d/freebsd-skq
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

NFC: Update stale comments.

Browse Source 
Reviewed by:	alc
This commit is contained in:
mini 2003-11-10 00:44:00 +00:00
parent f6f94ac7fc
commit 918610ef5e
2 changed files with 4 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
sys/vm/vm_fault.c
View File

@ -801,7 +801,7 @@ RetryFault:;
/*
* To avoid trying to write_lock the map while another process
* has it read_locked (in vm_map_pageable), we do not try for
* has it read_locked (in vm_map_wire), we do not try for
* write permission. If the page is still writable, we will
* get write permission. If it is not, or has been marked
* needs_copy, we enter the mapping without write permission,

6
sys/vm/vm_kern.c
View File

@ -181,12 +181,12 @@ kmem_alloc(map, size)
/*
* Guarantee that there are pages already in this object before
* calling vm_map_pageable. This is to prevent the following
* calling vm_map_wire. This is to prevent the following
* scenario:
*
* 1) Threads have swapped out, so that there is a pager for the
* kernel_object. 2) The kmsg zone is empty, and so we are
* kmem_allocing a new page for it. 3) vm_map_pageable calls vm_fault;
* kmem_allocing a new page for it. 3) vm_map_wire calls vm_fault;
* there is no page, but there is a pager, so we call
* pager_data_request. But the kmsg zone is empty, so we must
* kmem_alloc. 4) goto 1 5) Even if the kmsg zone is not empty: when
@ -194,7 +194,7 @@ kmem_alloc(map, size)
* non-zero data. kmem_alloc is defined to return zero-filled memory.
*
* We're intentionally not activating the pages we allocate to prevent a
* race with page-out. vm_map_pageable will wire the pages.
* race with page-out. vm_map_wire will wire the pages.
*/
VM_OBJECT_LOCK(kernel_object);
for (i = 0; i < size; i += PAGE_SIZE) {