored
This puts us back in sync with current U-Boot CVS tree board info definition. Please apply. Signed-off-by:
-
Eugene Surovegin authored
this trivial patch adds missing exports for "indirect" DCR helpers - __mtdcr & __mfcdr. Signed-off-by:
Eugene Surovegin <ebs@ebshome.net> Signed-off-by:
-
Gerhard Jaeger authored
While writing some BSP code for a 440GX custom board, I noticed, that the DCRN_SDR_PFC1_EPS and DCRN_SDR_PFC1_EPS_SHIFT definitions are wrong and therefore the functions ibm440gx_get_eth_grp() and ibm440gx_set_eth_grp() won't work correctly. Signed-off-by:
Gerhard Jaeger <gjaeger@sysgo.com> Signed-off-by:
-
Andrew Morton authored
Ingo points out that it's unusable anyway, because with some configs the get_cpu() is evaluated and with others it is not. Signed-off-by:
-
Alexander Viro authored
Signed-off-by:
Al Viro <viro@parcelfarce.linux.org.uk> Signed-off-by:
-
Alexander Viro authored
trivial endianness annotations in drivers/usb (apply after ohci and isd200 fixes). Note: drivers/usb is nearly endian-clean at that point; there are several very dubious places in there (in particular, rtl8150, pegasus and usbnet are almost certainly broken in mii-related code on big-endian hosts); I'm leaving them alone for now. Signed-off-by:
-
Alexander Viro authored
unsigned long is not a good type to use in declaration of structure we feed to hardware... Signed-off-by:
-
Alexander Viro authored
->dma can be a 64bit variable on 64bit boxen; its value will fit into 32 bits just fine (due to dma mask). However, cpu_to_le32p(&...) will break if we are on a 64bit big-endian; we'll end up up passing it the address of upper 32 bits and get 0 instead of correct value. Fix is trivial... Signed-off-by:
-
Alexander Viro authored
hfs_bnode_read_u8() always returns 0 on little-endian (cut'n'paste bug - function is almost exact copy of its u16 counterpart, but be16_to_cpu() should've been removed here). Signed-off-by:
-
Alexander Viro authored
Signed-off-by:
-
Alexander Viro authored
Signed-off-by:
-
Alexander Viro authored
Signed-off-by:
-
Alexander Viro authored
Signed-off-by:
-
Alexander Viro authored
Signed-off-by:
-
Alexander Viro authored
Signed-off-by:
-
Alexander Viro authored
missing conversions from on-disk to host-endian in a couple of places... Signed-off-by:
-
Alexander Viro authored
Signed-off-by:
-
Alexander Viro authored
Signed-off-by:
-
Alexander Viro authored
* remaining endiannes cleanups * don't mess with setting finfo.i.dataStreamSize when creating the root directory inode; that field is ignored when populating in-core directory inodes. * missing cpu_to_le16() in ncp_search_for_fileset() (for big-endian clients server sees 0xff7f instead of intended 0x7fff). Signed-off-by:
-
Alexander Viro authored
100% straightforward annotations - from gcc POV no code had been changed. Signed-off-by:
-
Alexander Viro authored
New helper: ncp_reply_le16() (decode 16bit little-endian). ConvertToNWfromDWORD() cleaned up and fixed (it used to have one too many le16_to_cpu() in arithmetics, on top of ugly tricks with memcpy() et.al.). ncp_reply_word() has no callers left; removed. Signed-off-by:
-
Alexander Viro authored
New marshalling helpers - ncp_add_be16() and ncp_reply_be16(). Encode and decode a big-endian 16bit field resp. A bunch of places in ncplib_kernel.c switched to usign these. Signed-off-by:
-
Alexander Viro authored
New helper in marshalling code: ncp_add_be32() Takes host-endian argument, converts to big-endian and stores into packet we are building. A bunch of places in ncplib_kernel.c converted to using it. Signed-off-by:
-
Alexander Viro authored
trivial cleanup: endianness conversions pulled into ncp_date_dos2unix() and ncp_date_unix2dos() from their callers. Signed-off-by:
-
Alexander Viro authored
That's the beginning of ncpfs endianness cleanup. * converted fixed-endian constants to little-endian (i.e. replaced htons(0xCDAB) with cpu_to_le16(0xABCD), etc.). These guys _are_ little-endian and make much more sense that way, even aside of annotation issues. Signed-off-by:
-
Alexander Viro authored
Signed-off-by:
-
Alexander Viro authored
Signed-off-by:
-
Alexander Viro authored
udf/balloc.c::find_next_one_bit() had been confused about types... Signed-off-by:
-
