Source-Changes-HG archive
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]
[src/netbsd-7]: src/sys Pull up following revision(s) (requested by mrg in ti...
details: https://anonhg.NetBSD.org/src/rev/0cbf52eb1d22
branches: netbsd-7
changeset: 799244:0cbf52eb1d22
user: snj <snj%NetBSD.org@localhost>
date: Thu Apr 23 07:31:16 2015 +0000
description:
Pull up following revision(s) (requested by mrg in ticket #718):
sys/arch/x86/include/pmap.h: revision 1.56
sys/arch/x86/x86/pmap.c: revision 1.188
sys/dev/pci/agp_amd64.c: revision 1.8
sys/dev/pci/agp_i810.c: revision 1.118
sys/external/bsd/drm2/dist/drm/i915/i915_dma.c: revision 1.16
sys/external/bsd/drm2/dist/drm/i915/i915_gem.c: revision 1.29
sys/external/bsd/drm2/dist/drm/nouveau/nouveau_agp.c: revision 1.3
sys/external/bsd/drm2/dist/drm/nouveau/nouveau_ttm.c: revision 1.4
sys/external/bsd/drm2/dist/drm/radeon/atombios_crtc.c: revision 1.3
sys/external/bsd/drm2/dist/drm/radeon/radeon_agp.c: revision 1.3
sys/external/bsd/drm2/dist/drm/radeon/radeon_display.c: revision 1.3
sys/external/bsd/drm2/dist/drm/radeon/radeon_legacy_crtc.c: revision 1.2
sys/external/bsd/drm2/dist/drm/radeon/radeon_object.c: revision 1.3
sys/external/bsd/drm2/dist/drm/radeon/radeon_ttm.c: revision 1.7
sys/external/bsd/drm2/dist/drm/ttm/ttm_bo.c: revisions 1.7-1.10
sys/external/bsd/drm2/dist/drm/ttm/ttm_bo_util.c: revision 1.5
sys/external/bsd/drm2/i915drm/intelfb.c: revision 1.13
sys/external/bsd/drm2/include/drm/drm_wait_netbsd.h: revisions 1.12, 1.13
sys/external/bsd/drm2/include/linux/mm.h: revision 1.5
sys/external/bsd/drm2/include/linux/pci.h: revisions 1.16, 1.17
sys/external/bsd/drm2/nouveau/nouveaufb.c: revision 1.2
sys/external/bsd/drm2/radeon/radeon_pci.c: revisions 1.8, 1.9
sys/uvm/uvm_init.c: revision 1.46
Hack against the blank console problem:
Leave the CLUT alone on ancient cards. At least this leaves us with a
semi working console (red and blue are flipped). Leave an example of what
seems to be happening but disable it because colors are better than 444 bit
greyscale.
--
Initialize P->V tracking for unmanaged device pages in uvm_init.
Conditional on __HAVE_PMAP_PV_TRACK until we add it to all pmaps.
MI part of pmap_pv(9) change proposed on tech-kern:
https://mail-index.netbsd.org/tech-kern/2015/03/26/msg018561.html
--
Implement pmap_pv(9) for x86 for P->V tracking of unmanaged pages.
Proposed on tech-kern with no objections:
https://mail-index.netbsd.org/tech-kern/2015/03/26/msg018561.html
--
Use pmap_pv(9) to remove mappings of Intel graphics aperture pages.
Proposed on tech-kern with no objections:
https://mail-index.netbsd.org/tech-kern/2015/03/26/msg018561.html
Further background at:
https://mail-index.netbsd.org/tech-kern/2014/07/23/msg017392.html
--
Use pmap_pv(9) to remove mappings of device pages in TTM.
Adapt nouveau and radeon to do pmap_pv_track for their device pages.
Proposed on tech-kern with no objections:
https://mail-index.netbsd.org/tech-kern/2015/03/26/msg018561.html
Further background at:
https://mail-index.netbsd.org/tech-kern/2014/07/23/msg017392.html
--
Fix error branches in agp_amd64.c.
- agp_generic_detach always.
- Free asc if it was allocated. (Found by Brainy, noted by maxv@.)
- Free the GATT if it was allocated.
--
pmf_device_register returns false on failure, not true
--
In DRM_SPIN_WAIT_ON, don't stop after waiting only one tick.
Continue the loop to recheck the condition and count the whole
duration.
--
Don't use the video BIOS memory as an i915 flush page!
--
Don't let anyone else allocate the video BIOS either.
--
Missed a zero: it's 0x100000, not 0x10000.
--
Don't reserve if atomic -- caller must have pre-pinned the buffer.
--
Don't reserve if atomic -- caller must have pre-pinned the buffer.
--
almost add radeondrmkms suspend/resume support. it unfortunately doesn't work.
--
Need the page's uvm object lock to do pmap_page_protect.
--
Use KASSERTMSG to show bad base/offset.
--
KASSERT about page-alignment on initialization too.
--
Don't break when hardclock_ticks wraps around.
Since we now only count time spent in wait, rather than determining
the end time and checking whether we've passed it, timeouts might be
marginally longer in effect. Unlikely to be an issue.
--
Remove broken drm2 vm_mmap stub. Can't possibly have ever worked.
--
apply some of the additional changes from Arto Huusko in PR#49645:
- call pmf_device_deregister on detach.
i've kept the "resume = true" for radeon_resume_kms() call as it
seems to work for me (indeed, code inspection shows it is unused
on netbsd :-)
my old nforce4 box that can resume old drm (or could, last i tried
several years ago) while X and GL apps were running, can at least
survive a resume if X hasn't started. my one attempt so far with
X exited, but having run, did not work.
--
First attempt to make ttm_buffer_object_transfer less bogus.
--
Make sure mem.bus.is_iomem is initialized. PR 49833
diffstat:
sys/arch/x86/include/pmap.h | 28 +-
sys/arch/x86/x86/pmap.c | 309 ++++++++++--
sys/dev/pci/agp_amd64.c | 45 +-
sys/dev/pci/agp_i810.c | 21 +-
sys/external/bsd/drm2/dist/drm/i915/i915_dma.c | 7 +
sys/external/bsd/drm2/dist/drm/i915/i915_gem.c | 39 +-
sys/external/bsd/drm2/dist/drm/nouveau/nouveau_agp.c | 13 +-
sys/external/bsd/drm2/dist/drm/nouveau/nouveau_ttm.c | 14 +-
sys/external/bsd/drm2/dist/drm/radeon/atombios_crtc.c | 17 +-
sys/external/bsd/drm2/dist/drm/radeon/radeon_agp.c | 7 +
sys/external/bsd/drm2/dist/drm/radeon/radeon_display.c | 36 +-
sys/external/bsd/drm2/dist/drm/radeon/radeon_legacy_crtc.c | 13 +-
sys/external/bsd/drm2/dist/drm/radeon/radeon_object.c | 8 +
sys/external/bsd/drm2/dist/drm/radeon/radeon_ttm.c | 8 +
sys/external/bsd/drm2/dist/drm/ttm/ttm_bo.c | 22 +-
sys/external/bsd/drm2/dist/drm/ttm/ttm_bo_util.c | 5 +
sys/external/bsd/drm2/i915drm/intelfb.c | 6 +-
sys/external/bsd/drm2/include/drm/drm_wait_netbsd.h | 77 +-
sys/external/bsd/drm2/include/linux/mm.h | 59 +--
sys/external/bsd/drm2/include/linux/pci.h | 4 +-
sys/external/bsd/drm2/nouveau/nouveaufb.c | 6 +-
sys/external/bsd/drm2/radeon/radeon_pci.c | 51 +-
sys/uvm/uvm_init.c | 7 +-
23 files changed, 571 insertions(+), 231 deletions(-)
diffs (truncated from 1521 to 300 lines):
diff -r c4e185269b7d -r 0cbf52eb1d22 sys/arch/x86/include/pmap.h
--- a/sys/arch/x86/include/pmap.h Thu Apr 23 07:03:49 2015 +0000
+++ b/sys/arch/x86/include/pmap.h Thu Apr 23 07:31:16 2015 +0000
@@ -1,4 +1,4 @@
-/* $NetBSD: pmap.h,v 1.55 2013/10/17 20:59:16 christos Exp $ */
+/* $NetBSD: pmap.h,v 1.55.4.1 2015/04/23 07:31:16 snj Exp $ */
/*
* Copyright (c) 1997 Charles D. Cranor and Washington University.
