| Age | Commit message (Collapse) | Author | |
|---|---|---|---|
| 2003-12-24 | Murphy's law applied to BSD: | Miod Vallat | |
| Every non-KNF file in the source tree contains at least one severe bug. KNF and a timid start at cleaning, but not very far, so there is no functional change at the moment. This code is still pathetic for now, but it sorta works and the price was right. | |||
| 2003-12-24 | Since we have real bus error faults, and handle them as such, always treat | Miod Vallat | |
| uvm_fault() returning EACCES as a segmentation fault rather than a bus error, whatever address the fault is at. As a result, this correctly delivers SIGSEGV, rather than SIGBUS, when attempting to write to a page with only PROT_READ permissions. | |||
| 2003-12-23 | Data access faults occuring inside copyin() or copyinstr() would get | Miod Vallat | |
| treated as usermode traps, because they reference a user space address. However, the tests for pcb_onfault being set were only present in the kernelmode traps handling. Since pcb_onfault is only set in those functions, move the associated recovery code from the kernelmode part to the usermode part, and only attempt to jump to pcb_onfault() if the access could not be resolved by uvm_fault() earlier. This lets things like setlogin(NULL) correctly return EFAULT, rather than killing the process with SIGSEGV, and incidentally lets sendmail in a non-default configuration run. | |||
| 2003-12-23 | style | Miod Vallat | |
| 2003-12-23 | Simplify the syscall() routines. Also put a really meaningful limit to the | Miod Vallat | |
| number of syscall arguments, and enforce it correctly. | |||
| 2003-12-22 | Do not play games with the instruction fetch pointers when preserving a | Miod Vallat | |
| process state in sendsig(), for it must be restored in the exact same fetching state by sigreturn(). mvme88k now passes the sigreturn regression test. | |||
| 2003-12-22 | Get rid of kvtop() usage in dma_cachectl(). | Miod Vallat | |
| 2003-12-22 | Correctly handle vme attachments which specify a larger ca_len size than | Miod Vallat | |
| initially expected. | |||
| 2003-12-21 | -fno-builtin-* finally tested | Marc Espie | |
| 2003-12-21 | Switch db_expr_t from int to long, per consistency with other platforms. | Miod Vallat | |
| 2003-12-19 | Pass -Wformat, and silence a few vme* debug printf. | Miod Vallat | |
| 2003-12-19 | The physical address of a pmap's segment table is only used to compute an | Miod Vallat | |
| apr value used in pmap_activate(). So, rather than storing the address in the pmap structure, store the precomputed apr value... | |||
| 2003-12-19 | When temporarily unlocking a pmap, do not release the spl as well. | Miod Vallat | |
| While there, get rid of that ugly PT_FREE() macro. | |||
| 2003-12-14 | Replace pmap_testbit/pmap_changebit sequences with a tailor-made pmap_unsetbit | Miod Vallat | |
| routine, in order to speed up pmap_clear_modify() and pmap_clear_reference(). | |||
| 2003-12-14 | Remove unnecessary trunc_page() calls, and change pmap_enter() logic to | Miod Vallat | |
| match m68k. | |||
| 2003-12-12 | Provide a better delay() resolution on 188. | Miod Vallat | |
| 2003-12-11 | Oops, forgot to commit this chunk of pcb_t removal. Spotted by deraadt. | Miod Vallat | |
| 2003-11-18 | Remove the duplicated code in ssh_checkintr() - checked against mvme68k. | Miod Vallat | |
| 2003-11-18 | In ssh_checkintr(), better check pointers for NULL before they | Miod Vallat | |
| are dereferenced. XXX Looks like there are duplicated sections in this function. This driver needs a huge cleanup anyway... | |||
| 2003-11-17 | Be sure to pmap_deactivate() pmaps when processes exit or are scheduled out. | Miod Vallat | |
