| Age | Commit message (Collapse) | Author | |
|---|---|---|---|
| 2011-09-18 | A few more fixes to the trap save/restore code. | Mark Kettenis | |
| 2011-09-18 | Disable space-ID hashing like we do on hppa. | Mark Kettenis | |
| 2011-09-18 | Use an uvm object to keep track of the page table pages, just like we do on | Mark Kettenis | |
| hppa to reduce the diffs between the two pmaps. Also add cache flushing in places where hppa does them. | |||
| 2011-09-18 | Take into account the PIE bits when masking PDEs. | Mark Kettenis | |
| 2011-09-18 | Add missing cache and TLB purges. | Mark Kettenis | |
| 2011-09-08 | Provide namespace-safe alignment macros in <machine/_types.h>, with | Philip Guenthe | |
| compat names kept in <machine/param.h>. In <sys/socket.h>, pull in <sys/_types.h> instead of the namespace polluting <machine/param.h> and completely eliminate __CMSG_ALIGN, replaced by _ALIGN ok deraadt@ | |||
| 2011-09-08 | Make the INT_FAST*_{MIN,MAX} macros match the types they're defined to. | Philip Guenthe | |
| Since the underlying types of the int_fast types are set by machine/_types.h, put internal macros in that same file and define the exposed INT_FAST*_{MIN,MAX} macros from those. ok millert@, kettenis@ | |||
| 2011-08-29 | Only provide FLT_EVAL_METHOD for C99. Add missing DECIMAL_DIG for C99. | Mark Kettenis | |
| ok guenther@ | |||
| 2011-08-19 | Bring libc and libm in line with the compiler now that we no longer have a | Mark Kettenis | |
| 128-bit long double. Diff committed on behalf of martynas@ | |||
| 2011-08-17 | Set the PSL_O bit, like we do for 64-bit capable CPUs on hppa. | Mark Kettenis | |
| 2011-08-17 | Don't set the PSL_S bit. We generally don't tend to hide CPU counter timers | Mark Kettenis | |
| from userland, and setting it upsets the sigreturn code. | |||
| 2011-08-16 | Use pitlbe and pdtlbe instructions when flushing the entire TLB. | Mark Kettenis | |
| 2011-08-16 | The gp value is stored in a different register on hppa64. | Mark Kettenis | |
| 2011-08-16 | Old proc is stored in %arg0, not in %arg2. | Mark Kettenis | |
| 2011-08-16 | Compile in serial console settings support if any com(4) device is included | Mark Kettenis | |
| in the kernel, not only when com @ gsc is there. | |||
| 2011-08-16 | Don't export 'struct pmap' to userland. | Mark Kettenis | |
| 2011-08-07 | A few more fixes to the lazy FPU context switching code. | Mark Kettenis | |
| 2011-08-07 | Avoid clobbering system call arguments upon system call entry. | Mark Kettenis | |
| 2011-08-07 | Enable ssio(4) and com(4). | Mark Kettenis | |
| 2011-08-07 | Bring over ssio(4) from hppa for proper serial console support. | Mark Kettenis | |
| 2011-08-07 | Recognize 64-bit store instructions. | Mark Kettenis | |
| 2011-08-07 | Make sure interrupts are properly masked at the level they're established at. | Mark Kettenis | |
| The old code left an interrupt enabled at that level, making spl-protection non functional. Remove __packed from struct hppa_iv. Instead re-order the members to pack things efficient on a 64-bit host. Remove explicit padding. | |||
| 2011-08-04 | Make this compile with gcc4. Same fix as for hppa. | Mark Kettenis | |
| 2011-08-01 | Synch with hppa. Seems to fix at least one of the remaining pmap bugs. | Mark Kettenis | |
| 2011-07-10 | If there is recognized MD disk partitioning info, use it to determine | Kenneth R Westerback | |
| where to write the OpenBSD disklabel even if there isn't yet a valid OpenBSD disklabel at that location. Alpha/Mac68k/Sparc/Sparc64 to come. | |||
| 2011-07-09 | syscall() can be a lot simpler on hpaa64. | Mark Kettenis | |
| 2011-07-08 | Don't accidentally mask off the T_USER bit before calling trap(). | Mark Kettenis | |
| At least mickey was as confused by the backwards hppa bit numbering as I am. | |||
| 2011-07-08 | When reading dos labels use DOS_LABELSECTOR and don't use LABELOFFSET. For ↵ | Kenneth R Westerback | |
| all these arch's LABELSECTOR == DOS_LABELSECTOR == 1, and LABELOFFSET == 0. Thus, to quote bob, "This is a no-op". Makes the expression used when writing label the same as the one used in readdoslabel(). | |||
| 2011-07-07 | Add missing bits to the syscall entry path. Not perfect yet, but it works | Mark Kettenis | |
| for at least the open(2), write(2) and exit(2) system calls. Map the syscall entry page publically. For now the physical address is hardcoded because of toolchain problems. It is somewhat likely to say the same as long as we don't add stuff to the start of locore.S. | |||
| 2011-07-07 | Take more care to write 'dos' labels in the same place readdoslabel() | Kenneth R Westerback | |
| will look for them. Without changing native label handling. Compile tested on hppa by kettenis@. "I see no issues" deraadt@ | |||
