- Jul 08, 2005
-
-
Roland Dreier authored
First of a series of patches which add support for direct userspace access to InfiniBand hardware -- so-called "userspace verbs." I believe these patches are ready to merge, but a final review would be useful. These patches should incorporate all of the feedback from the discussion when I posted an earlier version back in April (see http://lkml.org/lkml/2005/4/4/267 for the start of the thread). In particular, memory pinned for use by userspace is accounted for in current->mm->vm_locked and requests to pin memory are checked against RLIMIT_MEMLOCK. This patch: Modify the ib_verbs.h header file with changes required for InfiniBand userspace verbs support. We add a few structures to keep track of userspace context, and extend the driver API so that low-level drivers know when they're creating resources that will be used from userspace. Signed-off-by:
Roland Dreier <rolandd@cisco.com> Signed-off-by:
Andrew Morton <akpm@osdl.org> Signed-off-by:
Linus Torvalds <torvalds@osdl.org>
-
Andrew Morton authored
The infiniband code expects that the arch implements pgprot_noncached(). We're mapping PCI areas anyway, so this probabyl wasn't needed and we should make infiniband stop doing that.. Cc: Roland Dreier <rolandd@cisco.com> Cc: Richard Henderson <rth@twiddle.net> Signed-off-by:
-
KAMBAROV, ZAUR authored
In __xprt_lock_write() we check to see if `task' is NULL, but in other places we just go and dereference it. `task' shouldn't be NULL anyway, so remove this test. This defect was found automatically by Coverity Prevent, a static analysis tool. Signed-off-by:
Zaur Kambarov <zkambarov@coverity.com> Acked-by:
Trond Myklebust <trond.myklebust@fys.uio.no> Cc: Neil Brown <neilb@cse.unsw.edu.au> Signed-off-by:
-
KAMBAROV, ZAUR authored
We're dereferencing `flp' and then we're testing it for NULLness. Either the compiler accidentally saved us or the existing null-pointer checdk is redundant. This defect was found automatically by Coverity Prevent, a static analysis tool. Signed-off-by:
-
KAMBAROV, ZAUR authored
Correctly test for a null pointer before going and dereferencing it. This defect was found automatically by Coverity Prevent, a static analysis tool. Signed-off-by:
-
Ingo Molnar authored
The BKS might be reacquired before we have dropped PREEMPT_ACTIVE, which could trigger a second could trigger a second cond_resched() call. Bug found by Hirofumi Ogawa. Signed-off-by:
Ingo Molnar <mingo@elte.hu> Signed-off-by:
-
David Howells authored
The attached patch makes the keyring functions calculate the new size of a keyring's payload based on the size of pointer to the key struct, not the size of the key struct itself. Signed-Off-By:
David Howells <dhowells@redhat.com> Signed-off-by:
-
Ian Kent authored
Fix debugging printk. Signed-off-by:
Ian Kent <raven@themaw.net> Signed-off-by:
-
Jesse Millan authored
GCC 4 complains because the function put_compat_shminfo() can't get to its return statement if there is no error... If the function does not return -EFAULT, it doesn't return anything at all. Looks like a typo. Signed-off-by:
Jesse Millan <jessem@cs.pdx.edu> Signed-off-by:
Domen Puncer <domen@coderock.org> Signed-off-by:
-
Andreas Gruenbacher authored
We are not using the in-inode space for xattrs in reserved inodes because mkfs.ext3 doesn't initialize it properly. For those inodes, we set i_extra_isize to 0. Make sure that we also don't overwrite the i_extra_isize field when writing out the inode in that case. This is for cleanliness only, and doesn't fix an actual bug. Signed-off-by:
Andreas Gruenbacher <agruen@suse.de> Signed-off-by:
-
Christoph Lameter authored
Add a new section called ".data.read_mostly" for data items that are read frequently and rarely written to like cpumaps etc. If these maps are placed in the .data section then these frequenly read items may end up in cachelines with data is is frequently updated. In that case all processors in an SMP system must needlessly reload the cachelines again and again containing elements of those frequently used variables. The ability to share these cachelines will allow each cpu in an SMP system to keep local copies of those shared cachelines thereby optimizing performance. Signed-off-by:
Alok N Kataria <alokk@calsoftinc.com> Signed-off-by:
Shobhit Dayal <shobhit@calsoftinc.com> Signed-off-by:
Christoph Lameter <christoph@scalex86.org> Signed-off-by:
Shai Fultheim <shai@scalex86.org> Signed-off-by:
-
Alexey Dobriyan authored
Signed-off-by:
Alexey Dobriyan <adobriyan@gmail.com> Signed-off-by:
-
Andreas Gruenbacher authored
Original patch from Matt Mackall <mpm@selenic.com> Signed-off-by:
-
Nick Piggin authored
Use a bit spin lock in the first buffer of the page to synchronise asynch IO buffer completions, instead of the global page_uptodate_lock, which is showing some scalabilty problems. Signed-off-by:
Nick Piggin <nickpiggin@yahoo.com.au> Signed-off-by:
-
Roman Zippel authored