| 2003-11-16 | convert __attribute__((__packed__)) to __packed so that parsers unaware | Anil Madhavapeddy | |
| of gcc extensions have more of a chance. ok mcbride@, no objections from millert@, deraadt@ | |||
| 2003-11-14 | the the; rohee@ | Miod Vallat | |
| 2003-11-09 | Typos; Jonathan Gray | Miod Vallat | |
| 2003-11-08 | Inline invalidate_pte(). | Miod Vallat | |
| 2003-11-08 | Save the current floatingpoint state in the pcb. | Miod Vallat | |
| 2003-11-07 | adress -> address, and a few more; all from Jonathon Gray; | Jason McIntyre | |
| (mvme68k/mvme88k) vs.c and (vax) if_le.c ok miod@ isakmpd ones ok ho@ | |||
| 2003-11-06 | move netisr definition into md code to allow arch provide suitable ↵ | Michael Shalayeff | |
| allocation; tested on most archs | |||
| 2003-11-03 | spelling fixes (in the comments) | David Krause | |
| 2003-10-28 | Split pmap_remove_range() into its main loop and a new function, | Miod Vallat | |
| pmap_remove_pte(), which takes the va and a pte. Use pmap_remove_pte() instead of pmap_remove_range() in strategic places, in order to save some cycles by avoiding unnecessary pte address recomputations. pmap_remove_range() is still preferred for large amounts of memory, as it will skip empty segments. | |||
| 2003-10-28 | In pmap_remove_range(), be sure to propagate pvlist flags when removing | Miod Vallat | |
| the head entry. While there, invalidate the affected pte earlier for safety, and store its PG_U bit in the pvlist flags as well. And also get rid of a pdt table group leftover in pmap_collect(). | |||
| 2003-10-27 | Be sure to always have a valid faulting address when we send a signal | Miod Vallat | |
| to a process. | |||
| 2003-10-27 | Be sure to flush the entire va space in cpu_switch(). | Miod Vallat | |
| This used to work because the cmmu code is suboptimal at the moment. | |||
| 2003-10-27 | Remove dead code and polish comments | Miod Vallat | |
| 2003-10-24 | Remove dead code. | Miod Vallat | |
| 2003-10-24 | No \n in panic() messages... | Miod Vallat | |
| 2003-10-23 | Correctly recover from bus error during guarded_access() on 88110. | Miod Vallat | |
| 2003-10-23 | No dependency on NDART needed here. | Miod Vallat | |
| 2003-10-19 | Simplify pmap DEBUG test constructs. | Miod Vallat | |
| 2003-10-16 | Better DEBUG output. | Miod Vallat | |
| 2003-10-15 | Better reset code, even. | Miod Vallat | |
| 2003-10-15 | Do not assert SYSRST on 1x7 boards, if we are not a SYSCON, or unexpected | Miod Vallat | |
| collateral damage will happen. | |||
| 2003-10-13 | Get rid of the "page table group" stuff. "groups" used to be one logical page | Miod Vallat | |
| big anyways. However, getting rid of the various constants (expanding to 0 for logs and shifts, and 1 for size) allows us to do further simplifications in some pmap loops, as there is no risk anymore of address wrapround. While there, fix some typos and comments. | |||
| 2003-10-13 | Do not define LEDEBUG by default. | Miod Vallat | |
| 2003-10-11 | Better sync comment with reality after last commit. | Miod Vallat | |
| 2003-10-11 | Get rid of "union cpupid"; makes <machine/cpus.h> disappear as well. | Miod Vallat | |
| 2003-10-11 | Nuke trailing whitespace. | Miod Vallat | |
| 2003-10-11 | Page tables can be cached again on 187 and 188, as long as they are cached | Miod Vallat | |
| write through and global (for 188). | |||
| 2003-10-11 | Simplify DAE debug display. | Miod Vallat | |
| 2003-10-11 | Merge sclock and clock pseudo-devices. | Miod Vallat | |
| 2003-10-10 | In pmap_expand_kmap(), be sure to initialize new page tables. | Miod Vallat | |
 |
index : src | |
| OpenBSD base system |
| summaryrefslogtreecommitdiff |