FreeBSD source tree https://cgit-dev.freebsd.org/src/atom?h=releng%2F8.1 2010-05-23T21:57:45Z 2010-05-23T21:57:45Z Alan Cox alc@FreeBSD.org 2010-05-23T21:57:45Z urn:sha1:452e6e0d44a4186c41fc7e11e4c5d4b0b13c9cc4 Add a comment about the proper use of vm_object_page_remove(). 2010-05-20T16:21:19Z Alan Cox alc@FreeBSD.org 2010-05-20T16:21:19Z urn:sha1:b28c6ddbbcc319ef6379eaa4b4b52efc3f03959f Change vm_object_madvise() so that it checks whether the page is invalid or unmanaged before acquiring the page queues lock. Neither of these tests require that lock. Moreover, a better way of testing if the page is unmanaged is to test the type of vm object. This avoids a pointless vm_page_lookup(). 2010-05-13T20:31:24Z Konstantin Belousov kib@FreeBSD.org 2010-05-13T20:31:24Z urn:sha1:04b7b96b05d1907532c281dd67bd57e0157fe1cf Remove a nonsensical test from vm_pageout_clean(). A page can't be in the inactive queue and have a non-zero wire count. 2010-05-13T20:26:16Z Konstantin Belousov kib@FreeBSD.org 2010-05-13T20:26:16Z urn:sha1:1f93868d76bb8084bba7e49761d3f8f3f10145d1 do not make sense or are harmful. MFC r206761 (by alc): Setting PG_REFERENCED on the requested page in swap_pager_getpages() is either redundant or harmful, depending on the caller. MFC r206768 (by alc): In vm_object_backing_scan(), setting PG_REFERENCED on a page before sleeping on that page is nonsensical. MFC r206770 (by alc): In vm_object_madvise() setting PG_REFERENCED on a page before sleeping on that page only makes sense if the advice is MADV_WILLNEED. MFC r206801 (by alc): There is no justification for vm_object_split() setting PG_REFERENCED on a page that it is going to sleep on. 2010-05-13T18:17:01Z Konstantin Belousov kib@FreeBSD.org 2010-05-13T18:17:01Z urn:sha1:ef7b0ac47b0990b42a5576e5eb5249ca0b4fe658 When doing kstack swapin, read as much pages in one run as possible. 2010-05-13T18:12:44Z Konstantin Belousov kib@FreeBSD.org 2010-05-13T18:12:44Z urn:sha1:83d6c6d862028f152dd1a4b23853e1771e5d6949 Simplify vm_thread_swapin(). Approved by: alc 2010-05-13T09:26:31Z Konstantin Belousov kib@FreeBSD.org 2010-05-13T09:26:31Z urn:sha1:e6ca6764e8085549da7db59f6441268a53a2613a In swap pager, do not free the non-requested pages from the run if they are wired. Kstack pages are wired, this change prepares swap pager for handling of long runs of kstack pages. 2010-05-10T11:50:26Z Konstantin Belousov kib@FreeBSD.org 2010-05-10T11:50:26Z urn:sha1:70020ca01e1adb1cc2e043f5a4e225eba7150fae Handle busy status of the page in a way expected for pager_getpage(). Flush requested page, unbusy other pages, do not clear m->busy. 2010-05-04T05:14:43Z Konstantin Belousov kib@FreeBSD.org 2010-05-04T05:14:43Z urn:sha1:88389d6103a728e508f8090b71c0bfef0bc66f8f When OOM searches for a process to kill, ignore the processes already killed by OOM. When killed process waits for a page allocation, try to satisfy the request as fast as possible. 2010-04-14T16:31:59Z Alan Cox alc@FreeBSD.org 2010-04-14T16:31:59Z urn:sha1:311d363969d37b9eb431754cb82f2b592165860a Introduce the function kmem_alloc_attr(), which allocates kernel virtual memory with the specified physical attributes. Correct an error in the prototype for kmem_malloc().