The patch fixes a few corner cases around tty line editing with very long input lines: - n_tty_receive_char(): don't simply drop eol characters, otherwise canon_data isn't increased and the reader isn't woken up. - n_tty_receive_room(): If there is no newline pending and the edit buffer is full, allow only a single character to be written (until eol is found and the line is flushed), so characters from the next line aren't dropped. - write_chan(): if an incomplete line was written, continue writing until write() returns 0, otherwise it might not write the eol character to flush the line and the writer goes to sleep without ever being woken up. BTW the core problem is that part of this should be handled in the receive_buf path, but for this it has to return the number of written characters, as the amount of written characters may not be the same as the amount of characters going into the write buffer, so the receive_room() usage in pty_write() is not really reliable. Alan said: The problem looks valid. The behaviour of 'traditional unix' appears to be the following If you exceed the line limit then beep and drop the character Always allow EOL to complete a canonical line input Always do signal/control processing if enabled Signed-off-by:
-
Arnd Bergmann authored
xtensa is now in -rc1, with the obsolete syscalls still in there, so I guess this about the last chance to correct the ABI. Applying the patch obviously breaks all sorts of user space binaries and probably also requires the appropriate changes to be made to libc. On the other hand, if a decision is made to keep the broken interface, it should at least be a conscious one instead of an oversight. Signed-off-by:
Arnd Bergmann <arnd@arndb.de> Cc: Chris Zankel <chris@zankel.net> Signed-off-by:
-
Paolo 'Blaisorblade' Giarrusso authored
Replace a semaphore (winch_handler_sem) used in atomic code with a spinlock, and reduces as needed the amount of protected code to the bare minimum (for instance no kmalloc calls are needed). This fixes the last problems with spinlocking (in UP mode with DEBUG options); the semaphore, taken inside spinlocks, caused a "spin_lock was already locked" warning, without this patch. Signed-off-by:
Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it> Cc: Jeff Dike <jdike@addtoit.com> Signed-off-by:
-
Paolo 'Blaisorblade' Giarrusso authored
Some time ago a trivial patch broke HPPFS (one var became a pointer, not all uses were updated). It wasn't fixed at that time because not very used, now it's been requested so I've fixed this, and it has been tested positively (at least partially). Signed-off-by:
-
Bodo Stroesser authored
This patch implements the clone-stub mechanism, which allows skas0 to run with proc_mm==0, even if the clib in UML uses modify_ldt. Note: There is a bug in skas3.v7 host patch, that avoids UML-skas from running properly on a SMP-box. In full skas3, I never really saw problems, but in skas0 they showed up. More commentary by jdike - What this patch does is makes sure that the host parent of each new host process matches the UML parent of the corresponding UML process. This ensures that any changed LDTs are inherited. This is done by having clone actually called by the UML process from its stub, rather than by the kernel. We have special syscall stubs that are loaded onto the stub code page because that code must be completely self-contained. These stubs are given C interfaces, and used like normal C functions, but there are subtleties. Principally, we have to be careful about stack variables in stub_clone_handler after the clone. The code is written so that there aren't any - everything boils down to a fixed address. If there were any locals, references to them after the clone would be wrong because the stack just changed. Signed-off-by:
Bodo Stroesser <bstroesser@fujitsu-siemens.com> Signed-off-by:
Jeff Dike <jdike@addtoit.com> Signed-off-by:
-
Jeff Dike authored
UML has had two modes of operation - an insecure, slow mode (tt mode) in which the kernel is mapped into every process address space which requires no host kernel modifications, and a secure, faster mode (skas mode) in which the UML kernel is in a separate host address space, which requires a patch to the host kernel. This patch implements something very close to skas mode for hosts which don't support skas - I'm calling this skas0. It provides the security of the skas host patch, and some of the performance gains. The two main things that are provided by the skas patch, /proc/mm and PTRACE_FAULTINFO, are implemented in a way that require no host patch. For the remote address space changing stuff (mmap, munmap, and mprotect), we set aside two pages in the process above its stack, one of which contains a little bit of code which can call mmap et al. To update the address space, the system call information (system call number and arguments) are written to the stub page above the code. The %esp is set to the beginning of the data, the %eip is set the the start of the stub, and it repeatedly pops the information into its registers and makes the system call until it sees a system call number of zero. This is to amortize the cost of the context