kernel.git/arch/um/include, branch linux-2.6.25.y

uml: Fix boot crash

2008-08-20T18:15:41Z

commit 7c1fed03b9fa32d4323d5caa6a9c7dcdd7eba767 upstream My copying of linux/init.h didn't go far enough. The definition of __used singled out gcc minor version 3, but didn't care what the major version was. This broke when unit-at-a-time was added and gcc started throwing out initcalls. This results in an early boot crash when ptrace tries to initialize a process with an empty, uninitialized register set. Signed-off-by: Jeff Dike Signed-off-by: Linus Torvalds Signed-off-by: Greg Kroah-Hartman

uml: deal with inaccessible address space start

2008-08-20T18:15:37Z

commit 40fb16a360d9c6459afee91dc793c1e3374feb94 upstream This patch makes os_get_task_size locate the bottom of the address space, as well as the top. This is for systems which put a lower limit on mmap addresses. It works by manually scanning pages from zero onwards until a valid page is found. Because the bottom of the address space may not be zero, it's not sufficient to assume the top of the address space is the size of the address space. The size is the difference between the top address and bottom address. [jdike@addtoit.com: changed the name to reflect that this function is supposed to return the top of the process address space, not its size and changed the return value to reflect that. Also some minor formatting changes] Signed-off-by: Tom Spink Signed-off-by: Jeff Dike Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds Signed-off-by: Greg Kroah-Hartman

uml: track and make up lost ticks

2008-08-20T18:15:34Z

commit fe2cc53ee013a4d4d0317d418e7019fe6533a5a8 upstream Alarm delivery could be noticably late in the !CONFIG_NOHZ case because lost ticks weren't being taken into account. This is now treated more carefully, with the time between ticks being calculated and the appropriate number of ticks delivered to the timekeeping system. Cc: Nix Signed-off-by: Jeff Dike Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds Signed-off-by: Greg Kroah-Hartman

uml: fix build when SLOB is enabled

2008-08-20T18:15:32Z

commit 43f5b3085fdd27c4edf535d938b2cb0ccead4f75 upstream Reintroduce uml_kmalloc for the benefit of UML libc code. The previous tactic of declaring __kmalloc so it could be called directly from the libc side of the house turned out to be getting too intimate with slab, and it doesn't work with slob. So, the uml_kmalloc wrapper is back. It calls kmalloc or whatever that translates into, and libc code calls it. kfree is left alone since that still works, leaving a somewhat inconsistent API. Signed-off-by: Jeff Dike Cc: WANG Cong Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds Signed-off-by: Greg Kroah-Hartman

uml: fix FP register corruption

2008-02-24T01:12:15Z

Commit ee3d9bd4de1ed93d2a7ee41c331ed30a1c7b8acd ("uml: simplify SIGSEGV handling"), while greatly simplifying the kernel SIGSEGV handler that runs in the process address space, introduced a bug which corrupts FP state in the process. Previously, the SIGSEGV handler called the sigreturn system call by hand - it couldn't return through the restorer provided to it because that could try to call the libc restorer which likely wouldn't exist in the process address space. So, it blocked off some signals, including SIGUSR1, on entry to the SIGSEGV handler, queued a SIGUSR1 to itself, and invoked sigreturn. The SIGUSR1 was delivered, and was visible to the UML kernel after sigreturn finished. The commit eliminated the signal masking and the call to sigreturn. The handler simply hits itself with a SIGTRAP to let the UML kernel know that it is finished. UML then restores the process registers, which effectively longjmps the process out of the signal handler, skipping sigreturn's restoring of register state and the signal mask. The bug is that the host apparently sets used_fp to 0 when it saves the process FP state in the sigcontext on the process signal stack. Thus, when the process is longjmped out of the handler, its FP state is corrupt because it wasn't saved on the context switch to the UML kernel. This manifested itself as sleep hanging. For some reason, sleep uses floating point in order to calculate the sleep interval. When a page fault corrupts its FP state, it is faked into essentially sleeping forever. This patch saves the FP state before entering the SIGSEGV handler and restores it afterwards. Signed-off-by: Jeff Dike Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds

uml: fix mm_context memory leak

2008-02-08T17:22:43Z

[ Spotted by Miklos ] Fix a memory leak in init_new_context. The struct page ** buffer allocated for install_special_mapping was never recorded, and thus leaked when the mm_struct was freed. Fix it by saving the pointer in mm_context_t and freeing it in arch_exit_mmap. Signed-off-by: Jeff Dike Cc: Miklos Szeredi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds

uml: style fixes in arch/um/os-Linux

2008-02-08T17:22:42Z

Style changes under arch/um/os-Linux: include trimming CodingStyle fixes some printks needed severity indicators make_tempfile turns out not to be used outside of mem.c, so it is now static. Its declaration in tempfile.h is no longer needed, and tempfile.h itself is no longer needed. create_tmp_file was also made static. checkpatch moans about an EXPORT_SYMBOL in user_syms.c which is part of a macro definition - this is copying a bit of kernel infrastructure into the libc side of UML because the kernel headers can't be included there. Signed-off-by: Jeff Dike Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds

uml: runtime host VMSPLIT detection

2008-02-08T17:22:42Z

Calculate TASK_SIZE at run-time by figuring out the host's VMSPLIT - this is needed on i386 if UML is to run on hosts with varying VMSPLITs without recompilation. TASK_SIZE is now defined in terms of a variable, task_size. This gets rid of an include of pgtable.h from processor.h, which can cause include loops. On i386, task_size is calculated early in boot by probing the address space in a binary search to figure out where the boundary between usable and non-usable memory is. This tries to make sure that a page that is considered to be in userspace is, or can be made, read-write. I'm concerned about a system-global VDSO page in kernel memory being hit and considered to be a userspace page. On x86_64, task_size is just the old value of CONFIG_TOP_ADDR. A bunch of config variable are gone now. CONFIG_TOP_ADDR is directly replaced by TASK_SIZE. NEST_LEVEL is gone since the relocation of the stubs makes it irrelevant. All the HOST_VMSPLIT stuff is gone. All references to these in arch/um/Makefile are also gone. I noticed and fixed a missing extern in os.h when adding os_get_task_size. Note: This has been revised to fix the 32-bit UML on 64-bit host bug that Miklos ran into. Signed-off-by: Jeff Dike Cc: Miklos Szeredi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds

uml: redo the calculation of NR_syscalls

2008-02-05T17:44:32Z

Redo the calculation of NR_syscalls since that disappeared from i386 and use a similar mechanism on x86_64. We now figure out the size of the system call table in arch code and stick that in syscall_table_size. arch/um/kernel/skas/syscall.c defines NR_syscalls in terms of that since its the only thing that needs to know how many system calls there are. The old mechananism that was used on x86_64 is gone. arch/um/include/sysdep-i386/syscalls.h got some formatting since I was looking at it. Signed-off-by: Jeff Dike Cc: WANG Cong Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds

uml: remove unused fields from mm_context

2008-02-05T17:44:31Z

The 3-level page table fixes forgot to remove a couple now-unused fields from struct mm_context. Signed-off-by: Jeff Dike Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds