Memory-related changes sneak peek for 1.6
Subsystems (2)
cover Memory-related changes sneak peek for 1.6
Patches 1-10 are small changes, some of them bugfixes for the IOMMU portion of the patches.
Patches 11-13 introduce three new address_space APIs and clean up the dispatch logic a bit by removing memory_region_section_addr.
Patches 14-22 introduce IOMMU memory regions, convert sPAPR to the new API, and finally eliminate DMAContext in favor of AddressSpace.
Patches 24-37 set an owner on all regions owned by hot-unpluggable devices (and several others when that was easy), and add ref/unref calls that keep the regions alive even if the owner is hot-unplugged.
Patches 38-40 make memory_region_find safe for usage outside the BQL.
These patches are also at git://github.com/bonzini/qemu.git, branch iommu.
Please review!
Alexey Kardashevskiy (1):
memory: give name to every AddressSpace
Avi Kivity (5):
memory: fix address space initialization/destruction
memory: limit sections in the radix tree to the actual address space
size
memory: iommu support
vfio: abort if an emulated iommu is used
pci: use memory core for iommu support
Jan Kiszka (3):
memory: Replace open-coded memory_region_is_romd
memory: Rename readable flag to romd_mode
memory: Introduce address_space_lookup_region
Paolo Bonzini (31):
memory: assert that PhysPageEntry's ptr does not overflow
memory: allow memory_region_find() to run on non-root memory regions
memory: do not duplicate memory_region_destructor_none
memory: make memory_global_sync_dirty_bitmap take an AddressSpace
memory: create FlatView for new address spaces
exec: remove obsolete comment
memory: add address_space_valid
memory: add address_space_translate
spapr: convert TCE API to use an opaque type
spapr: make IOMMU translation go through IOMMUTLBEntry
spapr: use memory core for iommu support
dma: eliminate old-style IOMMU support
spapr_vio: take care of creating our own AddressSpace/DMAContext
dma: eliminate DMAContext
memory: add getter/setter for owner
memory: add ref/unref
memory: add ref/unref calls
pci: set owner for BARs
sysbus: set owner for MMIO regions
acpi: add memory_region_set_owner calls
misc: add memory_region_set_owner calls
isa/portio: allow setting an owner
vga: add memory_region_set_owner calls
pci-assign: add memory_region_set_owner calls
vfio: add memory_region_set_owner calls
exec: check MRU in qemu_ram_addr_from_host
memory: return MemoryRegion from qemu_ram_addr_from_host
memory: ref/unref memory across address_space_map/unmap
memory: access FlatView from a local variable
memory: use a new FlatView pointer on every topology update
memory: add reference counting to FlatView
arch_init.c | 2 +- cputlb.c | 20 ++- dma-helpers.c | 180 +------------------ exec.c | 315 ++++++++++++++++++++++----------- hw/acpi/ich9.c | 1 + hw/acpi/piix4.c | 5 + hw/block/pflash_cfi01.c | 6 +- hw/block/pflash_cfi02.c | 2 +- hw/char/serial-pci.c | 1 + hw/core/loader.c | 1 + hw/core/sysbus.c | 2 + hw/display/cirrus_vga.c | 19 ++- hw/display/exynos4210_fimd.c | 6 + hw/display/framebuffer.c | 10 +- hw/display/qxl.c | 5 +- hw/display/vga-isa-mm.c | 2 +- hw/display/vga-isa.c | 4 +- hw/display/vga-pci.c | 5 +- hw/display/vga.c | 19 ++- hw/display/vga_int.h | 9 +- hw/display/vmware_vga.c | 4 +- hw/dma/pl330.c | 8 +- hw/i386/kvm/ioapic.c | 2 + hw/i386/kvm/pci-assign.c | 11 ++ hw/i386/kvmvapic.c | 1 + hw/ide/ahci.c | 18 +- hw/ide/ahci.h | 4 +- hw/ide/ich.c | 2 +- hw/ide/macio.c | 4 +- hw/isa/apm.c | 1 + hw/isa/isa-bus.c | 2 + hw/misc/pc-testdev.c | 7 + hw/misc/vfio.c | 12 ++ hw/pci/pci.c | 53 +++--- hw/ppc/spapr_iommu.c | 108 +++++------- hw/ppc/spapr_pci.c | 18 ++- hw/ppc/spapr_vio.c | 13 +- hw/scsi/megasas.c | 4 +- hw/scsi/virtio-scsi.c | 2 +- hw/scsi/vmw_pvscsi.c | 2 +- hw/sd/sdhci.c | 22 ++-- hw/usb/hcd-ehci-pci.c | 4 +- hw/usb/hcd-ehci-sysbus.c | 2 +- hw/usb/hcd-ehci.c | 12 +- hw/usb/hcd-ehci.h | 2 +- hw/usb/hcd-ohci.c | 30 ++-- hw/usb/libhw.c | 4 +- hw/virtio/dataplane/hostmem.c | 7 + hw/virtio/vhost.c | 2 + hw/virtio/virtio-balloon.c | 1 + hw/xen/xen_pt.c | 4 + include/exec/cpu-common.h | 2 +- include/exec/cputlb.h | 12 +- include/exec/ioport.h | 3 + include/exec/memory.h | 186 +++++++++++++++---- include/hw/pci-host/spapr.h | 2 +- include/hw/pci/pci.h | 24 ++-- include/hw/pci/pci_bus.h | 5 +- include/hw/ppc/spapr.h | 12 +- include/hw/ppc/spapr_vio.h | 36 +++-- include/hw/virtio/dataplane/hostmem.h | 1 + include/sysemu/dma.h | 158 +++++------------ ioport.c | 10 + kvm-all.c | 2 + memory.c | 267 ++++++++++++++++++++++------ target-arm/kvm.c | 2 + target-i386/kvm.c | 4 +- target-sparc/mmu_helper.c | 1 + translate-all.c | 8 +- xen-all.c | 2 + 70 files changed, 983 insertions(+), 734 deletions(-)
PATCH 1/40 memory: assert that PhysPageEntry's ptr does not overflow
--- exec.c | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-)
exec.c +2 -0
@@ -719,6 +719,8 @@ static void destroy_all_mappings(AddressSpaceDispatch *d)static uint16_t phys_section_add (MemoryRegionSection *section) { +assert (phys_sections_nb <TARGET_PAGE_SIZE ); +if (phys_sections_nb == phys_sections_nb_alloc) { phys_sections_nb_alloc =MAX (phys_sections_nb_alloc *2 ,16 ); phys_sections = g_renew(MemoryRegionSection, phys_sections,
Signature · 1 line
1.7.1
Re: [Qemu-devel] [PATCH 01/40] memory: assert that PhysPageEntry's ptr does not overflow
Quoted 10 lines · Paolo Bonzini
Signed-off-by Paolo Bonzini <[email protected]>--- exec.c | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-)exec.c +0
-0
exec.c static void destroy_all_mappings(AddressSpaceDispatch *d) · line 726
Quoted 8 lines
exec.c +2
-0@@ -719,6 +719,8 @@ static void destroy_all_mappings(AddressSpaceDispatch *d)static uint16_t phys_section_add (MemoryRegionSection *section) { +assert (phys_sections_nb <TARGET_PAGE_SIZE ); +if (phys_sections_nb == phys_sections_nb_alloc) { phys_sections_nb_alloc =MAX (phys_sections_nb_alloc *2 ,16 ); phys_sections = g_renew(MemoryRegionSection, phys_sections,
Why is the limit we're asserting not the same as the maximum size that we pass to g_renew() below?
-- PMM
Re: [Qemu-devel] [PATCH 01/40] memory: assert that PhysPageEntry's ptr does not overflow
----- Messaggio originale -----
Quoted 20 lines
Da: "Peter Maydell" <[email protected]>
A: "Paolo Bonzini" <[email protected]>
Cc: [email protected], [email protected], "jan kiszka" <[email protected]>, [email protected], [email protected], [email protected]
Inviato: Martedì, 7 maggio 2013 17:44:59
Oggetto: Re: [Qemu-devel] [PATCH 01/40] memory: assert that PhysPageEntry's ptr does not overflowQuoted 12 lines · Paolo Bonzini
Signed-off-by Paolo Bonzini <[email protected]>--- exec.c | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-)exec.c +0
-0@@ -719,6 +719,8 @@ static void destroy_all_mappings(AddressSpaceDispatch *d)
Quoted 10 lines
Quoted 7 lines
exec.c +2
-0static uint16_t phys_section_add (MemoryRegionSection *section) { +assert (phys_sections_nb <TARGET_PAGE_SIZE ); +if (phys_sections_nb == phys_sections_nb_alloc) { phys_sections_nb_alloc =MAX (phys_sections_nb_alloc *2 ,16 );
phys_sections= g_renew(MemoryRegionSection,phys_sections,Why is the limit we're asserting not the same as the maximum size that we pass to
g_renew()below?
That's a minimum size, isn't it?
I'm asserting that the physical section number doesn't overflow into the page, since the TLB entries are stored as a combination of the two.
Paolo
Re: [Qemu-devel] [PATCH 01/40] memory: assert that PhysPageEntry's ptr does not overflow
Quoted 5 lines · Paolo Bonzini
Quoted 3 lines
Da: "Peter Maydell" <[email protected]>
Why is the limit we're asserting not the same as the maximum size that we pass tog_renew()below?That's a minimum size, isn't it?
Doh, so it is.
Quoted 2 lines
I'm asserting that the physical section number doesn't overflow into the page, since the TLB entries are stored as a combination of the two.
Maybe this could use a comment about where (in what data structure) the section number lives, because it's not entirely obvious (well, to me anyway).
-- PMM
Re: [Qemu-devel] [PATCH 01/40] memory: assert that PhysPageEntry's ptr does not overflow
In reply to Paolo BonziniQuoted 34 lines · Paolo Bonzini
----- Messaggio originale -----
Quoted 20 lines
Da: "Peter Maydell" <[email protected]>
A: "Paolo Bonzini" <[email protected]>
Cc: [email protected], [email protected], "jan kiszka" <[email protected]>, [email protected], [email protected], [email protected]
Inviato: Martedì, 7 maggio 2013 17:44:59
Oggetto: Re: [Qemu-devel] [PATCH 01/40] memory: assert that PhysPageEntry's ptr does not overflowQuoted 12 lines · Paolo Bonzini
Signed-off-by Paolo Bonzini <[email protected]>--- exec.c | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-)exec.c +0
-0@@ -719,6 +719,8 @@ static void destroy_all_mappings(AddressSpaceDispatch *d)Quoted 10 lines
Quoted 7 lines
exec.c +2
-0static uint16_t phys_section_add (MemoryRegionSection *section) { +assert (phys_sections_nb <TARGET_PAGE_SIZE ); +if (phys_sections_nb == phys_sections_nb_alloc) { phys_sections_nb_alloc =MAX (phys_sections_nb_alloc *2 ,16 );
phys_sections= g_renew(MemoryRegionSection,phys_sections,Why is the limit we're asserting not the same as the maximum size that we pass to
g_renew()below?
Quoted 5 lines
That's a minimum size, isn't it?
I'm asserting that the physical section number doesn't overflow into the page, since the TLB entries are stored as a combination of the two.
Could you explain more detail? Why < TARGET_PAGE_SIZE, not 2^15?
Thanks,
Pingfan
Quoted 1 line
Paolo
Re: [Qemu-devel] [PATCH 01/40] memory: assert that PhysPageEntry's ptr does not overflow
Il 09/05/2013 05:41, liu ping fan ha scritto:
Quoted 35 lines · Paolo Bonzini
Quoted 34 lines
----- Messaggio originale -----
Quoted 30 lines
Da: "Peter Maydell" <[email protected]>
A: "Paolo Bonzini" <[email protected]>
Cc: [email protected], [email protected], "jan kiszka" <[email protected]>, [email protected], [email protected], [email protected]
Inviato: Martedì, 7 maggio 2013 17:44:59
Oggetto: Re: [Qemu-devel] [PATCH 01/40] memory: assert that PhysPageEntry's ptr does not overflowQuoted 18 lines · Paolo Bonzini
Signed-off-by Paolo Bonzini <[email protected]>--- exec.c | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-)exec.c +2
-0@@ -719,6 +719,8 @@ static void destroy_all_mappings(AddressSpaceDispatch *d)static uint16_t phys_section_add (MemoryRegionSection *section) { +assert (phys_sections_nb <TARGET_PAGE_SIZE ); +if (phys_sections_nb == phys_sections_nb_alloc) { phys_sections_nb_alloc =MAX (phys_sections_nb_alloc *2 ,16 );Quoted 1 line
phys_sections= g_renew(MemoryRegionSection,phys_sections,Why is the limit we're asserting not the same as the maximum size that we pass to
g_renew()below?
Quoted 6 lines
Quoted 5 lines
That's a minimum size, isn't it?
I'm asserting that the physical section number doesn't overflow into the page, since the TLB entries are stored as a combination of the two.
Could you explain more detail? Why <
TARGET_PAGE_SIZE, not 2^15?
Because the TLB entry is the "or" of the page address and the phys_section. Look here:
hwaddr memory_region_section_get_iotlb(CPUArchState *env,
MemoryRegionSection *section, target_ulong vaddr, hwaddr paddr, int prot, target_ulong *address)
{
hwaddr iotlb;
CPUWatchpoint *wp;
if (memory_region_is_ram(section->mr)) {
/* Normal RAM. */
iotlb = (memory_region_get_ram_addr(section->mr) & TARGET_PAGE_MASK)
+ memory_region_section_addr(section, paddr);
if (!section->readonly) {
iotlb |= phys_section_notdirty;
} else {
iotlb |= phys_section_rom;
}
} else {
iotlb = section - phys_sections;
iotlb += memory_region_section_addr(section, paddr);
}where the else could be written better as:
iotlb = memory_region_section_addr(section, paddr);
iotlb |= section - phys_sections;memory_region_section_addr will return a page-aligned value.
Paolo
PATCH 2/40 memory: allow memory_region_find() to run on non-root memory regions
In reply to Paolo Bonzinimemory_region_find() is similar to registering a MemoryListener and checking for the MemoryRegionSections that come from a particular region. There is no reason for this to be limited to a root memory region.
--- include/exec/memory.h | 13 +++++++------ memory.c | 20 +++++++++++++++----- 2 files changed, 22 insertions(+), 11 deletions(-)
include/exec/memory.h +7 -6
@@ -725,17 +725,18 @@ void memory_region_set_alias_offset(MemoryRegion *mr,* * Returns a #MemoryRegionSection that describes a contiguous overlap. * It will have the following characteristics: -* .@offset_within_address_space >= @addr -* .@offset_within_address_space + .@size <= @addr + @size * .@size = 0 iff no overlap was found * .@mr is non-%NULL iff an overlap was found * -* @address_space: a top-level (i.e. parentless) region that contains -* the region to be found -* @addr: start of the area within @address_space to be searched +* If @mr is parent-less, +* .@offset_within_address_space >= @addr +* .@offset_within_address_space + .@size <= @addr + @size +* +* @mr: a (possibly indirect) parent that contains the region to be found +* @addr: start of the area within @as to be searched * @size: size of the area to be searched */ -MemoryRegionSection memory_region_find (MemoryRegion *address_space, +MemoryRegionSection memory_region_find (MemoryRegion *mr,hwaddr addr,uint64_t size);/**
memory.c +15 -5
@@ -1451,15 +1451,24 @@ static FlatRange *address_space_lookup(AddressSpace *as, AddrRange addr) sizeof(FlatRange), cmp_flatrange_addr); } -MemoryRegionSection memory_region_find (MemoryRegion *address_space, +MemoryRegionSection memory_region_find (MemoryRegion *mr,hwaddr addr,uint64_t size) { -AddressSpace *as =memory_region_to_address_space (address_space); -AddrRange range =addrrange_make (int128_make64 (addr), -int128_make64 (size)); -FlatRange *fr =address_space_lookup (as, range);MemoryRegionSection ret = { .mr =NULL , .size =0 }; +MemoryRegion *root; +AddressSpace *as; +AddrRange range; +FlatRange *fr; + + addr += mr->addr; +for (root = mr; root->parent; ) { + root = root->parent; + addr += root->addr; + } + as =memory_region_to_address_space (root); + range =addrrange_make (int128_make64 (addr),int128_make64 (size)); + fr =address_space_lookup (as, range);if (!fr) {return ret; }
@@ -1470,6 +1479,7 @@ MemoryRegionSection memory_region_find(MemoryRegion *address_space, } ret.mr = fr->mr; + ret.address_space = as; range =addrrange_intersection (range, fr->addr); ret.offset_within_region = fr->offset_in_region; ret.offset_within_region += int128_get64(int128_sub(range.start,
Signature · 1 line
1.7.1
Re: [Qemu-devel] [PATCH 02/40] memory: allow memory_region_find() to run on non-root memory regions
Quoted 20 lines · Paolo Bonzini
memory_region_find()is similar to registering a MemoryListener and checking for the MemoryRegionSections that come from a particular region. There is no reason for this to be limited to a root memory region.Signed-off-by Paolo Bonzini <[email protected]>--- include/exec/memory.h | 13 +++++++------ memory.c | 20 +++++++++++++++----- 2 files changed, 22 insertions(+), 11 deletions(-)include/exec/memory.h +0
-1@@ -725,17 +725,18 @@ void memory_region_set_alias_offset(MemoryRegion *mr,* * Returns a #MemoryRegionSection that describes a contiguous overlap. * It will have the following characteristics: -* .@offset_within_address_space >= @addr
include/exec/memory.h void memory_region_set_alias_offset(MemoryRegion *mr, · line 733
Quoted 10 lines
include/exec/memory.h +3
-4-* .@offset_within_address_space + .@size <= @addr + @size * .@size = 0 iff no overlap was found * .@mr is non-%NULL iff an overlap was found * -* @address_space: a top-level (i.e. parentless) region that contains -* the region to be found -* @addr: start of the area within @address_space to be searched +* If @mr is parent-less, +* .@offset_within_address_space >= @addr +* .@offset_within_address_space + .@size <= @addr + @size
Shouldn't we also say something about what you get back in the "@mr is not parentless" case?
include/exec/memory.h void memory_region_set_alias_offset(MemoryRegion *mr, · line 736
Quoted 3 lines
include/exec/memory.h +3
-0+* +* @mr: a (possibly indirect) parent that contains the region to be found +* @addr: start of the area within @as to be searched
There's no @as parameter...
include/exec/memory.h void memory_region_set_alias_offset(MemoryRegion *mr, · line 740
Quoted 5 lines
include/exec/memory.h +1
-1* @size: size of the area to be searched */ -MemoryRegionSection memory_region_find (MemoryRegion *address_space, +MemoryRegionSection memory_region_find (MemoryRegion *mr,hwaddr addr,uint64_t size);
thanks
-- PMM
Re: [Qemu-devel] [PATCH 02/40] memory: allow memory_region_find() to run on non-root memory regions
In reply to Paolo BonziniQuoted 1 line · Paolo Bonzini
memory_region_find()is similar to registering a MemoryListener and
But losing the priority of MemoryListener. Had better document to warn this, so the user will decide whether it is safe to use the interfaces based on this or not.
Quoted 85 lines
checking for the MemoryRegionSections that come from a particular region. There is no reason for this to be limited to a root memory region.
Signed-off-by Paolo Bonzini <[email protected]>--- include/exec/memory.h | 13 +++++++------ memory.c | 20 +++++++++++++++----- 2 files changed, 22 insertions(+), 11 deletions(-)include/exec/memory.h +7
-6@@ -725,17 +725,18 @@ void memory_region_set_alias_offset(MemoryRegion *mr,* * Returns a #MemoryRegionSection that describes a contiguous overlap. * It will have the following characteristics: -* .@offset_within_address_space >= @addr -* .@offset_within_address_space + .@size <= @addr + @size * .@size = 0 iff no overlap was found * .@mr is non-%NULL iff an overlap was found * -* @address_space: a top-level (i.e. parentless) region that contains -* the region to be found -* @addr: start of the area within @address_space to be searched +* If @mr is parent-less, +* .@offset_within_address_space >= @addr +* .@offset_within_address_space + .@size <= @addr + @size +* +* @mr: a (possibly indirect) parent that contains the region to be found +* @addr: start of the area within @as to be searched * @size: size of the area to be searched */ -MemoryRegionSection memory_region_find (MemoryRegion *address_space, +MemoryRegionSection memory_region_find (MemoryRegion *mr,hwaddr addr,uint64_t size);/** memory.c +15
-6@@ -1451,15 +1451,24 @@ static FlatRange *address_space_lookup(AddressSpace *as, AddrRange addr) sizeof(FlatRange), cmp_flatrange_addr); } -MemoryRegionSection memory_region_find (MemoryRegion *address_space, +MemoryRegionSection memory_region_find (MemoryRegion *mr,hwaddr addr,uint64_t size) { -AddressSpace *as =memory_region_to_address_space (address_space); -AddrRange range =addrrange_make (int128_make64 (addr), -int128_make64 (size)); -FlatRange *fr =address_space_lookup (as, range);MemoryRegionSection ret = { .mr =NULL , .size =0 }; +MemoryRegion *root; +AddressSpace *as; +AddrRange range; +FlatRange *fr; + + addr += mr->addr; +for (root = mr; root->parent; ) { + root = root->parent; + addr += root->addr; + } + as =memory_region_to_address_space (root); + range =addrrange_make (int128_make64 (addr),int128_make64 (size)); + fr =address_space_lookup (as, range);if (!fr) {return ret; }@@ -1470,6 +1479,7 @@ MemoryRegionSection memory_region_find(MemoryRegion *address_space, } ret.mr = fr->mr; + ret.address_space = as; range =addrrange_intersection (range, fr->addr); ret.offset_within_region = fr->offset_in_region; ret.offset_within_region += int128_get64(int128_sub(range.start, --1.7.1
PATCH 3/40 memory: Replace open-coded memory_region_is_romd
In reply to Paolo BonziniFrom: Jan Kiszka <[email protected]>
Improves readability.
--- translate-all.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-)
translate-all.c +1 -1
@@ -1359,7 +1359,7 @@ void tb_invalidate_phys_addr(hwaddr addr) section =phys_page_find (address_space_memory.dispatch, addr >>TARGET_PAGE_BITS );if (!(memory_region_is_ram (section->mr) - || (section->mr->rom_device && section->mr->readable))) { + ||memory_region_is_romd (section->mr))) {return ; } ram_addr = (memory_region_get_ram_addr (section->mr) &TARGET_PAGE_MASK )
Signature · 1 line
1.7.1
Re: [Qemu-devel] [PATCH 03/40] memory: Replace open-coded memory_region_is_romd
Quoted 5 lines · Paolo Bonzini
From: Jan Kiszka <[email protected]>
Improves readability.
Signed-off-by Jan Kiszka <[email protected]>
Still
:-)
-- PMM
PATCH 4/40 memory: Rename readable flag to romd_mode
In reply to Paolo BonziniFrom: Jan Kiszka <[email protected]>
"Readable" is a very unfortunate name for this flag because even a rom_device region will always be readable from the guest POV. What differs is the mapping, just like the comments had to explain already. Also, readable could currently be understood as being a generic region flag, but it only applies to rom_device regions.
So name the flag and the function to modify it after the original term "ROMD" which could also be interpreted as "ROM direct", i.e. ROM mode with direct access. In any case, the scope if the flag is clearer now.
--- hw/block/pflash_cfi01.c | 6 +++--- hw/block/pflash_cfi02.c | 2 +- include/exec/memory.h | 22 +++++++++++----------- memory.c | 30 +++++++++++++++--------------- 4 files changed, 30 insertions(+), 30 deletions(-)
hw/block/pflash_cfi01.c +3 -3
@@ -105,7 +105,7 @@ static void pflash_timer (void *opaque)DPRINTF ("%s: command %02x done\n" , __func__, pfl->cmd);/* Reset flash */ pfl->status ^=0x80 ; -memory_region_rom_device_set_readable (&pfl->mem, true); +memory_region_rom_device_set_romd (&pfl->mem, true); pfl->wcycle =0 ; pfl->cmd =0 ; }
@@ -281,7 +281,7 @@ static void pflash_write(pflash_t *pfl, hwaddr offset,if (!pfl->wcycle) {/* Set the device in I/O access mode */ -memory_region_rom_device_set_readable (&pfl->mem, false); +memory_region_rom_device_set_romd (&pfl->mem, false); }switch (pfl->wcycle) {
@@ -458,7 +458,7 @@ static void pflash_write(pflash_t *pfl, hwaddr offset, "\n", __func__, offset, pfl->wcycle, pfl->cmd, value); reset_flash: -memory_region_rom_device_set_readable (&pfl->mem, true); +memory_region_rom_device_set_romd (&pfl->mem, true); pfl->wcycle =0 ; pfl->cmd =0 ;
hw/block/pflash_cfi02.c +1 -1
@@ -111,7 +111,7 @@ static void pflash_setup_mappings(pflash_t *pfl)static void pflash_register_memory (pflash_t *pfl,int rom_mode) { -memory_region_rom_device_set_readable (&pfl->orig_mem, rom_mode); +memory_region_rom_device_set_romd (&pfl->orig_mem, rom_mode); pfl->rom_mode = rom_mode; }
include/exec/memory.h +11 -11
@@ -126,7 +126,7 @@ struct MemoryRegion {ram_addr_t ram_addr;bool subpage;bool terminates; -bool readable; +bool romd_mode;bool ram;bool readonly;/* For RAM regions */ bool enabled;
@@ -355,16 +355,16 @@ uint64_t memory_region_size(MemoryRegion *mr);bool memory_region_is_ram (MemoryRegion *mr);/** -* memory_region_is_romd: check whether a memory region is ROMD +* memory_region_is_romd: check whether a memory region is in ROMD mode * -* Returns %true is a memory region is ROMD and currently set to allow +* Returns %true is a memory region is a ROM device and currently set to allow * direct reads. * * @mr: the memory region being queried */ static inline bool memory_region_is_romd (MemoryRegion *mr) { -return mr->rom_device && mr->readable; +return mr->rom_device && mr->romd_mode; }/**
@@ -502,18 +502,18 @@ void memory_region_reset_dirty(MemoryRegion *mr, hwaddr addr,void memory_region_set_readonly (MemoryRegion *mr,bool readonly);/** -* memory_region_rom_device_set_readable: enable/disable ROM readability +* memory_region_rom_device_set_romd: enable/disable ROMD mode * * Allows a ROM device (initialized with memory_region_init_rom_device() to -* to be marked as readable (default) or not readable. When it is readable, -* the device is mapped to guest memory. When not readable, reads are -* forwarded to the #MemoryRegion.read function. +* set to ROMD mode (default) or MMIO mode. When it is in ROMD mode, the +* device is mapped to guest memory and satisfies read access directly. +* When in MMIO mode, reads are forwarded to the #MemoryRegion.read function. +* Writes are always handled by the #MemoryRegion.write function. * * @mr: the memory region to be updated -* @readable: whether reads are satisified directly (%true) or via callbacks -* (%false) +* @romd_mode: whether the region in in ROMD mode or not */ -void memory_region_rom_device_set_readable (MemoryRegion *mr,bool readable); +void memory_region_rom_device_set_romd (MemoryRegion *mr,bool romd_mode);/** * memory_region_set_coalescing: Enable memory coalescing for the region.
memory.c +15 -15
@@ -213,7 +213,7 @@ struct FlatRange {hwaddr offset_in_region;AddrRange addr;uint8_t dirty_log_mask; -bool readable; +bool romd_mode;bool readonly; };
@@ -236,7 +236,7 @@ static bool flatrange_equal(FlatRange *a, FlatRange *b)return a->mr == b->mr &&addrrange_equal (a->addr, b->addr) && a->offset_in_region == b->offset_in_region - && a->readable == b->readable + && a->romd_mode == b->romd_mode && a->readonly == b->readonly; }
@@ -276,7 +276,7 @@ static bool can_merge(FlatRange *r1, FlatRange *r2) r1->addr.size), int128_make64(r2->offset_in_region)) && r1->dirty_log_mask == r2->dirty_log_mask - && r1->readable == r2->readable + && r1->romd_mode == r2->romd_mode && r1->readonly == r2->readonly; }
@@ -532,7 +532,7 @@ static void render_memory_region(FlatView *view, fr.offset_in_region = offset_in_region; fr.addr =addrrange_make (base, now); fr.dirty_log_mask = mr->dirty_log_mask; - fr.readable = mr->readable; + fr.romd_mode = mr->romd_mode; fr.readonly = readonly;flatview_insert (view, i, &fr); ++i;
@@ -552,7 +552,7 @@ static void render_memory_region(FlatView *view, fr.offset_in_region = offset_in_region; fr.addr =addrrange_make (base, remain); fr.dirty_log_mask = mr->dirty_log_mask; - fr.readable = mr->readable; + fr.romd_mode = mr->romd_mode; fr.readonly = readonly;flatview_insert (view, i, &fr); }
@@ -801,7 +801,7 @@ void memory_region_init(MemoryRegion *mr, mr->enabled = true; mr->terminates = false; mr->ram = false; - mr->readable = true; + mr->romd_mode = true; mr->readonly = false; mr->rom_device = false; mr->destructor = memory_region_destructor_none;
@@ -1121,11 +1121,11 @@ void memory_region_set_readonly(MemoryRegion *mr, bool readonly) } } -void memory_region_rom_device_set_readable (MemoryRegion *mr,bool readable) +void memory_region_rom_device_set_romd (MemoryRegion *mr,bool romd_mode) { -if (mr->readable != readable) { +if (mr->romd_mode != romd_mode) {memory_region_transaction_begin (); - mr->readable = readable; + mr->romd_mode = romd_mode; memory_region_update_pending |= mr->enabled;memory_region_transaction_commit (); }
@@ -1659,9 +1659,9 @@ static void mtree_print_mr(fprintf_function mon_printf, void *f, base + mr->addr + (hwaddr )int128_get64 (mr->size) -1 , mr->priority, - mr->readable ?'R' :'-' , - !mr->readonly && !(mr->rom_device && mr->readable) ?'W' - :'-' , + mr->romd_mode ?'R' :'-' , + !mr->readonly && !(mr->rom_device && mr->romd_mode) ?'W' + :'-' , mr->name, mr->alias->name, mr->alias_offset,
@@ -1674,9 +1674,9 @@ static void mtree_print_mr(fprintf_function mon_printf, void *f, base + mr->addr + (hwaddr )int128_get64 (mr->size) -1 , mr->priority, - mr->readable ?'R' :'-' , - !mr->readonly && !(mr->rom_device && mr->readable) ?'W' - :'-' , + mr->romd_mode ?'R' :'-' , + !mr->readonly && !(mr->rom_device && mr->romd_mode) ?'W' + :'-' , mr->name); }
Signature · 1 line
1.7.1
Re: [Qemu-devel] [PATCH 04/40] memory: Rename readable flag to romd_mode
Quoted 9 lines · Paolo Bonzini
From: Jan Kiszka <[email protected]>
"Readable" is a very unfortunate name for this flag because even a
rom_deviceregion will always be readable from the guest POV. What differs is the mapping, just like the comments had to explain already. Also, readable could currently be understood as being a generic region flag, but it only applies torom_deviceregions.So name the flag and the function to modify it after the original term
"rename"
Quoted 2 lines
"ROMD" which could also be interpreted as "ROM direct", i.e. ROM mode with direct access. In any case, the scope if the flag is clearer now.
"scope of".
hw/block/pflash_cfi02.c static void pflash_setup_mappings(pflash_t *pfl) · line 114
Quoted 8 lines
hw/block/pflash_cfi02.c +1
-1@@ -111,7 +111,7 @@ static void pflash_setup_mappings(pflash_t *pfl)static void pflash_register_memory (pflash_t *pfl,int rom_mode) { -memory_region_rom_device_set_readable (&pfl->orig_mem, rom_mode); +memory_region_rom_device_set_romd (&pfl->orig_mem, rom_mode);
Obvious opportunity to change the name of the pflash_register_memory parameter to match (separate trivial patch would be fine).
Quoted 17 lines
hw/block/pflash_cfi02.c +0
-0pfl->rom_mode = rom_mode; }include/exec/memory.h +1
-1@@ -126,7 +126,7 @@ struct MemoryRegion {ram_addr_t ram_addr;bool subpage;bool terminates; -bool readable; +bool romd_mode;bool ram;bool readonly;/* For RAM regions */ bool enabled;
include/exec/memory.h uint64_t memory_region_size(MemoryRegion *mr); · line 360
Quoted 8 lines
include/exec/memory.h +2
-2@@ -355,16 +355,16 @@ uint64_t memory_region_size(MemoryRegion *mr);bool memory_region_is_ram (MemoryRegion *mr);/** -* memory_region_is_romd: check whether a memory region is ROMD +* memory_region_is_romd: check whether a memory region is in ROMD mode * -* Returns %true is a memory region is ROMD and currently set to allow +* Returns %true is a memory region is a ROM device and currently set to allow
"Returns %true if"
Quoted 21 lines
include/exec/memory.h +2
-4* direct reads. * * @mr: the memory region being queried */ static inline bool memory_region_is_romd (MemoryRegion *mr) { -return mr->rom_device && mr->readable; +return mr->rom_device && mr->romd_mode; }/** @@ -502,18 +502,18 @@ void memory_region_reset_dirty(MemoryRegion *mr, hwaddr addr,void memory_region_set_readonly (MemoryRegion *mr,bool readonly);/** -* memory_region_rom_device_set_readable: enable/disable ROM readability +* memory_region_rom_device_set_romd: enable/disable ROMD mode * * Allows a ROM device (initialized with memory_region_init_rom_device() to -* to be marked as readable (default) or not readable. When it is readable, -* the device is mapped to guest memory. When not readable, reads are
include/exec/memory.h void memory_region_reset_dirty(MemoryRegion *mr, hwaddr addr, · line 514
Quoted 10 lines
include/exec/memory.h +5
-3-* forwarded to the #MemoryRegion.read function. +* set to ROMD mode (default) or MMIO mode. When it is in ROMD mode, the +* device is mapped to guest memory and satisfies read access directly. +* When in MMIO mode, reads are forwarded to the #MemoryRegion.read function. +* Writes are always handled by the #MemoryRegion.write function. * * @mr: the memory region to be updated -* @readable: whether reads are satisified directly (%true) or via callbacks -* (%false) +* @romd_mode: whether the region in in ROMD mode or not
"is in", but that's not what the parameter means anyway.
"@romd_mode: %true to put the region into ROMD mode"
thanks
-- PMM
PATCH v2 1/1 memory: Rename readable flag to romd_mode
"Readable" is a very unfortunate name for this flag because even a rom_device region will always be readable from the guest POV. What differs is the mapping, just like the comments had to explain already. Also, readable could currently be understood as being a generic region flag, but it only applies to rom_device regions.
So rename the flag and the function to modify it after the original term "ROMD" which could also be interpreted as "ROM direct", i.e. ROM mode with direct access. In any case, the scope of the flag is clearer now.