switch across multiple address space updates. When the updates are done, it SIGSTOPs itself, and the kernel process continues what it was doing. For a PTRACE_FAULTINFO replacement, we set up a SIGSEGV handler in the child, and let it handle segfaults rather than nullifying them. The handler is in the same page as the mmap stub. The second page is used as the stack. The handler reads cr2 and err from the sigcontext, sticks them at the base of the stack in a faultinfo struct, and SIGSTOPs itself. The kernel then reads the faultinfo and handles the fault. A complication on x86_64 is that this involves resetting the registers to the segfault values when the process is inside the kill system call. This breaks on x86_64 because %rcx will contain %rip because you tell SYSRET where to return to by putting the value in %rcx. So, this corrupts $rcx on return from the segfault. To work around this, I added an arch_finish_segv, which on x86 does nothing, but which on x86_64 ptraces the child back through the sigreturn. This causes %rcx to be restored by sigreturn and avoids the corruption. Ultimately, I think I will replace this with the trick of having it send itself a blocked signal which will be unblocked by the sigreturn. This will allow it to be stopped just after the sigreturn, and PTRACE_SYSCALLed without all the back-and-forth of PTRACE_SYSCALLing it through sigreturn. This runs on a stock host, so theoretically (and hopefully), tt mode isn't needed any more. We need to make sure that this is better in every way than tt mode, though. I'm concerned about the speed of address space updates and page fault handling, since they involve extra round-trips to the child. We can amortize the round-trip cost for large address space updates by writing all of the operations to the data page and having the child execute them all at the same time. This will help fork and exec, but not page faults, since they involve only one page. I can't think of any way to help page faults, except to add something like PTRACE_FAULTINFO to the host. There is PTRACE_SIGINFO, but UML doesn't use siginfo for SIGSEGV (or anything else) because there isn't enough information in the siginfo struct to handle page faults (the faulting operation type is missing). Adding that would make PTRACE_SIGINFO a usable equivalent to PTRACE_FAULTINFO. As for the code itself: - The system call stub is in arch/um/kernel/sys-$(SUBARCH)/stub.S. It is put in its own section of the binary along with stub_segv_handler in arch/um/kernel/skas/process.c. This is manipulated with run_syscall_stub in arch/um/kernel/skas/mem_user.c. syscall_stub will execute any system call at all, but it's only used for mmap, munmap, and mprotect. - The x86_64 stub calls sigreturn by hand rather than allowing the normal sigreturn to happen, because the normal sigreturn is a SA_RESTORER in UML's address space provided by libc. Needless to say, this is not available in the child's address space. Also, it does a couple of odd pops before that which restore the stack to the state it was in at the time the signal handler was called. - There is a new field in the arch mmu_context, which is now a union. This is the pid to be manipulated rather than the /proc/mm file descriptor. Code which deals with this now checks proc_mm to see whether it should use the usual skas code or the new code. - userspace_tramp is now used to create a new host process for every UML process, rather than one per UML processor. It checks proc_mm and ptrace_faultinfo to decide whether to map in the pages above its stack. - start_userspace now makes CLONE_VM conditional on proc_mm since we need separate address spaces now. - switch_mm_skas now just sets userspace_pid[0] to the new pid rather than PTRACE_SWITCH_MM. There is an addition to userspace which updates its idea of the pid being manipulated each time around the loop. This is important on exec, when the pid will change underneath userspace(). - The stub page has a pte, but it can't be mapped in using tlb_flush because it is part of tlb_flush. This is why it's required for it to be mapped in by userspace_tramp. Other random things: - The stub section in uml.lds.S is page aligned. This page is written out to the backing vm file in setup_physmem because it is mapped from there into user processes. - There's some confusion with TASK_SIZE now that there are a couple of extra pages that the process can't use. TASK_SIZE is considered by the elf code to be the usable process memory, which is reasonable, so it is decreased by two pages. This confuses the definition of USER_PGDS_IN_LAST_PML4, making it too small because of the rounding down of the uneven division. So we round it to the nearest PGDIR_SIZE rather than the lower one. - I added a missing PT_SYSCALL_ARG6_OFFSET macro. - um_mmu.h was made into a userspace-usable file. - proc_mm and ptrace_faultinfo are globals which say whether the host supports these features. - There is a bad interaction between the mm.nr_ptes check at the end of exit_mmap, stack randomization, and skas0. exit_mmap will