EXEC_ELF64 is used for NetBSD/sparc64, and EXEC_ELF32 is used for NetBSD/sparc_elf.
We really need a way to turn these on and off depending on whether we're building
a 32-bit or 64-bit kernel...
When we put a page on the collection list, we must subtract NPVPPG from the
total free count: one for each pv_entry that's free in that page, and one for
each free pv_entry in other pages that we're going to eat by moving the ones
in the page being collected.
pmap_find_pv(), pmap_clean_page() and pmap_remove_all() are only called on
managed pages, after VM initialization. Panic if this invariant is violated.
Also, panic if we try to enter a PT page through pmap_enter(), rather than
silently patching it up.
pmap_initialized is now #ifdef DIAGNOSTIC.
and dec_3maxplus.c. The ERRSYN/CHKSYN register contains data, not an address.
Pass the address of the register rather than the contents to dec_mtasic_err()
instead of the register contents so it can read/write the register.
Correctable memory errors won't trap in dec_mtasic_err() anymore.
and netinet, currently only tested under netinet.
Disabled by default, enabled by compiling the kernel with option
IFA_STATS. Enabling this feature seems to make the ip_output function
take 13% longer than before, which should be OK for people that need
this feature.
* Map the message buffer with access_type = VM_PROT_READ|VM_PROT_WRITE `just
because'.
* Map the file system buffers with access_type = VM_PROT_READ|VM_PROT_WRITE to
avoid possible problems with pagemove().
* Do not use VM_PROT_EXEC with either of the above.
* Map pages for /dev/mem with access_type = prot. Also, DO NOT use
pmap_kenter() for this, as we DO NOT want to lose modification information.
* Map pages in dumpsys() with VM_PROT_READ.
* Map pages in m68k mappedcopyin()/mappedcopyout() and writeback() with
access_type = prot.
* For now, bus_dma*(), pmap_map(), vmapbuf(), and similar functions still use
access_type = 0. This should probably be revisited.
siop2.c. Add wide negotiation and Ultra support. Modify siop.c to match
the siop2.c sync negotiation changes. The CyberStorm MKIII driver now
supports 15 targets. Remove some old table-driven sync rate stuff from
the original Zeus driver.
emulation of managed pages. This required the following `interesting' changes:
* File system buffers must be entered with an access type of
VM_PROT_READ|VM_PROT_WRITE, so that the pages will be accessible immediately.
Otherwise we would have to teach pagemove() to update the R/M information.
Since they're never eligible for paging, the latter is overkill.
* We must insure that pages allocated before the pmap is completely set up
(that is, pages allocated early by the VM system) are not eligible for R/M
emulation, since the memory needed for this isn't available. We do this by
allocating the pmap's internal memory with uvm_pageboot_alloc(). This also
fixes an absolutely horrible hack where the pmap only worked because page 0
happened to be mapped.
to be mapped.
Also:
* Push the wired page counting into the p->v list maintenance functions. This
avoids code duplication, and fixes some cases where we were confused about
which pages to do it with.
* Fix lots of problems associated with pmap_nightmare() (and rename it to
pmap_vac_me_harder()).
* Since the early pages are no longer considered `managed', just make
pmap_*_pv() panic if !pmap_initialized.
memory access a mapping was caused by. This is passed through from uvm_fault()
and udv_fault(), and in most other cases is 0.
The pmap module may use this to preset R/M information. On MMUs which require
R/M emulation, the implementation may preset the bits and avoid taking another
fault. On MMUs which keep R/M information in hardware, the implementation may
preset its cached bits to speed up the next call to pmap_is_modified() or
pmap_is_referenced().
numerous pagedaemon improvements were needed to make this useful:
- don't bother waking up procs waiting for memory if there's none to be had.
- start 4 times as many pageouts as we need free pages.
this should reduce latency in low-memory situations.
- in inactive scanning, if we find dirty swap-backed pages when swap space
is full of non-resident pages, reactivate some number of these to flush
less active pages to the inactive queue so we can consider paging them out.
this replaces the previous scheme of inactivating pages beyond the
inactive target when we failed to free anything during inactive scanning.
- during both active and inactive scanning, free any swap resources from
dirty swap-backed pages if swap space is full. this allows other pages
be paged out into that swap space.
and ignore the error. Scanset 2 should be the default after reset, so
this allows some broken keyboards to work. (Reset is needed because at
least 1 keyboard locks up if the "set scanset" is attempted.)
allocated from a pool, and the MIPS and Alpha use KSEG to map pool
pages. So, mb_map wasn't actually being used. Saves around 4MB of
kernel virtual address space in a typical configuration.
Garbage-collect the related VM_MBUF_SIZE constant.
- ETHER_ADDR_LEN: length of Ethernet address (actually, we already defined
this).
- ETHER_TYPE_LEN: length of the Ethernet header `type' field.
- ETHER_CRC_LEN: length of the Ethernet CRC (explorer got this already, mostly
because I forgot to commit these changes earlier).
- ETHER_HDR_LEN: total length of the Ethernet header
- ETHER_MAX_LEN: maximum length of an Ethernet frame, including header and CRC
- ETHER_MIN_LEN: minimum length of an Ethernet frame, including header and CRC
Define ETHERMTU and ETHERMIN (payload sizes) in terms of the above constants.
namely, toggle whether vnodes loaded only for cleaning (as opposed to
normal filesystem use) are freed to the *head* of the vnode free list,
rather than the tail. This should avoid a possible cache flushing
effect, if the cleaner cleans a segment containing a large number of
live inodes.
dirop is completely written to disk. This means that ordinary calls to
ufs vnops which would ordinarily call VOP_INACTIVE through vrele/vput,
don't. This patch detects that condition after such vnops have been
run, and calls VOP_INACTIVE if it would ordinarily have been called by
the ufs call.
if we are short on vnodes, lfs_vflush from another process can grab a
vnode that lfs_markv has already processed but not yet written; but
lfs_markv holds the seglock. When lfs_vflush gets around to writing it,
the context for copyin is gone. So, now lfs_markv calls copyin itself,
rather than having lfs_writeseg do it.
dbj