---
Addresses Peter's findings except for the pflash_register_memory argument renaming (that would require renaming the state field as well - some other day).
hw/block/pflash_cfi01.c | 6 +++--- hw/block/pflash_cfi02.c | 2 +- include/exec/memory.h | 22 +++++++++++----------- memory.c | 30 +++++++++++++++--------------- 4 files changed, 30 insertions(+), 30 deletions(-)
hw/block/pflash_cfi01.c +3 -3
@@ -105,7 +105,7 @@ static void pflash_timer (void *opaque)DPRINTF ("%s: command %02x done\n" , __func__, pfl->cmd);/* Reset flash */ pfl->status ^=0x80 ; -memory_region_rom_device_set_readable (&pfl->mem, true); +memory_region_rom_device_set_romd (&pfl->mem, true); pfl->wcycle =0 ; pfl->cmd =0 ; }
@@ -281,7 +281,7 @@ static void pflash_write(pflash_t *pfl, hwaddr offset,if (!pfl->wcycle) {/* Set the device in I/O access mode */ -memory_region_rom_device_set_readable (&pfl->mem, false); +memory_region_rom_device_set_romd (&pfl->mem, false); }switch (pfl->wcycle) {
@@ -458,7 +458,7 @@ static void pflash_write(pflash_t *pfl, hwaddr offset, "\n", __func__, offset, pfl->wcycle, pfl->cmd, value); reset_flash: -memory_region_rom_device_set_readable (&pfl->mem, true); +memory_region_rom_device_set_romd (&pfl->mem, true); pfl->wcycle =0 ; pfl->cmd =0 ;
hw/block/pflash_cfi02.c +1 -1
@@ -111,7 +111,7 @@ static void pflash_setup_mappings(pflash_t *pfl)static void pflash_register_memory (pflash_t *pfl,int rom_mode) { -memory_region_rom_device_set_readable (&pfl->orig_mem, rom_mode); +memory_region_rom_device_set_romd (&pfl->orig_mem, rom_mode); pfl->rom_mode = rom_mode; }
include/exec/memory.h +11 -11
@@ -126,7 +126,7 @@ struct MemoryRegion {ram_addr_t ram_addr;bool subpage;bool terminates; -bool readable; +bool romd_mode;bool ram;bool readonly;/* For RAM regions */ bool enabled;
@@ -355,16 +355,16 @@ uint64_t memory_region_size(MemoryRegion *mr);bool memory_region_is_ram (MemoryRegion *mr);/** -* memory_region_is_romd: check whether a memory region is ROMD +* memory_region_is_romd: check whether a memory region is in ROMD mode * -* Returns %true is a memory region is ROMD and currently set to allow +* Returns %true if a memory region is a ROM device and currently set to allow * direct reads. * * @mr: the memory region being queried */ static inline bool memory_region_is_romd (MemoryRegion *mr) { -return mr->rom_device && mr->readable; +return mr->rom_device && mr->romd_mode; }/**
@@ -502,18 +502,18 @@ void memory_region_reset_dirty(MemoryRegion *mr, hwaddr addr,void memory_region_set_readonly (MemoryRegion *mr,bool readonly);/** -* memory_region_rom_device_set_readable: enable/disable ROM readability +* memory_region_rom_device_set_romd: enable/disable ROMD mode * * Allows a ROM device (initialized with memory_region_init_rom_device() to -* to be marked as readable (default) or not readable. When it is readable, -* the device is mapped to guest memory. When not readable, reads are -* forwarded to the #MemoryRegion.read function. +* set to ROMD mode (default) or MMIO mode. When it is in ROMD mode, the +* device is mapped to guest memory and satisfies read access directly. +* When in MMIO mode, reads are forwarded to the #MemoryRegion.read function. +* Writes are always handled by the #MemoryRegion.write function. * * @mr: the memory region to be updated -* @readable: whether reads are satisified directly (%true) or via callbacks -* (%false) +* @romd_mode: %true to put the region into ROMD mode */ -void memory_region_rom_device_set_readable (MemoryRegion *mr,bool readable); +void memory_region_rom_device_set_romd (MemoryRegion *mr,bool romd_mode);/** * memory_region_set_coalescing: Enable memory coalescing for the region.
memory.c +15 -15
@@ -213,7 +213,7 @@ struct FlatRange {hwaddr offset_in_region;AddrRange addr;uint8_t dirty_log_mask; -bool readable; +bool romd_mode;bool readonly; };
@@ -236,7 +236,7 @@ static bool flatrange_equal(FlatRange *a, FlatRange *b)return a->mr == b->mr &&addrrange_equal (a->addr, b->addr) && a->offset_in_region == b->offset_in_region - && a->readable == b->readable + && a->romd_mode == b->romd_mode && a->readonly == b->readonly; }
@@ -276,7 +276,7 @@ static bool can_merge(FlatRange *r1, FlatRange *r2) r1->addr.size), int128_make64(r2->offset_in_region)) && r1->dirty_log_mask == r2->dirty_log_mask - && r1->readable == r2->readable + && r1->romd_mode == r2->romd_mode && r1->readonly == r2->readonly; }
@@ -532,7 +532,7 @@ static void render_memory_region(FlatView *view, fr.offset_in_region = offset_in_region; fr.addr =addrrange_make (base, now); fr.dirty_log_mask = mr->dirty_log_mask; - fr.readable = mr->readable; + fr.romd_mode = mr->romd_mode; fr.readonly = readonly;flatview_insert (view, i, &fr); ++i;
@@ -552,7 +552,7 @@ static void render_memory_region(FlatView *view, fr.offset_in_region = offset_in_region; fr.addr =addrrange_make (base, remain); fr.dirty_log_mask = mr->dirty_log_mask; - fr.readable = mr->readable; + fr.romd_mode = mr->romd_mode; fr.readonly = readonly;flatview_insert (view, i, &fr); }
@@ -801,7 +801,7 @@ void memory_region_init(MemoryRegion *mr, mr->enabled = true; mr->terminates = false; mr->ram = false; - mr->readable = true; + mr->romd_mode = true; mr->readonly = false; mr->rom_device = false; mr->destructor = memory_region_destructor_none;
@@ -1121,11 +1121,11 @@ void memory_region_set_readonly(MemoryRegion *mr, bool readonly) } } -void memory_region_rom_device_set_readable (MemoryRegion *mr,bool readable) +void memory_region_rom_device_set_romd (MemoryRegion *mr,bool romd_mode) { -if (mr->readable != readable) { +if (mr->romd_mode != romd_mode) {memory_region_transaction_begin (); - mr->readable = readable; + mr->romd_mode = romd_mode; memory_region_update_pending |= mr->enabled;memory_region_transaction_commit (); }
@@ -1649,9 +1649,9 @@ static void mtree_print_mr(fprintf_function mon_printf, void *f, base + mr->addr + (hwaddr )int128_get64 (mr->size) -1 , mr->priority, - mr->readable ?'R' :'-' , - !mr->readonly && !(mr->rom_device && mr->readable) ?'W' - :'-' , + mr->romd_mode ?'R' :'-' , + !mr->readonly && !(mr->rom_device && mr->romd_mode) ?'W' + :'-' , mr->name, mr->alias->name, mr->alias_offset,
@@ -1664,9 +1664,9 @@ static void mtree_print_mr(fprintf_function mon_printf, void *f, base + mr->addr + (hwaddr )int128_get64 (mr->size) -1 , mr->priority, - mr->readable ?'R' :'-' , - !mr->readonly && !(mr->rom_device && mr->readable) ?'W' - :'-' , + mr->romd_mode ?'R' :'-' , + !mr->readonly && !(mr->rom_device && mr->romd_mode) ?'W' + :'-' , mr->name); }
Signature · 1 line
1.7.3.4
Re: [Qemu-devel] [PATCH v2] memory: Rename readable flag to romd_mode
Quoted 10 lines · Jan Kiszka
"Readable" is a very unfortunate name for this flag because even a
rom_deviceregion will always be readable from the guest POV. What differs is the mapping, just like the comments had to explain already. Also, readable could currently be understood as being a generic region flag, but it only applies torom_deviceregions.So rename the flag and the function to modify it after the original term "ROMD" which could also be interpreted as "ROM direct", i.e. ROM mode with direct access. In any case, the scope of the flag is clearer now.
Quoted 6 lines
Signed-off-by Jan Kiszka <[email protected]>---
Addresses Peter's findings except for the
pflash_register_memoryargument renaming (that would require renaming the state field as well - some other day).
thanks
-- PMM
PATCH 5/40 memory: do not duplicate memory_region_destructor_none
In reply to Paolo Bonzini--- memory.c | 6 +----- 1 files changed, 1 insertions(+), 5 deletions(-)
memory.c +1 -5
@@ -768,10 +768,6 @@ static void memory_region_destructor_ram_from_ptr(MemoryRegion *mr)qemu_ram_free_from_ptr (mr->ram_addr); } -static void memory_region_destructor_iomem (MemoryRegion *mr) -{ -} -static void memory_region_destructor_rom_device (MemoryRegion *mr) {qemu_ram_free (mr->ram_addr &TARGET_PAGE_MASK );
@@ -929,7 +925,7 @@ void memory_region_init_io(MemoryRegion *mr, mr->ops = ops; mr->opaque = opaque; mr->terminates = true; - mr->destructor = memory_region_destructor_iomem; + mr->destructor = memory_region_destructor_none; mr->ram_addr = ~(ram_addr_t )0 ; }
Signature · 1 line
1.7.1
Re: [Qemu-devel] [PATCH 05/40] memory: do not duplicate memory_region_destructor_none
Quoted 21 lines · Paolo Bonzini
Signed-off-by Paolo Bonzini <[email protected]>--- memory.c | 6 +----- 1 files changed, 1 insertions(+), 5 deletions(-)memory.c +0
-4@@ -768,10 +768,6 @@ static void memory_region_destructor_ram_from_ptr(MemoryRegion *mr)qemu_ram_free_from_ptr (mr->ram_addr); } -static void memory_region_destructor_iomem (MemoryRegion *mr) -{ -} -static void memory_region_destructor_rom_device (MemoryRegion *mr) {qemu_ram_free (mr->ram_addr &TARGET_PAGE_MASK );
memory.c void memory_region_init_io(MemoryRegion *mr, · line 929
Quoted 6 lines
memory.c +1
-1@@ -929,7 +925,7 @@ void memory_region_init_io(MemoryRegion *mr, mr->ops = ops; mr->opaque = opaque; mr->terminates = true; - mr->destructor = memory_region_destructor_iomem; + mr->destructor = memory_region_destructor_none; mr->ram_addr = ~(ram_addr_t )0 ;
We've just called memory_region_init() which sets the destructor to the 'none' function, so you don't need to re-set it again here.
-- PMM
PATCH 6/40 memory: make memory_global_sync_dirty_bitmap take an AddressSpace
In reply to Paolo Bonzini--- arch_init.c | 2 +- include/exec/memory.h | 7 +++---- memory.c | 3 +-- 3 files changed, 5 insertions(+), 7 deletions(-)
arch_init.c +1 -1
@@ -386,7 +386,7 @@ static void migration_bitmap_sync(void) }trace_migration_bitmap_sync_start (); -memory_global_sync_dirty_bitmap (get_system_memory ()); +address_space_sync_dirty_bitmap (&address_space_memory); QTAILQ_FOREACH(block, &ram_list.blocks, next) {for (addr =0 ; addr < block->length; addr +=TARGET_PAGE_SIZE ) {
include/exec/memory.h +3 -4
@@ -757,13 +757,12 @@ memory_region_section_addr(MemoryRegionSection *section, }/** -* memory_global_sync_dirty_bitmap: synchronize the dirty log for all memory +* address_space_sync_dirty_bitmap: synchronize the dirty log for all memory * * Synchronizes the dirty page log for an entire address space. -* @address_space: a top-level (i.e. parentless) region that contains the -* memory being synchronized +* @as: the address space that contains the memory being synchronized */ -void memory_global_sync_dirty_bitmap (MemoryRegion *address_space); +void address_space_sync_dirty_bitmap (AddressSpace *as);/** * memory_region_transaction_begin: Start a transaction.
memory.c +1 -2
@@ -1486,9 +1486,8 @@ MemoryRegionSection memory_region_find(MemoryRegion *mr,return ret; } -void memory_global_sync_dirty_bitmap (MemoryRegion *address_space) +void address_space_sync_dirty_bitmap (AddressSpace *as) { -AddressSpace *as =memory_region_to_address_space (address_space);FlatRange *fr; FOR_EACH_FLAT_RANGE(fr, as->current_map) {
Signature · 1 line
1.7.1
Re: [Qemu-devel] [PATCH 06/40] memory: make memory_global_sync_dirty_bitmap take an AddressSpace
Quoted 2 lines · Paolo Bonzini
Suggested-by Peter Maydell <[email protected]>Signed-off-by Paolo Bonzini <[email protected]>
-- PMM
PATCH 7/40 memory: fix address space initialization/destruction
In reply to Paolo BonziniFrom: Avi Kivity <[email protected]>
A couple of fields were left uninitialized. This was not observed earlier because all address spaces were statically allocated. Also free allocation for those fields.
--- memory.c | 3 +++ 1 files changed, 3 insertions(+), 0 deletions(-)
memory.c +3 -0
@@ -1573,6 +1573,8 @@ void address_space_init(AddressSpace *as, MemoryRegion *root) as->root = root; as->current_map =g_new (FlatView,1 );flatview_init (as->current_map); + as->ioeventfd_nb =0 ; + as->ioeventfds =NULL ;QTAILQ_INSERT_TAIL (&address_spaces, as, address_spaces_link); as->name =NULL ;memory_region_transaction_commit ();
@@ -1589,6 +1591,7 @@ void address_space_destroy(AddressSpace *as)address_space_destroy_dispatch (as);flatview_destroy (as->current_map);g_free (as->current_map); +g_free (as->ioeventfds); }uint64_t io_mem_read (MemoryRegion *mr,hwaddr addr,unsigned size)
Signature · 1 line
1.7.1
Re: [Qemu-devel] [PATCH 07/40] memory: fix address space initialization/destruction
Quoted 8 lines · Paolo Bonzini
From: Avi Kivity <[email protected]>
A couple of fields were left uninitialized. This was not observed earlier because all address spaces were statically allocated. Also free allocation for those fields.
Signed-off-by Avi Kivity <[email protected]>Signed-off-by Paolo Bonzini <[email protected]>
-- PMM
PATCH 8/40 memory: limit sections in the radix tree to the actual address space size
In reply to Paolo BonziniFrom: Avi Kivity <[email protected]>
The radix tree is statically sized to fit TARGET_PHYS_ADDR_SPACE_BITS. If a larger memory region is registered, it will overflow.
Fix by limiting any section in the radix tree to the supported size.
This problem was not observed earlier since artificial regions (containers and aliases) are eliminated by the memory core, leaving only device regions which have reasonable sizes. An IOMMU however cannot be eliminated by the memory core, and may have an artificial size.
--- exec.c | 15 ++++++++++++++- 1 files changed, 14 insertions(+), 1 deletions(-)
exec.c +14 -1
@@ -777,10 +777,23 @@ static void register_multipage(AddressSpaceDispatch *d, MemoryRegionSection *sec section_index); } +static MemoryRegionSection limit (MemoryRegionSection section) +{ +unsigned practical_as_bits =MIN (TARGET_PHYS_ADDR_SPACE_BITS ,62 ); +hwaddr as_limit; + + as_limit = (hwaddr )1 << practical_as_bits; + + section.size =MIN (section.offset_within_address_space + section.size, as_limit) + - section.offset_within_address_space; + +return section; +} +static void mem_add (MemoryListener *listener,MemoryRegionSection *section) {AddressSpaceDispatch *d =container_of (listener, AddressSpaceDispatch, listener); -MemoryRegionSection now = *section, remain = *section; +MemoryRegionSection now =limit (*section), remain =limit (*section);if ((now.offset_within_address_space & ~TARGET_PAGE_MASK ) || (now.size <TARGET_PAGE_SIZE )) {
Signature · 1 line
1.7.1
Re: [Qemu-devel] [PATCH 08/40] memory: limit sections in the radix tree to the actual address space size
Quoted 12 lines · Paolo Bonzini
From: Avi Kivity <[email protected]>
The radix tree is statically sized to fit
TARGET_PHYS_ADDR_SPACE_BITS. If a larger memory region is registered, it will overflow.Fix by limiting any section in the radix tree to the supported size.
This problem was not observed earlier since artificial regions (containers and aliases) are eliminated by the memory core, leaving only device regions which have reasonable sizes. An IOMMU however cannot be eliminated by the memory core, and may have an artificial size.
exec.c static void register_multipage(AddressSpaceDispatch *d, MemoryRegionSection *sec · line 788
Quoted 9 lines
exec.c +9
-0+static MemoryRegionSection limit (MemoryRegionSection section) +{ +unsigned practical_as_bits =MIN (TARGET_PHYS_ADDR_SPACE_BITS ,62 ); +hwaddr as_limit; + + as_limit = (hwaddr )1 << practical_as_bits; + + section.size =MIN (section.offset_within_address_space + section.size, as_limit) + - section.offset_within_address_space;
Isn't this going to give you a negative size for a section which is up at the top of physical memory in a CPU with a 63 or 64 bit physical address space? [ie one where the section.offset_within_address_space > as_limit]
(also, overly long lines)
thanks
-- PMM
Re: [Qemu-devel] [PATCH 08/40] memory: limit sections in the radix tree to the actual address space size
----- Messaggio originale -----
Quoted 20 lines
Da: "Peter Maydell" <[email protected]>
A: "Paolo Bonzini" <[email protected]>
Cc: [email protected], [email protected], "jan kiszka" <[email protected]>, [email protected], "Avi Kivity" <[email protected]>, [email protected], [email protected]
Inviato: Martedì, 7 maggio 2013 19:13:16
Oggetto: Re: [Qemu-devel] [PATCH 08/40] memory: limit sections in the radix tree to the actual address space sizeQuoted 11 lines · Paolo Bonzini
From: Avi Kivity <[email protected]>
The radix tree is statically sized to fit
TARGET_PHYS_ADDR_SPACE_BITS. If a larger memory region is registered, it will overflow.Fix by limiting any section in the radix tree to the supported size.
This problem was not observed earlier since artificial regions (containers and aliases) are eliminated by the memory core, leaving only device regions which have reasonable sizes. An IOMMU however cannot be eliminated by the memory core, and may have an artificial size.
Quoted 15 lines
Quoted 10 lines
exec.c +8
-0+static MemoryRegionSection limit (MemoryRegionSection section) +{ +unsigned practical_as_bits =MIN (TARGET_PHYS_ADDR_SPACE_BITS ,62 ); +hwaddr as_limit; + + as_limit = (hwaddr )1 << practical_as_bits; + + section.size = MIN(section.offset_within_address_space + section.size,
as_limit)
+ - section.offset_within_address_space;Isn't this going to give you a negative size for a section which is up at the top of physical memory in a CPU with a 63 or 64 bit physical address space? [ie one where the section.offset_within_address_space >
as_limit]
Yes, this assumes that TARGET_PHYS_ADDR_SPACE_BITS < 62 in practice, and that only artificial regions are larger. But perhaps we should have a BUILD_BUG_ON instead of the MIN.
The only target that has a bigger _physical_ address space is s390. Alex, is that definition correct?
Paolo
Re: [Qemu-devel] [PATCH 08/40] memory: limit sections in the radix tree to the actual address space size
Quoted 35 lines · Paolo Bonzini
----- Messaggio originale -----
Quoted 25 lines
Da: "Peter Maydell"<[email protected]>
A: "Paolo Bonzini"<[email protected]>
Cc: [email protected], [email protected], "jan kiszka"<[email protected]>, [email protected], "Avi Kivity" <[email protected]>, [email protected], [email protected]
Inviato: Martedì, 7 maggio 2013 19:13:16
Oggetto: Re: [Qemu-devel] [PATCH 08/40] memory: limit sections in the radix tree to the actual address space sizeQuoted 17 lines · Paolo Bonzini
From: Avi Kivity<[email protected]>
The radix tree is statically sized to fit
TARGET_PHYS_ADDR_SPACE_BITS. If a larger memory region is registered, it will overflow.Fix by limiting any section in the radix tree to the supported size.
This problem was not observed earlier since artificial regions (containers and aliases) are eliminated by the memory core, leaving only device regions which have reasonable sizes. An IOMMU however cannot be eliminated by the memory core, and may have an artificial size.
exec.c +6
-0+static MemoryRegionSection limit (MemoryRegionSection section) +{ +unsigned practical_as_bits =MIN (TARGET_PHYS_ADDR_SPACE_BITS ,62 ); +hwaddr as_limit; + + as_limit = (hwaddr )1 << practical_as_bits;Quoted 8 lines
Quoted 4 lines
(quoted diff) +2
-0+ + section.size = MIN(section.offset_within_address_space + section.size,
as_limit)
+ - section.offset_within_address_space;Isn't this going to give you a negative size for a section which is up at the top of physical memory in a CPU with a 63 or 64 bit physical address space? [ie one where the section.offset_within_address_space>
as_limit]
Quoted 6 lines
Yes, this assumes that TARGET_PHYS_ADDR_SPACE_BITS< 62 in practice, and that only artificial regions are larger. But perhaps we should have a
BUILD_BUG_ONinstead of the MIN.The only target that has a bigger _physical_ address space is s390. Alex, is that definition correct?
IIRC s390x can handle up to 64 bit of physical address space from its page tables. But we never use that much, so it's perfectly fine for me if you just decrease the value.
Alex
PATCH 9/40 memory: create FlatView for new address spaces
In reply to Paolo BonziniA new address space has an empty FlatView. In order to create it, we must (a) call memory_region_transaction_commit after the address space is inserted into the list; (b) force memory_region_transaction_commit to do something.
This bug was latent so far because all address spaces started empty, including the PCI address space where the bus master region is initially disabled. However, the target address space of an IOMMU is usually rooted at get_system_memory(), which might not be empty at the time the IOMMU is created.
--- memory.c | 3 ++- 1 files changed, 2 insertions(+), 1 deletions(-)
memory.c +2 -1
@@ -1577,8 +1577,9 @@ void address_space_init(AddressSpace *as, MemoryRegion *root) as->ioeventfds =NULL ;QTAILQ_INSERT_TAIL (&address_spaces, as, address_spaces_link); as->name =NULL ; -memory_region_transaction_commit ();address_space_init_dispatch (as); + memory_region_update_pending |= root->enabled; +memory_region_transaction_commit (); }void address_space_destroy (AddressSpace *as)
Signature · 1 line
1.7.1
Re: [Qemu-devel] [PATCH 09/40] memory: create FlatView for new address spaces
Quoted 9 lines · Paolo Bonzini
A new address space has an empty FlatView. In order to create it, we must (a) call
memory_region_transaction_commitafter the address space is inserted into the list; (b) forcememory_region_transaction_committo do something.This bug was latent so far because all address spaces started empty, including the PCI address space where the bus master region is initially disabled. However, the target address space of an IOMMU is usually rooted at
get_system_memory(), which might not be empty at the time the IOMMU is created.
The code looks ok but I don't entirely understand the commit message. Surely we're definitely already creating the FlatView -- that's what the g_new and flatview_init do. What we're not doing is calling notifiers for the initial contents of the address space, right? (and thus in turn we're not setting up the data structures those notifiers control, but as far as I can see from the code that's not the FlatView.)
thanks
-- PMM
Re: [Qemu-devel] [PATCH 09/40] memory: create FlatView for new address spaces
Quoted 14 lines · Paolo Bonzini
Quoted 11 lines
A new address space has an empty FlatView. In order to create it, we must (a) call
memory_region_transaction_commitafter the address space is inserted into the list; (b) forcememory_region_transaction_committo do something.This bug was latent so far because all address spaces started empty, including
the PCI address space where the bus master region is initially disabled. However, the target address space of an IOMMU is usually rooted atget_system_memory(), which might not be empty at the time the IOMMU is created.
Quoted 8 lines
The code looks ok but I don't entirely understand the commit message. Surely we're definitely already creating the FlatView -- that's what the
g_newandflatview_initdo. What we're not doing is calling notifiers for the initial contents of the address space, right? (and thus in turn we're not setting up the data structures those notifiers control, but as far as I can see from the code that's not the FlatView.)
Right, it should be s/create/populate/g.
Paolo
PATCH 10/40 exec: remove obsolete comment
In reply to Paolo BonziniSee how we call memory_region_section_addr two lines below to convert a physical address to a base address in the region.
--- exec.c | 6 ------ 1 files changed, 0 insertions(+), 6 deletions(-)
exec.c +0 -6
@@ -639,12 +639,6 @@ hwaddr memory_region_section_get_iotlb(CPUArchState *env, iotlb |= phys_section_rom; } }else { -/* IO handlers are currently passed a physical address. -It would be nice to pass an offset from the base address -of that region. This would avoid having to special case RAM, -and avoid full address decoding in every device. -We can't use the high bits of pd for this because -IO_MEM_ROMD uses these as a ram address. */ iotlb = section - phys_sections; iotlb +=memory_region_section_addr (section, paddr); }
Signature · 1 line
1.7.1
Re: [Qemu-devel] [PATCH 10/40] exec: remove obsolete comment
Quoted 4 lines · Paolo Bonzini
See how we call
memory_region_section_addrtwo lines below to convert a physical address to a base address in the region.Signed-off-by Paolo Bonzini <[email protected]>
Yeah, this comment is years out of date...
-- PMM
PATCH 11/40 memory: add address_space_valid
In reply to Paolo BonziniChecking whether an address space is possible in the old-style IOMMU implementation, but there is no equivalent in the memory API. Implement it with a lookup of the dispatch tree.
--- dma-helpers.c | 5 +++++ exec.c | 24 ++++++++++++++++++++++++ include/exec/memory.h | 12 ++++++++++++ include/sysemu/dma.h | 3 ++- 4 files changed, 43 insertions(+), 1 deletions(-)
dma-helpers.c +5 -0
@@ -298,6 +298,11 @@ bool iommu_dma_memory_valid(DMAContext *dma, dma_addr_t addr, dma_addr_t len, plen = len; } +if (!address_space_valid (dma->as, paddr, len, + dir ==DMA_DIRECTION_FROM_DEVICE )) { +return false; + } + len -= plen; addr += plen; }
exec.c +24 -0
@@ -2093,6 +2093,30 @@ static void cpu_notify_map_clients(void) } } +bool address_space_valid (AddressSpace *as,hwaddr addr,int len,bool is_write) +{ +AddressSpaceDispatch *d = as->dispatch; +MemoryRegionSection *section; +int l; +hwaddr page; + +while (len >0 ) { + page = addr &TARGET_PAGE_MASK ; + l = (page +TARGET_PAGE_SIZE ) - addr; +if (l > len) { + l = len; + } + section =phys_page_find (d, addr >>TARGET_PAGE_BITS ); +if (section->mr == &io_mem_unassigned) { +return false; + } + + len -= l; + addr += l; + } +return true; +} +/* Map a physical memory region into a host virtual address. * May map a subset of the requested range, given by and returned in *plen. * May return NULL if resources needed to perform the mapping are exhausted.
include/exec/memory.h +12 -0
@@ -857,6 +857,18 @@ void address_space_write(AddressSpace *as, hwaddr addr,*/ void address_space_read (AddressSpace *as,hwaddr addr,uint8_t *buf,int len); +/* address_space_valid: check for validity of an address space range +* +* Check whether access to the given address space range is permitted by +* any IOMMU regions that are active for the address space. +* +* @as: #AddressSpace to be accessed +* @addr: address within that address space +* @len: pointer to length +* @is_write: indicates the transfer direction +*/ +bool address_space_valid (AddressSpace *as,hwaddr addr,int len,bool is_write); +/* address_space_map: map a physical memory region into a host virtual address * * May map a subset of the requested range, given by and returned in @plen.
include/sysemu/dma.h +2 -1
@@ -113,7 +113,8 @@ static inline bool dma_memory_valid(DMAContext *dma, DMADirection dir) {if (!dma_has_iommu (dma)) { -return true; +return address_space_valid (dma->as, addr, len, + dir ==DMA_DIRECTION_FROM_DEVICE ); }else {return iommu_dma_memory_valid (dma, addr, len, dir); }
Signature · 1 line
1.7.1
Re: [Qemu-devel] [PATCH 11/40] memory: add address_space_valid
Quoted 2 lines · Paolo Bonzini
Checking whether an address space is possible in the old-style IOMMU implementation, but there is no equivalent in the memory API.
This sentence appears to be missing a clause ("whether an address space is valid" ?)
Quoted 41 lines
Implement it with a lookup of the dispatch tree.
Signed-off-by Paolo Bonzini <[email protected]>--- dma-helpers.c | 5 +++++ exec.c | 24 ++++++++++++++++++++++++ include/exec/memory.h | 12 ++++++++++++ include/sysemu/dma.h | 3 ++- 4 files changed, 43 insertions(+), 1 deletions(-)dma-helpers.c +5
-0@@ -298,6 +298,11 @@ bool iommu_dma_memory_valid(DMAContext *dma, dma_addr_t addr, dma_addr_t len, plen = len; } +if (!address_space_valid (dma->as, paddr, len, + dir ==DMA_DIRECTION_FROM_DEVICE )) { +return false; + } + len -= plen; addr += plen; }exec.c +7
-0@@ -2093,6 +2093,30 @@ static void cpu_notify_map_clients(void) } } +bool address_space_valid (AddressSpace *as,hwaddr addr,int len,bool is_write) +{ +AddressSpaceDispatch *d = as->dispatch; +MemoryRegionSection *section; +int l; +hwaddr page; +
exec.c static void cpu_notify_map_clients(void) · line 2112
Quoted 10 lines
exec.c +10
-0+while (len >0 ) { + page = addr &TARGET_PAGE_MASK ; + l = (page +TARGET_PAGE_SIZE ) - addr; +if (l > len) { + l = len; + } + section =phys_page_find (d, addr >>TARGET_PAGE_BITS ); +if (section->mr == &io_mem_unassigned) { +return false; + }
Shouldn't we also be failing attempts to write to read-only memory regions?
What about the case where there's a subpage-sized unassigned region in the middle of the area we want to access?
Quoted 16 lines
exec.c +7
-0+ + len -= l; + addr += l; + } +return true; +} +/* Map a physical memory region into a host virtual address. * May map a subset of the requested range, given by and returned in *plen. * May return NULL if resources needed to perform the mapping are exhausted. include/exec/memory.h +0
-0@@ -857,6 +857,18 @@ void address_space_write(AddressSpace *as, hwaddr addr,*/
include/exec/memory.h void address_space_write(AddressSpace *as, hwaddr addr, · line 867
Quoted 10 lines
include/exec/memory.h +8
-0void address_space_read (AddressSpace *as,hwaddr addr,uint8_t *buf,int len); +/* address_space_valid: check for validity of an address space range +* +* Check whether access to the given address space range is permitted by +* any IOMMU regions that are active for the address space. +* +* @as: #AddressSpace to be accessed +* @addr: address within that address space +* @len: pointer to length
Really a pointer? Function prototype suggests not.
include/exec/memory.h void address_space_write(AddressSpace *as, hwaddr addr, · line 870
Quoted 3 lines
include/exec/memory.h +3
-0+* @is_write: indicates the transfer direction +*/ +bool address_space_valid (AddressSpace *as,hwaddr addr,int len,bool is_write);
The function name suggests that the functionality ought to be "check whether this AddressSpace is valid" (whatever that would mean), rather than "check whether this access to this memory range is permitted in this AddressSpace".
address_space_access_ok() ?
(Aside: I notice that address_space_{rw,read,write} don't document their 'len' parameters.)
thanks
-- PMM
Re: [Qemu-devel] [PATCH 11/40] memory: add address_space_valid
Il 07/05/2013 19:40, Peter Maydell ha scritto:
Quoted 6 lines · Paolo Bonzini
Quoted 2 lines
Checking whether an address space is possible in the old-style IOMMU implementation, but there is no equivalent in the memory API.
This sentence appears to be missing a clause ("whether an address space is valid" ?)
The old-style IOMMU lets you check whether an access is valid in a given DMAContext. There is no equivalent for AddressSpace in the memory API, implement it with a lookup of the dispatch tree.
Quoted 46 lines
Quoted 46 lines
Implement it with a lookup of the dispatch tree.
Signed-off-by Paolo Bonzini <[email protected]>--- dma-helpers.c | 5 +++++ exec.c | 24 ++++++++++++++++++++++++ include/exec/memory.h | 12 ++++++++++++ include/sysemu/dma.h | 3 ++- 4 files changed, 43 insertions(+), 1 deletions(-)dma-helpers.c +5
-0@@ -298,6 +298,11 @@ bool iommu_dma_memory_valid(DMAContext *dma, dma_addr_t addr, dma_addr_t len, plen = len; } +if (!address_space_valid (dma->as, paddr, len, + dir ==DMA_DIRECTION_FROM_DEVICE )) { +return false; + } + len -= plen; addr += plen; }exec.c +12
-0@@ -2093,6 +2093,30 @@ static void cpu_notify_map_clients(void) } } +bool address_space_valid (AddressSpace *as,hwaddr addr,int len,bool is_write) +{ +AddressSpaceDispatch *d = as->dispatch; +MemoryRegionSection *section; +int l; +hwaddr page; + +while (len >0 ) { + page = addr &TARGET_PAGE_MASK ; + l = (page +TARGET_PAGE_SIZE ) - addr; +if (l > len) { + l = len;
Quoted 8 lines
exec.c static void cpu_notify_map_clients(void) · line 2112
Quoted 5 lines
exec.c +5
-0+ } + section =phys_page_find (d, addr >>TARGET_PAGE_BITS ); +if (section->mr == &io_mem_unassigned) { +return false; + }Shouldn't we also be failing attempts to write to read-only memory regions?
Yes.
Quoted 2 lines
What about the case where there's a subpage-sized unassigned region in the middle of the area we want to access?
Indeed subpage ranges are not supported yet. I noticed it when reviewing Jan's patch (which, if salvaged would let us implement that too). I'll document the limitation.
Quoted 29 lines
Quoted 16 lines
exec.c +7
-0+ + len -= l; + addr += l; + } +return true; +} +/* Map a physical memory region into a host virtual address. * May map a subset of the requested range, given by and returned in *plen. * May return NULL if resources needed to perform the mapping are exhausted. include/exec/memory.h +0
-0@@ -857,6 +857,18 @@ void address_space_write(AddressSpace *as, hwaddr addr,*/ include/exec/memory.h void address_space_write(AddressSpace *as, hwaddr addr, · line 867
Quoted 10 lines
include/exec/memory.h +8
-0void address_space_read (AddressSpace *as,hwaddr addr,uint8_t *buf,int len); +/* address_space_valid: check for validity of an address space range +* +* Check whether access to the given address space range is permitted by +* any IOMMU regions that are active for the address space. +* +* @as: #AddressSpace to be accessed +* @addr: address within that address space +* @len: pointer to length Really a pointer? Function prototype suggests not.
Quoted 8 lines
include/exec/memory.h void address_space_write(AddressSpace *as, hwaddr addr, · line 870
Quoted 3 lines
include/exec/memory.h +3
-0+* @is_write: indicates the transfer direction +*/ +bool address_space_valid (AddressSpace *as,hwaddr addr,int len,bool is_write);The function name suggests that the functionality ought to be "check whether this AddressSpace is valid" (whatever that would mean), rather than "check whether this access to this memory range is permitted in this AddressSpace".
I was mimicking dma_memory_valid, but it's not a good example to follow.
Paolo
Quoted 8 lines
address_space_access_ok()?(Aside: I notice that address_space_{rw,read,write} don't document their 'len' parameters.)
thanks
-- PMM
PATCH 12/40 memory: add address_space_translate
In reply to Paolo BonziniUsing phys_page_find to translate an AddressSpace to a MemoryRegionSection is unwieldy. It requires to pass the page index rather than the address, and later memory_region_section_addr has to be called. Replace memory_region_section_addr with a function that does all of it: call phys_page_find, compute the offset within the region, and check how big the current mapping is. This way, a large flat region can be written with a single lookup rather than a page at a time.
address_space_translate will also provide a single point where IOMMU forwarding is implemented.
--- cputlb.c | 20 +++--- exec.c | 184 ++++++++++++++++++++++++++---------------------- include/exec/cputlb.h | 12 ++-- include/exec/memory.h | 31 ++++----- translate-all.c | 6 +- 5 files changed, 133 insertions(+), 120 deletions(-)
cputlb.c +11 -9
@@ -248,13 +248,18 @@ void tlb_set_page(CPUArchState *env, target_ulong vaddr,target_ulong code_address;uintptr_t addend;CPUTLBEntry *te; -hwaddr iotlb; +hwaddr iotlb, xlat, sz;assert (size >=TARGET_PAGE_SIZE );if (size !=TARGET_PAGE_SIZE ) {tlb_add_large_page (env, vaddr, size); } - section =phys_page_find (address_space_memory.dispatch, paddr >>TARGET_PAGE_BITS ); + + sz = size; + section =address_space_translate (&address_space_memory, paddr, &xlat, &sz, + false); +assert (sz >=TARGET_PAGE_SIZE ); +#if defined(DEBUG_TLB ) printf("tlb_set_page: vaddr=" TARGET_FMT_lx " paddr=0x" TARGET_FMT_plx " prot=%x idx=%d pd=0x%08lx\n",
@@ -269,15 +274,14 @@ void tlb_set_page(CPUArchState *env, target_ulong vaddr, }if (memory_region_is_ram (section->mr) ||memory_region_is_romd (section->mr)) { - addend = (uintptr_t )memory_region_get_ram_ptr (section->mr) - +memory_region_section_addr (section, paddr); + addend = (uintptr_t )memory_region_get_ram_ptr (section->mr) + xlat; }else { addend =0 ; } code_address = address; - iotlb =memory_region_section_get_iotlb (env, section, vaddr, paddr, prot, - &address); + iotlb =memory_region_section_get_iotlb (env, section, vaddr, paddr, xlat, + prot, &address); index = (vaddr >>TARGET_PAGE_BITS ) & (CPU_TLB_SIZE -1 ); env->iotlb[mmu_idx][index] = iotlb - vaddr;
@@ -300,9 +304,7 @@ void tlb_set_page(CPUArchState *env, target_ulong vaddr,/* Write access calls the I/O callback. */ te->addr_write = address |TLB_MMIO ; }else if (memory_region_is_ram (section->mr) - && !cpu_physical_memory_is_dirty ( - section->mr->ram_addr - +memory_region_section_addr (section, paddr))) { + && !cpu_physical_memory_is_dirty (section->mr->ram_addr + xlat)) { te->addr_write = address |TLB_NOTDIRTY ; }else { te->addr_write = address;
exec.c +100 -84
415 lines not on this page · raw message
include/exec/cputlb.h +5 -7
@@ -26,8 +26,6 @@ void tlb_unprotect_code_phys(CPUArchState *env, ram_addr_t ram_addr, target_ulong vaddr);void tlb_reset_dirty_range (CPUTLBEntry *tlb_entry,uintptr_t start,uintptr_t length); -MemoryRegionSection *phys_page_find (struct AddressSpaceDispatch *d, -hwaddr index);void cpu_tlb_reset_dirty_all (ram_addr_t start1,ram_addr_t length);void tlb_set_dirty (CPUArchState *env,target_ulong vaddr);extern int tlb_flush_count;
@@ -35,11 +33,11 @@ extern int tlb_flush_count;/* exec.c */ void tb_flush_jmp_cache (CPUArchState *env,target_ulong addr);hwaddr memory_region_section_get_iotlb (CPUArchState *env, -MemoryRegionSection *section, -target_ulong vaddr, -hwaddr paddr, -int prot, -target_ulong *address); +MemoryRegionSection *section, +target_ulong vaddr, +hwaddr paddr,hwaddr xlat, +int prot, +target_ulong *address);bool memory_region_is_unassigned (MemoryRegion *mr);#endif
include/exec/memory.h +14 -17
@@ -740,23 +740,6 @@ MemoryRegionSection memory_region_find(MemoryRegion *mr, hwaddr addr, uint64_t size);/** -* memory_region_section_addr: get offset within MemoryRegionSection -* -* Returns offset within MemoryRegionSection -* -* @section: the memory region section being queried -* @addr: address in address space -*/ -static inline hwaddr -memory_region_section_addr (MemoryRegionSection *section, -hwaddr addr) -{ - addr -= section->offset_within_address_space; - addr += section->offset_within_region; -return addr; -} - -/** * address_space_sync_dirty_bitmap: synchronize the dirty log for all memory * * Synchronizes the dirty page log for an entire address space.