@@ -243,8 +243,10 @@
void pmap_activate(struct lwp *);
void pmap_bootstrap(vaddr_t);
bool pmap_clear_attrs(struct vm_page *, unsigned);
+bool pmap_pv_clear_attrs(paddr_t, unsigned);
void pmap_deactivate(struct lwp *);
-void pmap_page_remove (struct vm_page *);
+void pmap_page_remove(struct vm_page *);
+void pmap_pv_remove(paddr_t);
void pmap_remove(struct pmap *, vaddr_t, vaddr_t);
bool pmap_test_attrs(struct vm_page *, unsigned);
void pmap_write_protect(struct pmap *, vaddr_t, vaddr_t, vm_prot_t);
@@ -258,6 +260,11 @@
void pmap_emap_remove(vaddr_t, vsize_t);
void pmap_emap_sync(bool);
+#define __HAVE_PMAP_PV_TRACK 1
+void pmap_pv_init(void);
+void pmap_pv_track(paddr_t, psize_t);
+void pmap_pv_untrack(paddr_t, psize_t);
+
void pmap_map_ptes(struct pmap *, struct pmap **, pd_entry_t **,
pd_entry_t * const **);
void pmap_unmap_ptes(struct pmap *, struct pmap *);
@@ -359,6 +366,23 @@
}
/*
+ * pmap_pv_protect: change the protection of all recorded mappings
+ * of an unmanaged page
+ */
+
+__inline static void __unused
+pmap_pv_protect(paddr_t pa, vm_prot_t prot)
+{
+ if ((prot & VM_PROT_WRITE) == 0) {
+ if (prot & (VM_PROT_READ|VM_PROT_EXECUTE)) {
+ (void) pmap_pv_clear_attrs(pa, PG_RW);
+ } else {
+ pmap_pv_remove(pa);
+ }
+ }
+}
+
+/*
* pmap_protect: change the protection of pages in a pmap
*
* => this function is a frontend for pmap_remove/pmap_write_protect
diff -r c4e185269b7d -r 0cbf52eb1d22 sys/arch/x86/x86/pmap.c
--- a/sys/arch/x86/x86/pmap.c Thu Apr 23 07:03:49 2015 +0000
+++ b/sys/arch/x86/x86/pmap.c Thu Apr 23 07:31:16 2015 +0000
@@ -1,4 +1,4 @@
-/* $NetBSD: pmap.c,v 1.183.2.1 2014/10/14 07:37:37 martin Exp $ */
+/* $NetBSD: pmap.c,v 1.183.2.2 2015/04/23 07:31:16 snj Exp $ */
/*-
* Copyright (c) 2008, 2010 The NetBSD Foundation, Inc.
@@ -171,7 +171,7 @@
*/
#include <sys/cdefs.h>
-__KERNEL_RCSID(0, "$NetBSD: pmap.c,v 1.183.2.1 2014/10/14 07:37:37 martin Exp $");
+__KERNEL_RCSID(0, "$NetBSD: pmap.c,v 1.183.2.2 2015/04/23 07:31:16 snj Exp $");
#include "opt_user_ldt.h"
#include "opt_lockdebug.h"
@@ -191,6 +191,8 @@
#include <sys/intr.h>
#include <sys/xcall.h>
#include <sys/kcore.h>
+#include <sys/kmem.h>
+#include <sys/pserialize.h>
#include <uvm/uvm.h>
@@ -248,8 +250,10 @@
* data structures we use include:
*
* - struct pmap: describes the address space of one thread
+ * - struct pmap_page: describes one pv-tracked page, without
+ * necessarily a corresponding vm_page
* - struct pv_entry: describes one <PMAP,VA> mapping of a PA
- * - struct pv_head: there is one pv_head per managed page of
+ * - struct pv_head: there is one pv_head per pv-tracked page of
* physical memory. the pv_head points to a list of pv_entry
* structures which describe all the <PMAP,VA> pairs that this
* page is mapped in. this is critical for page based operations
@@ -303,7 +307,7 @@
*
* - pvh_lock (per pv_head)
* this lock protects the pv_entry list which is chained off the
- * pv_head structure for a specific managed PA. it is locked
+ * pv_head structure for a specific pv-tracked PA. it is locked
* when traversing the list (e.g. adding/removing mappings,
* syncing R/M bits, etc.)