| 2011-07-07 | Stuff arg0-arg7 in the right slots of the trap frame. | Mark Kettenis | |
| 2011-07-07 | The drahn memorial bad kernel build fix: prevent blood pressure | Philip Guenthe | |
| spikes in other developers by making it so that removal of a .d file without removing the corresponding object will result in the latter being treated as out of date. ok beck@ art@ drahn@ | |||
| 2011-07-07 | Raise the cache aliasing boundary to 4MB like we did on hppa a couple of years | Mark Kettenis | |
| ago. | |||
| 2011-07-07 | Bring over a few more cache flushing and TLB purging fixes from hppa. | Mark Kettenis | |
| 2011-07-07 | Don't panic if uvm_fault fails, but enter ddb such that we can inspect the | Mark Kettenis | |
| state that caused the fault. | |||
| 2011-07-07 | Set PSL_W bit for forked processes such that they execute in 64-bit mode | Mark Kettenis | |
| instead of 32-bit mode. | |||
| 2011-07-06 | make clean should clean .d files, so as to leave a fresh canvas. | Ted Unangst | |
| ok beck deraadt | |||
| 2011-07-05 | Recommit the reverted sigacts change now that the NFS use-after-free | Philip Guenthe | |
| problem has been tracked down. This fixes the sharing of the signal handling state: shared bits go in sigacts, per-rthread bits goes in struct proc. ok deraadt@ | |||
| 2011-07-04 | Finish ansification. | Mark Kettenis | |
| 2011-07-04 | Give each pmap its own space ID. The current algorithm is silly, as we'll wrap | Mark Kettenis | |
| and might end up with duplicates, but that's not an issue until we're multi-user. Change the TLB miss code to enter mappings with the right protection ID. Properly switch pmaps on context switches. This makes the copyins we do before starting init actually work instead of failing with EFAULT. | |||
| 2011-06-29 | Replace all instances of "scsibus* at hba?" with a single "scsibus* at | Matthew Dempsky | |
| scsi?" rule, similar to how ethernet PHY drivers attach at mii. Discussed on icb. | |||
| 2011-06-26 | if_downall() cannot be done MI in sys_reboot() because vfs_shutdown() | Theo de Raadt | |
| might need network (ie. nfs). Move the call to the MD boot() routines. This cause for boot hangs diagnosed by kettenis. | |||
| 2011-06-24 | When reading the value from a simple lock, we do not need to ldcws into a | Joel Sing | |
| __cpu_simple_lock_t - gcc most likely ignores the alignment for a stack variable anyway. Also remove unnecessary initialisation. ok kettenis@ miod@ | |||
| 2011-06-05 | Move the bufcachepercent setting code to MI locations -- set it to 42% | Theo de Raadt | |
| for now; that is unlikely to hit some of the remaining starvation bugs. Repair the bufpages calculation too; i386 was doing it ahead of time (incorrectly) and then re-calculating it. ok thib | |||
| 2011-05-30 | Remove the freelist member from vm_physseg | Owain Ainsworth | |
| The new world order of pmemrange makes this data completely redundant (being dealt with by the pmemrange constraints instead). Remove all code that messes with the freelist. While touching every caller of uvm_page_physload() anyway, add the flags argument to all callers (all but one is 0 and that one already used PHYSLOAD_DEVICE) and remove the macro magic to allow callers to continue without it. Should shrink the code a bit, as well. matthew@ pointed out some mistakes i'd made. ``freelist death, I like. Ok.' ariane@ `I agree with the general direction, go ahead and i'll fix any fallout shortly'' miod@ (68k 88k and vax i could not check would build) | |||
| 2011-05-25 | Also enclose variable names with external linkage in __BEGIN_DECLS. | Martynas Venckus | |
| Doesn't matter much since C++ ABI used by GCC doesn't mangle variable names; however technically is required by Section 7.5 of the C++ spec. Discussed with/OK guenther@, matthew@. | |||
| 2011-05-24 | On OpenBSD, we've always attached scsibus(4) to umass(4), even for | Matthew Dempsky | |
| ATAPI devices. atapiscsi(4) is only for handling ATAPI devices on an ATA bus, so umass(4) shouldn't care about it. ok krw@, dlg@; no objections from deraadt@ | |||
| 2011-05-21 | Add pci_intr_map_msi() stub. | Mark Kettenis | |
| 2011-05-07 | So long, uvm_pglist.h | Owain Ainsworth | |
| This header defined three thing. two of which are unused throughout the tree, the final one was the definition of the pagq head type, move that to uvm_page.h and nuke the header ok thib@. Thanks to krw@ for testing the hppa build for me. | |||
| 2011-05-05 | The gateway page (which contains the system call entry point) needs to be | Mark Kettenis | |
| public such that it is accessable from all userland processes. | |||
 |
index : src | |
| OpenBSD base system |
| summaryrefslogtreecommitdiff |