@@ -857,6 +840,20 @@ void address_space_write(AddressSpace *as, hwaddr addr,*/ void address_space_read (AddressSpace *as,hwaddr addr,uint8_t *buf,int len); +/* address_space_translate: translate an address range into an address space +* into a MemoryRegionSection and an address range into that section +* +* @as: #AddressSpace to be accessed +* @addr: address within that address space +* @xlat: pointer to address within the returned memory region section's +* #MemoryRegion. +* @len: pointer to length +* @is_write: indicates the transfer direction +*/ +MemoryRegionSection *address_space_translate (AddressSpace *as,hwaddr addr, +hwaddr *xlat,hwaddr *len, +bool is_write); +/* address_space_valid: check for validity of an address space range * * Check whether access to the given address space range is permitted by
translate-all.c +3 -3
@@ -1355,15 +1355,15 @@ void tb_invalidate_phys_addr(hwaddr addr) {ram_addr_t ram_addr;MemoryRegionSection *section; +hwaddr l =1 ; - section =phys_page_find (address_space_memory.dispatch, - addr >>TARGET_PAGE_BITS ); + section =address_space_translate (&address_space_memory, addr, &addr, &l, false);if (!(memory_region_is_ram (section->mr) ||memory_region_is_romd (section->mr))) {return ; } ram_addr = (memory_region_get_ram_addr (section->mr) &TARGET_PAGE_MASK ) - +memory_region_section_addr (section, addr); + + addr;tb_invalidate_phys_page_range (ram_addr, ram_addr +1 ,0 ); }#endif /* TARGET_HAS_ICE && !defined(CONFIG_USER_ONLY) */
Signature · 1 line
1.7.1
Re: [Qemu-devel] [PATCH 12/40] memory: add address_space_translate
Quoted 10 lines · Paolo Bonzini
Using
phys_page_findto translate an AddressSpace to a MemoryRegionSection is unwieldy. It requires to pass the page index rather than the address, and latermemory_region_section_addrhas to be called. Replacememory_region_section_addrwith a function that does all of it: callphys_page_find, compute the offset within the region, and check how big the current mapping is. This way, a large flat region can be written with a single lookup rather than a page at a time.
address_space_translatewill also provide a single point where IOMMU forwarding is implemented.
This generally looks right, so the below is really all nitpicks.
Quoted 80 lines
Signed-off-by Paolo Bonzini <[email protected]>--- cputlb.c | 20 +++--- exec.c | 184 ++++++++++++++++++++++++++---------------------- include/exec/cputlb.h | 12 ++-- include/exec/memory.h | 31 ++++----- translate-all.c | 6 +- 5 files changed, 133 insertions(+), 120 deletions(-)cputlb.c +11
-9@@ -248,13 +248,18 @@ void tlb_set_page(CPUArchState *env, target_ulong vaddr,target_ulong code_address;uintptr_t addend;CPUTLBEntry *te; -hwaddr iotlb; +hwaddr iotlb, xlat, sz;assert (size >=TARGET_PAGE_SIZE );if (size !=TARGET_PAGE_SIZE ) {tlb_add_large_page (env, vaddr, size); } - section =phys_page_find (address_space_memory.dispatch, paddr >>TARGET_PAGE_BITS ); + + sz = size; + section =address_space_translate (&address_space_memory, paddr, &xlat, &sz, + false); +assert (sz >=TARGET_PAGE_SIZE ); +#if defined(DEBUG_TLB ) printf("tlb_set_page: vaddr=" TARGET_FMT_lx " paddr=0x" TARGET_FMT_plx " prot=%x idx=%d pd=0x%08lx\n",@@ -269,15 +274,14 @@ void tlb_set_page(CPUArchState *env, target_ulong vaddr, }if (memory_region_is_ram (section->mr) ||memory_region_is_romd (section->mr)) { - addend = (uintptr_t )memory_region_get_ram_ptr (section->mr) - +memory_region_section_addr (section, paddr); + addend = (uintptr_t )memory_region_get_ram_ptr (section->mr) + xlat; }else { addend =0 ; } code_address = address; - iotlb =memory_region_section_get_iotlb (env, section, vaddr, paddr, prot, - &address); + iotlb =memory_region_section_get_iotlb (env, section, vaddr, paddr, xlat, + prot, &address); index = (vaddr >>TARGET_PAGE_BITS ) & (CPU_TLB_SIZE -1 ); env->iotlb[mmu_idx][index] = iotlb - vaddr;@@ -300,9 +304,7 @@ void tlb_set_page(CPUArchState *env, target_ulong vaddr,/* Write access calls the I/O callback. */ te->addr_write = address |TLB_MMIO ; }else if (memory_region_is_ram (section->mr) - && !cpu_physical_memory_is_dirty ( - section->mr->ram_addr - +memory_region_section_addr (section, paddr))) { + && !cpu_physical_memory_is_dirty (section->mr->ram_addr + xlat)) { te->addr_write = address |TLB_NOTDIRTY ; }else { te->addr_write = address;exec.c +1
-2@@ -182,24 +182,36 @@ static void phys_page_set(AddressSpaceDispatch *d,phys_page_set_level (&d->phys_map, &index, &nb, leaf,P_L2_LEVELS -1 ); } -MemoryRegionSection *phys_page_find (AddressSpaceDispatch *d,hwaddr index) +static MemoryRegionSection *phys_page_find (AddressSpaceDispatch *d,hwaddr index) {PhysPageEntry lp = d->phys_map;PhysPageEntry *p;int i; -uint16_t s_index = phys_section_unassigned;
exec.c static void phys_page_set(AddressSpaceDispatch *d, · line 199
Quoted 10 lines
exec.c +3
-1for (i =P_L2_LEVELS -1 ; i >=0 && !lp.is_leaf; i--) {if (lp.ptr ==PHYS_MAP_NODE_NIL ) { - goto not_found; +return &phys_sections[phys_section_unassigned]; } p = phys_map_nodes[lp.ptr]; lp = p[(index >> (i *L2_BITS )) & (L2_SIZE -1 )]; } +return &phys_sections[lp.ptr]; +}
The changes to this function are just minor refactoring tweaks, right? (mostly, removing the goto). If they were a separate patch that would help, especially since diff has interlaced this hunk with the new function below.
exec.c static void phys_page_set(AddressSpaceDispatch *d, · line 208
Quoted 9 lines
exec.c +9
-0+ +MemoryRegionSection *address_space_translate (AddressSpace *as,hwaddr addr, +hwaddr *xlat,hwaddr *plen, +bool is_write) +{ +MemoryRegionSection *section; + + section =phys_page_find (as->dispatch, addr >>TARGET_PAGE_BITS ); +/* Compute offset with MemoryRegionSection */
"within" ?
exec.c static void phys_page_set(AddressSpaceDispatch *d, · line 212
Quoted 7 lines
exec.c +3
-3+ addr -= section->offset_within_address_space; + *plen =MIN (section->size - addr, *plen); - s_index = lp.ptr; -not_found: -return &phys_sections[s_index]; +/* Compute offset with MemoryRegion */
also "within"
Quoted 41 lines
exec.c +9
-7+ *xlat = addr + section->offset_within_region; +return section; }bool memory_region_is_unassigned (MemoryRegion *mr)@@ -620,11 +632,11 @@ static int cpu_physical_memory_set_dirty_tracking(int enable) }hwaddr memory_region_section_get_iotlb (CPUArchState *env, -MemoryRegionSection *section, -target_ulong vaddr, -hwaddr paddr, -int prot, -target_ulong *address) +MemoryRegionSection *section, +target_ulong vaddr, +hwaddr paddr,hwaddr xlat, +int prot, +target_ulong *address) {hwaddr iotlb;CPUWatchpoint *wp;@@ -632,7 +644,7 @@ hwaddr memory_region_section_get_iotlb(CPUArchState *env,if (memory_region_is_ram (section->mr)) {/* Normal RAM. */ iotlb = (memory_region_get_ram_addr (section->mr) &TARGET_PAGE_MASK ) - +memory_region_section_addr (section, paddr); + + xlat;if (!section->readonly) { iotlb |= phys_section_notdirty; }else {@@ -640,7 +652,7 @@ hwaddr memory_region_section_get_iotlb(CPUArchState *env, } }else { iotlb = section - phys_sections; - iotlb +=memory_region_section_addr (section, paddr); + iotlb += xlat; }/* Make accesses to pages with watchpoints go via the
exec.c static void invalidate_and_set_dirty(hwaddr addr, · line 1918
Quoted 6 lines
exec.c +1
-2@@ -1903,24 +1915,18 @@ static void invalidate_and_set_dirty(hwaddr addr,void address_space_rw (AddressSpace *as,hwaddr addr,uint8_t *buf,int len,bool is_write) { -AddressSpaceDispatch *d = as->dispatch; -int l; +hwaddr l;
...maybe the 'len' argument to this function should really be a hwaddr too?
exec.c static void invalidate_and_set_dirty(hwaddr addr, · line 1921
Quoted 4 lines
exec.c +1
-1uint8_t *ptr;uint32_t val; -hwaddr page; +hwaddr addr1;
(not your fault, but) 'addr1' is a really awful variable name...
Quoted 146 lines
exec.c +21
-45MemoryRegionSection *section;while (len >0 ) { - page = addr &TARGET_PAGE_MASK ; - l = (page +TARGET_PAGE_SIZE ) - addr; -if (l > len) - l = len; - section =phys_page_find (d, page >>TARGET_PAGE_BITS ); + l = len; + section =address_space_translate (as, addr, &addr1, &l, is_write);if (is_write) {if (!memory_region_is_ram (section->mr)) { -hwaddr addr1; - addr1 =memory_region_section_addr (section, addr);/* XXX: could force cpu_single_env to NULL to avoid potential bugs */ if (l >=4 && ((addr1 &3 ) ==0 )) {@@ -1940,9 +1946,7 @@ void address_space_rw(AddressSpace *as, hwaddr addr, uint8_t *buf, l =1 ; } }else if (!section->readonly) { -ram_addr_t addr1; - addr1 =memory_region_get_ram_addr (section->mr) - +memory_region_section_addr (section, addr); + addr1 +=memory_region_get_ram_addr (section->mr);/* RAM case */ ptr =qemu_get_ram_ptr (addr1);memcpy (ptr, buf, l);@@ -1952,9 +1956,7 @@ void address_space_rw(AddressSpace *as, hwaddr addr, uint8_t *buf, }else {if (!(memory_region_is_ram (section->mr) ||memory_region_is_romd (section->mr))) { -hwaddr addr1;/* I/O case */ - addr1 =memory_region_section_addr (section, addr);if (l >=4 && ((addr1 &3 ) ==0 )) {/* 32 bit read access */ val =io_mem_read (section->mr, addr1,4 );@@ -1973,9 +1975,7 @@ void address_space_rw(AddressSpace *as, hwaddr addr, uint8_t *buf, } }else {/* RAM case */ - ptr =qemu_get_ram_ptr (section->mr->ram_addr - +memory_region_section_addr (section, - addr)); + ptr =qemu_get_ram_ptr (section->mr->ram_addr + addr1);memcpy (buf, ptr, l);qemu_put_ram_ptr (ptr); }@@ -2015,26 +2015,21 @@ void cpu_physical_memory_rw(hwaddr addr, uint8_t *buf,void cpu_physical_memory_write_rom (hwaddr addr,const uint8_t *buf,int len) { -AddressSpaceDispatch *d = address_space_memory.dispatch; -int l; +hwaddr l;uint8_t *ptr; -hwaddr page; +hwaddr addr1;MemoryRegionSection *section;while (len >0 ) { - page = addr &TARGET_PAGE_MASK ; - l = (page +TARGET_PAGE_SIZE ) - addr; -if (l > len) - l = len; - section =phys_page_find (d, page >>TARGET_PAGE_BITS ); + l = len; + section =address_space_translate (&address_space_memory, + addr, &addr1, &l, true);if (!(memory_region_is_ram (section->mr) ||memory_region_is_romd (section->mr))) {/* do nothing */ }else { -unsigned long addr1; - addr1 =memory_region_get_ram_addr (section->mr) - +memory_region_section_addr (section, addr); + addr1 +=memory_region_get_ram_addr (section->mr);/* ROM/RAM case */ ptr =qemu_get_ram_ptr (addr1);memcpy (ptr, buf, l);@@ -2095,18 +2090,12 @@ static void cpu_notify_map_clients(void)bool address_space_valid (AddressSpace *as,hwaddr addr,int len,bool is_write) { -AddressSpaceDispatch *d = as->dispatch;MemoryRegionSection *section; -int l; -hwaddr page; +hwaddr l, xlat;while (len >0 ) { - page = addr &TARGET_PAGE_MASK ; - l = (page +TARGET_PAGE_SIZE ) - addr; -if (l > len) { - l = len; - } - section =phys_page_find (d, addr >>TARGET_PAGE_BITS ); + l = len; + section =address_space_translate (as, addr, &xlat, &l, is_write);if (section->mr == &io_mem_unassigned) {return false; }@@ -2129,22 +2118,17 @@ void *address_space_map(AddressSpace *as, hwaddr *plen, bool is_write) { -AddressSpaceDispatch *d = as->dispatch;hwaddr len = *plen;hwaddr todo =0 ; -int l; -hwaddr page; +hwaddr l, xlat;MemoryRegionSection *section;ram_addr_t raddr =RAM_ADDR_MAX ;ram_addr_t rlen;void *ret;while (len >0 ) { - page = addr &TARGET_PAGE_MASK ; - l = (page +TARGET_PAGE_SIZE ) - addr; -if (l > len) - l = len; - section =phys_page_find (d, page >>TARGET_PAGE_BITS ); + l = len; + section =address_space_translate (as, addr, &xlat, &l, is_write);if (!(memory_region_is_ram (section->mr) && !section->readonly)) {if (todo || bounce.buffer) {@@ -2161,8 +2145,11 @@ void *address_space_map(AddressSpace *as,return bounce.buffer; }if (!todo) { - raddr =memory_region_get_ram_addr (section->mr) - +memory_region_section_addr (section, addr); + raddr =memory_region_get_ram_addr (section->mr) + xlat; + }else { +if (memory_region_get_ram_addr (section->mr) + xlat != raddr + todo) { +break ; + } } len -= l;
exec.c static inline uint32_t ldl_phys_internal(hwaddr addr, · line 2221
Quoted 8 lines
exec.c +3
-1@@ -2228,13 +2215,17 @@ static inline uint32_t ldl_phys_internal(hwaddr addr,uint8_t *ptr;uint32_t val;MemoryRegionSection *section; +hwaddr l =4 ; - section =phys_page_find (address_space_memory.dispatch, addr >>TARGET_PAGE_BITS ); + section =address_space_translate (&address_space_memory, addr, &addr, &l, + false);
I find it a little confusing reusing 'addr' for the returned offset from address_space_translate(); maybe using a different variable would be clearer? (don't feel very strongly about it)
Quoted 51 lines
exec.c +11
-5+if (l <4 ) { +return -1 ; + }if (!(memory_region_is_ram (section->mr) ||memory_region_is_romd (section->mr))) {/* I/O case */ - addr =memory_region_section_addr (section, addr); val =io_mem_read (section->mr, addr,4 );#if defined(TARGET_WORDS_BIGENDIAN )if (endian ==DEVICE_LITTLE_ENDIAN ) {@@ -2249,7 +2240,7 @@ static inline uint32_t ldl_phys_internal(hwaddr addr,/* RAM case */ ptr =qemu_get_ram_ptr ((memory_region_get_ram_addr (section->mr) &TARGET_PAGE_MASK ) - +memory_region_section_addr (section, addr)); + + addr);switch (endian) {case DEVICE_LITTLE_ENDIAN : val =ldl_le_p (ptr);@@ -2287,13 +2278,17 @@ static inline uint64_t ldq_phys_internal(hwaddr addr,uint8_t *ptr;uint64_t val;MemoryRegionSection *section; +hwaddr l =8 ; - section =phys_page_find (address_space_memory.dispatch, addr >>TARGET_PAGE_BITS ); + section =address_space_translate (&address_space_memory, addr, &addr, &l, + false); +if (l <8 ) { +return -1 ; + }if (!(memory_region_is_ram (section->mr) ||memory_region_is_romd (section->mr))) {/* I/O case */ - addr =memory_region_section_addr (section, addr);/* XXX This is broken when device endian != cpu endian. Fix and add "endian" variable check */ @@ -2308,7 +2303,7 @@ static inline uint64_t ldq_phys_internal(hwaddr addr,/* RAM case */ ptr =qemu_get_ram_ptr ((memory_region_get_ram_addr (section->mr) &TARGET_PAGE_MASK ) - +memory_region_section_addr (section, addr)); + + addr);switch (endian) {case DEVICE_LITTLE_ENDIAN : val =ldq_le_p (ptr);@@ -2354,13 +2349,17 @@ static inline uint32_t lduw_phys_internal(hwaddr addr,uint8_t *ptr;
exec.c static inline uint32_t lduw_phys_internal(hwaddr addr, · line 2358
Quoted 10 lines
exec.c +6
-1uint64_t val;MemoryRegionSection *section; +hwaddr l =2 ; - section =phys_page_find (address_space_memory.dispatch, addr >>TARGET_PAGE_BITS ); + section =address_space_translate (&address_space_memory, addr, &addr, &l, + false); +if (l <2 ) { +return -1 ; + }
This kind of "let's just return -1 if the read failed" kind of bothers me, because "memory access aborted" is really completely different from "memory access succeeded and returned -1". But there are a lot of APIs which we'd need to fix to properly communicate the problem back to the caller, so let it slide for now. (What used to happen in the old code in this case?)
Quoted 154 lines
exec.c +32
-16if (!(memory_region_is_ram (section->mr) ||memory_region_is_romd (section->mr))) {/* I/O case */ - addr =memory_region_section_addr (section, addr); val =io_mem_read (section->mr, addr,2 );#if defined(TARGET_WORDS_BIGENDIAN )if (endian ==DEVICE_LITTLE_ENDIAN ) {@@ -2375,7 +2374,7 @@ static inline uint32_t lduw_phys_internal(hwaddr addr,/* RAM case */ ptr =qemu_get_ram_ptr ((memory_region_get_ram_addr (section->mr) &TARGET_PAGE_MASK ) - +memory_region_section_addr (section, addr)); + + addr);switch (endian) {case DEVICE_LITTLE_ENDIAN : val =lduw_le_p (ptr);@@ -2413,11 +2412,15 @@ void stl_phys_notdirty(hwaddr addr, uint32_t val) {uint8_t *ptr;MemoryRegionSection *section; +hwaddr l =2 ; - section =phys_page_find (address_space_memory.dispatch, addr >>TARGET_PAGE_BITS ); + section =address_space_translate (&address_space_memory, addr, &addr, &l, + true); +if (l <2 ) { +return ; + }if (!memory_region_is_ram (section->mr) || section->readonly) { - addr =memory_region_section_addr (section, addr);if (memory_region_is_ram (section->mr)) { section = &phys_sections[phys_section_rom]; }@@ -2425,7 +2428,7 @@ void stl_phys_notdirty(hwaddr addr, uint32_t val) }else {unsigned long addr1 = (memory_region_get_ram_addr (section->mr) &TARGET_PAGE_MASK ) - +memory_region_section_addr (section, addr); + + addr; ptr =qemu_get_ram_ptr (addr1);stl_p (ptr, val);@@ -2445,11 +2448,15 @@ void stq_phys_notdirty(hwaddr addr, uint64_t val) {uint8_t *ptr;MemoryRegionSection *section; +hwaddr l =4 ; - section =phys_page_find (address_space_memory.dispatch, addr >>TARGET_PAGE_BITS ); + section =address_space_translate (&address_space_memory, addr, &addr, &l, + true); +if (l <4 ) { +return ; + }if (!memory_region_is_ram (section->mr) || section->readonly) { - addr =memory_region_section_addr (section, addr);if (memory_region_is_ram (section->mr)) { section = &phys_sections[phys_section_rom]; }@@ -2463,7 +2470,7 @@ void stq_phys_notdirty(hwaddr addr, uint64_t val) }else { ptr =qemu_get_ram_ptr ((memory_region_get_ram_addr (section->mr) &TARGET_PAGE_MASK ) - +memory_region_section_addr (section, addr)); + + addr);stq_p (ptr, val); } }@@ -2474,11 +2481,15 @@ static inline void stl_phys_internal(hwaddr addr, uint32_t val, {uint8_t *ptr;MemoryRegionSection *section; +hwaddr l =8 ; - section =phys_page_find (address_space_memory.dispatch, addr >>TARGET_PAGE_BITS ); + section =address_space_translate (&address_space_memory, addr, &addr, &l, + true); +if (l <8 ) { +return ; + }if (!memory_region_is_ram (section->mr) || section->readonly) { - addr =memory_region_section_addr (section, addr);if (memory_region_is_ram (section->mr)) { section = &phys_sections[phys_section_rom]; }@@ -2495,7 +2506,7 @@ static inline void stl_phys_internal(hwaddr addr, uint32_t val, }else {unsigned long addr1; addr1 = (memory_region_get_ram_addr (section->mr) &TARGET_PAGE_MASK ) - +memory_region_section_addr (section, addr); + + addr;/* RAM case */ ptr =qemu_get_ram_ptr (addr1);switch (endian) {@@ -2541,11 +2552,15 @@ static inline void stw_phys_internal(hwaddr addr, uint32_t val, {uint8_t *ptr;MemoryRegionSection *section; +hwaddr l =4 ; - section =phys_page_find (address_space_memory.dispatch, addr >>TARGET_PAGE_BITS ); + section =address_space_translate (&address_space_memory, addr, &addr, &l, + true); +if (l <4 ) { +return ; + }if (!memory_region_is_ram (section->mr) || section->readonly) { - addr =memory_region_section_addr (section, addr);if (memory_region_is_ram (section->mr)) { section = &phys_sections[phys_section_rom]; }@@ -2562,7 +2577,7 @@ static inline void stw_phys_internal(hwaddr addr, uint32_t val, }else {unsigned long addr1; addr1 = (memory_region_get_ram_addr (section->mr) &TARGET_PAGE_MASK ) - +memory_region_section_addr (section, addr); + + addr;/* RAM case */ ptr =qemu_get_ram_ptr (addr1);switch (endian) {@@ -2666,9 +2681,10 @@ bool virtio_is_big_endian(void)bool cpu_physical_memory_is_io (hwaddr phys_addr) {MemoryRegionSection *section; +hwaddr l =1 ; - section =phys_page_find (address_space_memory.dispatch, - phys_addr >>TARGET_PAGE_BITS ); + section =address_space_translate (&address_space_memory, + phys_addr, &phys_addr, &l, false);return !(memory_region_is_ram (section->mr) ||memory_region_is_romd (section->mr));include/exec/cputlb.h +0
-2@@ -26,8 +26,6 @@ void tlb_unprotect_code_phys(CPUArchState *env, ram_addr_t ram_addr, target_ulong vaddr);void tlb_reset_dirty_range (CPUTLBEntry *tlb_entry,uintptr_t start,uintptr_t length); -MemoryRegionSection *phys_page_find (struct AddressSpaceDispatch *d, -hwaddr index);void cpu_tlb_reset_dirty_all (ram_addr_t start1,ram_addr_t length);void tlb_set_dirty (CPUArchState *env,target_ulong vaddr);extern int tlb_flush_count;@@ -35,11 +33,11 @@ extern int tlb_flush_count;/* exec.c */ void tb_flush_jmp_cache (CPUArchState *env,target_ulong addr); hwaddr memory_region_section_get_iotlb(CPUArchState *env,
include/exec/cputlb.h extern int tlb_flush_count; · line 40
Quoted 10 lines
include/exec/cputlb.h +5
-5-MemoryRegionSection *section, - target_ulong vaddr, - hwaddr paddr, - int prot, - target_ulong *address); +MemoryRegionSection *section, + target_ulong vaddr, + hwaddr paddr, hwaddr xlat, + int prot, + target_ulong *address);
Any chance of a documentation comment? the purpose of the arguments to this function was mostly guessable until you added xlat.
Quoted 32 lines
include/exec/cputlb.h +0
-0bool memory_region_is_unassigned (MemoryRegion *mr);#endif include/exec/memory.h +0
-17@@ -740,23 +740,6 @@ MemoryRegionSection memory_region_find(MemoryRegion *mr, hwaddr addr, uint64_t size);/** -* memory_region_section_addr: get offset within MemoryRegionSection -* -* Returns offset within MemoryRegionSection -* -* @section: the memory region section being queried -* @addr: address in address space -*/ -static inline hwaddr -memory_region_section_addr (MemoryRegionSection *section, -hwaddr addr) -{ - addr -= section->offset_within_address_space; - addr += section->offset_within_region; -return addr; -} - -/** * address_space_sync_dirty_bitmap: synchronize the dirty log for all memory * * Synchronizes the dirty page log for an entire address space.
include/exec/memory.h void address_space_write(AddressSpace *as, hwaddr addr, · line 843
Quoted 4 lines
include/exec/memory.h +1
-0@@ -857,6 +840,20 @@ void address_space_write(AddressSpace *as, hwaddr addr,*/ void address_space_read (AddressSpace *as,hwaddr addr,uint8_t *buf,int len); +/* address_space_translate: translate an address range into an address space
"within an address space"
include/exec/memory.h void address_space_write(AddressSpace *as, hwaddr addr, · line 852
Quoted 9 lines
include/exec/memory.h +9
-0+* into a MemoryRegionSection and an address range into that section +* +* @as: #AddressSpace to be accessed +* @addr: address within that address space +* @xlat: pointer to address within the returned memory region section's +* #MemoryRegion. +* @len: pointer to length +* @is_write: indicates the transfer direction +*/
*len is used both as an input and an output, right? I think the doc comment could probably use expansion (ie a descriptive paragraph after the short parameter list). Then you can say "On entry *@len should be the length of the address range. On return *@len is updated with [whatever] and *@xlat is set to [whatever]." without compromising clarity by trying to squish it all into a one-sentence description of the param.
include/exec/memory.h void address_space_write(AddressSpace *as, hwaddr addr, · line 855
Quoted 3 lines
include/exec/memory.h +3
-0+MemoryRegionSection *address_space_translate (AddressSpace *as,hwaddr addr, +hwaddr *xlat,hwaddr *len, +bool is_write);
Your parameter names here don't match the ones in the function definition (len vs plen).
Quoted 28 lines
include/exec/memory.h +1
-0+/* address_space_valid: check for validity of an address space range * * Check whether access to the given address space range is permitted by translate-all.c +3
-4@@ -1355,15 +1355,15 @@ void tb_invalidate_phys_addr(hwaddr addr) {ram_addr_t ram_addr;MemoryRegionSection *section; +hwaddr l =1 ; - section =phys_page_find (address_space_memory.dispatch, - addr >>TARGET_PAGE_BITS ); + section =address_space_translate (&address_space_memory, addr, &addr, &l, false);if (!(memory_region_is_ram (section->mr) ||memory_region_is_romd (section->mr))) {return ; } ram_addr = (memory_region_get_ram_addr (section->mr) &TARGET_PAGE_MASK ) - +memory_region_section_addr (section, addr); + + addr;tb_invalidate_phys_page_range (ram_addr, ram_addr +1 ,0 ); }#endif /* TARGET_HAS_ICE && !defined(CONFIG_USER_ONLY) */ --
Quoted 4 lines
1.7.1
thanks
-- PMM
Re: [Qemu-devel] [PATCH 12/40] memory: add address_space_translate
Il 07/05/2013 20:08, Peter Maydell ha scritto:
Quoted 7 lines
Quoted 4 lines
exec.c static inline uint32_t ldl_phys_internal(hwaddr addr, · line 2221
Quoted 4 lines
exec.c +2
-1- section =phys_page_find (address_space_memory.dispatch, addr >>TARGET_PAGE_BITS ); + section =address_space_translate (&address_space_memory, addr, &addr, &l, + false);I find it a little confusing reusing 'addr' for the returned offset from
address_space_translate(); maybe using a different variable would be clearer? (don't feel very strongly about it)
True, on the other hand this matches usage before this patch.
- addr = memory_region_section_addr(section, addr);
Quoted 26 lines
Quoted 19 lines
Quoted 17 lines
exec.c +4
-2+if (l <4 ) { +return -1 ; + }if (!(memory_region_is_ram (section->mr) ||memory_region_is_romd (section->mr))) {/* I/O case */ - addr =memory_region_section_addr (section, addr); val =io_mem_read (section->mr, addr,4 );#if defined(TARGET_WORDS_BIGENDIAN )if (endian ==DEVICE_LITTLE_ENDIAN ) {@@ -2249,7 +2240,7 @@ static inline uint32_t ldl_phys_internal(hwaddr addr,/* RAM case */ ptr =qemu_get_ram_ptr ((memory_region_get_ram_addr (section->mr) &TARGET_PAGE_MASK ) - +memory_region_section_addr (section, addr)); + + addr);exec.c static inline uint32_t lduw_phys_internal(hwaddr addr, · line 2358
Quoted 6 lines
exec.c +5
-1- section =phys_page_find (address_space_memory.dispatch, addr >>TARGET_PAGE_BITS ); + section =address_space_translate (&address_space_memory, addr, &addr, &l, + false); +if (l <2 ) { +return -1 ; + }
Quoted 7 lines
This kind of "let's just return -1 if the read failed" kind of bothers me, because "memory access aborted" is really completely different from "memory access succeeded and returned -1". But there are a lot of APIs which we'd need to fix to properly communicate the problem back to the caller, so let it slide for now. (What used to happen in the old code in this case?)
It would call unassigned_mem_read and return 0. I'll change this to call unassigned_mem_read, which also requires fixing the "double use" of addr that you pointed out above.
Paolo
PATCH 13/40 memory: Introduce address_space_lookup_region
In reply to Paolo BonziniFrom: Jan Kiszka <[email protected]>
This introduces a wrapper for phys_page_find (before we complicate address_space_translate with IOMMU translation). This function will also encapsulate locking and reference counting when we introduce BQL-free dispatching.
--- exec.c | 8 +++++++- 1 files changed, 7 insertions(+), 1 deletions(-)
exec.c +7 -1
@@ -198,13 +198,19 @@ static MemoryRegionSection *phys_page_find(AddressSpaceDispatch *d, hwaddr indexreturn &phys_sections[lp.ptr]; } +static MemoryRegionSection *address_space_lookup_region (AddressSpace *as, +hwaddr addr) +{ +return phys_page_find (as->dispatch, addr >>TARGET_PAGE_BITS ); +} +MemoryRegionSection *address_space_translate (AddressSpace *as,hwaddr addr,hwaddr *xlat,hwaddr *plen,bool is_write) {MemoryRegionSection *section; - section =phys_page_find (as->dispatch, addr >>TARGET_PAGE_BITS ); + section =address_space_lookup_region (as, addr);/* Compute offset with MemoryRegionSection */ addr -= section->offset_within_address_space; *plen =MIN (section->size - addr, *plen);
Signature · 1 line
1.7.1
PATCH 14/40 memory: iommu support
In reply to Paolo BonziniFrom: Avi Kivity <[email protected]>
Add a new memory region type that translates addresses it is given, then forwards them to a target address space. This is similar to an alias, except that the mapping is more flexible than a linear translation and trucation, and also less efficient since the translation happens at runtime.
The implementation uses an AddressSpace mapping the target region to avoid hierarchical dispatch all the way to the resolved region; only iommu regions are looked up dynamically.
--- exec.c | 35 +++++++++++++++++++++++++++++------ include/exec/memory.h | 44 ++++++++++++++++++++++++++++++++++++++++++++ memory.c | 21 +++++++++++++++++++++ 3 files changed, 94 insertions(+), 6 deletions(-)
exec.c +29 -6
@@ -208,15 +208,38 @@ MemoryRegionSection *address_space_translate(AddressSpace *as, hwaddr addr, hwaddr *xlat, hwaddr *plen, bool is_write) { +IOMMUTLBEntry iotlb;MemoryRegionSection *section; +hwaddr len = *plen; - section =address_space_lookup_region (as, addr); -/* Compute offset with MemoryRegionSection */ - addr -= section->offset_within_address_space; - *plen =MIN (section->size - addr, *plen); +for (;;) { + section =address_space_lookup_region (as, addr); + +/* Compute offset with MemoryRegionSection */ + addr -= section->offset_within_address_space; + len =MIN (section->size - addr, len); + +/* Compute offset with MemoryRegion */ + addr += section->offset_within_region; + +if (!section->mr->iommu_ops) { +break ; + } + + iotlb = section->mr->iommu_ops->translate (section->mr, addr); + addr = ((iotlb.translated_addr & ~iotlb.addr_mask) + | (addr & iotlb.addr_mask)); + len =MIN (len, (addr | iotlb.addr_mask) - addr +1 ); +if (!iotlb.perm[is_write]) { + section = &phys_sections[phys_section_unassigned]; +break ; + } + + as = section->mr->iommu_target_as; + } -/* Compute offset with MemoryRegion */ - *xlat = addr + section->offset_within_region; + *plen = len; + *xlat = addr;return section; }
include/exec/memory.h +44 -0
74 lines not on this page · raw message
memory.c +21 -0
@@ -787,6 +787,7 @@ void memory_region_init(MemoryRegion *mr, uint64_t size) { mr->ops =NULL ; + mr->iommu_ops =NULL ; mr->parent =NULL ; mr->size =int128_make64 (size);if (size ==UINT64_MAX ) {
@@ -978,6 +979,21 @@ void memory_region_init_rom_device(MemoryRegion *mr, mr->ram_addr =qemu_ram_alloc (size, mr); } +void memory_region_init_iommu (MemoryRegion *mr, +MemoryRegionIOMMUOps *ops, +AddressSpace *target_as, +const char *name, +uint64_t size) +{ +memory_region_init (mr, name, size); + mr->ops =NULL ; + mr->iommu_ops = ops, + mr->opaque = mr; + mr->terminates = true;/* then re-forwards */ + mr->destructor = memory_region_destructor_none; + mr->iommu_target_as = target_as; +} +static uint64_t invalid_read (void *opaque,hwaddr addr,unsigned size) {
@@ -1052,6 +1068,11 @@ bool memory_region_is_rom(MemoryRegion *mr)return mr->ram && mr->readonly; } +bool memory_region_is_iommu (MemoryRegion *mr) +{ +return mr->iommu_ops; +} +void memory_region_set_log (MemoryRegion *mr,bool log,unsigned client) {uint8_t mask =1 << client;
Signature · 1 line
1.7.1
Re: [Qemu-devel] [PATCH 14/40] memory: iommu support
Quoted 47 lines · Paolo Bonzini
From: Avi Kivity <[email protected]>
Add a new memory region type that translates addresses it is given, then forwards them to a target address space. This is similar to an alias, except that the mapping is more flexible than a linear translation and trucation, and also less efficient since the translation happens at runtime.
The implementation uses an AddressSpace mapping the target region to avoid hierarchical dispatch all the way to the resolved region; only iommu regions are looked up dynamically.