*
@@ -459,6 +463,120 @@
}
/*
+ * unmanaged pv-tracked ranges
+ *
+ * This is a linear list for now because the only user are the DRM
+ * graphics drivers, with a single tracked range per device, for the
+ * graphics aperture, so there are expected to be few of them.
+ *
+ * This is used only after the VM system is initialized well enough
+ * that we can use kmem_alloc.
+ */
+
+struct pv_track {
+ paddr_t pvt_start;
+ psize_t pvt_size;
+ struct pv_track *pvt_next;
+ struct pmap_page pvt_pages[];
+};
+
+static struct {
+ kmutex_t lock;
+ pserialize_t psz;
+ struct pv_track *list;
+} pv_unmanaged __cacheline_aligned;
+
+void
+pmap_pv_init(void)
+{
+
+ mutex_init(&pv_unmanaged.lock, MUTEX_DEFAULT, IPL_VM);
+ pv_unmanaged.psz = pserialize_create();
+ pv_unmanaged.list = NULL;
+}
+
+void
+pmap_pv_track(paddr_t start, psize_t size)
+{
+ struct pv_track *pvt;
+ size_t npages;
+
+ KASSERT(start == trunc_page(start));
+ KASSERT(size == trunc_page(size));
+
+ npages = size >> PAGE_SHIFT;
+ pvt = kmem_zalloc(offsetof(struct pv_track, pvt_pages[npages]),
+ KM_SLEEP);
+ pvt->pvt_start = start;
+ pvt->pvt_size = size;
+
+ mutex_enter(&pv_unmanaged.lock);
+ pvt->pvt_next = pv_unmanaged.list;
+ membar_producer();
+ pv_unmanaged.list = pvt;
+ mutex_exit(&pv_unmanaged.lock);
+}
+
+void
+pmap_pv_untrack(paddr_t start, psize_t size)
+{
+ struct pv_track **pvtp, *pvt;
+ size_t npages;
+
+ KASSERT(start == trunc_page(start));
+ KASSERT(size == trunc_page(size));
+
+ mutex_enter(&pv_unmanaged.lock);
+ for (pvtp = &pv_unmanaged.list;
+ (pvt = *pvtp) != NULL;
+ pvtp = &pvt->pvt_next) {
+ if (pvt->pvt_start != start)
+ continue;
+ if (pvt->pvt_size != size)
+ panic("pmap_pv_untrack: pv-tracking at 0x%"PRIxPADDR
+ ": 0x%"PRIxPSIZE" bytes, not 0x%"PRIxPSIZE" bytes",
+ pvt->pvt_start, pvt->pvt_size, size);
+ *pvtp = pvt->pvt_next;
+ pserialize_perform(pv_unmanaged.psz);
+ pvt->pvt_next = NULL;
+ goto out;
+ }
+ panic("pmap_pv_untrack: pages not pv-tracked at 0x%"PRIxPADDR
+ " (0x%"PRIxPSIZE" bytes)",
+ start, size);
+out: mutex_exit(&pv_unmanaged.lock);
+
+ npages = size >> PAGE_SHIFT;
+ kmem_free(pvt, offsetof(struct pv_track, pvt_pages[npages]));
+}
+
+static struct pmap_page *
+pmap_pv_tracked(paddr_t pa)
+{
+ struct pv_track *pvt;
+ size_t pgno;
+ int s;
+
+ KASSERT(pa == trunc_page(pa));
+
+ s = pserialize_read_enter();