stop freeing pages at the PGDIR_SIZE boundary after the last vma. If the stack isn't on the last page table page, the last pte page won't be freed, as it should be since the stub ptes are there, and exit_mmap will BUG because there is an unfreed page. To get around this, TASK_SIZE is set to the next lowest PGDIR_SIZE boundary and mm->nr_ptes is decremented after the calls to init_stub_pte. This ensures that we know the process stack (and all other process mappings) will be below the top page table page, and thus we know that mm->nr_ptes will be one too many, and can be decremented. Things that need fixing: - We may need better assurrences that the stub code is PIC. - The stub pte is set up in init_new_context_skas. - alloc_pgdir is probably the right place. Signed-off-by:
-
Pavel Machek authored
freezeable() already tests for TRACED/STOPPED processes, no need to do it twice. Signed-off-by:
Pavel Machek <pavel@suse.cz> Signed-off-by:
-
Pavel Machek authored
Fix error handling and whitespace in swsusp.c. swsusp_free() was called when there was nothing allocating, leading to oops. Signed-off-by:
-
Pavel Machek authored
Move device name resolution code around so that it is not called from resume-from-initrd. name_to_dev_t may be unavailable at that point. Signed-off-by:
-
Bernard Blackham authored
Fix u32 vs pm_message_t confusion in cpufreq. Signed-off-by:
Bernard Blackham <bernard@blackham.com.au> Signed-off-by:
-
Pavel Machek authored
Few more u32 vs. pm_message_t fixes. Signed-off-by:
-
Dave Jones authored
If CONFIG_NUMA isn't set, we use the define in <linux/mmzone.h> for early_pfn_to_nid (which defines it to 0). Because of this, the prototype needs to move inside the CONFIG_NUMA too, or anal gcc's get really confused. Signed-off-by:
Dave Jones <davej@redhat.com> Signed-off-by:
-
Dave Jones authored
The recent cleanups to asm-i386/mmzone.h were suboptimal nesting an ifdef of the same symbol. This patch removes some of the ifdef'ery to make things more readable again. Signed-off-by:
-
Shaohua Li authored
There has been some discuss about solving the SMP MTRR suspend/resume breakage, but I didn't find a patch for it. This is an intent for it. The basic idea is moving mtrr initializing into cpu_identify for all APs (so it works for cpu hotplug). For BP, restore_processor_state is responsible for restoring MTRR. Signed-off-by:
Shaohua Li <shaohua.li@intel.com> Acked-by:
Andi Kleen <ak@muc.de> Signed-off-by:
-
Adrian Bunk authored
This patch by Yoshihiro MATSUYAMA (already ACK'ed by David Howells) adds a defconfig for the frv arch. Signed-Off-By:
Yoshihiro MATSUYAMA <y.matsu@jp.fujitsu.com> Signed-off-by:
Adrian Bunk <bunk@stusta.de> Cc: David Howells <dhowells@redhat.com> Signed-off-by:
-
Anton Blanchard authored
We dont need to use the PERFMON exception on POWER5, in fact the firmware returns an error. Due to this just remove the warning. Also now that we have proper runlatch support we can remove the bootup hack. Signed-off-by:
Anton Blanchard <anton@samba.org> Signed-off-by:
-
Michael Ellerman authored
Not sure if we really need this, but it was handy to know which iSeries loop I was testing. Be consistent about printing which idle loop we're using, with this patch we cover all cases. Signed-off-by:
Michael Ellerman <michael@ellerman.id.au> Signed-off-by:
-
Anton Blanchard authored
Fix a compile warning introduced by the previous patches. Signed-off-by:
-
Anton Blanchard authored
- remove some unnecessary includes - add runlatch support - no need to use raw_smp_processor_id any more, current preempt debug logic checks for processes that are bound to one cpu. Signed-off-by:
-
Anton Blanchard authored
- separate out sleep logic in dedicated_idle, it was so far indented that it got squashed against the right side of the screen. - add runlatch support, looping on runlatch disable. Signed-off-by:
-
Anton Blanchard authored
- remove min/max yield time, we dont use the values anywhere - separate shared and dedicated idle loops - check need_resched again with irqs off to avoid sleeping with pending work - continually set runlatch off in idle loop, this means we dont need to turn the runlatch off on exception exit and suffer that associated cost for all exceptions. (A future patch will turn the runlatch on at exception entry) Signed-off-by:
-
Michael Ellerman authored
Now that the idle loop is configured by each platform we don't need idle_setup() anymore. Signed-off-by:
-
Michael Ellerman authored
This patch fixes up iSeries, pSeries, pmac and maple to set the correct idle function for each platform. Signed-off-by:
-
Michael Ellerman authored
dedicated_idle() and shared_idle() are only used by pSeries, so move them into pSeries_setup.c Signed-off-by:
-
Michael Ellerman authored
Move iSeries_idle() into iSeries_setup.c, no one else needs to know about it. Signed-off-by:
-
Michael Ellerman authored
This patch adds an idle member to the ppc_md structure and calls it from cpu_idle(). If a platform leaves ppc_md.idle as null it will get the default idle loop default_idle(). Signed-off-by:
-