Signed-off-by Avi Kivity <[email protected]>[Modified to put translation in address_space_translate - Paolo]Signed-off-by Paolo Bonzini <[email protected]>--- exec.c | 35 +++++++++++++++++++++++++++++------ include/exec/memory.h | 44 ++++++++++++++++++++++++++++++++++++++++++++ memory.c | 21 +++++++++++++++++++++ 3 files changed, 94 insertions(+), 6 deletions(-)exec.c +12
-4@@ -208,15 +208,38 @@ MemoryRegionSection *address_space_translate(AddressSpace *as, hwaddr addr, hwaddr *xlat, hwaddr *plen, bool is_write) { +IOMMUTLBEntry iotlb;MemoryRegionSection *section; +hwaddr len = *plen; - section =address_space_lookup_region (as, addr); -/* Compute offset with MemoryRegionSection */ - addr -= section->offset_within_address_space; - *plen =MIN (section->size - addr, *plen); +for (;;) { + section =address_space_lookup_region (as, addr); + +/* Compute offset with MemoryRegionSection */ + addr -= section->offset_within_address_space; + len =MIN (section->size - addr, len); + +/* Compute offset with MemoryRegion */ + addr += section->offset_within_region; +
exec.c MemoryRegionSection *address_space_translate(AddressSpace *as, hwaddr addr, · line 234
Quoted 10 lines
exec.c +10
-0+if (!section->mr->iommu_ops) { +break ; + } + + iotlb = section->mr->iommu_ops->translate (section->mr, addr); + addr = ((iotlb.translated_addr & ~iotlb.addr_mask) + | (addr & iotlb.addr_mask)); + len =MIN (len, (addr | iotlb.addr_mask) - addr +1 ); +if (!iotlb.perm[is_write]) { + section = &phys_sections[phys_section_unassigned];
Why is "unwritable" handled the same as "no memory"? They are likely to result in different fault information, surely?
Quoted 19 lines
exec.c +7
-2+break ; + } + + as = section->mr->iommu_target_as; + } -/* Compute offset with MemoryRegion */ - *xlat = addr + section->offset_within_region; + *plen = len; + *xlat = addr;return section; }include/exec/memory.h +0
-0@@ -112,12 +112,27 @@ struct MemoryRegionOps {const MemoryRegionMmio old_mmio;
include/exec/memory.h struct MemoryRegionOps { · line 122
Quoted 10 lines
include/exec/memory.h +8
-0}; +typedef struct IOMMUTLBEntry IOMMUTLBEntry ; +typedef struct MemoryRegionIOMMUOps MemoryRegionIOMMUOps ; + +struct IOMMUTLBEntry { +hwaddr translated_addr; +hwaddr addr_mask;/* 0xfff = 4k translation */ +bool perm[2 ];/* permissions, [0] for read, [1] for write */ +};
This could use a proper doc comment.
Quoted 26 lines
include/exec/memory.h +8
-0+ +struct MemoryRegionIOMMUOps { +/* Returns a TLB entry that contains a given address. */ +IOMMUTLBEntry (*translate)(MemoryRegion *iommu,hwaddr addr); +}; +typedef struct CoalescedMemoryRange CoalescedMemoryRange ;typedef struct MemoryRegionIoeventfd MemoryRegionIoeventfd ;struct MemoryRegion {/* All fields are private - violators will be prosecuted */ const MemoryRegionOps *ops; +const MemoryRegionIOMMUOps *iommu_ops;void *opaque;MemoryRegion *parent;Int128 size;@@ -144,6 +159,7 @@ struct MemoryRegion {uint8_t dirty_log_mask;unsigned ioeventfd_nb;MemoryRegionIoeventfd *ioeventfds; +struct AddressSpace *iommu_target_as; };struct MemoryRegionPortio {@@ -329,6 +345,25 @@ void memory_region_init_rom_device(MemoryRegion *mr, void memory_region_init_reservation(MemoryRegion *mr,
include/exec/memory.h void memory_region_init_rom_device(MemoryRegion *mr, · line 355
Quoted 10 lines
include/exec/memory.h +8
-0const char *name, uint64_t size); + +/** +* memory_region_init_iommu: Initialize a memory region that translates addresses +* +* An IOMMU region translates addresses and forwards accesses to a target memory region. +* +* @mr: the #MemoryRegion to be initialized +* @ops: a function that translates addresses into the @target region
This is an ops pointer, not a function pointer.
Quoted 15 lines
include/exec/memory.h +11
-0+* @target_as: the #AddressSpace that will be used to satisfy accesses to translated +* addresses +* @name: used for debugging; not visible to the user or ABI +* @size: size of the region. +*/ +void memory_region_init_iommu (MemoryRegion *mr, +MemoryRegionIOMMUOps *ops, +AddressSpace *target_as, +const char *name, +uint64_t size); +/** * memory_region_destroy: Destroy a memory region and reclaim all resources. *
include/exec/memory.h static inline bool memory_region_is_romd(MemoryRegion *mr) · line 408
Quoted 6 lines
include/exec/memory.h +3
-0@@ -368,6 +403,15 @@ static inline bool memory_region_is_romd(MemoryRegion *mr) }/** +* memory_region_is_iommu: check whether a memory region is an iommu +* +* Returns %true is a memory region is an iommu.
"%true if this memory region is an iommu"
(should we be consistently capitalising IOMMU?)
Quoted 56 lines
include/exec/memory.h +6
-0+* +* @mr: the memory region being queried +*/ +bool memory_region_is_iommu (MemoryRegion *mr); + +/** * memory_region_name: get a memory region's name * * Returns the string that was used to initialize the memory region. memory.c +21
-1@@ -787,6 +787,7 @@ void memory_region_init(MemoryRegion *mr, uint64_t size) { mr->ops =NULL ; + mr->iommu_ops =NULL ; mr->parent =NULL ; mr->size =int128_make64 (size);if (size ==UINT64_MAX ) {@@ -978,6 +979,21 @@ void memory_region_init_rom_device(MemoryRegion *mr, mr->ram_addr =qemu_ram_alloc (size, mr); } +void memory_region_init_iommu (MemoryRegion *mr, +MemoryRegionIOMMUOps *ops, +AddressSpace *target_as, +const char *name, +uint64_t size) +{ +memory_region_init (mr, name, size); + mr->ops =NULL ; + mr->iommu_ops = ops, + mr->opaque = mr; + mr->terminates = true;/* then re-forwards */ + mr->destructor = memory_region_destructor_none; + mr->iommu_target_as = target_as; +} +static uint64_t invalid_read (void *opaque,hwaddr addr,unsigned size) {@@ -1052,6 +1068,11 @@ bool memory_region_is_rom(MemoryRegion *mr)return mr->ram && mr->readonly; } +bool memory_region_is_iommu (MemoryRegion *mr) +{ +return mr->iommu_ops; +} +void memory_region_set_log (MemoryRegion *mr,bool log,unsigned client) {uint8_t mask =1 << client; --
Quoted 2 lines
1.7.1
thanks
-- PMM
PATCH 15/40 vfio: abort if an emulated iommu is used
In reply to Paolo BonziniFrom: Avi Kivity <[email protected]>
vfio doesn't support guest iommus yet, indicate it to the user by gently depositing a core on their disk.
--- hw/misc/vfio.c | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-)
hw/misc/vfio.c +2 -0
@@ -1939,6 +1939,8 @@ static void vfio_listener_region_add(MemoryListener *listener,void *vaddr;int ret; +assert (!memory_region_is_iommu (section->mr)); +if (vfio_listener_skipped_section (section)) { DPRINTF("SKIPPING region_add %"HWADDR_PRIx" - %"PRIx64"\n", section->offset_within_address_space,
Signature · 1 line
1.7.1
PATCH 16/40 spapr: convert TCE API to use an opaque type
In reply to Paolo BonziniThe TCE table is currently returned as a DMAContext, and non-type-safe APIs are called later passing back the DMAContext. Since we want to move away from DMAContext, use an opaque type instead, and add an accessor to retrieve the DMAContext from it.
--- hw/ppc/spapr_iommu.c | 54 +++++++++++++++++------------------------- hw/ppc/spapr_pci.c | 8 +++--- hw/ppc/spapr_vio.c | 13 +++++---- include/hw/pci-host/spapr.h | 2 +- include/hw/ppc/spapr.h | 12 +++++---- include/hw/ppc/spapr_vio.h | 1 + 6 files changed, 42 insertions(+), 48 deletions(-)
hw/ppc/spapr_iommu.c +22 -32
@@ -36,8 +36,6 @@ enum sPAPRTCEAccess { SPAPR_TCE_RW = 3, }; -typedef struct sPAPRTCETable sPAPRTCETable ; -struct sPAPRTCETable {DMAContext dma;uint32_t liobn;
@@ -122,7 +120,7 @@ static int spapr_tce_translate(DMAContext *dma,return 0 ; } -DMAContext *spapr_tce_new_dma_context (uint32_t liobn,size_t window_size) +sPAPRTCETable *spapr_tce_new_table (uint32_t liobn,size_t window_size) {sPAPRTCETable *tcet;
@@ -155,43 +153,40 @@ DMAContext *spapr_tce_new_dma_context(uint32_t liobn, size_t window_size) }#ifdef DEBUG_TCE -fprintf (stderr,"spapr_iommu: New TCE table, liobn=0x%x, context @ %p, " -"table @ %p, fd=%d\n" , liobn, &tcet->dma, tcet->table, tcet->fd); +fprintf (stderr,"spapr_iommu: New TCE table @ %p, liobn=0x%x, " +"table @ %p, fd=%d\n" , tcet, liobn, tcet->table, tcet->fd);#endif QLIST_INSERT_HEAD (&spapr_tce_tables, tcet, list); -return &tcet->dma; +return tcet; } -void spapr_tce_free (DMAContext *dma) +void spapr_tce_free (sPAPRTCETable *tcet) { +QLIST_REMOVE (tcet, list); -if (dma) { -sPAPRTCETable *tcet =DO_UPCAST (sPAPRTCETable, dma, dma); - -QLIST_REMOVE (tcet, list); - -if (!kvm_enabled () || - (kvmppc_remove_spapr_tce (tcet->table, tcet->fd, - tcet->window_size) !=0 )) { -g_free (tcet->table); - } - -g_free (tcet); +if (!kvm_enabled () || + (kvmppc_remove_spapr_tce (tcet->table, tcet->fd, + tcet->window_size) !=0 )) { +g_free (tcet->table); } + +g_free (tcet); } -void spapr_tce_set_bypass (DMAContext *dma,bool bypass) +DMAContext *spapr_tce_get_dma (sPAPRTCETable *tcet) { -sPAPRTCETable *tcet =DO_UPCAST (sPAPRTCETable, dma, dma); +return &tcet->dma; +} +void spapr_tce_set_bypass (sPAPRTCETable *tcet,bool bypass) +{ tcet->bypass = bypass; } -void spapr_tce_reset (DMAContext *dma) +void spapr_tce_reset (sPAPRTCETable *tcet) { -sPAPRTCETable *tcet =DO_UPCAST (sPAPRTCETable, dma, dma);size_t table_size = (tcet->window_size >>SPAPR_TCE_PAGE_SHIFT ) *sizeof (sPAPRTCE);
@@ -277,17 +272,12 @@ int spapr_dma_dt(void *fdt, int node_off, const char *propname, }int spapr_tcet_dma_dt (void *fdt,int node_off,const char *propname, -DMAContext *iommu) +sPAPRTCETable *tcet) { -if (!iommu) { +if (!tcet) {return 0 ; } -if (iommu->translate == spapr_tce_translate) { -sPAPRTCETable *tcet =DO_UPCAST (sPAPRTCETable, dma, iommu); -return spapr_dma_dt (fdt, node_off, propname, - tcet->liobn,0 , tcet->window_size); - } - -return -1 ; +return spapr_dma_dt (fdt, node_off, propname, + tcet->liobn,0 , tcet->window_size); }
hw/ppc/spapr_pci.c +4 -4
@@ -511,7 +511,7 @@ static DMAContext *spapr_pci_dma_context_fn(PCIBus *bus, void *opaque, {sPAPRPHBState *phb = opaque; -return phb->dma; +return spapr_tce_get_dma (phb->tcet); }static int spapr_phb_init (SysBusDevice *s)
@@ -646,8 +646,8 @@ static int spapr_phb_init(SysBusDevice *s) sphb->dma_window_start =0 ; sphb->dma_window_size =0x40000000 ; - sphb->dma =spapr_tce_new_dma_context (sphb->dma_liobn, sphb->dma_window_size); -if (!sphb->dma) { + sphb->tcet =spapr_tce_new_table (sphb->dma_liobn, sphb->dma_window_size); +if (!sphb->tcet) {fprintf (stderr,"Unable to create TCE table for %s\n" , sphb->dtbusname);return -1 ; }
@@ -676,7 +676,7 @@ static void spapr_phb_reset(DeviceState *qdev)sPAPRPHBState *sphb =SPAPR_PCI_HOST_BRIDGE (s);/* Reset the IOMMU state */ -spapr_tce_reset (sphb->dma); +spapr_tce_reset (sphb->tcet); }static Property spapr_phb_properties[] = {
hw/ppc/spapr_vio.c +7 -6
@@ -145,7 +145,7 @@ static int vio_make_devnode(VIOsPAPRDevice *dev, } } - ret =spapr_tcet_dma_dt (fdt, node_off,"ibm,my-dma-window" , dev->dma); + ret =spapr_tcet_dma_dt (fdt, node_off,"ibm,my-dma-window" , dev->tcet);if (ret <0 ) {return ret; }
@@ -319,8 +319,8 @@ int spapr_vio_send_crq(VIOsPAPRDevice *dev, uint8_t *crq)static void spapr_vio_quiesce_one (VIOsPAPRDevice *dev) { -if (dev->dma) { -spapr_tce_reset (dev->dma); +if (dev->tcet) { +spapr_tce_reset (dev->tcet); }free_crq (dev); }
@@ -345,12 +345,12 @@ static void rtas_set_tce_bypass(sPAPREnvironment *spapr, uint32_t token,return ; } -if (!dev->dma) { +if (!dev->tcet) {rtas_st (rets,0 ,-3 );return ; } -spapr_tce_set_bypass (dev->dma, !!enable); +spapr_tce_set_bypass (dev->tcet, !!enable);rtas_st (rets,0 ,0 ); }
@@ -457,7 +457,8 @@ static int spapr_vio_busdev_init(DeviceState *qdev)if (pc->rtce_window_size) {uint32_t liobn =SPAPR_VIO_BASE_LIOBN | dev->reg; - dev->dma =spapr_tce_new_dma_context (liobn, pc->rtce_window_size); + dev->tcet =spapr_tce_new_table (liobn, pc->rtce_window_size); + dev->dma =spapr_tce_get_dma (dev->tcet); }return pc->init (dev);
include/hw/pci-host/spapr.h +1 -1
@@ -49,7 +49,7 @@ typedef struct sPAPRPHBState {uint32_t dma_liobn;uint64_t dma_window_start;uint64_t dma_window_size; -DMAContext *dma; +sPAPRTCETable *tcet;struct {uint32_t irq;
include/hw/ppc/spapr.h +7 -5
@@ -342,17 +342,19 @@ typedef struct sPAPRTCE {#define RTAS_ERROR_LOG_MAX 2048 +typedef struct sPAPRTCETable sPAPRTCETable ;void spapr_iommu_init (void );void spapr_events_init (sPAPREnvironment *spapr);void spapr_events_fdt_skel (void *fdt,uint32_t epow_irq); -DMAContext *spapr_tce_new_dma_context (uint32_t liobn,size_t window_size); -void spapr_tce_free (DMAContext *dma); -void spapr_tce_reset (DMAContext *dma); -void spapr_tce_set_bypass (DMAContext *dma,bool bypass); +sPAPRTCETable *spapr_tce_new_table (uint32_t liobn,size_t window_size); +DMAContext *spapr_tce_get_dma (sPAPRTCETable *tcet); +void spapr_tce_free (sPAPRTCETable *tcet); +void spapr_tce_reset (sPAPRTCETable *tcet); +void spapr_tce_set_bypass (sPAPRTCETable *tcet,bool bypass);int spapr_dma_dt (void *fdt,int node_off,const char *propname,uint32_t liobn,uint64_t window,uint32_t size);int spapr_tcet_dma_dt (void *fdt,int node_off,const char *propname, -DMAContext *dma); +sPAPRTCETable *tcet);#endif /* !defined (__HW_SPAPR_H__) */
include/hw/ppc/spapr_vio.h +1 -0
@@ -63,6 +63,7 @@ struct VIOsPAPRDevice {uint32_t irq;target_ulong signal_state;VIOsPAPR_CRQ crq; +sPAPRTCETable *tcet;DMAContext *dma; };
Signature · 1 line
1.7.1
PATCH 17/40 spapr: make IOMMU translation go through IOMMUTLBEntry
In reply to Paolo BonziniThe next step is to introduce the translation code that will be used for IOMMU MemoryRegions, but still do the actual translation in a DMAContext.
--- hw/ppc/spapr_iommu.c | 57 ++++++++++++++++++++++++++++---------------------- 1 files changed, 32 insertions(+), 25 deletions(-)
hw/ppc/spapr_iommu.c +32 -25
@@ -68,15 +68,8 @@ static sPAPRTCETable *spapr_tce_find_by_liobn(uint32_t liobn)return NULL ; } -static int spapr_tce_translate (DMAContext *dma, -dma_addr_t addr, -hwaddr *paddr, -hwaddr *len, -DMADirection dir) +static IOMMUTLBEntry spapr_tce_translate_iommu (sPAPRTCETable *tcet,hwaddr addr) { -sPAPRTCETable *tcet =DO_UPCAST (sPAPRTCETable, dma, dma); -enum sPAPRTCEAccess access = (dir ==DMA_DIRECTION_FROM_DEVICE ) - ?SPAPR_TCE_WO :SPAPR_TCE_RO ;uint64_t tce;#ifdef DEBUG_TCE
@@ -85,9 +78,11 @@ static int spapr_tce_translate(DMAContext *dma,#endif if (tcet->bypass) { - *paddr = addr; - *len = (hwaddr)-1 ; -return 0 ; +return (IOMMUTLBEntry ) { + .translated_addr =0 , + .addr_mask = ~(hwaddr )0 , + .perm = { true, true }, + }; }/* Check if we are in bound */
@@ -95,28 +90,40 @@ static int spapr_tce_translate(DMAContext *dma,#ifdef DEBUG_TCE fprintf (stderr,"spapr_tce_translate out of bounds\n" );#endif -return -EFAULT ; +return (IOMMUTLBEntry ) { .perm = { false, false } }; } tce = tcet->table[addr >>SPAPR_TCE_PAGE_SHIFT ].tce; -/* Check TCE */ -if (!(tce & access)) { -return -EPERM ; - } +#ifdef DEBUG_TCE +fprintf (stderr," -> *paddr=0x%llx, *len=0x%llx\n" , + (tce & ~SPAPR_TCE_PAGE_MASK ),SPAPR_TCE_PAGE_MASK +1 ); +#endif -/* How much til end of page ? */ - *len = ((~addr) &SPAPR_TCE_PAGE_MASK ) +1 ; +return (IOMMUTLBEntry ) { + .translated_addr = tce & ~SPAPR_TCE_PAGE_MASK , + .addr_mask =SPAPR_TCE_PAGE_MASK , + .perm = { [0 ] = tce &SPAPR_TCE_RO , [1 ] = tce &SPAPR_TCE_WO }, + }; +} -/* Translate */ - *paddr = (tce & ~SPAPR_TCE_PAGE_MASK ) | - (addr &SPAPR_TCE_PAGE_MASK ); +static int spapr_tce_translate (DMAContext *dma, +dma_addr_t addr, +hwaddr *paddr, +hwaddr *len, +DMADirection dir) + { +sPAPRTCETable *tcet =DO_UPCAST (sPAPRTCETable, dma, dma); +bool is_write = (dir ==DMA_DIRECTION_FROM_DEVICE ); +IOMMUTLBEntry entry =spapr_tce_translate_iommu (tcet, addr); -#ifdef DEBUG_TCE -fprintf (stderr," -> *paddr=0x" TARGET_FMT_plx", *len=0x" - TARGET_FMT_plx"\n" , *paddr, *len); -#endif +if (!entry.perm[is_write]) { +return -EPERM ; + } +/* Translate */ + *paddr = entry.translated_addr | (addr & entry.addr_mask); + *len = (addr | entry.addr_mask) - addr +1 ;return 0 ; }
Signature · 1 line
1.7.1
PATCH 18/40 spapr: use memory core for iommu support
In reply to Paolo BonziniNow we can stop using a "translating" DMAContext, but we do not yet modify the sPAPRTCETable users to get an AddressSpace; they keep using the table via a DMAContext.
--- hw/ppc/spapr_iommu.c | 42 ++++++++++++++++++++---------------------- include/hw/ppc/spapr.h | 1 + 2 files changed, 21 insertions(+), 22 deletions(-)
hw/ppc/spapr_iommu.c +20 -22
@@ -37,12 +37,16 @@ enum sPAPRTCEAccess { };struct sPAPRTCETable { +/* temporary until everyone has its own AddressSpace */ DMAContext dma; +AddressSpace as; +uint32_t liobn;uint32_t window_size;sPAPRTCE *table;bool bypass;int fd; +MemoryRegion iommu;QLIST_ENTRY (sPAPRTCETable) list; };
@@ -68,8 +72,9 @@ static sPAPRTCETable *spapr_tce_find_by_liobn(uint32_t liobn)return NULL ; } -static IOMMUTLBEntry spapr_tce_translate_iommu (sPAPRTCETable *tcet,hwaddr addr) +static IOMMUTLBEntry spapr_tce_translate_iommu (MemoryRegion *iommu,hwaddr addr) { +sPAPRTCETable *tcet =container_of (iommu, sPAPRTCETable, iommu);uint64_t tce;#ifdef DEBUG_TCE
@@ -107,25 +112,9 @@ static IOMMUTLBEntry spapr_tce_translate_iommu(sPAPRTCETable *tcet, hwaddr addr) }; } -static int spapr_tce_translate (DMAContext *dma, -dma_addr_t addr, -hwaddr *paddr, -hwaddr *len, -DMADirection dir) - { -sPAPRTCETable *tcet =DO_UPCAST (sPAPRTCETable, dma, dma); -bool is_write = (dir ==DMA_DIRECTION_FROM_DEVICE ); -IOMMUTLBEntry entry =spapr_tce_translate_iommu (tcet, addr); - -if (!entry.perm[is_write]) { -return -EPERM ; - } - -/* Translate */ - *paddr = entry.translated_addr | (addr & entry.addr_mask); - *len = (addr | entry.addr_mask) - addr +1 ; -return 0 ; -} +static MemoryRegionIOMMUOps spapr_iommu_ops = { + .translate = spapr_tce_translate_iommu, +};sPAPRTCETable *spapr_tce_new_table (uint32_t liobn,size_t window_size) {
@@ -142,8 +131,6 @@ sPAPRTCETable *spapr_tce_new_table(uint32_t liobn, size_t window_size) } tcet =g_malloc0 (sizeof (*tcet)); -dma_context_init (&tcet->dma, &address_space_memory, spapr_tce_translate,NULL ,NULL ); - tcet->liobn = liobn; tcet->window_size = window_size;
@@ -164,6 +151,12 @@ sPAPRTCETable *spapr_tce_new_table(uint32_t liobn, size_t window_size) "table @ %p, fd=%d\n", tcet, liobn, tcet->table, tcet->fd);#endif +memory_region_init_iommu (&tcet->iommu, &spapr_iommu_ops, + &address_space_memory, +"iommu-spapr" ,INT64_MAX ); +address_space_init (&tcet->as, &tcet->iommu); +dma_context_init (&tcet->dma, &tcet->as,NULL ,NULL ,NULL ); +QLIST_INSERT_HEAD (&spapr_tce_tables, tcet, list);return tcet;
@@ -187,6 +180,11 @@ DMAContext *spapr_tce_get_dma(sPAPRTCETable *tcet)return &tcet->dma; } +MemoryRegion *spapr_tce_get_iommu (sPAPRTCETable *tcet) +{ +return &tcet->iommu; +} +void spapr_tce_set_bypass (sPAPRTCETable *tcet,bool bypass) { tcet->bypass = bypass;
include/hw/ppc/spapr.h +1 -0
@@ -349,6 +349,7 @@ void spapr_events_init(sPAPREnvironment *spapr);void spapr_events_fdt_skel (void *fdt,uint32_t epow_irq);sPAPRTCETable *spapr_tce_new_table (uint32_t liobn,size_t window_size);DMAContext *spapr_tce_get_dma (sPAPRTCETable *tcet); +MemoryRegion *spapr_tce_get_iommu (sPAPRTCETable *tcet);void spapr_tce_free (sPAPRTCETable *tcet);void spapr_tce_reset (sPAPRTCETable *tcet);void spapr_tce_set_bypass (sPAPRTCETable *tcet,bool bypass);
Signature · 1 line
1.7.1
PATCH 19/40 dma: eliminate old-style IOMMU support
In reply to Paolo BonziniThe translate function in the DMAContext is now always NULL. Remove every reference to it.
--- dma-helpers.c | 173 ++------------------------------------------------ exec.c | 3 +- hw/pci/pci.c | 3 +- hw/ppc/spapr_iommu.c | 2 +- include/sysemu/dma.h | 118 +++++++++------------------------- 5 files changed, 42 insertions(+), 257 deletions(-)
dma-helpers.c +7 -166
@@ -14,9 +14,12 @@/* #define DEBUG_IOMMU */ -static void do_dma_memory_set (AddressSpace *as, -dma_addr_t addr,uint8_t c,dma_addr_t len) +int dma_memory_set (DMAContext *dma,dma_addr_t addr,uint8_t c,dma_addr_t len) { +AddressSpace *as = dma->as; + +dma_barrier (dma,DMA_DIRECTION_FROM_DEVICE ); +#define FILLBUF_SIZE 512uint8_t fillbuf[FILLBUF_SIZE ];int l;
@@ -28,16 +31,6 @@ static void do_dma_memory_set(AddressSpace *as, len -= l; addr += l; } -} - -int dma_memory_set (DMAContext *dma,dma_addr_t addr,uint8_t c,dma_addr_t len) -{ -dma_barrier (dma,DMA_DIRECTION_FROM_DEVICE ); - -if (dma_has_iommu (dma)) { -return iommu_dma_memory_set (dma, addr, c, len); - } -do_dma_memory_set (dma->as, addr, c, len);return 0 ; }
@@ -278,162 +271,10 @@ void dma_acct_start(BlockDriverState *bs, BlockAcctCookie *cookie,bdrv_acct_start (bs, cookie, sg->size, type); } -bool iommu_dma_memory_valid (DMAContext *dma,dma_addr_t addr,dma_addr_t len, -DMADirection dir) -{ -hwaddr paddr, plen; - -#ifdef DEBUG_IOMMU -fprintf (stderr,"dma_memory_check context=%p addr=0x" DMA_ADDR_FMT -" len=0x" DMA_ADDR_FMT " dir=%d\n" , dma, addr, len, dir); -#endif - -while (len) { -if (dma->translate (dma, addr, &paddr, &plen, dir) !=0 ) { -return false; - } - -/* The translation might be valid for larger regions. */ -if (plen > len) { - plen = len; - } - -if (!address_space_valid (dma->as, paddr, len, - dir ==DMA_DIRECTION_FROM_DEVICE )) { -return false; - } - - len -= plen; - addr += plen; - } - -return true; -} - -int iommu_dma_memory_rw (DMAContext *dma,dma_addr_t addr, -void *buf,dma_addr_t len,DMADirection dir) +void dma_context_init (DMAContext *dma,AddressSpace *as) { -hwaddr paddr, plen; -int err; -#ifdef DEBUG_IOMMU -fprintf (stderr,"dma_memory_rw context=%p addr=0x" DMA_ADDR_FMT " len=0x" -DMA_ADDR_FMT " dir=%d\n" , dma, addr, len, dir); -#endif - -while (len) { - err = dma->translate (dma, addr, &paddr, &plen, dir); -if (err) { -/* -* In case of failure on reads from the guest, we clean the -* destination buffer so that a device that doesn't test -* for errors will not expose qemu internal memory. -*/ -memset (buf,0 , len); -return -1 ; - } - -/* The translation might be valid for larger regions. */ -if (plen > len) { - plen = len; - } - -address_space_rw (dma->as, paddr, buf, plen, dir ==DMA_DIRECTION_FROM_DEVICE ); - - len -= plen; - addr += plen; - buf += plen; - } - -return 0 ; -} - -int iommu_dma_memory_set (DMAContext *dma,dma_addr_t addr,uint8_t c, -dma_addr_t len) -{ -hwaddr paddr, plen; -int err; - -#ifdef DEBUG_IOMMU -fprintf (stderr,"dma_memory_set context=%p addr=0x" DMA_ADDR_FMT -" len=0x" DMA_ADDR_FMT "\n" , dma, addr, len); -#endif - -while (len) { - err = dma->translate (dma, addr, &paddr, &plen, -DMA_DIRECTION_FROM_DEVICE ); -if (err) { -return err; - } - -/* The translation might be valid for larger regions. */ -if (plen > len) { - plen = len; - } - -do_dma_memory_set (dma->as, paddr, c, plen); - - len -= plen; - addr += plen; - } - -return 0 ; -} - -void dma_context_init (DMAContext *dma,AddressSpace *as,DMATranslateFunc translate, -DMAMapFunc map,DMAUnmapFunc unmap) -{ -#ifdef DEBUG_IOMMU -fprintf (stderr,"dma_context_init(%p, %p, %p, %p)\n" , - dma, translate, map, unmap); +fprintf (stderr,"dma_context_init(%p -> %p)\n" , dma, as);#endif dma->as = as; - dma->translate = translate; - dma->map = map; - dma->unmap = unmap; -} - -void *iommu_dma_memory_map (DMAContext *dma,dma_addr_t addr,dma_addr_t *len, -DMADirection dir) -{ -int err; -hwaddr paddr, plen; -void *buf; - -if (dma->map) { -return dma->map (dma, addr, len, dir); - } - - plen = *len; - err = dma->translate (dma, addr, &paddr, &plen, dir); -if (err) { -return NULL ; - } - -/* -* If this is true, the virtual region is contiguous, -* but the translated physical region isn't. We just -* clamp *len, much like address_space_map() does. -*/ -if (plen < *len) { - *len = plen; - } - - buf =address_space_map (dma->as, paddr, &plen, dir ==DMA_DIRECTION_FROM_DEVICE ); - *len = plen; - -return buf; -} - -void iommu_dma_memory_unmap (DMAContext *dma,void *buffer,dma_addr_t len, -DMADirection dir,dma_addr_t access_len) -{ -if (dma->unmap) { - dma->unmap (dma, buffer, len, dir, access_len); -return ; - } - -address_space_unmap (dma->as, buffer, len, dir ==DMA_DIRECTION_FROM_DEVICE , - access_len); - }
exec.c +1 -2
@@ -1870,8 +1870,7 @@ static void memory_map_init(void)memory_listener_register (&io_memory_listener, &address_space_io);memory_listener_register (&tcg_memory_listener, &address_space_memory); -dma_context_init (&dma_context_memory, &address_space_memory, -NULL ,NULL ,NULL ); +dma_context_init (&dma_context_memory, &address_space_memory); }MemoryRegion *get_system_memory (void )
hw/pci/pci.c +2 -1
@@ -814,8 +814,9 @@ static PCIDevice *do_pci_register_device(PCIDevice *pci_dev, PCIBus *bus,memory_region_set_enabled (&pci_dev->bus_master_enable_region, false);address_space_init (&pci_dev->bus_master_as, &pci_dev->bus_master_enable_region); pci_dev->dma =g_new (DMAContext,1 ); -dma_context_init (pci_dev->dma, &pci_dev->bus_master_as,NULL ,NULL ,NULL ); +dma_context_init (pci_dev->dma, &pci_dev->bus_master_as); } + pci_dev->devfn = devfn;pstrcpy (pci_dev->name,sizeof (pci_dev->name), name); pci_dev->irq_state =0 ;
hw/ppc/spapr_iommu.c +1 -1
@@ -155,7 +155,7 @@ sPAPRTCETable *spapr_tce_new_table(uint32_t liobn, size_t window_size) &address_space_memory, "iommu-spapr", INT64_MAX);address_space_init (&tcet->as, &tcet->iommu); -dma_context_init (&tcet->dma, &tcet->as,NULL ,NULL ,NULL ); +dma_context_init (&tcet->dma, &tcet->as);QLIST_INSERT_HEAD (&spapr_tce_tables, tcet, list);
include/sysemu/dma.h +31 -87
180 lines not on this page · raw message
Signature · 1 line
1.7.1
Re: [Qemu-devel] [PATCH 19/40] dma: eliminate old-style IOMMU support
Quoted 12 lines · Paolo Bonzini
The translate function in the DMAContext is now always NULL. Remove every reference to it.
include/sysemu/dma.h +3
-3-static inline int dma_memory_rw_relaxed (DMAContext *dma,dma_addr_t addr, -void *buf,dma_addr_t len, -DMADirection dir) +static inline void dma_memory_rw_relaxed (DMAContext *dma,dma_addr_t addr, +void *buf,dma_addr_t len, +DMADirection dir)
include/sysemu/dma.h static inline void dma_barrier(DMAContext *dma, DMADirection dir) · line 99
Quoted 10 lines
include/sysemu/dma.h +1
-7{ -if (!dma_has_iommu (dma)) { -/* Fast-path for no IOMMU */ -address_space_rw (dma->as, addr, buf, len, dir ==DMA_DIRECTION_FROM_DEVICE ); -return 0 ; - }else { -return iommu_dma_memory_rw (dma, addr, buf, len, dir); - } +address_space_rw (dma->as, addr, buf, len, dir ==DMA_DIRECTION_FROM_DEVICE ); }
Should we really be removing the success-or-failure indication from the read/write functions?
thanks
-- PMM
Re: [Qemu-devel] [PATCH 19/40] dma: eliminate old-style IOMMU support
Il 07/05/2013 20:20, Peter Maydell ha scritto:
Quoted 18 lines · Paolo Bonzini
Quoted 17 lines
The translate function in the DMAContext is now always NULL. Remove every reference to it.
include/sysemu/dma.h +3
-7-static inline int dma_memory_rw_relaxed (DMAContext *dma,dma_addr_t addr, -void *buf,dma_addr_t len, -DMADirection dir) +static inline void dma_memory_rw_relaxed (DMAContext *dma,dma_addr_t addr, +void *buf,dma_addr_t len, +DMADirection dir) { -if (!dma_has_iommu (dma)) { -/* Fast-path for no IOMMU */ -address_space_rw (dma->as, addr, buf, len, dir ==DMA_DIRECTION_FROM_DEVICE ); -return 0 ;
Quoted 8 lines
include/sysemu/dma.h static inline void dma_barrier(DMAContext *dma, DMADirection dir) · line 99
Quoted 5 lines
include/sysemu/dma.h +1
-3- }else { -return iommu_dma_memory_rw (dma, addr, buf, len, dir); - } +address_space_rw (dma->as, addr, buf, len, dir ==DMA_DIRECTION_FROM_DEVICE ); }Should we really be removing the success-or-failure indication from the read/write functions?
For now yes. In the future I want to add it back to address_space_rw.
Paolo
PATCH 20/40 pci: use memory core for iommu support
In reply to Paolo BonziniFrom: Avi Kivity <[email protected]>
Use the new iommu support in the memory core for iommu support. The only user, spapr, is also converted, but it still provides a DMAContext interface until the non-PCI bits switch to AddressSpace.