+ for (pvt = pv_unmanaged.list; pvt != NULL; pvt = pvt->pvt_next) {
+ membar_datadep_consumer();
+ if ((pvt->pvt_start <= pa) &&
+ ((pa - pvt->pvt_start) < pvt->pvt_size))
+ break;
+ }
+ pserialize_read_exit(s);
+
+ if (pvt == NULL)
+ return NULL;
+ KASSERT(pvt->pvt_start <= pa);
+ KASSERT((pa - pvt->pvt_start) < pvt->pvt_size);
+ pgno = (pa - pvt->pvt_start) >> PAGE_SHIFT;
+ return &pvt->pvt_pages[pgno];
+}
+
+/*
* other data structures
*/
@@ -3300,27 +3418,30 @@
*/
if ((opte & PG_PVLIST) == 0) {
#if defined(DIAGNOSTIC) && !defined(DOM0OPS)
- if (PHYS_TO_VM_PAGE(pmap_pte2pa(opte)) != NULL)
- panic("pmap_remove_pte: managed page without "
- "PG_PVLIST for %#" PRIxVADDR, va);
+ if (PHYS_TO_VM_PAGE(pmap_pte2pa(opte)) != NULL ||
+ pmap_pv_tracked(pmap_pte2pa(opte)) != NULL)
+ panic("pmap_remove_pte: managed or pv-tracked page"
+ " without PG_PVLIST for %#"PRIxVADDR, va);
#endif
return true;
}
- pg = PHYS_TO_VM_PAGE(pmap_pte2pa(opte));
-
- KASSERTMSG(pg != NULL, "pmap_remove_pte: unmanaged page marked "
- "PG_PVLIST, va = %#" PRIxVADDR ", pa = %#" PRIxPADDR,
- va, (paddr_t)pmap_pte2pa(opte));
-
- KASSERT(uvm_page_locked_p(pg));
+ if ((pg = PHYS_TO_VM_PAGE(pmap_pte2pa(opte))) != NULL) {
+ KASSERT(uvm_page_locked_p(pg));
+ pp = VM_PAGE_TO_PP(pg);
+ } else if ((pp = pmap_pv_tracked(pmap_pte2pa(opte))) == NULL) {
+ paddr_t pa = pmap_pte2pa(opte);
+ panic("pmap_remove_pte: PG_PVLIST with pv-untracked page"
+ " va = 0x%"PRIxVADDR
+ " pa = 0x%"PRIxPADDR" (0x%"PRIxPADDR")",
+ va, pa, atop(pa));
+ }
/* Sync R/M bits. */
- pp = VM_PAGE_TO_PP(pg);
pp->pp_attrs |= opte;
pve = pmap_remove_pv(pp, ptp, va);
- if (pve) {
+ if (pve) {
pve->pve_next = *pv_tofree;
*pv_tofree = pve;
}
@@ -3545,26 +3666,16 @@
return 0;
}
-/*
- * pmap_page_remove: remove a managed vm_page from all pmaps that map it
- *
- * => R/M bits are sync'd back to attrs
- */
-
-void
-pmap_page_remove(struct vm_page *pg)
+static void
+pmap_pp_remove(struct pmap_page *pp, paddr_t pa)
{
- struct pmap_page *pp;
struct pv_pte *pvpte;
struct pv_entry *killlist = NULL;
struct vm_page *ptp;
pt_entry_t expect;
int count;
- KASSERT(uvm_page_locked_p(pg));
-
- pp = VM_PAGE_TO_PP(pg);
- expect = pmap_pa2pte(VM_PAGE_TO_PHYS(pg)) | PG_V;
+ expect = pmap_pa2pte(pa) | PG_V;
count = SPINLOCK_BACKOFF_MIN;
kpreempt_disable();
startover:
@@ -3637,6 +3748,42 @@
Home |
Main Index |
Thread Index |
Old Index