--- hw/pci/pci.c | 53 ++++++++++++++++++++++++++-------------------- hw/ppc/spapr_pci.c | 12 +++++++--- include/hw/pci/pci.h | 7 ++++- include/hw/pci/pci_bus.h | 5 ++- 4 files changed, 46 insertions(+), 31 deletions(-)
hw/pci/pci.c +30 -23
@@ -279,6 +279,16 @@ int pci_find_domain(const PCIBus *bus)return -1 ; } +static MemoryRegion *pci_default_iommu (PCIBus *bus,void *opaque,int devfn) +{ +/* FIXME: inherit memory region from bus creator */ +return get_system_memory (); +} + +static void pci_default_iommu_dtor (MemoryRegion *mr) +{ +} + static void pci_bus_init(PCIBus *bus, DeviceState *parent, const char *name, MemoryRegion *address_space_mem,
@@ -289,6 +299,7 @@ static void pci_bus_init(PCIBus *bus, DeviceState *parent, bus->devfn_min = devfn_min; bus->address_space_mem = address_space_mem; bus->address_space_io = address_space_io; +pci_setup_iommu (bus, pci_default_iommu,NULL ,NULL );/* host bridge */ QLIST_INIT (&bus->child);
@@ -801,21 +812,15 @@ static PCIDevice *do_pci_register_device(PCIDevice *pci_dev, PCIBus *bus, PCI_SLOT(devfn), PCI_FUNC(devfn), name, bus->devices[devfn]->name);return NULL ; } + pci_dev->bus = bus; -if (bus->dma_context_fn) { - pci_dev->dma = bus->dma_context_fn (bus, bus->dma_context_opaque, devfn); - }else { -/* FIXME: Make dma_context_fn use MemoryRegions instead, so this path is -* taken unconditionally */ -/* FIXME: inherit memory region from bus creator */ -memory_region_init_alias (&pci_dev->bus_master_enable_region,"bus master" , -get_system_memory (),0 , -memory_region_size (get_system_memory ())); -memory_region_set_enabled (&pci_dev->bus_master_enable_region, false); -address_space_init (&pci_dev->bus_master_as, &pci_dev->bus_master_enable_region); - pci_dev->dma =g_new (DMAContext,1 ); -dma_context_init (pci_dev->dma, &pci_dev->bus_master_as); - } + pci_dev->iommu = bus->iommu_fn (bus, bus->iommu_opaque, devfn); +memory_region_init_alias (&pci_dev->bus_master_enable_region,"bus master" , + pci_dev->iommu,0 ,memory_region_size (pci_dev->iommu)); +memory_region_set_enabled (&pci_dev->bus_master_enable_region, false); +address_space_init (&pci_dev->bus_master_as, &pci_dev->bus_master_enable_region); + pci_dev->dma =g_new (DMAContext,1 ); +dma_context_init (pci_dev->dma, &pci_dev->bus_master_as); pci_dev->devfn = devfn;pstrcpy (pci_dev->name,sizeof (pci_dev->name), name);
@@ -870,12 +875,12 @@ static void do_pci_unregister_device(PCIDevice *pci_dev) pci_dev->bus->devices[pci_dev->devfn] =NULL ;pci_config_free (pci_dev); -if (!pci_dev->bus->dma_context_fn) { -address_space_destroy (&pci_dev->bus_master_as); -memory_region_destroy (&pci_dev->bus_master_enable_region); -g_free (pci_dev->dma); - pci_dev->dma =NULL ; - } +address_space_destroy (&pci_dev->bus_master_as); +memory_region_del_subregion (&pci_dev->bus_master_enable_region, pci_dev->iommu); + pci_dev->bus->iommu_dtor_fn (pci_dev->iommu); +memory_region_destroy (&pci_dev->bus_master_enable_region); +g_free (pci_dev->dma); + pci_dev->dma =NULL ; }static void pci_unregister_io_regions (PCIDevice *pci_dev)
@@ -2234,10 +2239,12 @@ static void pci_device_class_init(ObjectClass *klass, void *data) k->props = pci_props; } -void pci_setup_iommu (PCIBus *bus,PCIDMAContextFunc fn,void *opaque) +void pci_setup_iommu (PCIBus *bus,PCIIOMMUFunc fn,PCIIOMMUDestructorFunc dtor, +void *opaque) { - bus->dma_context_fn = fn; - bus->dma_context_opaque = opaque; + bus->iommu_fn = fn; + bus->iommu_dtor_fn = dtor ? dtor : pci_default_iommu_dtor; + bus->iommu_opaque = opaque; }static const TypeInfo pci_device_type_info = {
hw/ppc/spapr_pci.c +3 -4
@@ -506,12 +506,11 @@ static const MemoryRegionOps spapr_msi_ops = {/* * PHB PCI device */ -static DMAContext *spapr_pci_dma_context_fn (PCIBus *bus,void *opaque, -int devfn) +static MemoryRegion *spapr_pci_dma_iommu (PCIBus *bus,void *opaque,int devfn) {sPAPRPHBState *phb = opaque; -return spapr_tce_get_dma (phb->tcet); +return spapr_tce_get_iommu (phb->tcet); }static int spapr_phb_init (SysBusDevice *s)
@@ -651,7 +655,7 @@ static int spapr_phb_init(SysBusDevice *s)fprintf (stderr,"Unable to create TCE table for %s\n" , sphb->dtbusname);return -1 ; } -pci_setup_iommu (bus, spapr_pci_dma_context_fn, sphb); +pci_setup_iommu (bus, spapr_pci_dma_iommu,NULL , sphb);QLIST_INSERT_HEAD (&spapr->phbs, sphb, list);
include/hw/pci/pci.h +5 -2
@@ -242,6 +242,7 @@ struct PCIDevice {PCIIORegion io_regions[PCI_NUM_REGIONS ];AddressSpace bus_master_as;MemoryRegion bus_master_enable_region; +MemoryRegion *iommu;DMAContext *dma;/* do not access the following fields */
@@ -401,9 +402,11 @@ int pci_read_devaddr(Monitor *mon, const char *addr, int *domp, int *busp,void pci_device_deassert_intx (PCIDevice *dev); -typedef DMAContext *(*PCIDMAContextFunc )(PCIBus *,void *,int ); +typedef MemoryRegion *(*PCIIOMMUFunc )(PCIBus *,void *,int ); +typedef void (*PCIIOMMUDestructorFunc )(MemoryRegion *mr); -void pci_setup_iommu (PCIBus *bus,PCIDMAContextFunc fn,void *opaque); +void pci_setup_iommu (PCIBus *bus,PCIIOMMUFunc fn,PCIIOMMUDestructorFunc dtor, +void *opaque);static inline void pci_set_byte (uint8_t *config,uint8_t val)
include/hw/pci/pci_bus.h +3 -2
@@ -10,8 +10,9 @@struct PCIBus {BusState qbus; -PCIDMAContextFunc dma_context_fn; -void *dma_context_opaque; +PCIIOMMUFunc iommu_fn; +PCIIOMMUDestructorFunc iommu_dtor_fn; +void *iommu_opaque;uint8_t devfn_min;pci_set_irq_fn set_irq;pci_map_irq_fn map_irq;
Signature · 1 line
1.7.1
Re: [Qemu-devel] [PATCH 20/40] pci: use memory core for iommu support
Quoted 21 lines · Paolo Bonzini
From: Avi Kivity <[email protected]>
Use the new iommu support in the memory core for iommu support. The only user, spapr, is also converted, but it still provides a DMAContext interface until the non-PCI bits switch to AddressSpace.
Cc Michael S. Tsirkin <[email protected]>Signed-off-by Avi Kivity <[email protected]>Signed-off-by Paolo Bonzini <[email protected]>--- hw/pci/pci.c | 53 ++++++++++++++++++++++++++-------------------- hw/ppc/spapr_pci.c | 12 +++++++--- include/hw/pci/pci.h | 7 ++++- include/hw/pci/pci_bus.h | 5 ++- 4 files changed, 46 insertions(+), 31 deletions(-)hw/pci/pci.c +0
-0
hw/pci/pci.c int pci_find_domain(const PCIBus *bus) · line 286
Quoted 8 lines
hw/pci/pci.c +5
-0@@ -279,6 +279,16 @@ int pci_find_domain(const PCIBus *bus)return -1 ; } +static MemoryRegion *pci_default_iommu (PCIBus *bus,void *opaque,int devfn) +{ +/* FIXME: inherit memory region from bus creator */ +return get_system_memory (); +}
This seems a bit misnamed since it doesn't actually need to return an iommu MemoryRegion (and in fact in most cases it won't). What it's actually returning is "this is the memory region representing the view for bus master DMA devices to DMA into", I think.
Also, technically speaking get_system_memory() is never the right answer, though in practice it's good enough for our purposes. (returning get_system_memory() would allow a bus master DMA device to access back into the PCI bus by DMAing to the address in system memory where the PCI host controller is mapped, which I'm guessing is not possible on real hardware.)
As you kind of imply with the FIXME comment here, I think what we probably actually eventually want is for pci_bus_init() and friends to have a parameter for the DMA MemoryRegion (but what this patch does is fine for now).
Once these patches go in I could use this to do the versatile_pci SMAP registers, though that's more for completeness than anything else.
Quoted 122 lines
hw/pci/pci.c +25
-23+ +static void pci_default_iommu_dtor (MemoryRegion *mr) +{ +} + static void pci_bus_init(PCIBus *bus, DeviceState *parent, const char *name, MemoryRegion *address_space_mem,@@ -289,6 +299,7 @@ static void pci_bus_init(PCIBus *bus, DeviceState *parent, bus->devfn_min = devfn_min; bus->address_space_mem = address_space_mem; bus->address_space_io = address_space_io; +pci_setup_iommu (bus, pci_default_iommu,NULL ,NULL );/* host bridge */ QLIST_INIT (&bus->child);@@ -801,21 +812,15 @@ static PCIDevice *do_pci_register_device(PCIDevice *pci_dev, PCIBus *bus, PCI_SLOT(devfn), PCI_FUNC(devfn), name, bus->devices[devfn]->name);return NULL ; } + pci_dev->bus = bus; -if (bus->dma_context_fn) { - pci_dev->dma = bus->dma_context_fn (bus, bus->dma_context_opaque, devfn); - }else { -/* FIXME: Make dma_context_fn use MemoryRegions instead, so this path is -* taken unconditionally */ -/* FIXME: inherit memory region from bus creator */ -memory_region_init_alias (&pci_dev->bus_master_enable_region,"bus master" , -get_system_memory (),0 , -memory_region_size (get_system_memory ())); -memory_region_set_enabled (&pci_dev->bus_master_enable_region, false); -address_space_init (&pci_dev->bus_master_as, &pci_dev->bus_master_enable_region); - pci_dev->dma =g_new (DMAContext,1 ); -dma_context_init (pci_dev->dma, &pci_dev->bus_master_as); - } + pci_dev->iommu = bus->iommu_fn (bus, bus->iommu_opaque, devfn); +memory_region_init_alias (&pci_dev->bus_master_enable_region,"bus master" , + pci_dev->iommu,0 ,memory_region_size (pci_dev->iommu)); +memory_region_set_enabled (&pci_dev->bus_master_enable_region, false); +address_space_init (&pci_dev->bus_master_as, &pci_dev->bus_master_enable_region); + pci_dev->dma =g_new (DMAContext,1 ); +dma_context_init (pci_dev->dma, &pci_dev->bus_master_as); pci_dev->devfn = devfn;pstrcpy (pci_dev->name,sizeof (pci_dev->name), name);@@ -870,12 +875,12 @@ static void do_pci_unregister_device(PCIDevice *pci_dev) pci_dev->bus->devices[pci_dev->devfn] =NULL ;pci_config_free (pci_dev); -if (!pci_dev->bus->dma_context_fn) { -address_space_destroy (&pci_dev->bus_master_as); -memory_region_destroy (&pci_dev->bus_master_enable_region); -g_free (pci_dev->dma); - pci_dev->dma =NULL ; - } +address_space_destroy (&pci_dev->bus_master_as); +memory_region_del_subregion (&pci_dev->bus_master_enable_region, pci_dev->iommu); + pci_dev->bus->iommu_dtor_fn (pci_dev->iommu); +memory_region_destroy (&pci_dev->bus_master_enable_region); +g_free (pci_dev->dma); + pci_dev->dma =NULL ; }static void pci_unregister_io_regions (PCIDevice *pci_dev)@@ -2234,10 +2239,12 @@ static void pci_device_class_init(ObjectClass *klass, void *data) k->props = pci_props; } -void pci_setup_iommu (PCIBus *bus,PCIDMAContextFunc fn,void *opaque) +void pci_setup_iommu (PCIBus *bus,PCIIOMMUFunc fn,PCIIOMMUDestructorFunc dtor, +void *opaque) { - bus->dma_context_fn = fn; - bus->dma_context_opaque = opaque; + bus->iommu_fn = fn; + bus->iommu_dtor_fn = dtor ? dtor : pci_default_iommu_dtor; + bus->iommu_opaque = opaque; }static const TypeInfo pci_device_type_info = {hw/ppc/spapr_pci.c +3
-4@@ -506,12 +506,11 @@ static const MemoryRegionOps spapr_msi_ops = {/* * PHB PCI device */ -static DMAContext *spapr_pci_dma_context_fn (PCIBus *bus,void *opaque, -int devfn) +static MemoryRegion *spapr_pci_dma_iommu (PCIBus *bus,void *opaque,int devfn) {sPAPRPHBState *phb = opaque; -return spapr_tce_get_dma (phb->tcet); +return spapr_tce_get_iommu (phb->tcet); }static int spapr_phb_init (SysBusDevice *s)@@ -651,7 +655,7 @@ static int spapr_phb_init(SysBusDevice *s)fprintf (stderr,"Unable to create TCE table for %s\n" , sphb->dtbusname);return -1 ; } -pci_setup_iommu (bus, spapr_pci_dma_context_fn, sphb); +pci_setup_iommu (bus, spapr_pci_dma_iommu,NULL , sphb);QLIST_INSERT_HEAD (&spapr->phbs, sphb, list);include/hw/pci/pci.h +1
-0@@ -242,6 +242,7 @@ struct PCIDevice {PCIIORegion io_regions[PCI_NUM_REGIONS ];AddressSpace bus_master_as;MemoryRegion bus_master_enable_region; +MemoryRegion *iommu;DMAContext *dma;/* do not access the following fields */
include/hw/pci/pci.h int pci_read_devaddr(Monitor *mon, const char *addr, int *domp, int *busp, · line 409
Quoted 10 lines
include/hw/pci/pci.h +4
-2@@ -401,9 +402,11 @@ int pci_read_devaddr(Monitor *mon, const char *addr, int *domp, int *busp,void pci_device_deassert_intx (PCIDevice *dev); -typedef DMAContext *(*PCIDMAContextFunc )(PCIBus *,void *,int ); +typedef MemoryRegion *(*PCIIOMMUFunc )(PCIBus *,void *,int ); +typedef void (*PCIIOMMUDestructorFunc )(MemoryRegion *mr); -void pci_setup_iommu (PCIBus *bus,PCIDMAContextFunc fn,void *opaque); +void pci_setup_iommu (PCIBus *bus,PCIIOMMUFunc fn,PCIIOMMUDestructorFunc dtor, +void *opaque);
I'm not entirely convinced by this API. Some doc comments might help...
thanks
-- PMM
Re: [Qemu-devel] [PATCH 20/40] pci: use memory core for iommu support
Quoted 25 lines · Peter Maydell
Quoted 24 lines · Paolo Bonzini
From: Avi Kivity <[email protected]>
Use the new iommu support in the memory core for iommu support. The only user, spapr, is also converted, but it still provides a DMAContext interface until the non-PCI bits switch to AddressSpace.
Cc Michael S. Tsirkin <[email protected]>Signed-off-by Avi Kivity <[email protected]>Signed-off-by Paolo Bonzini <[email protected]>--- hw/pci/pci.c | 53 ++++++++++++++++++++++++++-------------------- hw/ppc/spapr_pci.c | 12 +++++++--- include/hw/pci/pci.h | 7 ++++- include/hw/pci/pci_bus.h | 5 ++- 4 files changed, 46 insertions(+), 31 deletions(-)hw/pci/pci.c +0
-0@@ -279,6 +279,16 @@ int pci_find_domain(const PCIBus *bus)return -1 ; }
Quoted 12 lines
hw/pci/pci.c int pci_find_domain(const PCIBus *bus) · line 286
Quoted 5 lines
hw/pci/pci.c +5
-0+static MemoryRegion *pci_default_iommu (PCIBus *bus,void *opaque,int devfn) +{ +/* FIXME: inherit memory region from bus creator */ +return get_system_memory (); +}This seems a bit misnamed since it doesn't actually need to return an iommu MemoryRegion (and in fact in most cases it won't). What it's actually returning is "this is the memory region representing the view for bus master DMA devices to DMA into", I think.
Maybe pci_fake_iommu ?
Quoted 8 lines
Also, technically speaking
get_system_memory()is never the right answer, though in practice it's good enough for our purposes. (returningget_system_memory()would allow a bus master DMA device to access back into the PCI bus by DMAing to the address in system memory where the PCI host controller is mapped, which I'm guessing is not possible on real hardware.)
I think although it is rare case, but PCI device can fetch another's data on its pci-ram, expecially pci-spec does not forbid it.
Regards,
Pingfan
Quoted 147 lines
As you kind of imply with the FIXME comment here, I think what we probably actually eventually want is for
pci_bus_init()and friends to have a parameter for the DMA MemoryRegion (but what this patch does is fine for now).Once these patches go in I could use this to do the
versatile_pciSMAP registers, though that's more for completeness than anything else.Quoted 122 lines
hw/pci/pci.c +25
-23+ +static void pci_default_iommu_dtor (MemoryRegion *mr) +{ +} + static void pci_bus_init(PCIBus *bus, DeviceState *parent, const char *name, MemoryRegion *address_space_mem,@@ -289,6 +299,7 @@ static void pci_bus_init(PCIBus *bus, DeviceState *parent, bus->devfn_min = devfn_min; bus->address_space_mem = address_space_mem; bus->address_space_io = address_space_io; +pci_setup_iommu (bus, pci_default_iommu,NULL ,NULL );/* host bridge */ QLIST_INIT (&bus->child);@@ -801,21 +812,15 @@ static PCIDevice *do_pci_register_device(PCIDevice *pci_dev, PCIBus *bus, PCI_SLOT(devfn), PCI_FUNC(devfn), name, bus->devices[devfn]->name);return NULL ; } + pci_dev->bus = bus; -if (bus->dma_context_fn) { - pci_dev->dma = bus->dma_context_fn (bus, bus->dma_context_opaque, devfn); - }else { -/* FIXME: Make dma_context_fn use MemoryRegions instead, so this path is -* taken unconditionally */ -/* FIXME: inherit memory region from bus creator */ -memory_region_init_alias (&pci_dev->bus_master_enable_region,"bus master" , -get_system_memory (),0 , -memory_region_size (get_system_memory ())); -memory_region_set_enabled (&pci_dev->bus_master_enable_region, false); -address_space_init (&pci_dev->bus_master_as, &pci_dev->bus_master_enable_region); - pci_dev->dma =g_new (DMAContext,1 ); -dma_context_init (pci_dev->dma, &pci_dev->bus_master_as); - } + pci_dev->iommu = bus->iommu_fn (bus, bus->iommu_opaque, devfn); +memory_region_init_alias (&pci_dev->bus_master_enable_region,"bus master" , + pci_dev->iommu,0 ,memory_region_size (pci_dev->iommu)); +memory_region_set_enabled (&pci_dev->bus_master_enable_region, false); +address_space_init (&pci_dev->bus_master_as, &pci_dev->bus_master_enable_region); + pci_dev->dma =g_new (DMAContext,1 ); +dma_context_init (pci_dev->dma, &pci_dev->bus_master_as); pci_dev->devfn = devfn;pstrcpy (pci_dev->name,sizeof (pci_dev->name), name);@@ -870,12 +875,12 @@ static void do_pci_unregister_device(PCIDevice *pci_dev) pci_dev->bus->devices[pci_dev->devfn] =NULL ;pci_config_free (pci_dev); -if (!pci_dev->bus->dma_context_fn) { -address_space_destroy (&pci_dev->bus_master_as); -memory_region_destroy (&pci_dev->bus_master_enable_region); -g_free (pci_dev->dma); - pci_dev->dma =NULL ; - } +address_space_destroy (&pci_dev->bus_master_as); +memory_region_del_subregion (&pci_dev->bus_master_enable_region, pci_dev->iommu); + pci_dev->bus->iommu_dtor_fn (pci_dev->iommu); +memory_region_destroy (&pci_dev->bus_master_enable_region); +g_free (pci_dev->dma); + pci_dev->dma =NULL ; }static void pci_unregister_io_regions (PCIDevice *pci_dev)@@ -2234,10 +2239,12 @@ static void pci_device_class_init(ObjectClass *klass, void *data) k->props = pci_props; } -void pci_setup_iommu (PCIBus *bus,PCIDMAContextFunc fn,void *opaque) +void pci_setup_iommu (PCIBus *bus,PCIIOMMUFunc fn,PCIIOMMUDestructorFunc dtor, +void *opaque) { - bus->dma_context_fn = fn; - bus->dma_context_opaque = opaque; + bus->iommu_fn = fn; + bus->iommu_dtor_fn = dtor ? dtor : pci_default_iommu_dtor; + bus->iommu_opaque = opaque; }static const TypeInfo pci_device_type_info = {hw/ppc/spapr_pci.c +3
-4@@ -506,12 +506,11 @@ static const MemoryRegionOps spapr_msi_ops = {/* * PHB PCI device */ -static DMAContext *spapr_pci_dma_context_fn (PCIBus *bus,void *opaque, -int devfn) +static MemoryRegion *spapr_pci_dma_iommu (PCIBus *bus,void *opaque,int devfn) {sPAPRPHBState *phb = opaque; -return spapr_tce_get_dma (phb->tcet); +return spapr_tce_get_iommu (phb->tcet); }static int spapr_phb_init (SysBusDevice *s)@@ -651,7 +655,7 @@ static int spapr_phb_init(SysBusDevice *s)fprintf (stderr,"Unable to create TCE table for %s\n" , sphb->dtbusname);return -1 ; } -pci_setup_iommu (bus, spapr_pci_dma_context_fn, sphb); +pci_setup_iommu (bus, spapr_pci_dma_iommu,NULL , sphb);QLIST_INSERT_HEAD (&spapr->phbs, sphb, list);include/hw/pci/pci.h +1
-0@@ -242,6 +242,7 @@ struct PCIDevice {PCIIORegion io_regions[PCI_NUM_REGIONS ];AddressSpace bus_master_as;MemoryRegion bus_master_enable_region; +MemoryRegion *iommu;DMAContext *dma;/* do not access the following fields */ include/hw/pci/pci.h int pci_read_devaddr(Monitor *mon, const char *addr, int *domp, int *busp, · line 409
Quoted 10 lines
include/hw/pci/pci.h +4
-2@@ -401,9 +402,11 @@ int pci_read_devaddr(Monitor *mon, const char *addr, int *domp, int *busp,void pci_device_deassert_intx (PCIDevice *dev); -typedef DMAContext *(*PCIDMAContextFunc )(PCIBus *,void *,int ); +typedef MemoryRegion *(*PCIIOMMUFunc )(PCIBus *,void *,int ); +typedef void (*PCIIOMMUDestructorFunc )(MemoryRegion *mr); -void pci_setup_iommu (PCIBus *bus,PCIDMAContextFunc fn,void *opaque); +void pci_setup_iommu (PCIBus *bus,PCIIOMMUFunc fn,PCIIOMMUDestructorFunc dtor, +void *opaque);I'm not entirely convinced by this API. Some doc comments might help...
thanks
-- PMM
Re: [Qemu-devel] [PATCH 20/40] pci: use memory core for iommu support
Quoted 11 lines · liu ping fan
Quoted 8 lines · Peter Maydell
Also, technically speaking
get_system_memory()is never the right answer, though in practice it's good enough for our purposes. (returningget_system_memory()would allow a bus master DMA device to access back into the PCI bus by DMAing to the address in system memory where the PCI host controller is mapped, which I'm guessing is not possible on real hardware.)I think although it is rare case, but PCI device can fetch another's data on its pci-ram, expecially pci-spec does not forbid it.
Yes, certainly, but this happens because the other pci device is present in the PCI address space, not because PCI device A talks to the host controller's CPU-facing interface and accesses PCI device B that way...
-- PMM
Re: [Qemu-devel] [PATCH 20/40] pci: use memory core for iommu support
In reply to Paolo BonziniQuoted 56 lines · Paolo Bonzini
From: Avi Kivity <[email protected]>
Use the new iommu support in the memory core for iommu support. The only user, spapr, is also converted, but it still provides a DMAContext interface until the non-PCI bits switch to AddressSpace.
Cc Michael S. Tsirkin <[email protected]>Signed-off-by Avi Kivity <[email protected]>Signed-off-by Paolo Bonzini <[email protected]>--- hw/pci/pci.c | 53 ++++++++++++++++++++++++++-------------------- hw/ppc/spapr_pci.c | 12 +++++++--- include/hw/pci/pci.h | 7 ++++- include/hw/pci/pci_bus.h | 5 ++- 4 files changed, 46 insertions(+), 31 deletions(-)hw/pci/pci.c +12
-5@@ -279,6 +279,16 @@ int pci_find_domain(const PCIBus *bus)return -1 ; } +static MemoryRegion *pci_default_iommu (PCIBus *bus,void *opaque,int devfn) +{ +/* FIXME: inherit memory region from bus creator */ +return get_system_memory (); +} + +static void pci_default_iommu_dtor (MemoryRegion *mr) +{ +} + static void pci_bus_init(PCIBus *bus, DeviceState *parent, const char *name, MemoryRegion *address_space_mem,@@ -289,6 +299,7 @@ static void pci_bus_init(PCIBus *bus, DeviceState *parent, bus->devfn_min = devfn_min; bus->address_space_mem = address_space_mem; bus->address_space_io = address_space_io; +pci_setup_iommu (bus, pci_default_iommu,NULL ,NULL );/* host bridge */ QLIST_INIT (&bus->child);@@ -801,21 +812,15 @@ static PCIDevice *do_pci_register_device(PCIDevice *pci_dev, PCIBus *bus, PCI_SLOT(devfn), PCI_FUNC(devfn), name, bus->devices[devfn]->name);return NULL ; } + pci_dev->bus = bus; -if (bus->dma_context_fn) { - pci_dev->dma = bus->dma_context_fn (bus, bus->dma_context_opaque, devfn); - }else { -/* FIXME: Make dma_context_fn use MemoryRegions instead, so this path is -* taken unconditionally */
hw/pci/pci.c static PCIDevice *do_pci_register_device(PCIDevice *pci_dev, PCIBus *bus, · line 817
Quoted 10 lines
hw/pci/pci.c +1
-9-/* FIXME: inherit memory region from bus creator */ -memory_region_init_alias (&pci_dev->bus_master_enable_region,"bus master" , -get_system_memory (),0 , -memory_region_size (get_system_memory ())); -memory_region_set_enabled (&pci_dev->bus_master_enable_region, false); -address_space_init (&pci_dev->bus_master_as, &pci_dev->bus_master_enable_region); - pci_dev->dma =g_new (DMAContext,1 ); -dma_context_init (pci_dev->dma, &pci_dev->bus_master_as); - } + pci_dev->iommu = bus->iommu_fn (bus, bus->iommu_opaque, devfn);
I tried to create a PCI bus on the default PHB and put e1000 onto it like this:
-device pci-bridge,chassis_nr=1 -device e1000,bus=pci.0,addr=1.0
And I got a crash as bus->iommu_fn==NULL.
Is it because we need this bridge to use parent's iommu_fn/iommu or pci-bridge is not supposed to be created this way or ...?
I made small patch just to get one step further, the info qree is below. Now our system firmware fails (enters infinite loop) but this is a different story.
hw/pci/pci.c +0 -0
@@ -811,7 +811,11 @@ static PCIDevice *do_pci_register_device(PCIDevice
*pci_dev, PCIBus *bus,
}
pci_dev->bus = bus;- pci_dev->iommu = bus->iommu_fn(bus, bus->iommu_opaque, devfn);
+ if (bus->iommu_fn)
+ pci_dev->iommu = bus->iommu_fn(bus, bus->iommu_opaque, devfn);
+ else
+ pci_dev->iommu = bus->parent_dev->iommu;
+
memory_region_init_alias(&pci_dev->bus_master_enable_region, "bus master",
pci_dev->iommu, 0,
memory_region_size(pci_dev->iommu));
memory_region_set_enabled(&pci_dev->bus_master_enable_region, false);dev: spapr-pci-host-bridge, id ""
index = 0
buid = 0x800000020000000
liobn = 0x80000000
mem_win_addr = 0x100a0000000
mem_win_size = 0x20000000
io_win_addr = 0x10080000000
io_win_size = 0x10000
msi_win_addr = 0x10090000000
irq 0
bus: pci
type PCI
dev: pci-bridge, id ""
chassis_nr = 1
msi = on
addr = 00.0
romfile = <null>
rombar = 1
multifunction = off
command_serr_enable = on
class PCI bridge, addr 00:00.0, pci id 1b36:0001 (sub 0000:0000)
bar 0: mem at 0xffffffffffffffff [0xfe]
bus: pci.0
type PCI
dev: e1000, id ""
mac = 52:54:00:12:34:57
vlan = <null>
netdev = <null>
bootindex = -1
autonegotiation = on
addr = 01.0
romfile = "efi-e1000.rom"
rombar = 1
multifunction = off
command_serr_enable = on
class Ethernet controller, addr 00:01.0, pci id 8086:100e (sub
1af4:1100)
bar 0: mem at 0xffffffffffffffff [0x1fffe]
bar 1: i/o at 0xffffffffffffffff [0x3e]Quoted 115 lines
hw/pci/pci.c +17
-9+memory_region_init_alias (&pci_dev->bus_master_enable_region,"bus master" , + pci_dev->iommu,0 ,memory_region_size (pci_dev->iommu)); +memory_region_set_enabled (&pci_dev->bus_master_enable_region, false); +address_space_init (&pci_dev->bus_master_as, &pci_dev->bus_master_enable_region); + pci_dev->dma =g_new (DMAContext,1 ); +dma_context_init (pci_dev->dma, &pci_dev->bus_master_as); pci_dev->devfn = devfn;pstrcpy (pci_dev->name,sizeof (pci_dev->name), name);@@ -870,12 +875,12 @@ static void do_pci_unregister_device(PCIDevice *pci_dev) pci_dev->bus->devices[pci_dev->devfn] =NULL ;pci_config_free (pci_dev); -if (!pci_dev->bus->dma_context_fn) { -address_space_destroy (&pci_dev->bus_master_as); -memory_region_destroy (&pci_dev->bus_master_enable_region); -g_free (pci_dev->dma); - pci_dev->dma =NULL ; - } +address_space_destroy (&pci_dev->bus_master_as); +memory_region_del_subregion (&pci_dev->bus_master_enable_region, pci_dev->iommu); + pci_dev->bus->iommu_dtor_fn (pci_dev->iommu); +memory_region_destroy (&pci_dev->bus_master_enable_region); +g_free (pci_dev->dma); + pci_dev->dma =NULL ; }static void pci_unregister_io_regions (PCIDevice *pci_dev)@@ -2234,10 +2239,12 @@ static void pci_device_class_init(ObjectClass *klass, void *data) k->props = pci_props; } -void pci_setup_iommu (PCIBus *bus,PCIDMAContextFunc fn,void *opaque) +void pci_setup_iommu (PCIBus *bus,PCIIOMMUFunc fn,PCIIOMMUDestructorFunc dtor, +void *opaque) { - bus->dma_context_fn = fn; - bus->dma_context_opaque = opaque; + bus->iommu_fn = fn; + bus->iommu_dtor_fn = dtor ? dtor : pci_default_iommu_dtor; + bus->iommu_opaque = opaque; }static const TypeInfo pci_device_type_info = {hw/ppc/spapr_pci.c +3
-4@@ -506,12 +506,11 @@ static const MemoryRegionOps spapr_msi_ops = {/* * PHB PCI device */ -static DMAContext *spapr_pci_dma_context_fn (PCIBus *bus,void *opaque, -int devfn) +static MemoryRegion *spapr_pci_dma_iommu (PCIBus *bus,void *opaque,int devfn) {sPAPRPHBState *phb = opaque; -return spapr_tce_get_dma (phb->tcet); +return spapr_tce_get_iommu (phb->tcet); }static int spapr_phb_init (SysBusDevice *s)@@ -651,7 +655,7 @@ static int spapr_phb_init(SysBusDevice *s)fprintf (stderr,"Unable to create TCE table for %s\n" , sphb->dtbusname);return -1 ; } -pci_setup_iommu (bus, spapr_pci_dma_context_fn, sphb); +pci_setup_iommu (bus, spapr_pci_dma_iommu,NULL , sphb);QLIST_INSERT_HEAD (&spapr->phbs, sphb, list);include/hw/pci/pci.h +5
-2@@ -242,6 +242,7 @@ struct PCIDevice {PCIIORegion io_regions[PCI_NUM_REGIONS ];AddressSpace bus_master_as;MemoryRegion bus_master_enable_region; +MemoryRegion *iommu;DMAContext *dma;/* do not access the following fields */ @@ -401,9 +402,11 @@ int pci_read_devaddr(Monitor *mon, const char *addr, int *domp, int *busp,void pci_device_deassert_intx (PCIDevice *dev); -typedef DMAContext *(*PCIDMAContextFunc )(PCIBus *,void *,int ); +typedef MemoryRegion *(*PCIIOMMUFunc )(PCIBus *,void *,int ); +typedef void (*PCIIOMMUDestructorFunc )(MemoryRegion *mr); -void pci_setup_iommu (PCIBus *bus,PCIDMAContextFunc fn,void *opaque); +void pci_setup_iommu (PCIBus *bus,PCIIOMMUFunc fn,PCIIOMMUDestructorFunc dtor, +void *opaque);static inline void pci_set_byte (uint8_t *config,uint8_t val)include/hw/pci/pci_bus.h +3
-2@@ -10,8 +10,9 @@struct PCIBus {BusState qbus; -PCIDMAContextFunc dma_context_fn; -void *dma_context_opaque; +PCIIOMMUFunc iommu_fn; +PCIIOMMUDestructorFunc iommu_dtor_fn; +void *iommu_opaque;uint8_t devfn_min;pci_set_irq_fn set_irq;pci_map_irq_fn map_irq;
Signature · 1 line
Alexey
Re: [Qemu-devel] [PATCH 20/40] pci: use memory core for iommu support
Il 10/05/2013 15:07, Alexey Kardashevskiy ha scritto:
Quoted 11 lines
Review static void pci_device_class_init(ObjectClass *klass, void *data) · line 2245
Quoted 11 lines
(quoted diff) +3
-3@@ -2234,10 +2239,12 @@ static void pci_device_class_init(ObjectClass *klass, void *data) k->props = pci_props; } -void pci_setup_iommu(PCIBus *bus, PCIDMAContextFunc fn, void *opaque) +void pci_setup_iommu(PCIBus *bus, PCIIOMMUFunc fn, PCIIOMMUDestructorFunc dtor, + void *opaque) { - bus->dma_context_fn = fn; - bus->dma_context_opaque = opaque; + bus->iommu_fn = fn;
Or perhaps just use pci_default_iommu if fn == NULL.
Paolo
PATCH 21/40 spapr_vio: take care of creating our own AddressSpace/DMAContext
In reply to Paolo BonziniFetch the root region from the sPAPRTCETable, and use it to build an AddressSpace and DMAContext.
Now, everywhere we have a DMAContext we also have access to the corresponding AddressSpace (either because we create it just before the DMAContext, or because dma_context_memory's AddressSpace is trivially address_space_memory).
--- hw/ppc/spapr_iommu.c | 11 ----------- hw/ppc/spapr_vio.c | 3 ++- include/hw/ppc/spapr.h | 1 - include/hw/ppc/spapr_vio.h | 36 +++++++++++++++++++++++------------- 4 files changed, 25 insertions(+), 26 deletions(-)
hw/ppc/spapr_iommu.c +0 -11
@@ -37,10 +37,6 @@ enum sPAPRTCEAccess { };struct sPAPRTCETable { -/* temporary until everyone has its own AddressSpace */ -DMAContext dma; -AddressSpace as; -uint32_t liobn;uint32_t window_size;sPAPRTCE *table;
@@ -154,8 +150,6 @@ sPAPRTCETable *spapr_tce_new_table(uint32_t liobn, size_t window_size)memory_region_init_iommu (&tcet->iommu, &spapr_iommu_ops, &address_space_memory,"iommu-spapr" ,INT64_MAX ); -address_space_init (&tcet->as, &tcet->iommu); -dma_context_init (&tcet->dma, &tcet->as);QLIST_INSERT_HEAD (&spapr_tce_tables, tcet, list);
@@ -175,11 +169,6 @@ void spapr_tce_free(sPAPRTCETable *tcet)g_free (tcet); } -DMAContext *spapr_tce_get_dma (sPAPRTCETable *tcet) -{ -return &tcet->dma; -} -MemoryRegion *spapr_tce_get_iommu (sPAPRTCETable *tcet) {return &tcet->iommu;
hw/ppc/spapr_vio.c +2 -1
@@ -458,7 +458,8 @@ static int spapr_vio_busdev_init(DeviceState *qdev)if (pc->rtce_window_size) {uint32_t liobn =SPAPR_VIO_BASE_LIOBN | dev->reg; dev->tcet =spapr_tce_new_table (liobn, pc->rtce_window_size); - dev->dma =spapr_tce_get_dma (dev->tcet); +address_space_init (&dev->as,spapr_tce_get_iommu (dev->tcet)); +dma_context_init (&dev->dma, &dev->as); }return pc->init (dev);
include/hw/ppc/spapr.h +0 -1
@@ -348,7 +348,6 @@ void spapr_iommu_init(void);void spapr_events_init (sPAPREnvironment *spapr);void spapr_events_fdt_skel (void *fdt,uint32_t epow_irq);sPAPRTCETable *spapr_tce_new_table (uint32_t liobn,size_t window_size); -DMAContext *spapr_tce_get_dma (sPAPRTCETable *tcet);MemoryRegion *spapr_tce_get_iommu (sPAPRTCETable *tcet);void spapr_tce_free (sPAPRTCETable *tcet);void spapr_tce_reset (sPAPRTCETable *tcet);
include/hw/ppc/spapr_vio.h +23 -13
@@ -63,8 +63,9 @@ struct VIOsPAPRDevice {uint32_t irq;target_ulong signal_state;VIOsPAPR_CRQ crq; +AddressSpace as; +DMAContext dma;sPAPRTCETable *tcet; -DMAContext *dma; };#define DEFINE_SPAPR_PROPERTIES (type, field) \
@@ -92,35 +93,44 @@ static inline qemu_irq spapr_vio_qirq(VIOsPAPRDevice *dev)static inline bool spapr_vio_dma_valid (VIOsPAPRDevice *dev,uint64_t taddr,uint32_t size,DMADirection dir) { -return dma_memory_valid (dev->dma, taddr, size, dir); +return dma_memory_valid (&dev->dma, taddr, size, dir); }static inline int spapr_vio_dma_read (VIOsPAPRDevice *dev,uint64_t taddr,void *buf,uint32_t size) { -return (dma_memory_read (dev->dma, taddr, buf, size) !=0 ) ? -H_DEST_PARM :H_SUCCESS ; +if (!dma_memory_valid (&dev->dma, taddr, size,DMA_DIRECTION_TO_DEVICE )) { +return H_DEST_PARM ; + } +dma_memory_read (&dev->dma, taddr, buf, size); +return H_SUCCESS ; }static inline int spapr_vio_dma_write (VIOsPAPRDevice *dev,uint64_t taddr,const void *buf,uint32_t size) { -return (dma_memory_write (dev->dma, taddr, buf, size) !=0 ) ? -H_DEST_PARM :H_SUCCESS ; +if (!dma_memory_valid (&dev->dma, taddr, size,DMA_DIRECTION_FROM_DEVICE )) { +return H_DEST_PARM ; + } +dma_memory_write (&dev->dma, taddr, buf, size); +return H_SUCCESS ; }static inline int spapr_vio_dma_set (VIOsPAPRDevice *dev,uint64_t taddr,uint8_t c,uint32_t size) { -return (dma_memory_set (dev->dma, taddr, c, size) !=0 ) ? -H_DEST_PARM :H_SUCCESS ; +if (!dma_memory_valid (&dev->dma, taddr, size,DMA_DIRECTION_FROM_DEVICE )) { +return H_DEST_PARM ; + } +dma_memory_set (&dev->dma, taddr, c, size); +return H_SUCCESS ; } -#define vio_stb (_dev, _addr, _val) (stb_dma((_dev)->dma, (_addr), (_val))) -#define vio_sth (_dev, _addr, _val) (stw_be_dma((_dev)->dma, (_addr), (_val))) -#define vio_stl (_dev, _addr, _val) (stl_be_dma((_dev)->dma, (_addr), (_val))) -#define vio_stq (_dev, _addr, _val) (stq_be_dma((_dev)->dma, (_addr), (_val))) -#define vio_ldq (_dev, _addr) (ldq_be_dma((_dev)->dma, (_addr))) +#define vio_stb (_dev, _addr, _val) (stb_dma(&(_dev)->dma, (_addr), (_val))) +#define vio_sth (_dev, _addr, _val) (stw_be_dma(&(_dev)->dma, (_addr), (_val))) +#define vio_stl (_dev, _addr, _val) (stl_be_dma(&(_dev)->dma, (_addr), (_val))) +#define vio_stq (_dev, _addr, _val) (stq_be_dma(&(_dev)->dma, (_addr), (_val))) +#define vio_ldq (_dev, _addr) (ldq_be_dma(&(_dev)->dma, (_addr)))int spapr_vio_send_crq (VIOsPAPRDevice *dev,uint8_t *crq);
Signature · 1 line
1.7.1
PATCH 22/40 dma: eliminate DMAContext
In reply to Paolo BonziniThe DMAContext is a simple pointer to an AddressSpace that is now always already available. Make everyone hold the address space directly, and clean up the DMA API to use the AddressSpace directly.
--- dma-helpers.c | 24 ++++---------- exec.c | 3 -- hw/dma/pl330.c | 8 ++-- hw/ide/ahci.c | 18 +++++----- hw/ide/ahci.h | 4 +- hw/ide/ich.c | 2 +- hw/ide/macio.c | 4 +- hw/pci/pci.c | 4 -- hw/ppc/spapr_vio.c | 1 - hw/scsi/megasas.c | 4 +- hw/scsi/virtio-scsi.c | 2 +- hw/scsi/vmw_pvscsi.c | 2 +- hw/sd/sdhci.c | 22 ++++++------ hw/usb/hcd-ehci-pci.c | 4 +- hw/usb/hcd-ehci-sysbus.c | 2 +- hw/usb/hcd-ehci.c | 12 +++--- hw/usb/hcd-ehci.h | 2 +- hw/usb/hcd-ohci.c | 30 +++++++++--------- hw/usb/libhw.c | 4 +- include/hw/pci/pci.h | 17 +++++----- include/hw/ppc/spapr_vio.h | 25 +++++++------- include/sysemu/dma.h | 75 +++++++++++++++++++------------------------- 22 files changed, 119 insertions(+), 150 deletions(-)
dma-helpers.c +7 -17
@@ -14,11 +14,9 @@/* #define DEBUG_IOMMU */ -int dma_memory_set (DMAContext *dma,dma_addr_t addr,uint8_t c,dma_addr_t len) +int dma_memory_set (AddressSpace *as,dma_addr_t addr,uint8_t c,dma_addr_t len) { -AddressSpace *as = dma->as; - -dma_barrier (dma,DMA_DIRECTION_FROM_DEVICE ); +dma_barrier (as,DMA_DIRECTION_FROM_DEVICE );#define FILLBUF_SIZE 512uint8_t fillbuf[FILLBUF_SIZE ];
@@ -35,13 +33,13 @@ int dma_memory_set(DMAContext *dma, dma_addr_t addr, uint8_t c, dma_addr_t len)return 0 ; } -void qemu_sglist_init (QEMUSGList *qsg,int alloc_hint,DMAContext *dma) +void qemu_sglist_init (QEMUSGList *qsg,int alloc_hint,AddressSpace *as) { qsg->sg =g_malloc (alloc_hint *sizeof (ScatterGatherEntry)); qsg->nsg =0 ; qsg->nalloc = alloc_hint; qsg->size =0 ; - qsg->dma = dma; + qsg->as = as; }void qemu_sglist_add (QEMUSGList *qsg,dma_addr_t base,dma_addr_t len)
@@ -101,7 +99,7 @@ static void dma_bdrv_unmap(DMAAIOCB *dbs)int i;for (i =0 ; i < dbs->iov.niov; ++i) { -dma_memory_unmap (dbs->sg->dma, dbs->iov.iov[i].iov_base, +dma_memory_unmap (dbs->sg->as, dbs->iov.iov[i].iov_base, dbs->iov.iov[i].iov_len, dbs->dir, dbs->iov.iov[i].iov_len); }
@@ -149,7 +147,7 @@ static void dma_bdrv_cb(void *opaque, int ret)while (dbs->sg_cur_index < dbs->sg->nsg) { cur_addr = dbs->sg->sg[dbs->sg_cur_index].base + dbs->sg_cur_byte; cur_len = dbs->sg->sg[dbs->sg_cur_index].len - dbs->sg_cur_byte; - mem =dma_memory_map (dbs->sg->dma, cur_addr, &cur_len, dbs->dir); + mem =dma_memory_map (dbs->sg->as, cur_addr, &cur_len, dbs->dir);if (!mem)break ;qemu_iovec_add (&dbs->iov, mem, cur_len);
@@ -246,7 +244,7 @@ static uint64_t dma_buf_rw(uint8_t *ptr, int32_t len, QEMUSGList *sg,while (len >0 ) {ScatterGatherEntry entry = sg->sg[sg_cur_index++];int32_t xfer =MIN (len, entry.len); -dma_memory_rw (sg->dma, entry.base, ptr, xfer, dir); +dma_memory_rw (sg->as, entry.base, ptr, xfer, dir); ptr += xfer; len -= xfer; resid -= xfer;
@@ -270,11 +268,3 @@ void dma_acct_start(BlockDriverState *bs, BlockAcctCookie *cookie, {bdrv_acct_start (bs, cookie, sg->size, type); } - -void dma_context_init (DMAContext *dma,AddressSpace *as) -{ -#ifdef DEBUG_IOMMU -fprintf (stderr,"dma_context_init(%p -> %p)\n" , dma, as); -#endif - dma->as = as; -}
exec.c +0 -3
@@ -64,7 +64,6 @@ static MemoryRegion *system_io;AddressSpace address_space_io;AddressSpace address_space_memory; -DMAContext dma_context_memory;MemoryRegion io_mem_ram, io_mem_rom, io_mem_unassigned, io_mem_notdirty;static MemoryRegion io_mem_subpage_ram;
@@ -1869,8 +1868,6 @@ static void memory_map_init(void)memory_listener_register (&core_memory_listener, &address_space_memory);memory_listener_register (&io_memory_listener, &address_space_io);memory_listener_register (&tcg_memory_listener, &address_space_memory); - -dma_context_init (&dma_context_memory, &address_space_memory); }MemoryRegion *get_system_memory (void )
hw/dma/pl330.c +4 -4
@@ -1074,7 +1074,7 @@ static inline const PL330InsnDesc *pl330_fetch_insn(PL330Chan *ch)uint8_t opcode;int i; -dma_memory_read (&dma_context_memory, ch->pc, &opcode,1 ); +dma_memory_read (&address_space_memory, ch->pc, &opcode,1 );for (i =0 ; insn_desc[i].size; i++) {if ((opcode & insn_desc[i].opmask) == insn_desc[i].opcode) {return &insn_desc[i];
@@ -1088,7 +1088,7 @@ static inline void pl330_exec_insn(PL330Chan *ch, const PL330InsnDesc *insn)uint8_t buf[PL330_INSN_MAXSIZE ];assert (insn->size <=PL330_INSN_MAXSIZE ); -dma_memory_read (&dma_context_memory, ch->pc, buf, insn->size); +dma_memory_read (&address_space_memory, ch->pc, buf, insn->size); insn->exec (ch, buf[0 ], &buf[1 ], insn->size -1 ); }
@@ -1153,7 +1153,7 @@ static int pl330_exec_cycle(PL330Chan *channel)if (q !=NULL && q->len <=pl330_fifo_num_free (&s->fifo)) {int len = q->len - (q->addr & (q->len -1 )); -dma_memory_read (&dma_context_memory, q->addr, buf, len); +dma_memory_read (&address_space_memory, q->addr, buf, len);if (PL330_ERR_DEBUG >1 ) {DB_PRINT ("PL330 read from memory @%08x (size = %08x):\n" , q->addr, len);
@@ -1185,7 +1185,7 @@ static int pl330_exec_cycle(PL330Chan *channel) fifo_res =pl330_fifo_get (&s->fifo, buf, len, q->tag); }if (fifo_res ==PL330_FIFO_OK || q->z) { -dma_memory_write (&dma_context_memory, q->addr, buf, len); +dma_memory_write (&address_space_memory, q->addr, buf, len);if (PL330_ERR_DEBUG >1 ) {DB_PRINT ("PL330 read from memory @%08x (size = %08x):\n" , q->addr, len);
hw/ide/ahci.c +9 -9
@@ -597,7 +597,7 @@ static void ahci_write_fis_d2h(AHCIDevice *ad, uint8_t *cmd_fis)if (!cmd_fis) {/* map cmd_fis */ uint64_t tbl_addr =le64_to_cpu (ad->cur_cmd->tbl_addr); - cmd_fis =dma_memory_map (ad->hba->dma, tbl_addr, &cmd_len, + cmd_fis =dma_memory_map (ad->hba->as, tbl_addr, &cmd_len,DMA_DIRECTION_TO_DEVICE ); cmd_mapped =1 ; }
@@ -630,7 +630,7 @@ static void ahci_write_fis_d2h(AHCIDevice *ad, uint8_t *cmd_fis)ahci_trigger_irq (ad->hba, ad,PORT_IRQ_D2H_REG_FIS );if (cmd_mapped) { -dma_memory_unmap (ad->hba->dma, cmd_fis, cmd_len, +dma_memory_unmap (ad->hba->as, cmd_fis, cmd_len,DMA_DIRECTION_TO_DEVICE , cmd_len); } }
@@ -657,7 +657,7 @@ static int ahci_populate_sglist(AHCIDevice *ad, QEMUSGList *sglist, int offset) }/* map PRDT */ -if (!(prdt =dma_memory_map (ad->hba->dma, prdt_addr, &prdt_len, +if (!(prdt =dma_memory_map (ad->hba->as, prdt_addr, &prdt_len,DMA_DIRECTION_TO_DEVICE ))){DPRINTF (ad->port_no,"map failed\n" );return -1 ;
@@ -691,7 +691,7 @@ static int ahci_populate_sglist(AHCIDevice *ad, QEMUSGList *sglist, int offset) goto out; } -qemu_sglist_init (sglist, (sglist_alloc_hint - off_idx), ad->hba->dma); +qemu_sglist_init (sglist, (sglist_alloc_hint - off_idx), ad->hba->as);qemu_sglist_add (sglist,le64_to_cpu (tbl[off_idx].addr + off_pos),le32_to_cpu (tbl[off_idx].flags_size) +1 - off_pos);
@@ -703,7 +703,7 @@ static int ahci_populate_sglist(AHCIDevice *ad, QEMUSGList *sglist, int offset) } out: -dma_memory_unmap (ad->hba->dma, prdt, prdt_len, +dma_memory_unmap (ad->hba->as, prdt, prdt_len,DMA_DIRECTION_TO_DEVICE , prdt_len);return r; }
@@ -836,7 +836,7 @@ static int handle_cmd(AHCIState *s, int port, int slot) tbl_addr =le64_to_cpu (cmd->tbl_addr); cmd_len =0x80 ; - cmd_fis =dma_memory_map (s->dma, tbl_addr, &cmd_len, + cmd_fis =dma_memory_map (s->as, tbl_addr, &cmd_len,DMA_DIRECTION_FROM_DEVICE );if (!cmd_fis) {
@@ -963,7 +963,7 @@ static int handle_cmd(AHCIState *s, int port, int slot) } out: -dma_memory_unmap (s->dma, cmd_fis, cmd_len,DMA_DIRECTION_FROM_DEVICE , +dma_memory_unmap (s->as, cmd_fis, cmd_len,DMA_DIRECTION_FROM_DEVICE , cmd_len);if (s->dev[port].port.ifs[0 ].status & (BUSY_STAT |DRQ_STAT )) {
@@ -1145,12 +1145,12 @@ static const IDEDMAOps ahci_dma_ops = { .reset = ahci_dma_reset, }; -void ahci_init (AHCIState *s,DeviceState *qdev,DMAContext *dma,int ports) +void ahci_init (AHCIState *s,DeviceState *qdev,AddressSpace *as,int ports) {qemu_irq *irqs;int i; - s->dma = dma; + s->as = as; s->ports = ports; s->dev =g_malloc0 (sizeof (AHCIDevice) * ports);ahci_reg_init (s);
hw/ide/ahci.h +2 -2
@@ -297,7 +297,7 @@ typedef struct AHCIState {uint32_t idp_index;/* Current IDP index */ int32_t ports;qemu_irq irq; -DMAContext *dma; +AddressSpace *as; } AHCIState; typedef struct AHCIPCIState {
@@ -338,7 +338,7 @@ typedef struct NCQFrame {uint8_t reserved10; }QEMU_PACKED NCQFrame; -void ahci_init (AHCIState *s,DeviceState *qdev,DMAContext *dma,int ports); +void ahci_init (AHCIState *s,DeviceState *qdev,AddressSpace *as,int ports);void ahci_uninit (AHCIState *s);void ahci_reset (AHCIState *s);
hw/ide/ich.c +1 -1
@@ -104,7 +104,7 @@ static int pci_ich9_ahci_init(PCIDevice *dev)uint8_t *sata_cap; d = DO_UPCAST(struct AHCIPCIState, card, dev); -ahci_init (&d->ahci, &dev->qdev,pci_dma_context (dev),6 ); +ahci_init (&d->ahci, &dev->qdev,pci_get_address_space (dev),6 );pci_config_set_prog_interface (d->card.config,AHCI_PROGMODE_MAJOR_REV_1 );
hw/ide/macio.c +2 -2
@@ -71,7 +71,7 @@ static void pmac_ide_atapi_transfer_cb(void *opaque, int ret) s->io_buffer_size = io->len;qemu_sglist_init (&s->sg, io->len /MACIO_PAGE_SIZE +1 , - &dma_context_memory); + &address_space_memory);qemu_sglist_add (&s->sg, io->addr, io->len); io->addr += io->len; io->len =0 ;
@@ -128,7 +128,7 @@ static void pmac_ide_transfer_cb(void *opaque, int ret) s->io_buffer_size = io->len;qemu_sglist_init (&s->sg, io->len /MACIO_PAGE_SIZE +1 , - &dma_context_memory); + &address_space_memory);qemu_sglist_add (&s->sg, io->addr, io->len); io->addr += io->len; io->len =0 ;
hw/pci/pci.c +0 -4
@@ -819,8 +819,6 @@ static PCIDevice *do_pci_register_device(PCIDevice *pci_dev, PCIBus *bus, pci_dev->iommu, 0, memory_region_size(pci_dev->iommu));memory_region_set_enabled (&pci_dev->bus_master_enable_region, false);address_space_init (&pci_dev->bus_master_as, &pci_dev->bus_master_enable_region); - pci_dev->dma =g_new (DMAContext,1 ); -dma_context_init (pci_dev->dma, &pci_dev->bus_master_as); pci_dev->devfn = devfn;pstrcpy (pci_dev->name,sizeof (pci_dev->name), name);
@@ -879,8 +877,6 @@ static void do_pci_unregister_device(PCIDevice *pci_dev)memory_region_del_subregion (&pci_dev->bus_master_enable_region, pci_dev->iommu); pci_dev->bus->iommu_dtor_fn (pci_dev->iommu);memory_region_destroy (&pci_dev->bus_master_enable_region); -g_free (pci_dev->dma); - pci_dev->dma =NULL ; }static void pci_unregister_io_regions (PCIDevice *pci_dev)
hw/ppc/spapr_vio.c +0 -1
@@ -459,7 +459,6 @@ static int spapr_vio_busdev_init(DeviceState *qdev)uint32_t liobn =SPAPR_VIO_BASE_LIOBN | dev->reg; dev->tcet =spapr_tce_new_table (liobn, pc->rtce_window_size);address_space_init (&dev->as,spapr_tce_get_iommu (dev->tcet)); -dma_context_init (&dev->dma, &dev->as); }return pc->init (dev);
hw/scsi/megasas.c +2 -2
@@ -232,7 +232,7 @@ static int megasas_map_sgl(MegasasState *s, MegasasCmd *cmd, union mfi_sgl *sgl) MEGASAS_MAX_SGE);return iov_count; } -qemu_sglist_init (&cmd->qsg, iov_count,pci_dma_context (&s->dev)); +qemu_sglist_init (&cmd->qsg, iov_count,pci_get_address_space (&s->dev));for (i =0 ; i < iov_count; i++) {dma_addr_t iov_pa, iov_size_p;
@@ -628,7 +628,7 @@ static int megasas_map_dcmd(MegasasState *s, MegasasCmd *cmd) } iov_pa =megasas_sgl_get_addr (cmd, &cmd->frame->dcmd.sgl); iov_size =megasas_sgl_get_len (cmd, &cmd->frame->dcmd.sgl); -qemu_sglist_init (&cmd->qsg,1 ,pci_dma_context (&s->dev)); +qemu_sglist_init (&cmd->qsg,1 ,pci_get_address_space (&s->dev));qemu_sglist_add (&cmd->qsg, iov_pa, iov_size); cmd->iov_size = iov_size;return cmd->iov_size;
hw/scsi/virtio-scsi.c +1 -1
@@ -80,7 +80,7 @@ static void virtio_scsi_bad_req(void)static void qemu_sgl_init_external (QEMUSGList *qsgl,struct iovec *sg,hwaddr *addr,int num) { -qemu_sglist_init (qsgl, num, &dma_context_memory); +qemu_sglist_init (qsgl, num, &address_space_memory);while (num--) {qemu_sglist_add (qsgl, *(addr++), (sg++)->iov_len); }
hw/scsi/vmw_pvscsi.c +1 -1
@@ -617,7 +617,7 @@ pvscsi_build_sglist(PVSCSIState *s, PVSCSIRequest *r) {PCIDevice *d =PCI_DEVICE (s); -qemu_sglist_init (&r->sgl,1 ,pci_dma_context (d)); +pci_dma_sglist_init (&r->sgl, d,1 );if (r->req.flags &PVSCSI_FLAG_CMD_WITH_SG_LIST ) {pvscsi_convert_sglist (r); }else {
hw/sd/sdhci.c +11 -11
@@ -496,7 +496,7 @@ static void sdhci_sdma_transfer_multi_blocks(SDHCIState *s) s->blkcnt--; } } -dma_memory_write (&dma_context_memory, s->sdmasysad, +dma_memory_write (&address_space_memory, s->sdmasysad, &s->fifo_buffer[begin], s->data_count - begin); s->sdmasysad += s->data_count - begin;if (s->data_count == block_size) {
@@ -518,7 +518,7 @@ static void sdhci_sdma_transfer_multi_blocks(SDHCIState *s) s->data_count = block_size; boundary_count -= block_size - begin; } -dma_memory_read (&dma_context_memory, s->sdmasysad, +dma_memory_read (&address_space_memory, s->sdmasysad, &s->fifo_buffer[begin], s->data_count); s->sdmasysad += s->data_count - begin;if (s->data_count == block_size) {
@@ -557,10 +557,10 @@ static void sdhci_sdma_transfer_single_block(SDHCIState *s)for (n =0 ; n < datacnt; n++) { s->fifo_buffer[n] =sd_read_data (s->card); } -dma_memory_write (&dma_context_memory, s->sdmasysad, s->fifo_buffer, +dma_memory_write (&address_space_memory, s->sdmasysad, s->fifo_buffer, datacnt); }else { -dma_memory_read (&dma_context_memory, s->sdmasysad, s->fifo_buffer, +dma_memory_read (&address_space_memory, s->sdmasysad, s->fifo_buffer, datacnt);for (n =0 ; n < datacnt; n++) {sd_write_data (s->card, s->fifo_buffer[n]);
@@ -588,7 +588,7 @@ static void get_adma_description(SDHCIState *s, ADMADescr *dscr)hwaddr entry_addr = (hwaddr )s->admasysaddr;switch (SDHC_DMA_TYPE (s->hostctl)) {case SDHC_CTRL_ADMA2_32 : -dma_memory_read (&dma_context_memory, entry_addr, (uint8_t *)&adma2, +dma_memory_read (&address_space_memory, entry_addr, (uint8_t *)&adma2,sizeof (adma2)); adma2 =le64_to_cpu (adma2);/* The spec does not specify endianness of descriptor table.
@@ -600,7 +600,7 @@ static void get_adma_description(SDHCIState *s, ADMADescr *dscr) dscr->incr =8 ;break ;case SDHC_CTRL_ADMA1_32 : -dma_memory_read (&dma_context_memory, entry_addr, (uint8_t *)&adma1, +dma_memory_read (&address_space_memory, entry_addr, (uint8_t *)&adma1,sizeof (adma1)); adma1 =le32_to_cpu (adma1); dscr->addr = (hwaddr)(adma1 &0xFFFFF000 );
@@ -613,12 +613,12 @@ static void get_adma_description(SDHCIState *s, ADMADescr *dscr) }break ;case SDHC_CTRL_ADMA2_64 : -dma_memory_read (&dma_context_memory, entry_addr, +dma_memory_read (&address_space_memory, entry_addr, (uint8_t *)(&dscr->attr),1 ); -dma_memory_read (&dma_context_memory, entry_addr +2 , +dma_memory_read (&address_space_memory, entry_addr +2 , (uint8_t *)(&dscr->length),2 ); dscr->length =le16_to_cpu (dscr->length); -dma_memory_read (&dma_context_memory, entry_addr +4 , +dma_memory_read (&address_space_memory, entry_addr +4 , (uint8_t *)(&dscr->addr),8 ); dscr->attr =le64_to_cpu (dscr->attr); dscr->attr &=0xfffffff8 ;
@@ -678,7 +678,7 @@ static void sdhci_do_adma(SDHCIState *s) s->data_count = block_size; length -= block_size - begin; } -dma_memory_write (&dma_context_memory, dscr.addr, +dma_memory_write (&address_space_memory, dscr.addr, &s->fifo_buffer[begin], s->data_count - begin); dscr.addr += s->data_count - begin;
@@ -702,7 +702,7 @@ static void sdhci_do_adma(SDHCIState *s) s->data_count = block_size; length -= block_size - begin; } -dma_memory_read (&dma_context_memory, dscr.addr, +dma_memory_read (&address_space_memory, dscr.addr, &s->fifo_buffer[begin], s->data_count); dscr.addr += s->data_count - begin;if (s->data_count == block_size) {
hw/usb/hcd-ehci-pci.c +2 -2
@@ -63,7 +63,7 @@ static int usb_ehci_pci_initfn(PCIDevice *dev) s->caps[0x09 ] =0x68 ;/* EECP */ s->irq = dev->irq[3 ]; - s->dma =pci_dma_context (dev); + s->as =pci_get_address_space (dev); s->capsbase =0x00 ; s->opregbase =0x20 ;
@@ -86,7 +86,7 @@ static void usb_ehci_pci_write_config(PCIDevice *dev, uint32_t addr,return ; } busmaster =pci_get_word (dev->config +PCI_COMMAND ) &PCI_COMMAND_MASTER ; - i->ehci.dma = busmaster ?pci_dma_context (dev) :NULL ; + i->ehci.as = busmaster ?pci_get_address_space (dev) : &address_space_memory; }static Property ehci_pci_properties[] = {
hw/usb/hcd-ehci-sysbus.c +1 -1
@@ -40,7 +40,7 @@ static int usb_ehci_sysbus_initfn(SysBusDevice *dev) s->capsbase = sec->capsbase; s->opregbase = sec->opregbase; - s->dma = &dma_context_memory; + s->as = &address_space_memory;usb_ehci_initfn (s,DEVICE (dev));sysbus_init_irq (dev, &s->irq);
hw/usb/hcd-ehci.c +6 -6
@@ -446,7 +446,7 @@ static inline int get_dwords(EHCIState *ehci, uint32_t addr, {int i; -if (!ehci->dma) { +if (!ehci->as) {ehci_raise_irq (ehci,USBSTS_HSE ); ehci->usbcmd &= ~USBCMD_RUNSTOP ;trace_usb_ehci_dma_error ();
@@ -454,7 +454,7 @@ static inline int get_dwords(EHCIState *ehci, uint32_t addr, }for (i =0 ; i < num; i++, buf++, addr +=sizeof (*buf)) { -dma_memory_read (ehci->dma, addr, buf,sizeof (*buf)); +dma_memory_read (ehci->as, addr, buf,sizeof (*buf)); *buf =le32_to_cpu (*buf); }
@@ -467,7 +467,7 @@ static inline int put_dwords(EHCIState *ehci, uint32_t addr, {int i; -if (!ehci->dma) { +if (!ehci->as) {ehci_raise_irq (ehci,USBSTS_HSE ); ehci->usbcmd &= ~USBCMD_RUNSTOP ;trace_usb_ehci_dma_error ();
@@ -476,7 +476,7 @@ static inline int put_dwords(EHCIState *ehci, uint32_t addr,for (i =0 ; i < num; i++, buf++, addr +=sizeof (*buf)) {uint32_t tmp =cpu_to_le32 (*buf); -dma_memory_write (ehci->dma, addr, &tmp,sizeof (tmp)); +dma_memory_write (ehci->as, addr, &tmp,sizeof (tmp)); }return num;
@@ -1245,7 +1245,7 @@ static int ehci_init_transfer(EHCIPacket *p) cpage =get_field (p->qtd.token,QTD_TOKEN_CPAGE ); bytes =get_field (p->qtd.token,QTD_TOKEN_TBYTES ); offset = p->qtd.bufptr[0 ] & ~QTD_BUFPTR_MASK ; -qemu_sglist_init (&p->sgl,5 , p->queue->ehci->dma); +qemu_sglist_init (&p->sgl,5 , p->queue->ehci->as);while (bytes >0 ) {if (cpage >4 ) {
@@ -1484,7 +1484,7 @@ static int ehci_process_itd(EHCIState *ehci,return -1 ; } -qemu_sglist_init (&ehci->isgl,2 , ehci->dma); +qemu_sglist_init (&ehci->isgl,2 , ehci->as);if (off + len >4096 ) {/* transfer crosses page border */ uint32_t len2 = off + len -4096 ;
hw/usb/hcd-ehci.h +1 -1
@@ -261,7 +261,7 @@ struct EHCIState {USBBus bus;qemu_irq irq;MemoryRegion mem; -DMAContext *dma; +AddressSpace *as;MemoryRegion mem_caps;MemoryRegion mem_opreg;MemoryRegion mem_ports;
hw/usb/hcd-ohci.c +15 -15
@@ -62,7 +62,7 @@ typedef struct {USBBus bus;qemu_irq irq;MemoryRegion mem; -DMAContext *dma; +AddressSpace *as;int num_ports;const char *name;
@@ -508,7 +508,7 @@ static inline int get_dwords(OHCIState *ohci, addr += ohci->localmem_base;for (i =0 ; i < num; i++, buf++, addr +=sizeof (*buf)) { -dma_memory_read (ohci->dma, addr, buf,sizeof (*buf)); +dma_memory_read (ohci->as, addr, buf,sizeof (*buf)); *buf =le32_to_cpu (*buf); }
@@ -525,7 +525,7 @@ static inline int put_dwords(OHCIState *ohci,for (i =0 ; i < num; i++, buf++, addr +=sizeof (*buf)) {uint32_t tmp =cpu_to_le32 (*buf); -dma_memory_write (ohci->dma, addr, &tmp,sizeof (tmp)); +dma_memory_write (ohci->as, addr, &tmp,sizeof (tmp)); }return 1 ;
@@ -540,7 +540,7 @@ static inline int get_words(OHCIState *ohci, addr += ohci->localmem_base;for (i =0 ; i < num; i++, buf++, addr +=sizeof (*buf)) { -dma_memory_read (ohci->dma, addr, buf,sizeof (*buf)); +dma_memory_read (ohci->as, addr, buf,sizeof (*buf)); *buf =le16_to_cpu (*buf); }
@@ -557,7 +557,7 @@ static inline int put_words(OHCIState *ohci,for (i =0 ; i < num; i++, buf++, addr +=sizeof (*buf)) {uint16_t tmp =cpu_to_le16 (*buf); -dma_memory_write (ohci->dma, addr, &tmp,sizeof (tmp)); +dma_memory_write (ohci->as, addr, &tmp,sizeof (tmp)); }return 1 ;
@@ -585,7 +585,7 @@ static inline int ohci_read_iso_td(OHCIState *ohci,static inline int ohci_read_hcca (OHCIState *ohci,dma_addr_t addr,struct ohci_hcca *hcca) { -dma_memory_read (ohci->dma, addr + ohci->localmem_base, hcca,sizeof (*hcca)); +dma_memory_read (ohci->as, addr + ohci->localmem_base, hcca,sizeof (*hcca));return 1 ; }
@@ -617,7 +617,7 @@ static inline int ohci_put_iso_td(OHCIState *ohci,static inline int ohci_put_hcca (OHCIState *ohci,dma_addr_t addr,struct ohci_hcca *hcca) { -dma_memory_write (ohci->dma, +dma_memory_write (ohci->as, addr + ohci->localmem_base +HCCA_WRITEBACK_OFFSET , (char *)hcca +HCCA_WRITEBACK_OFFSET ,HCCA_WRITEBACK_SIZE );
@@ -634,12 +634,12 @@ static void ohci_copy_td(OHCIState *ohci, struct ohci_td *td, n =0x1000 - (ptr &0xfff );if (n > len) n = len; -dma_memory_rw (ohci->dma, ptr + ohci->localmem_base, buf, n, dir); +dma_memory_rw (ohci->as, ptr + ohci->localmem_base, buf, n, dir);if (n == len)return ; ptr = td->be & ~0xfffu ; buf += n; -dma_memory_rw (ohci->dma, ptr + ohci->localmem_base, buf, len - n, dir); +dma_memory_rw (ohci->as, ptr + ohci->localmem_base, buf, len - n, dir); }/* Read/Write the contents of an ISO TD from/to main memory. */
@@ -653,12 +653,12 @@ static void ohci_copy_iso_td(OHCIState *ohci, n =0x1000 - (ptr &0xfff );if (n > len) n = len; -dma_memory_rw (ohci->dma, ptr + ohci->localmem_base, buf, n, dir); +dma_memory_rw (ohci->as, ptr + ohci->localmem_base, buf, n, dir);if (n == len)return ; ptr = end_addr & ~0xfffu ; buf += n; -dma_memory_rw (ohci->dma, ptr + ohci->localmem_base, buf, len - n, dir); +dma_memory_rw (ohci->as, ptr + ohci->localmem_base, buf, len - n, dir); }static void ohci_process_lists (OHCIState *ohci,int completion);
@@ -1788,11 +1788,11 @@ static USBBusOps ohci_bus_ops = {static int usb_ohci_init (OHCIState *ohci,DeviceState *dev,int num_ports,dma_addr_t localmem_base,char *masterbus,uint32_t firstport, -DMAContext *dma) +AddressSpace *as) {int i; - ohci->dma = dma; + ohci->as = as;if (usb_frame_time ==0 ) {#ifdef OHCI_TIME_WARP
@@ -1859,7 +1859,7 @@ static int usb_ohci_initfn_pci(struct PCIDevice *dev)if (usb_ohci_init (&ohci->state, &dev->qdev, ohci->num_ports,0 , ohci->masterbus, ohci->firstport, -pci_dma_context (dev)) !=0 ) { +pci_get_address_space (dev)) !=0 ) {return -1 ; } ohci->state.irq = ohci->pci_dev.irq[0 ];
@@ -1882,7 +1882,7 @@ static int ohci_init_pxa(SysBusDevice *dev)/* Cannot fail as we pass NULL for masterbus */ usb_ohci_init (&s->ohci, &dev->qdev, s->num_ports, s->dma_offset,NULL ,0 , - &dma_context_memory); + &address_space_memory);sysbus_init_irq (dev, &s->ohci.irq);sysbus_init_mmio (dev, &s->ohci.mem);
hw/usb/libhw.c +2 -2
@@ -37,7 +37,7 @@ int usb_packet_map(USBPacket *p, QEMUSGList *sgl)while (len) {dma_addr_t xlen = len; - mem =dma_memory_map (sgl->dma, base, &xlen, dir); + mem =dma_memory_map (sgl->as, base, &xlen, dir);if (!mem) { goto err; }
@@ -63,7 +63,7 @@ void usb_packet_unmap(USBPacket *p, QEMUSGList *sgl)int i;for (i =0 ; i < p->iov.niov; i++) { -dma_memory_unmap (sgl->dma, p->iov.iov[i].iov_base, +dma_memory_unmap (sgl->as, p->iov.iov[i].iov_base, p->iov.iov[i].iov_len, dir, p->iov.iov[i].iov_len); }
include/hw/pci/pci.h +8 -9
@@ -243,7 +243,6 @@ struct PCIDevice {AddressSpace bus_master_as;MemoryRegion bus_master_enable_region;MemoryRegion *iommu; -DMAContext *dma;/* do not access the following fields */ PCIConfigReadFunc *config_read;
@@ -642,15 +641,15 @@ static inline uint32_t pci_config_size(const PCIDevice *d) }/* DMA access functions */ -static inline DMAContext *pci_dma_context (PCIDevice *dev) +static inline AddressSpace *pci_get_address_space (PCIDevice *dev) { -return dev->dma; +return &dev->bus_master_as; }static inline int pci_dma_rw (PCIDevice *dev,dma_addr_t addr,void *buf,dma_addr_t len,DMADirection dir) { -dma_memory_rw (pci_dma_context (dev), addr, buf, len, dir); +dma_memory_rw (pci_get_address_space (dev), addr, buf, len, dir);return 0 ; }
@@ -670,12 +669,12 @@ static inline int pci_dma_write(PCIDevice *dev, dma_addr_t addr, static inline uint##_bits##_t ld##_l##_pci_dma(PCIDevice *dev, \dma_addr_t addr) \ { \ - return ld##_l##_dma(pci_dma_context(dev), addr); \ + return ld##_l##_dma(pci_get_address_space(dev), addr); \ } \ static inline void st##_s##_pci_dma(PCIDevice *dev, \ dma_addr_t addr, uint##_bits##_t val) \ { \ - st##_s##_dma(pci_dma_context(dev), addr, val); \ + st##_s##_dma(pci_get_address_space(dev), addr, val); \ }PCI_DMA_DEFINE_LDST (ub, b,8 );
@@ -693,20 +692,20 @@ static inline void *pci_dma_map(PCIDevice *dev, dma_addr_t addr, {void *buf; - buf =dma_memory_map (pci_dma_context (dev), addr, plen, dir); + buf =dma_memory_map (pci_get_address_space (dev), addr, plen, dir);return buf; }static inline void pci_dma_unmap (PCIDevice *dev,void *buffer,dma_addr_t len,DMADirection dir,dma_addr_t access_len) { -dma_memory_unmap (pci_dma_context (dev), buffer, len, dir, access_len); +dma_memory_unmap (pci_get_address_space (dev), buffer, len, dir, access_len); }static inline void pci_dma_sglist_init (QEMUSGList *qsg,PCIDevice *dev,int alloc_hint) { -qemu_sglist_init (qsg, alloc_hint,pci_dma_context (dev)); +qemu_sglist_init (qsg, alloc_hint,pci_get_address_space (dev)); }extern const VMStateDescription vmstate_pci_device;
include/hw/ppc/spapr_vio.h +12 -13
@@ -64,7 +64,6 @@ struct VIOsPAPRDevice {target_ulong signal_state;VIOsPAPR_CRQ crq;AddressSpace as; -DMAContext dma;sPAPRTCETable *tcet; };
@@ -93,44 +92,44 @@ static inline qemu_irq spapr_vio_qirq(VIOsPAPRDevice *dev)static inline bool spapr_vio_dma_valid (VIOsPAPRDevice *dev,uint64_t taddr,uint32_t size,DMADirection dir) { -return dma_memory_valid (&dev->dma, taddr, size, dir); +return dma_memory_valid (&dev->as, taddr, size, dir); }static inline int spapr_vio_dma_read (VIOsPAPRDevice *dev,uint64_t taddr,void *buf,uint32_t size) { -if (!dma_memory_valid (&dev->dma, taddr, size,DMA_DIRECTION_TO_DEVICE )) { +if (!dma_memory_valid (&dev->as, taddr, size,DMA_DIRECTION_TO_DEVICE )) {return H_DEST_PARM ; } -dma_memory_read (&dev->dma, taddr, buf, size); +dma_memory_read (&dev->as, taddr, buf, size);return H_SUCCESS ; }static inline int spapr_vio_dma_write (VIOsPAPRDevice *dev,uint64_t taddr,const void *buf,uint32_t size) { -if (!dma_memory_valid (&dev->dma, taddr, size,DMA_DIRECTION_FROM_DEVICE )) { +if (!dma_memory_valid (&dev->as, taddr, size,DMA_DIRECTION_FROM_DEVICE )) {return H_DEST_PARM ; } -dma_memory_write (&dev->dma, taddr, buf, size); +dma_memory_write (&dev->as, taddr, buf, size);return H_SUCCESS ; }static inline int spapr_vio_dma_set (VIOsPAPRDevice *dev,uint64_t taddr,uint8_t c,uint32_t size) { -if (!dma_memory_valid (&dev->dma, taddr, size,DMA_DIRECTION_FROM_DEVICE )) { +if (!dma_memory_valid (&dev->as, taddr, size,DMA_DIRECTION_FROM_DEVICE )) {return H_DEST_PARM ; } -dma_memory_set (&dev->dma, taddr, c, size); +dma_memory_set (&dev->as, taddr, c, size);return H_SUCCESS ; } -#define vio_stb (_dev, _addr, _val) (stb_dma(&(_dev)->dma, (_addr), (_val))) -#define vio_sth (_dev, _addr, _val) (stw_be_dma(&(_dev)->dma, (_addr), (_val))) -#define vio_stl (_dev, _addr, _val) (stl_be_dma(&(_dev)->dma, (_addr), (_val))) -#define vio_stq (_dev, _addr, _val) (stq_be_dma(&(_dev)->dma, (_addr), (_val))) -#define vio_ldq (_dev, _addr) (ldq_be_dma(&(_dev)->dma, (_addr))) +#define vio_stb (_dev, _addr, _val) (stb_dma(&(_dev)->as, (_addr), (_val))) +#define vio_sth (_dev, _addr, _val) (stw_be_dma(&(_dev)->as, (_addr), (_val))) +#define vio_stl (_dev, _addr, _val) (stl_be_dma(&(_dev)->as, (_addr), (_val))) +#define vio_stq (_dev, _addr, _val) (stq_be_dma(&(_dev)->as, (_addr), (_val))) +#define vio_ldq (_dev, _addr) (ldq_be_dma(&(_dev)->as, (_addr)))int spapr_vio_send_crq (VIOsPAPRDevice *dev,uint8_t *crq);
include/sysemu/dma.h +32 -43
184 lines not on this page · raw message
Signature · 1 line
1.7.1
PATCH 23/40 memory: give name to every AddressSpace
In reply to Paolo BonziniFrom: Alexey Kardashevskiy <[email protected]>
The "info mtree" command in QEMU console prints only "memory" and "I/O" address spaces while there are actually a lot more other AddressSpace structs created by PCI and VIO devices. Those devices do not normally have names and therefore not present in "info mtree" output.
The patch fixes this.
--- exec.c | 6 ++---- hw/pci/pci.c | 3 ++- hw/ppc/spapr_vio.c | 2 +- include/exec/memory.h | 5 +++-- memory.c | 8 +++----- 5 files changed, 11 insertions(+), 13 deletions(-)
exec.c +2 -4
@@ -1857,13 +1857,11 @@ static void memory_map_init(void) { system_memory =g_malloc (sizeof (*system_memory));memory_region_init (system_memory,"system" ,INT64_MAX ); -address_space_init (&address_space_memory, system_memory); - address_space_memory.name ="memory" ; +address_space_init (&address_space_memory, system_memory,"memory" ); system_io =g_malloc (sizeof (*system_io));memory_region_init (system_io,"io" ,65536 ); -address_space_init (&address_space_io, system_io); - address_space_io.name ="I/O" ; +address_space_init (&address_space_io, system_io,"I/O" );memory_listener_register (&core_memory_listener, &address_space_memory);memory_listener_register (&io_memory_listener, &address_space_io);
hw/pci/pci.c +2 -1
@@ -818,7 +818,8 @@ static PCIDevice *do_pci_register_device(PCIDevice *pci_dev, PCIBus *bus,memory_region_init_alias (&pci_dev->bus_master_enable_region,"bus master" , pci_dev->iommu,0 ,memory_region_size (pci_dev->iommu));memory_region_set_enabled (&pci_dev->bus_master_enable_region, false); -address_space_init (&pci_dev->bus_master_as, &pci_dev->bus_master_enable_region); +address_space_init (&pci_dev->bus_master_as, &pci_dev->bus_master_enable_region, + name); pci_dev->devfn = devfn;pstrcpy (pci_dev->name,sizeof (pci_dev->name), name);
hw/ppc/spapr_vio.c +1 -1
@@ -458,7 +458,7 @@ static int spapr_vio_busdev_init(DeviceState *qdev)if (pc->rtce_window_size) {uint32_t liobn =SPAPR_VIO_BASE_LIOBN | dev->reg; dev->tcet =spapr_tce_new_table (liobn, pc->rtce_window_size); -address_space_init (&dev->as,spapr_tce_get_iommu (dev->tcet)); +address_space_init (&dev->as,spapr_tce_get_iommu (dev->tcet), qdev->id); }return pc->init (dev);
include/exec/memory.h +3 -2
@@ -177,7 +177,7 @@ struct MemoryRegionPortio {*/ struct AddressSpace {/* All fields are private. */ -const char *name; +char *name;MemoryRegion *root;struct FlatView *current_map;int ioeventfd_nb;
@@ -839,8 +839,9 @@ void mtree_info(fprintf_function mon_printf, void *f);* * @as: an uninitialized #AddressSpace * @root: a #MemoryRegion that routes addesses for the address space +* @name: an address space name */ -void address_space_init (AddressSpace *as,MemoryRegion *root); +void address_space_init (AddressSpace *as,MemoryRegion *root,const char *name);/**
memory.c +3 -5
@@ -1588,7 +1588,7 @@ void memory_listener_unregister(MemoryListener *listener)QTAILQ_REMOVE (&memory_listeners, listener, link); } -void address_space_init (AddressSpace *as,MemoryRegion *root) +void address_space_init (AddressSpace *as,MemoryRegion *root,const char *name) {memory_region_transaction_begin (); as->root = root;
@@ -1597,7 +1597,7 @@ void address_space_init(AddressSpace *as, MemoryRegion *root) as->ioeventfd_nb =0 ; as->ioeventfds =NULL ;QTAILQ_INSERT_TAIL (&address_spaces, as, address_spaces_link); - as->name =NULL ; + as->name =g_strdup (name?name:"anonymous" );address_space_init_dispatch (as); memory_region_update_pending |= root->enabled;memory_region_transaction_commit ();
@@ -1612,6 +1612,7 @@ void address_space_destroy(AddressSpace *as)QTAILQ_REMOVE (&address_spaces, as, address_spaces_link);address_space_destroy_dispatch (as);flatview_destroy (as->current_map); +g_free (as->name);g_free (as->current_map);g_free (as->ioeventfds); }
@@ -1738,9 +1739,6 @@ void mtree_info(fprintf_function mon_printf, void *f)QTAILQ_INIT (&ml_head); QTAILQ_FOREACH(as, &address_spaces, address_spaces_link) { -if (!as->name) { -continue ; - }mon_printf (f,"%s\n" , as->name);mtree_print_mr (mon_printf, f, as->root,0 ,0 , &ml_head); }
Signature · 1 line
1.7.1
PATCH 24/40 memory: add getter/setter for owner
In reply to Paolo BonziniWhenever memory regions are accessed outside the BQL, they need to be preserved against hot-unplug. MemoryRegions actually do not have their own reference count; they piggyback on a QOM object, their "owner". Add two functions to retrieve and specify the owner.
The setter function will affect the owner recursively on a whole tree of contained regions, but without crossing (a) aliases (b) regions that are already owned by another device. This is so that a device can create a complex tree of regions and a single call to memory_region_set_owner (perhaps even within a bus-specific function, e.g. pci_register_bar) will affect the entire tree.
--- include/exec/memory.h | 18 ++++++++++++++++++ memory.c | 21 +++++++++++++++++++++ 2 files changed, 39 insertions(+), 0 deletions(-)
include/exec/memory.h +18 -0
@@ -135,6 +135,7 @@ struct MemoryRegion {const MemoryRegionIOMMUOps *iommu_ops;void *opaque;MemoryRegion *parent; +struct Object *owner;Int128 size;hwaddr addr;void (*destructor)(MemoryRegion *mr);
@@ -374,6 +375,23 @@ void memory_region_init_iommu(MemoryRegion *mr,void memory_region_destroy (MemoryRegion *mr);/** +* memory_region_owner: get a memory region's owner. +* +* @mr: the memory region being queried. +*/ +struct Object *memory_region_owner (MemoryRegion *mr); + +/** +* memory_region_set_owner: set the owner for a memory region and all +* the unowned regions below it. +* +* @mr: the memory region being set. +* @owner: the object that acts as the owner +*/ +void memory_region_set_owner (MemoryRegion *mr, +struct Object *owner); + +/** * memory_region_size: get a memory region's size. * * @mr: the memory region being queried.
memory.c +21 -0
@@ -789,6 +789,7 @@ void memory_region_init(MemoryRegion *mr, mr->ops =NULL ; mr->iommu_ops =NULL ; mr->parent =NULL ; + mr->owner =NULL ; mr->size =int128_make64 (size);if (size ==UINT64_MAX ) { mr->size =int128_2_64 ();
@@ -1040,6 +1041,26 @@ void memory_region_destroy(MemoryRegion *mr)g_free (mr->ioeventfds); } +Object *memory_region_owner (MemoryRegion *mr) +{ +return mr->owner; +} + +void memory_region_set_owner (MemoryRegion *mr, +Object *owner) +{ +MemoryRegion *child; + +assert (mr->owner ==NULL || mr->owner == owner); + mr->owner = owner; + + QTAILQ_FOREACH(child, &mr->subregions, subregions_link) { +if (child->owner ==NULL || child->owner == owner) { +memory_region_set_owner (child, owner); + } + } +} +uint64_t memory_region_size (MemoryRegion *mr) {if (int128_eq (mr->size,int128_2_64 ())) {
Signature · 1 line
1.7.1
PATCH 25/40 memory: add ref/unref
In reply to Paolo Bonzini--- include/exec/memory.h | 34 ++++++++++++++++++++++++++++++++++ memory.c | 14 ++++++++++++++ 2 files changed, 48 insertions(+), 0 deletions(-)
include/exec/memory.h +34 -0
@@ -253,6 +253,40 @@ struct MemoryListener {void memory_region_init (MemoryRegion *mr,const char *name,uint64_t size); + +/** +* memory_region_ref: Add 1 to a memory region's reference count +* +* Whenever memory regions are accessed outside the BQL, they need to be +* preserved against hot-unplug. MemoryRegions actually do not have their +* own reference count; they piggyback on a QOM object, their "owner". +* This function adds a reference to the owner. +* +* All MemoryRegions must have an owner if they can disappear, even if the +* device they belong to operates exclusively under the BQL. This is because +* the region could be returned at any time by memory_region_find, and this +* is usually under guest control. +* +* @mr: the #MemoryRegion to be initialized +* @name: used for debugging; not visible to the user or ABI +* @size: size of the region; any subregions beyond this size will be clipped +*/ +void memory_region_ref (MemoryRegion *mr); + +/** +* memory_region_ref: Remove 1 to a memory region's reference count +* +* Whenever memory regions are accessed outside the BQL, they need to be +* preserved against hot-unplug. MemoryRegions actually do not have their +* own reference count; they piggyback on a QOM object, their "owner". +* This function removes a reference to the owner and possibly destroys it. +* +* @mr: the #MemoryRegion to be initialized +* @name: used for debugging; not visible to the user or ABI +* @size: size of the region; any subregions beyond this size will be clipped +*/ +void memory_region_unref (MemoryRegion *mr); +/** * memory_region_init_io: Initialize an I/O memory region. *
memory.c +14 -0
@@ -1061,6 +1061,20 @@ void memory_region_set_owner(MemoryRegion *mr, } } +void memory_region_ref (MemoryRegion *mr) +{ +if (mr && mr->owner) { +object_ref (mr->owner); + } +} + +void memory_region_unref (MemoryRegion *mr) +{ +if (mr && mr->owner) { +object_unref (mr->owner); + } +} +uint64_t memory_region_size (MemoryRegion *mr) {if (int128_eq (mr->size,int128_2_64 ())) {
Signature · 1 line
1.7.1
Re: [Qemu-devel] [PATCH 25/40] memory: add ref/unref
Quoted 19 lines · Paolo Bonzini
Signed-off-by Paolo Bonzini <[email protected]>--- include/exec/memory.h | 34 ++++++++++++++++++++++++++++++++++ memory.c | 14 ++++++++++++++ 2 files changed, 48 insertions(+), 0 deletions(-)include/exec/memory.h +5
-0@@ -253,6 +253,40 @@ struct MemoryListener {void memory_region_init (MemoryRegion *mr,const char *name,uint64_t size); + +/** +* memory_region_ref: Add 1 to a memory region's reference count +* +* Whenever memory regions are accessed outside the BQL, they need to be
include/exec/memory.h struct MemoryListener { · line 270
Quoted 10 lines
include/exec/memory.h +10
-0+* preserved against hot-unplug. MemoryRegions actually do not have their +* own reference count; they piggyback on a QOM object, their "owner". +* This function adds a reference to the owner. +* +* All MemoryRegions must have an owner if they can disappear, even if the +* device they belong to operates exclusively under the BQL. This is because +* the region could be returned at any time by memory_region_find, and this +* is usually under guest control. +* +* @mr: the #MemoryRegion to be initialized
s/to be initialized//
include/exec/memory.h struct MemoryListener { · line 272
Quoted 2 lines
include/exec/memory.h +2
-0+* @name: used for debugging; not visible to the user or ABI +* @size: size of the region; any subregions beyond this size will be clipped
Copy-pasted. These can be deleted.
include/exec/memory.h struct MemoryListener { · line 277
Quoted 5 lines
include/exec/memory.h +5
-0+*/ +void memory_region_ref (MemoryRegion *mr); + +/** +* memory_region_ref: Remove 1 to a memory region's reference count
s/memory_region_ref/memory_region_unref/
include/exec/memory.h struct MemoryListener { · line 286
Quoted 9 lines
include/exec/memory.h +9
-0+* +* Whenever memory regions are accessed outside the BQL, they need to be +* preserved against hot-unplug. MemoryRegions actually do not have their +* own reference count; they piggyback on a QOM object, their "owner". +* This function removes a reference to the owner and possibly destroys it. +* +* @mr: the #MemoryRegion to be initialized +* @name: used for debugging; not visible to the user or ABI +* @size: size of the region; any subregions beyond this size will be clipped
Same here.
PATCH 26/40 memory: add ref/unref calls
In reply to Paolo BonziniAdd ref/unref calls at the following places:
- places where memory regions are stashed by a listener and used outside the BQL (including in Xen or KVM).
- memory_region_find callsites
--- exec.c | 6 +++++- hw/core/loader.c | 1 + hw/display/exynos4210_fimd.c | 6 ++++++ hw/display/framebuffer.c | 10 ++++++---- hw/i386/kvm/ioapic.c | 2 ++ hw/i386/kvmvapic.c | 1 + hw/misc/vfio.c | 2 ++ hw/virtio/dataplane/hostmem.c | 7 +++++++ hw/virtio/vhost.c | 2 ++ hw/virtio/virtio-balloon.c | 1 + hw/xen/xen_pt.c | 4 ++++ include/hw/virtio/dataplane/hostmem.h | 1 + kvm-all.c | 2 ++ memory.c | 16 ++++++++++++++++ target-arm/kvm.c | 2 ++ target-sparc/mmu_helper.c | 1 + xen-all.c | 2 ++ 17 files changed, 61 insertions(+), 5 deletions(-)
exec.c +5 -1
@@ -761,12 +761,16 @@ static uint16_t phys_section_add(MemoryRegionSection *section) phys_sections_nb_alloc); } phys_sections[phys_sections_nb] = *section; +memory_region_ref (section->mr);return phys_sections_nb++; }static void phys_sections_clear (void ) { - phys_sections_nb =0 ; +while (phys_sections_nb >0 ) { +MemoryRegionSection *section = &phys_sections[--phys_sections_nb]; +memory_region_unref (section->mr); + } }static void register_subpage (AddressSpaceDispatch *d,MemoryRegionSection *section)
hw/core/loader.c +1 -0
@@ -727,6 +727,7 @@ int rom_load_all(void) addr += rom->romsize; section =memory_region_find (get_system_memory (), rom->addr,1 ); rom->isrom = section.size &&memory_region_is_rom (section.mr); +memory_region_unref (section.mr); }qemu_register_reset (rom_reset,NULL ); roms_loaded =1 ;
hw/display/exynos4210_fimd.c +6 -0
@@ -1126,6 +1126,11 @@ static void fimd_update_memory_section(Exynos4210fimdState *s, unsigned win)/* Total number of bytes of virtual screen used by current window */ w->fb_len = fb_mapped_len = (w->virtpage_width + w->virtpage_offsize) * (w->rightbot_y - w->lefttop_y +1 ); + +/* TODO: add .exit and unref the region there. Not needed yet since sysbus +* does not support hot-unplug. +*/ +memory_region_unref (w->mem_section.mr); w->mem_section =memory_region_find (sysbus_address_space (&s->busdev), fb_start_addr, w->fb_len);assert (w->mem_section.mr);
@@ -1154,6 +1159,7 @@ static void fimd_update_memory_section(Exynos4210fimdState *s, unsigned win)return ; error_return: +memory_region_unref (w->mem_section.mr); w->mem_section.mr =NULL ; w->mem_section.size =0 ; w->host_fb_addr =NULL ;
hw/display/framebuffer.c +6 -4
@@ -54,10 +54,10 @@ void framebuffer_update_display( src_len = src_width * rows; mem_section =memory_region_find (address_space, base, src_len); + mem = mem_section.mr;if (mem_section.size != src_len || !memory_region_is_ram (mem_section.mr)) { -return ; + goto out; } - mem = mem_section.mr;assert (mem);assert (mem_section.offset_within_address_space == base);
@@ -67,10 +67,10 @@ void framebuffer_update_display(but it's not really worth it as dirty flag tracking will probably already have failed above. */ if (!src_base) -return ; + goto out;if (src_len != src_width * rows) {cpu_physical_memory_unmap (src_base, src_len,0 ,0 ); -return ; + goto out; } src = src_base; dest =surface_data (ds);
@@ -107,4 +107,6 @@ void framebuffer_update_display( DIRTY_MEMORY_VGA); *first_row = first; *last_row = last; +out: +memory_region_unref (mem); }
hw/i386/kvm/ioapic.c +2 -0
@@ -107,6 +107,8 @@ static void kvm_ioapic_put(IOAPICCommonState *s)fprintf (stderr,"KVM_GET_IRQCHIP failed: %s\n" ,strerror (ret));abort (); } + +memory_region_unref (mrs.mr); }static void kvm_ioapic_reset (DeviceState *dev)
hw/i386/kvmvapic.c +1 -0
@@ -605,6 +605,7 @@ static void vapic_map_rom_writable(VAPICROMState *s) rom_size);memory_region_add_subregion_overlap (as, rom_paddr, &s->rom,1000 ); s->rom_mapped_writable = true; +memory_region_unref (section.mr); }static int vapic_prepare (VAPICROMState *s)
hw/misc/vfio.c +2 -0
@@ -1969,6 +1969,7 @@ static void vfio_listener_region_add(MemoryListener *listener,DPRINTF ("region_add %" HWADDR_PRIx" - %" HWADDR_PRIx" [%p]\n" , iova, end -1 , vaddr); +memory_region_ref (section->mr); ret =vfio_dma_map (container, iova, end - iova, vaddr, section->readonly);if (ret) { error_report("vfio_dma_map(%p, 0x%"HWADDR_PRIx", "
@@ -2010,6 +2011,7 @@ static void vfio_listener_region_del(MemoryListener *listener, iova, end - 1); ret =vfio_dma_unmap (container, iova, end - iova); +memory_region_unref (section->mr);if (ret) { error_report("vfio_dma_unmap(%p, 0x%"HWADDR_PRIx", " "0x%"HWADDR_PRIx") = %d (%m)",
hw/virtio/dataplane/hostmem.c +7 -0
@@ -64,8 +64,12 @@ out:static void hostmem_listener_commit (MemoryListener *listener) {HostMem *hostmem =container_of (listener, HostMem, listener); +int i;qemu_mutex_lock (&hostmem->current_regions_lock); +for (i =0 ; i < hostmem->num_current_regions; i++) { +memory_region_unref (hostmem->current_regions[i].mr); + }g_free (hostmem->current_regions); hostmem->current_regions = hostmem->new_regions; hostmem->num_current_regions = hostmem->num_new_regions;
@@ -92,8 +96,11 @@ static void hostmem_append_new_region(HostMem *hostmem, .guest_addr = section->offset_within_address_space, .size = section->size, .readonly = section->readonly, + .mr = section->mr, }; hostmem->num_new_regions++; + +memory_region_ref (section->mr); } static void hostmem_listener_append_region(MemoryListener *listener,
hw/virtio/vhost.c +2 -0
@@ -497,6 +497,7 @@ static void vhost_region_add(MemoryListener *listener, dev->mem_sections =g_renew (MemoryRegionSection, dev->mem_sections, dev->n_mem_sections); dev->mem_sections[dev->n_mem_sections -1 ] = *section; +memory_region_ref (section->mr);vhost_set_memory (listener, section, true); }
@@ -512,6 +513,7 @@ static void vhost_region_del(MemoryListener *listener, }vhost_set_memory (listener, section, false); +memory_region_unref (section->mr);for (i =0 ; i < dev->n_mem_sections; ++i) {if (dev->mem_sections[i].offset_within_address_space == section->offset_within_address_space) {
hw/virtio/virtio-balloon.c +1 -0
@@ -205,6 +205,7 @@ static void virtio_balloon_handle_output(VirtIODevice *vdev, VirtQueue *vq) addr = section.offset_within_region;balloon_page (memory_region_get_ram_ptr (section.mr) + addr, !!(vq == s->dvq)); +memory_region_unref (section.mr); }virtqueue_push (vq, &elem, offset);
hw/xen/xen_pt.c +4 -0
@@ -606,6 +606,7 @@ static void xen_pt_region_add(MemoryListener *l, MemoryRegionSection *sec)XenPCIPassthroughState *s =container_of (l, XenPCIPassthroughState, memory_listener); +memory_region_ref (sec->mr);xen_pt_region_update (s, sec, true); }
@@ -615,6 +616,7 @@ static void xen_pt_region_del(MemoryListener *l, MemoryRegionSection *sec) memory_listener);xen_pt_region_update (s, sec, false); +memory_region_unref (sec->mr); }static void xen_pt_io_region_add (MemoryListener *l,MemoryRegionSection *sec)
@@ -622,6 +624,7 @@ static void xen_pt_io_region_add(MemoryListener *l, MemoryRegionSection *sec)XenPCIPassthroughState *s =container_of (l, XenPCIPassthroughState, io_listener); +memory_region_ref (sec->mr);xen_pt_region_update (s, sec, true); }
@@ -631,6 +634,7 @@ static void xen_pt_io_region_del(MemoryListener *l, MemoryRegionSection *sec) io_listener);xen_pt_region_update (s, sec, false); +memory_region_unref (sec->mr); }static const MemoryListener xen_pt_memory_listener = {
include/hw/virtio/dataplane/hostmem.h +1 -0
@@ -18,6 +18,7 @@#include "qemu/thread.h" typedef struct { + MemoryRegion *mr; void *host_addr; hwaddr guest_addr; uint64_t size;
kvm-all.c +2 -0
@@ -766,6 +766,7 @@ static void kvm_set_phys_mem(MemoryRegionSection *section, bool add)static void kvm_region_add (MemoryListener *listener,MemoryRegionSection *section) { +memory_region_ref (section->mr);kvm_set_phys_mem (section, true); }
@@ -773,6 +774,7 @@ static void kvm_region_del(MemoryListener *listener, MemoryRegionSection *section) {kvm_set_phys_mem (section, false); +memory_region_unref (section->mr); } static void kvm_log_sync(MemoryListener *listener,
memory.c +16 -0
@@ -145,6 +145,7 @@ static bool memory_listener_match(MemoryListener *listener, } \ }while (0 ) +/* No need to ref/unref .mr, the FlatRange keeps it alive. */#define MEMORY_LISTENER_UPDATE_REGION (fr, as, dir, callback) \ MEMORY_LISTENER_CALL(callback, dir, (&(MemoryRegionSection) { \ .mr = (fr)->mr, \
@@ -260,11 +261,17 @@ static void flatview_insert(FlatView *view, unsigned pos, FlatRange *range)memmove (view->ranges + pos +1 , view->ranges + pos, (view->nr - pos) *sizeof (FlatRange)); view->ranges[pos] = *range; +memory_region_ref (range->mr); ++view->nr; }static void flatview_destroy (FlatView *view) { +int i; + +for (i =0 ; i < view->nr; i++) { +memory_region_unref (view->ranges[i].mr); + }g_free (view->ranges); }
@@ -763,6 +770,11 @@ static void memory_region_destructor_ram(MemoryRegion *mr)qemu_ram_free (mr->ram_addr); } +static void memory_region_destructor_alias (MemoryRegion *mr) +{ +memory_region_unref (mr->alias); +} +static void memory_region_destructor_ram_from_ptr (MemoryRegion *mr) {qemu_ram_free_from_ptr (mr->ram_addr);
@@ -961,6 +973,8 @@ void memory_region_init_alias(MemoryRegion *mr, uint64_t size) {memory_region_init (mr, name, size); +memory_region_ref (orig); + mr->destructor = memory_region_destructor_alias; mr->alias = orig; mr->alias_offset = offset; }
@@ -1539,6 +1553,8 @@ MemoryRegionSection memory_region_find(MemoryRegion *mr, ret.size =int128_get64 (range.size); ret.offset_within_address_space =int128_get64 (range.start); ret.readonly = fr->readonly; +memory_region_ref (ret.mr); +return ret; }
target-arm/kvm.c +2 -0
@@ -127,6 +127,7 @@ static void kvm_arm_machine_init_done(Notifier *notifier, void *data)abort (); } } +memory_region_unref (kd->mr);g_free (kd); } }
@@ -152,6 +153,7 @@ void kvm_arm_register_device(MemoryRegion *mr, uint64_t devid) kd->kda.id = devid; kd->kda.addr =-1 ;QSLIST_INSERT_HEAD (&kvm_devices_head, kd, entries); +memory_region_ref (kd->mr); } typedef struct Reg {
target-sparc/mmu_helper.c +1 -0
@@ -845,6 +845,7 @@ hwaddr cpu_get_phys_page_debug(CPUSPARCState *env, target_ulong addr) } } section =memory_region_find (get_system_memory (), phys_addr,1 ); +memory_region_unref (section.mr);if (!section.size) {return -1 ; }
xen-all.c +2 -0
@@ -459,6 +459,7 @@ static void xen_set_memory(struct MemoryListener *listener,static void xen_region_add (MemoryListener *listener,MemoryRegionSection *section) { +memory_region_ref (section->mr);xen_set_memory (listener, section, true); }
@@ -466,6 +467,7 @@ static void xen_region_del(MemoryListener *listener, MemoryRegionSection *section) {xen_set_memory (listener, section, false); +memory_region_unref (section->mr); } static void xen_sync_dirty_bitmap(XenIOState *state,
Signature · 1 line
1.7.1
PATCH 27/40 pci: set owner for BARs
In reply to Paolo Bonzini--- hw/pci/pci.c | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-)
hw/pci/pci.c +2 -0
@@ -919,6 +919,8 @@ void pci_register_bar(PCIDevice *pci_dev, int region_num,uint64_t wmask;pcibus_t size =memory_region_size (memory); +memory_region_set_owner (memory,OBJECT (pci_dev)); +assert (region_num >=0 );assert (region_num <PCI_NUM_REGIONS );if (size & (size-1 )) {
Signature · 1 line
1.7.1
PATCH 28/40 sysbus: set owner for MMIO regions
In reply to Paolo Bonzini--- hw/core/sysbus.c | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-)
hw/core/sysbus.c +2 -0
@@ -115,6 +115,8 @@ void sysbus_init_mmio(SysBusDevice *dev, MemoryRegion *memory) n = dev->num_mmio++; dev->mmio[n].addr =-1 ; dev->mmio[n].memory = memory; + +memory_region_set_owner (dev->mmio[n].memory,OBJECT (dev)); }MemoryRegion *sysbus_mmio_get_region (SysBusDevice *dev,int n)
Signature · 1 line
1.7.1
PATCH 29/40 acpi: add memory_region_set_owner calls
In reply to Paolo BonziniACPI regions are added directly to the I/O address space, without going through BARs. Thus they need the owner to be set directly.
--- hw/acpi/ich9.c | 1 + hw/acpi/piix4.c | 5 +++++ hw/isa/apm.c | 1 + 3 files changed, 7 insertions(+), 0 deletions(-)
hw/acpi/ich9.c +1 -0
@@ -223,6 +223,7 @@ void ich9_pm_init(PCIDevice *lpc_pci, ICH9LPCPMRegs *pm, 8);memory_region_add_subregion (&pm->io,ICH9_PMIO_SMI_EN , &pm->io_smi); +memory_region_set_owner (&pm->io,OBJECT (lpc_pci)); pm->irq = sci_irq;qemu_register_reset (pm_reset, pm); pm->powerdown_notifier.notify = pm_powerdown_req;
hw/acpi/piix4.c +5 -0
@@ -433,6 +433,8 @@ static int piix4_pm_initfn(PCIDevice *dev)acpi_pm1_cnt_init (&s->ar, &s->io, s->s4_val);acpi_gpe_init (&s->ar,GPE_LEN ); +memory_region_set_owner (&s->io,OBJECT (s)); + s->powerdown_notifier.notify = piix4_pm_powerdown_req;qemu_register_powerdown_notifier (&s->powerdown_notifier);
@@ -672,10 +674,12 @@ static void piix4_acpi_system_hot_add_init(MemoryRegion *parent, {memory_region_init_io (&s->io_gpe, &piix4_gpe_ops, s,"apci-gpe0" ,GPE_LEN ); +memory_region_set_owner (&s->io_gpe,OBJECT (s));memory_region_add_subregion (parent,GPE_BASE , &s->io_gpe);memory_region_init_io (&s->io_pci, &piix4_pci_ops, s,"apci-pci-hotplug" ,PCI_HOTPLUG_SIZE ); +memory_region_set_owner (&s->io_pci,OBJECT (s));memory_region_add_subregion (parent,PCI_HOTPLUG_ADDR , &s->io_pci);pci_bus_hotplug (bus, piix4_device_hotplug, &s->dev.qdev);
@@ -683,6 +687,7 @@ static void piix4_acpi_system_hot_add_init(MemoryRegion *parent,qemu_for_each_cpu (piix4_init_cpu_status, &s->gpe_cpu);memory_region_init_io (&s->io_cpu, &cpu_hotplug_ops, s,"apci-cpu-hotplug" ,PIIX4_PROC_LEN ); +memory_region_set_owner (&s->io_cpu,OBJECT (s));memory_region_add_subregion (parent,PIIX4_PROC_BASE , &s->io_cpu); s->cpu_added_notifier.notify = piix4_cpu_added_req;qemu_register_cpu_added_notifier (&s->cpu_added_notifier);
hw/isa/apm.c +1 -0
@@ -97,6 +97,7 @@ void apm_init(PCIDevice *dev, APMState *apm, apm_ctrl_changed_t callback,/* ioport 0xb2, 0xb3 */ memory_region_init_io (&apm->io, &apm_ops, apm,"apm-io" ,2 ); +memory_region_set_owner (&apm->io,OBJECT (dev));memory_region_add_subregion (pci_address_space_io (dev),APM_CNT_IOPORT , &apm->io); }
Signature · 1 line
1.7.1
PATCH 30/40 misc: add memory_region_set_owner calls
In reply to Paolo Bonzini--- hw/char/serial-pci.c | 1 + hw/misc/pc-testdev.c | 7 +++++++ 2 files changed, 8 insertions(+), 0 deletions(-)
hw/char/serial-pci.c +1 -0
@@ -106,6 +106,7 @@ static int multi_serial_pci_init(PCIDevice *dev) s->irq = pci->irqs[i]; pci->name[i] =g_strdup_printf ("uart #%d" , i+1 );memory_region_init_io (&s->io, &serial_io_ops, s, pci->name[i],8 ); +memory_region_set_owner (&s->io,OBJECT (pci));memory_region_add_subregion (&pci->iobar,8 * i, &s->io); }return 0 ;
hw/misc/pc-testdev.c +7 -0
@@ -150,12 +150,19 @@ static int init_test_device(ISADevice *isa)memory_region_init_io (&dev->ioport, &test_ioport_ops, dev,"pc-testdev-ioport" ,4 ); +memory_region_set_owner (&dev->ioport,OBJECT (dev)); +memory_region_init_io (&dev->flush, &test_flush_ops, dev,"pc-testdev-flush-page" ,4 ); +memory_region_set_owner (&dev->flush,OBJECT (dev)); +memory_region_init_io (&dev->irq, &test_irq_ops, dev,"pc-testdev-irq-line" ,24 ); +memory_region_set_owner (&dev->irq,OBJECT (dev)); +memory_region_init_io (&dev->iomem, &test_iomem_ops, dev,"pc-testdev-iomem" ,IOMEM_LEN ); +memory_region_set_owner (&dev->iomem,OBJECT (dev));memory_region_add_subregion (io,0xe0 , &dev->ioport);memory_region_add_subregion (io,0xe4 , &dev->flush);
Signature · 1 line
1.7.1
PATCH 31/40 isa/portio: allow setting an owner
In reply to Paolo Bonzini--- hw/isa/isa-bus.c | 2 ++ include/exec/ioport.h | 3 +++ ioport.c | 10 ++++++++++ 3 files changed, 15 insertions(+), 0 deletions(-)
hw/isa/isa-bus.c +2 -0
@@ -100,6 +100,7 @@ static inline void isa_init_ioport(ISADevice *dev, uint16_t ioport)void isa_register_ioport (ISADevice *dev,MemoryRegion *io,uint16_t start) { +memory_region_set_owner (io,OBJECT (dev));memory_region_add_subregion (isabus->address_space_io, start, io);isa_init_ioport (dev, start); }
@@ -116,6 +117,7 @@ void isa_register_portio_list(ISADevice *dev, uint16_t start,isa_init_ioport (dev, start);portio_list_init (piolist, pio_start, opaque, name); +portio_list_set_owner (piolist,OBJECT (dev));portio_list_add (piolist, isabus->address_space_io, start); }
include/exec/ioport.h +3 -0
@@ -62,6 +62,7 @@ typedef struct PortioList {unsigned nr;struct MemoryRegion **regions;struct MemoryRegion **aliases; +struct Object *owner;void *opaque;const char *name; } PortioList;
@@ -69,6 +70,8 @@ typedef struct PortioList {void portio_list_init (PortioList *piolist,const struct MemoryRegionPortio *callbacks,void *opaque,const char *name); +void portio_list_set_owner (PortioList *piolist, +struct Object *owner);void portio_list_destroy (PortioList *piolist); void portio_list_add(PortioList *piolist, struct MemoryRegion *address_space,
ioport.c +10 -0
@@ -347,6 +347,12 @@ void portio_list_init(PortioList *piolist, piolist->address_space =NULL ; piolist->opaque = opaque; piolist->name = name; + piolist->owner =NULL ; +} + +void portio_list_set_owner (PortioList *piolist,Object *owner) +{ + piolist->owner = owner; }void portio_list_destroy (PortioList *piolist)
@@ -386,8 +392,12 @@ static void portio_list_add_1(PortioList *piolist,*/ memory_region_init_io (region, ops, piolist->opaque, piolist->name,INT64_MAX ); +memory_region_set_owner (region, piolist->owner); +memory_region_init_alias (alias, piolist->name, region, start + off_low, off_high - off_low); +memory_region_set_owner (alias, piolist->owner); +memory_region_add_subregion (piolist->address_space, start + off_low, alias); piolist->regions[piolist->nr] = region;
Signature · 1 line
1.7.1
PATCH 32/40 vga: add memory_region_set_owner calls
In reply to Paolo Bonzini--- hw/display/cirrus_vga.c | 19 ++++++++++++++----- hw/display/qxl.c | 5 +++-- hw/display/vga-isa-mm.c | 2 +- hw/display/vga-isa.c | 4 ++-- hw/display/vga-pci.c | 5 +++-- hw/display/vga.c | 19 ++++++++++++++----- hw/display/vga_int.h | 9 +++++---- hw/display/vmware_vga.c | 4 ++-- 8 files changed, 44 insertions(+), 23 deletions(-)
hw/display/cirrus_vga.c +14 -5
@@ -2806,12 +2806,14 @@ static const MemoryRegionOps cirrus_vga_io_ops = { }, }; -static void cirrus_init_common (CirrusVGAState * s,int device_id,int is_pci, +static void cirrus_init_common (CirrusVGAState * s,int device_id, +DeviceState *owner,MemoryRegion *system_memory,MemoryRegion *system_io) {int i;static int inited; +int is_pci = !!object_dynamic_cast (OBJECT (owner),TYPE_PCI_DEVICE );if (!inited) { inited =1 ;
@@ -2843,19 +2845,23 @@ static void cirrus_init_common(CirrusVGAState * s, int device_id, int is_pci,/* Register ioport 0x3b0 - 0x3df */ memory_region_init_io (&s->cirrus_vga_io, &cirrus_vga_io_ops, s,"cirrus-io" ,0x30 ); +memory_region_set_owner (&s->cirrus_vga_io,OBJECT (owner));memory_region_add_subregion (system_io,0x3b0 , &s->cirrus_vga_io);memory_region_init (&s->low_mem_container,"cirrus-lowmem-container" ,0x20000 ); +memory_region_set_owner (&s->low_mem_container,OBJECT (owner));memory_region_init_io (&s->low_mem, &cirrus_vga_mem_ops, s,"cirrus-low-memory" ,0x20000 ); +memory_region_set_owner (&s->low_mem,OBJECT (owner));memory_region_add_subregion (&s->low_mem_container,0 , &s->low_mem);for (i =0 ; i <2 ; ++i) {static const char *names[] = {"vga.bank0" ,"vga.bank1" };MemoryRegion *bank = &s->cirrus_bank[i];memory_region_init_alias (bank, names[i], &s->vga.vram,0 ,0x8000 ); +memory_region_set_owner (bank,OBJECT (owner));memory_region_set_enabled (bank, false);memory_region_add_subregion_overlap (&s->low_mem_container, i *0x8000 , bank,1 );
@@ -2870,6 +2876,7 @@ static void cirrus_init_common(CirrusVGAState * s, int device_id, int is_pci,memory_region_init_io (&s->cirrus_linear_io, &cirrus_linear_io_ops, s,"cirrus-linear-io" , s->vga.vram_size_mb *1024 *1024 ); +memory_region_set_owner (&s->cirrus_linear_io,OBJECT (owner));memory_region_set_flush_coalesced (&s->cirrus_linear_io);/* I/O handler for LFB */
@@ -2878,11 +2885,13 @@ static void cirrus_init_common(CirrusVGAState * s, int device_id, int is_pci, s,"cirrus-bitblt-mmio" , 0x400000); +memory_region_set_owner (&s->cirrus_linear_bitblt_io,OBJECT (owner));memory_region_set_flush_coalesced (&s->cirrus_linear_bitblt_io);/* I/O handler for memory-mapped I/O */ memory_region_init_io (&s->cirrus_mmio_io, &cirrus_mmio_io_ops, s,"cirrus-mmio" ,CIRRUS_PNPMMIO_SIZE ); +memory_region_set_owner (&s->cirrus_mmio_io,OBJECT (owner));memory_region_set_flush_coalesced (&s->cirrus_mmio_io); s->real_vram_size =
@@ -2912,8 +2921,8 @@ static int vga_initfn(ISADevice *dev)ISACirrusVGAState *d =ISA_CIRRUS_VGA (dev);VGACommonState *s = &d->cirrus_vga.vga; -vga_common_init (s); -cirrus_init_common (&d->cirrus_vga,CIRRUS_ID_CLGD5430 ,0 , +vga_common_init (s,DEVICE (dev)); +cirrus_init_common (&d->cirrus_vga,CIRRUS_ID_CLGD5430 ,DEVICE (d),isa_address_space (dev),isa_address_space_io (dev)); s->con =graphic_console_init (DEVICE (dev), s->hw_ops, s);rom_add_vga (VGABIOS_CIRRUS_FILENAME );
@@ -2959,8 +2968,8 @@ static int pci_cirrus_vga_initfn(PCIDevice *dev)int16_t device_id = pc->device_id;/* setup VGA */ -vga_common_init (&s->vga); -cirrus_init_common (s, device_id,1 ,pci_address_space (dev), +vga_common_init (&s->vga,DEVICE (dev)); +cirrus_init_common (s, device_id,DEVICE (dev),pci_address_space (dev),pci_address_space_io (dev)); s->vga.con =graphic_console_init (DEVICE (dev), s->vga.hw_ops, &s->vga);
hw/display/qxl.c +3 -2
@@ -2063,9 +2063,10 @@ static int qxl_init_primary(PCIDevice *dev) qxl->id =0 ;qxl_init_ramsize (qxl); vga->vram_size_mb = qxl->vga.vram_size >>20 ; -vga_common_init (vga); -vga_init (vga,pci_address_space (dev),pci_address_space_io (dev), false); +vga_common_init (vga, dev); +vga_init (vga, dev,pci_address_space (dev),pci_address_space_io (dev), false);portio_list_init (qxl_vga_port_list, qxl_vga_portio_list, vga,"vga" ); +portio_list_set_owner (vbe_port_list,OBJECT (dev));portio_list_add (qxl_vga_port_list,pci_address_space_io (dev),0x3b0 ); vga->con =graphic_console_init (DEVICE (dev), &qxl_ops, qxl);
hw/display/vga-isa-mm.c +1 -1
@@ -132,7 +132,7 @@ int isa_vga_mm_init(hwaddr vram_base, s =g_malloc0 (sizeof (*s)); s->vga.vram_size_mb =VGA_RAM_SIZE >>20 ; -vga_common_init (&s->vga); +vga_common_init (&s->vga,NULL );vga_mm_init (s, vram_base, ctrl_base, it_shift, address_space); s->vga.con =graphic_console_init (NULL , s->vga.hw_ops, s);
hw/display/vga-isa.c +2 -2
@@ -55,9 +55,9 @@ static int vga_initfn(ISADevice *dev)MemoryRegion *vga_io_memory;const MemoryRegionPortio *vga_ports, *vbe_ports; -vga_common_init (s); +vga_common_init (s,DEVICE (dev)); s->legacy_address_space =isa_address_space (dev); - vga_io_memory =vga_init_io (s, &vga_ports, &vbe_ports); + vga_io_memory =vga_init_io (s,DEVICE (dev), &vga_ports, &vbe_ports);isa_register_portio_list (dev,0x3b0 , vga_ports, s,"vga" );if (vbe_ports) {isa_register_portio_list (dev,0x1ce , vbe_ports, s,"vbe" );
hw/display/vga-pci.c +3 -2
@@ -147,8 +147,9 @@ static int pci_std_vga_initfn(PCIDevice *dev)VGACommonState *s = &d->vga;/* vga + console init */ -vga_common_init (s); -vga_init (s,pci_address_space (dev),pci_address_space_io (dev), true); +vga_common_init (s,DEVICE (dev)); +vga_init (s,DEVICE (dev),pci_address_space (dev),pci_address_space_io (dev), + true); s->con =graphic_console_init (DEVICE (dev), s->hw_ops, s);
hw/display/vga.c +14 -5
@@ -199,6 +199,8 @@ static void vga_update_memory_access(VGACommonState *s) base += isa_mem_base; region =g_malloc (sizeof (*region));memory_region_init_alias (region,"vga.chain4" , &s->vram, offset, size); +memory_region_set_owner (region,memory_region_owner (&s->vram)); +memory_region_add_subregion_overlap (s->legacy_address_space, base, region,2 ); s->chain4_alias = region;
@@ -2256,7 +2258,7 @@ static const GraphicHwOps vga_ops = { .text_update = vga_update_text, }; -void vga_common_init (VGACommonState *s) +void vga_common_init (VGACommonState *s,DeviceState *owner) {int i, j, v, b;
@@ -2293,6 +2295,7 @@ void vga_common_init(VGACommonState *s) s->is_vbe_vmstate =1 ;memory_region_init_ram (&s->vram,"vga.vram" , s->vram_size); +memory_region_set_owner (&s->vram,OBJECT (owner));vmstate_register_ram_global (&s->vram);xen_register_framebuffer (&s->vram); s->vram_ptr =memory_region_get_ram_ptr (&s->vram);
@@ -2333,7 +2336,7 @@ static const MemoryRegionPortio vbe_portio_list[] = { };/* Used by both ISA and PCI */ -MemoryRegion *vga_init_io (VGACommonState *s, +MemoryRegion *vga_init_io (VGACommonState *s,DeviceState *owner,const MemoryRegionPortio **vga_ports,const MemoryRegionPortio **vbe_ports) {
@@ -2345,13 +2348,15 @@ MemoryRegion *vga_init_io(VGACommonState *s, vga_mem =g_malloc (sizeof (*vga_mem));memory_region_init_io (vga_mem, &vga_mem_ops, s,"vga-lowmem" ,0x20000 ); +memory_region_set_owner (vga_mem,OBJECT (owner));memory_region_set_flush_coalesced (vga_mem);return vga_mem; } -void vga_init (VGACommonState *s,MemoryRegion *address_space, -MemoryRegion *address_space_io,bool init_vga_ports) +void vga_init (VGACommonState *s,DeviceState *owner, +MemoryRegion *address_space,MemoryRegion *address_space_io, +bool init_vga_ports) {MemoryRegion *vga_io_memory;const MemoryRegionPortio *vga_ports, *vbe_ports;
@@ -2364,7 +2369,7 @@ void vga_init(VGACommonState *s, MemoryRegion *address_space, s->legacy_address_space = address_space; - vga_io_memory =vga_init_io (s, &vga_ports, &vbe_ports); + vga_io_memory =vga_init_io (s, owner, &vga_ports, &vbe_ports); memory_region_add_subregion_overlap(address_space, isa_mem_base + 0x000a0000, vga_io_memory,
@@ -2372,10 +2377,12 @@ void vga_init(VGACommonState *s, MemoryRegion *address_space,memory_region_set_coalescing (vga_io_memory);if (init_vga_ports) {portio_list_init (vga_port_list, vga_ports, s,"vga" ); +portio_list_set_owner (vga_port_list,OBJECT (owner));portio_list_add (vga_port_list, address_space_io,0x3b0 ); }if (vbe_ports) {portio_list_init (vbe_port_list, vbe_ports, s,"vbe" ); +portio_list_set_owner (vbe_port_list,OBJECT (owner));portio_list_add (vbe_port_list, address_space_io,0x1ce ); } }
@@ -2387,6 +2394,8 @@ void vga_init_vbe(VGACommonState *s, MemoryRegion *system_memory)*/ memory_region_init_alias (&s->vram_vbe,"vram.vbe" , &s->vram,0 ,memory_region_size (&s->vram)); +memory_region_set_owner (&s->vram_vbe,memory_region_owner (&s->vram)); +/* XXX: use optimized standard vga accesses */ memory_region_add_subregion(system_memory, VBE_DISPI_LFB_PHYSICAL_ADDRESS,
hw/display/vga_int.h +5 -4
@@ -177,10 +177,11 @@ static inline int c6_to_8(int v)return (v <<2 ) | (b <<1 ) | b; } -void vga_common_init (VGACommonState *s); -void vga_init (VGACommonState *s,MemoryRegion *address_space, -MemoryRegion *address_space_io,bool init_vga_ports); -MemoryRegion *vga_init_io (VGACommonState *s, +void vga_common_init (VGACommonState *s,DeviceState *owner); +void vga_init (VGACommonState *s,DeviceState *owner, +MemoryRegion *address_space,MemoryRegion *address_space_io, +bool init_vga_ports); +MemoryRegion *vga_init_io (VGACommonState *s,DeviceState *owner,const MemoryRegionPortio **vga_ports,const MemoryRegionPortio **vbe_ports);void vga_common_reset (VGACommonState *s);
hw/display/vmware_vga.c +2 -2
@@ -1198,8 +1198,8 @@ static void vmsvga_init(DeviceState *dev, struct vmsvga_state_s *s,vmstate_register_ram_global (&s->fifo_ram); s->fifo_ptr =memory_region_get_ram_ptr (&s->fifo_ram); -vga_common_init (&s->vga); -vga_init (&s->vga, address_space, io, true); +vga_common_init (&s->vga, dev); +vga_init (&s->vga, dev, address_space, io, true);vmstate_register (NULL ,0 , &vmstate_vga_common, &s->vga); s->new_depth =32 ; }
Signature · 1 line
1.7.1
PATCH 33/40 pci-assign: add memory_region_set_owner calls
In reply to Paolo Bonzini--- hw/i386/kvm/pci-assign.c | 11 +++++++++++ 1 files changed, 11 insertions(+), 0 deletions(-)
hw/i386/kvm/pci-assign.c +11 -0
@@ -300,6 +300,7 @@ static void assigned_dev_iomem_setup(PCIDevice *pci_dev, int region_num,if (e_size >0 ) {memory_region_init (®ion->container,"assigned-dev-container" , e_size); +memory_region_set_owner (®ion->container,OBJECT (pci_dev));memory_region_add_subregion (®ion->container,0 , ®ion->real_iomem);/* deal with MSI-X MMIO page */
@@ -330,9 +331,12 @@ static void assigned_dev_ioport_setup(PCIDevice *pci_dev, int region_num, region->e_size = size;memory_region_init (®ion->container,"assigned-dev-container" , size); +memory_region_set_owner (®ion->container,OBJECT (pci_dev)); +memory_region_init_io (®ion->real_iomem, &assigned_dev_ioport_ops, r_dev->v_addrs + region_num,"assigned-dev-iomem" , size); +memory_region_set_owner (®ion->real_iomem,OBJECT (pci_dev));memory_region_add_subregion (®ion->container,0 , ®ion->real_iomem); }
@@ -482,6 +486,8 @@ static int assigned_dev_register_regions(PCIRegion *io_regions, &slow_bar_ops, &pci_dev->v_addrs[i],"assigned-dev-slow-bar" , cur_region->size); +memory_region_set_owner (&pci_dev->v_addrs[i].real_iomem, +OBJECT (pci_dev)); }else {void *virtbase = pci_dev->v_addrs[i].u.r_virtbase;char name[32 ];
@@ -490,6 +496,9 @@ static int assigned_dev_register_regions(PCIRegion *io_regions,memory_region_init_ram_ptr (&pci_dev->v_addrs[i].real_iomem, name, cur_region->size, virtbase); +memory_region_set_owner (&pci_dev->v_addrs[i].real_iomem, +OBJECT (pci_dev)); +vmstate_register_ram (&pci_dev->v_addrs[i].real_iomem, &pci_dev->dev.qdev); }
@@ -1633,6 +1642,7 @@ static int assigned_dev_register_msix_mmio(AssignedDevice *dev)memory_region_init_io (&dev->mmio, &assigned_dev_msix_mmio_ops, dev,"assigned-dev-msix" ,MSIX_PAGE_SIZE ); +memory_region_set_owner (&dev->mmio,OBJECT (dev));return 0 ; }
@@ -1898,6 +1908,7 @@ static void assigned_dev_load_option_rom(AssignedDevice *dev)snprintf (name,sizeof (name),"%s.rom" ,object_get_typename (OBJECT (dev)));memory_region_init_ram (&dev->dev.rom, name, st.st_size); +memory_region_set_owner (&dev->dev.rom,OBJECT (dev));vmstate_register_ram (&dev->dev.rom, &dev->dev.qdev); ptr =memory_region_get_ram_ptr (&dev->dev.rom);memset (ptr,0xff , st.st_size);
Signature · 1 line
1.7.1
PATCH 34/40 vfio: add memory_region_set_owner calls
In reply to Paolo Bonzini--- hw/misc/vfio.c | 8 ++++++++ 1 files changed, 8 insertions(+), 0 deletions(-)
hw/misc/vfio.c +8 -0
@@ -1156,6 +1156,7 @@ static void vfio_vga_probe_ati_3c3_quirk(VFIODevice *vdev)memory_region_init_io (&quirk->mem, &vfio_ati_3c3_quirk, quirk,"vfio-ati-3c3-quirk" ,1 ); +memory_region_set_owner (&quirk->mem,OBJECT (vdev));memory_region_add_subregion (&vdev->vga.region[QEMU_PCI_VGA_IO_HI ].mem,3 , &quirk->mem);
@@ -1247,6 +1248,7 @@ static void vfio_probe_ati_4010_quirk(VFIODevice *vdev, int nr)memory_region_init_io (&quirk->mem, &vfio_ati_4010_quirk, quirk,"vfio-ati-4010-quirk" ,8 ); +memory_region_set_owner (&quirk->mem,OBJECT (vdev));memory_region_add_subregion_overlap (&vdev->bars[nr].mem,0 , &quirk->mem,1 );QLIST_INSERT_HEAD (&vdev->bars[nr].quirks, quirk, next);
@@ -1333,6 +1335,7 @@ static void vfio_probe_ati_f10_quirk(VFIODevice *vdev, int nr)memory_region_init_io (&quirk->mem, &vfio_ati_f10_quirk, quirk,"vfio-ati-f10-quirk" ,8 ); +memory_region_set_owner (&quirk->mem,OBJECT (vdev));memory_region_add_subregion_overlap (&vdev->bars[nr].mem,0 , &quirk->mem,1 );QLIST_INSERT_HEAD (&vdev->bars[nr].quirks, quirk, next);
@@ -1453,6 +1456,7 @@ static void vfio_vga_probe_nvidia_3d0_quirk(VFIODevice *vdev)memory_region_init_io (&quirk->mem, &vfio_nvidia_3d0_quirk, quirk,"vfio-nvidia-3d0-quirk" ,6 ); +memory_region_set_owner (&quirk->mem,OBJECT (vdev));memory_region_add_subregion (&vdev->vga.region[QEMU_PCI_VGA_IO_HI ].mem,0x10 , &quirk->mem);
@@ -1568,6 +1572,7 @@ static void vfio_probe_nvidia_bar5_window_quirk(VFIODevice *vdev, int nr)memory_region_init_io (&quirk->mem, &vfio_nvidia_bar5_window_quirk, quirk,"vfio-nvidia-bar5-window-quirk" ,16 ); +memory_region_set_owner (&quirk->mem,OBJECT (vdev));memory_region_add_subregion_overlap (&vdev->bars[nr].mem,0 , &quirk->mem,1 );QLIST_INSERT_HEAD (&vdev->bars[nr].quirks, quirk, next);
@@ -1647,6 +1652,7 @@ static void vfio_probe_nvidia_bar0_88000_quirk(VFIODevice *vdev, int nr)memory_region_init_io (&quirk->mem, &vfio_nvidia_bar0_88000_quirk, quirk,"vfio-nvidia-bar0-88000-quirk" ,TARGET_PAGE_ALIGN (PCIE_CONFIG_SPACE_SIZE )); +memory_region_set_owner (&quirk->mem,OBJECT (vdev));memory_region_add_subregion_overlap (&vdev->bars[nr].mem,0x88000 &TARGET_PAGE_MASK , &quirk->mem,1 );
@@ -1726,6 +1732,7 @@ static void vfio_probe_nvidia_bar0_1800_quirk(VFIODevice *vdev, int nr)memory_region_init_io (&quirk->mem, &vfio_nvidia_bar0_1800_quirk, quirk,"vfio-nvidia-bar0-1800-quirk" ,TARGET_PAGE_ALIGN (PCI_CONFIG_SPACE_SIZE )); +memory_region_set_owner (&quirk->mem,OBJECT (vdev));memory_region_add_subregion_overlap (&vdev->bars[nr].mem,0x1800 &TARGET_PAGE_MASK , &quirk->mem,1 );
@@ -2237,6 +2244,7 @@ empty_region:memory_region_init (submem, name,0 ); } +memory_region_set_owner (submem,memory_region_owner (mem));memory_region_add_subregion (mem, offset, submem);return ret;
Signature · 1 line
1.7.1
PATCH 35/40 exec: check MRU in qemu_ram_addr_from_host
In reply to Paolo BonziniThis function is not used outside the iothread mutex, so it can use ram_list.mru_block.
--- exec.c | 12 ++++++++++-- 1 files changed, 10 insertions(+), 2 deletions(-)
exec.c +10 -2
@@ -1412,18 +1412,26 @@ int qemu_ram_addr_from_host(void *ptr, ram_addr_t *ram_addr)return 0 ; } + block = ram_list.mru_block; +if (block && block->host && host - block->host < block->length) { + goto found; + } + QTAILQ_FOREACH(block, &ram_list.blocks, next) {/* This case append when the block is not mapped. */ if (block->host ==NULL ) {continue ; }if (host - block->host < block->length) { - *ram_addr = block->offset + (host - block->host); -return 0 ; + goto found; } }return -1 ; + +found: + *ram_addr = block->offset + (host - block->host); +return 0 ; }/* Some of the softmmu routines need to translate from a host pointer
Signature · 1 line
1.7.1
PATCH 36/40 memory: return MemoryRegion from qemu_ram_addr_from_host
In reply to Paolo BonziniIt will be needed in the next patch.
--- exec.c | 35 +++++++++++++++++++++-------------- include/exec/cpu-common.h | 2 +- target-i386/kvm.c | 4 ++-- 3 files changed, 24 insertions(+), 17 deletions(-)
exec.c +21 -14
@@ -1295,15 +1295,7 @@ void qemu_ram_remap(ram_addr_t addr, ram_addr_t length) }#endif /* !_WIN32 */ -/* Return a host pointer to ram allocated with qemu_ram_alloc. -With the exception of the softmmu code in this file, this should -only be used for local memory (e.g. video ram) that the device owns, -and knows it isn't going to access beyond the end of the block. - -It should not be used for general purpose DMA. -Use cpu_physical_memory_map/cpu_physical_memory_rw instead. -*/ -void *qemu_get_ram_ptr (ram_addr_t addr) +static RAMBlock *qemu_get_ram_block (ram_addr_t addr) {RAMBlock *block;
@@ -1323,6 +1315,21 @@ void *qemu_get_ram_ptr(ram_addr_t addr) found: ram_list.mru_block = block; +return block; +} + +/* Return a host pointer to ram allocated with qemu_ram_alloc. +With the exception of the softmmu code in this file, this should +only be used for local memory (e.g. video ram) that the device owns, +and knows it isn't going to access beyond the end of the block. + +It should not be used for general purpose DMA. +Use cpu_physical_memory_map/cpu_physical_memory_rw instead. +*/ +void *qemu_get_ram_ptr (ram_addr_t addr) +{ +RAMBlock *block =qemu_get_ram_block (addr); +if (xen_enabled ()) {/* We need to check if the requested address is in the RAM * because we don't want to map the entire memory in QEMU.
@@ -1402,14 +1409,14 @@ void qemu_put_ram_ptr(void *addr)trace_qemu_put_ram_ptr (addr); } -int qemu_ram_addr_from_host (void *ptr,ram_addr_t *ram_addr) +MemoryRegion *qemu_ram_addr_from_host (void *ptr,ram_addr_t *ram_addr) {RAMBlock *block;uint8_t *host = ptr;if (xen_enabled ()) { *ram_addr =xen_ram_addr_from_mapcache (ptr); -return 0 ; +return qemu_get_ram_block (*ram_addr)->mr; } block = ram_list.mru_block;
@@ -1427,11 +1434,11 @@ int qemu_ram_addr_from_host(void *ptr, ram_addr_t *ram_addr) } } -return -1 ; +return NULL ; found: *ram_addr = block->offset + (host - block->host); -return 0 ; +return block->mr; }/* Some of the softmmu routines need to translate from a host pointer
@@ -1440,7 +1447,7 @@ ram_addr_t qemu_ram_addr_from_host_nofail(void *ptr) {ram_addr_t ram_addr; -if (qemu_ram_addr_from_host (ptr, &ram_addr)) { +if (qemu_ram_addr_from_host (ptr, &ram_addr) ==NULL ) {fprintf (stderr,"Bad ram pointer %p\n" , ptr);abort (); }
include/exec/cpu-common.h +1 -1
@@ -53,7 +53,7 @@ void qemu_ram_remap(ram_addr_t addr, ram_addr_t length);void *qemu_get_ram_ptr (ram_addr_t addr);void qemu_put_ram_ptr (void *addr);/* This should not be used by devices. */ -int qemu_ram_addr_from_host (void *ptr,ram_addr_t *ram_addr); +MemoryRegion *qemu_ram_addr_from_host (void *ptr,ram_addr_t *ram_addr);ram_addr_t qemu_ram_addr_from_host_nofail (void *ptr);void qemu_ram_set_idstr (ram_addr_t addr,const char *name,DeviceState *dev);
target-i386/kvm.c +2 -2
@@ -318,7 +318,7 @@ int kvm_arch_on_sigbus_vcpu(CPUState *c, int code, void *addr)if ((env->mcg_cap &MCG_SER_P ) && addr && (code ==BUS_MCEERR_AR || code ==BUS_MCEERR_AO )) { -if (qemu_ram_addr_from_host (addr, &ram_addr) || +if (qemu_ram_addr_from_host (addr, &ram_addr) ==NULL || !kvm_physical_memory_addr_from_host (c->kvm_state, addr, &paddr)) {fprintf (stderr,"Hardware memory error for memory used by " "QEMU itself instead of guest system!\n" );
@@ -350,7 +350,7 @@ int kvm_arch_on_sigbus(int code, void *addr)hwaddr paddr;/* Hope we are lucky for AO MCE */ -if (qemu_ram_addr_from_host (addr, &ram_addr) || +if (qemu_ram_addr_from_host (addr, &ram_addr) ==NULL || !kvm_physical_memory_addr_from_host (CPU (first_cpu)->kvm_state, addr, &paddr)) { fprintf(stderr, "Hardware memory error for memory used by "
Signature · 1 line
1.7.1
PATCH 37/40 memory: ref/unref memory across address_space_map/unmap
In reply to Paolo BonziniThe iothread mutex might be released between map and unmap, so the mapped region might disappear.
--- exec.c | 12 +++++++++++- 1 files changed, 11 insertions(+), 1 deletions(-)
exec.c +11 -1
@@ -2085,6 +2085,7 @@ void cpu_physical_memory_write_rom(hwaddr addr, } typedef struct { + MemoryRegion *mr; void *buffer; hwaddr addr; hwaddr len;
@@ -2179,15 +2180,18 @@ void *address_space_map(AddressSpace *as, bounce.buffer =qemu_memalign (TARGET_PAGE_SIZE ,TARGET_PAGE_SIZE ); bounce.addr = addr; bounce.len = l; + bounce.mr = section->mr;if (!is_write) {address_space_read (as, addr, bounce.buffer, l); } *plen = l; +memory_region_ref (section->mr);return bounce.buffer; }if (!todo) { raddr =memory_region_get_ram_addr (section->mr) + xlat; +memory_region_ref (section->mr); }else {if (memory_region_get_ram_addr (section->mr) + xlat != raddr + todo) {break ;
@@ -2212,8 +2216,12 @@ void address_space_unmap(AddressSpace *as, void *buffer, hwaddr len, int is_write, hwaddr access_len) {if (buffer != bounce.buffer) { +MemoryRegion *mr; +ram_addr_t addr1; + + mr =qemu_ram_addr_from_host (buffer, &addr1); +assert (mr);if (is_write) { -ram_addr_t addr1 =qemu_ram_addr_from_host_nofail (buffer);while (access_len) {unsigned l; l =TARGET_PAGE_SIZE ;
@@ -2227,6 +2235,7 @@ void address_space_unmap(AddressSpace *as, void *buffer, hwaddr len,if (xen_enabled ()) {xen_invalidate_map_cache_entry (buffer); } +memory_region_unref (mr);return ; }if (is_write) {
@@ -2234,6 +2243,7 @@ void address_space_unmap(AddressSpace *as, void *buffer, hwaddr len, }qemu_vfree (bounce.buffer); bounce.buffer =NULL ; +memory_region_unref (bounce.mr);cpu_notify_map_clients (); }
Signature · 1 line
1.7.1
PATCH 38/40 memory: access FlatView from a local variable
In reply to Paolo BonziniWe will soon require accesses to as->current_map to be placed under a lock (with reference counting so as to keep the critical section small). To simplify this change, always fetch as->current_map into a local variable and access it through that variable.
--- memory.c | 31 +++++++++++++++++++++---------- 1 files changed, 21 insertions(+), 10 deletions(-)
memory.c +21 -10
@@ -632,13 +632,15 @@ static void address_space_add_del_ioeventfds(AddressSpace *as,static void address_space_update_ioeventfds (AddressSpace *as) { +FlatView *view;FlatRange *fr;unsigned ioeventfd_nb =0 ;MemoryRegionIoeventfd *ioeventfds =NULL ;AddrRange tmp;unsigned i; - FOR_EACH_FLAT_RANGE(fr, as->current_map) { + view = as->current_map; + FOR_EACH_FLAT_RANGE(fr, view) {for (i =0 ; i < fr->mr->ioeventfd_nb; ++i) { tmp = addrrange_shift(fr->mr->ioeventfds[i].addr, int128_sub(fr->addr.start,
@@ -1169,7 +1171,8 @@ void memory_region_sync_dirty_bitmap(MemoryRegion *mr)FlatRange *fr; QTAILQ_FOREACH(as, &address_spaces, address_spaces_link) { - FOR_EACH_FLAT_RANGE(fr, as->current_map) { +FlatView *view = as->current_map; + FOR_EACH_FLAT_RANGE(fr, view) {if (fr->mr == mr) {MEMORY_LISTENER_UPDATE_REGION (fr, as, Forward, log_sync); }
@@ -1219,12 +1222,14 @@ void *memory_region_get_ram_ptr(MemoryRegion *mr)static void memory_region_update_coalesced_range_as (MemoryRegion *mr,AddressSpace *as) { +FlatView *view;FlatRange *fr;CoalescedMemoryRange *cmr;AddrRange tmp;MemoryRegionSection section; - FOR_EACH_FLAT_RANGE(fr, as->current_map) { + view = as->current_map; + FOR_EACH_FLAT_RANGE(fr, view) {if (fr->mr == mr) { section = (MemoryRegionSection ) { .address_space = as,
@@ -1511,9 +1516,9 @@ static int cmp_flatrange_addr(const void *addr_, const void *fr_)return 0 ; } -static FlatRange *address_space_lookup (AddressSpace *as,AddrRange addr) +static FlatRange *flatview_lookup (FlatView *view,AddrRange addr) { -return bsearch (&addr, as->current_map->ranges, as->current_map->nr, +return bsearch (&addr, view->ranges, view->nr,sizeof (FlatRange), cmp_flatrange_addr); }
@@ -1524,6 +1529,7 @@ MemoryRegionSection memory_region_find(MemoryRegion *mr,MemoryRegion *root;AddressSpace *as;AddrRange range; +FlatView *view;FlatRange *fr; addr += mr->addr;
@@ -1534,13 +1540,14 @@ MemoryRegionSection memory_region_find(MemoryRegion *mr, as =memory_region_to_address_space (root); range =addrrange_make (int128_make64 (addr),int128_make64 (size)); - fr =address_space_lookup (as, range); + + view = as->current_map; + fr =flatview_lookup (view, range);if (!fr) {return ret; } -while (fr > as->current_map->ranges - &&addrrange_intersects (fr[-1 ].addr, range)) { +while (fr > view->ranges &&addrrange_intersects (fr[-1 ].addr, range)) { --fr; }
@@ -1560,9 +1567,11 @@ MemoryRegionSection memory_region_find(MemoryRegion *mr,void address_space_sync_dirty_bitmap (AddressSpace *as) { +FlatView *view;FlatRange *fr; - FOR_EACH_FLAT_RANGE(fr, as->current_map) { + view = as->current_map; + FOR_EACH_FLAT_RANGE(fr, view) {MEMORY_LISTENER_UPDATE_REGION (fr, as, Forward, log_sync); } }
@@ -1582,6 +1591,7 @@ void memory_global_dirty_log_stop(void)static void listener_add_address_space (MemoryListener *listener,AddressSpace *as) { +FlatView *view;FlatRange *fr; if (listener->address_space_filter
@@ -1595,7 +1605,8 @@ static void listener_add_address_space(MemoryListener *listener, } } - FOR_EACH_FLAT_RANGE(fr, as->current_map) { + view = as->current_map; + FOR_EACH_FLAT_RANGE(fr, view) {MemoryRegionSection section = { .mr = fr->mr, .address_space = as,
Signature · 1 line
1.7.1
PATCH 39/40 memory: use a new FlatView pointer on every topology update
In reply to Paolo BonziniThis is the first step towards converting as->current_map to RCU-style updates, where the FlatView updates run concurrently with uses of an old FlatView.
--- memory.c | 34 ++++++++++++++++++---------------- 1 files changed, 18 insertions(+), 16 deletions(-)
memory.c +18 -16
@@ -273,6 +273,7 @@ static void flatview_destroy(FlatView *view)memory_region_unref (view->ranges[i].mr); }g_free (view->ranges); +g_free (view); }static bool can_merge (FlatRange *r1,FlatRange *r2)
@@ -566,17 +567,18 @@ static void render_memory_region(FlatView *view, }/* Render a memory topology into a list of disjoint absolute ranges. */ -static FlatView generate_memory_topology (MemoryRegion *mr) +static FlatView *generate_memory_topology (MemoryRegion *mr) { -FlatView view; +FlatView *view; -flatview_init (&view); + view =g_new (FlatView,1 ); +flatview_init (view);if (mr) { -render_memory_region (&view, mr,int128_zero (), +render_memory_region (view, mr,int128_zero (),addrrange_make (int128_zero (),int128_2_64 ()), false); } -flatview_simplify (&view); +flatview_simplify (view);return view; }
@@ -664,8 +666,8 @@ static void address_space_update_ioeventfds(AddressSpace *as) }static void address_space_update_topology_pass (AddressSpace *as, -FlatView old_view, -FlatView new_view, +const FlatView *old_view, +const FlatView *new_view,bool adding) {unsigned iold, inew;
@@ -675,14 +677,14 @@ static void address_space_update_topology_pass(AddressSpace *as,* Kill ranges in the old map, and instantiate ranges in the new map. */ iold = inew =0 ; -while (iold < old_view.nr || inew < new_view.nr) { -if (iold < old_view.nr) { - frold = &old_view.ranges[iold]; +while (iold < old_view->nr || inew < new_view->nr) { +if (iold < old_view->nr) { + frold = &old_view->ranges[iold]; }else { frold =NULL ; } -if (inew < new_view.nr) { - frnew = &new_view.ranges[inew]; +if (inew < new_view->nr) { + frnew = &new_view->ranges[inew]; }else { frnew =NULL ; }
@@ -728,14 +730,14 @@ static void address_space_update_topology_pass(AddressSpace *as,static void address_space_update_topology (AddressSpace *as) { -FlatView old_view = *as->current_map; -FlatView new_view =generate_memory_topology (as->root); +FlatView *old_view = as->current_map; +FlatView *new_view =generate_memory_topology (as->root);address_space_update_topology_pass (as, old_view, new_view, false);address_space_update_topology_pass (as, old_view, new_view, true); - *as->current_map = new_view; -flatview_destroy (&old_view); + as->current_map = new_view; +flatview_destroy (old_view);address_space_update_ioeventfds (as); }
Signature · 1 line
1.7.1
PATCH 40/40 memory: add reference counting to FlatView
In reply to Paolo BonziniWith this change, a FlatView can be used even after a concurrent update has replaced it. Because we do not have RCU, we use a mutex to protect the small critical sections that read/write the as->current_map pointer. Accesses to the FlatView can be done outside the mutex.
If a MemoryRegion will be used after the FlatView is unref-ed (or after a MemoryListener callback is returned), a reference has to be added to that MemoryRegion. For example, memory_region_find adds a reference to the MemoryRegion that it returns.
--- memory.c | 75 +++++++++++++++++++++++++++++++++++++++++++++++++++++++------ 1 files changed, 67 insertions(+), 8 deletions(-)
memory.c +67 -8
@@ -26,12 +26,26 @@ static unsigned memory_region_transaction_depth;static bool memory_region_update_pending;static bool global_dirty_log = false; +/* flat_view_mutex is taken around reading as->current_map; the critical +* section is extremely short, so I'm using a single mutex for every AS. +* We could also RCU for the read-side. +* +* The BQL is taken around transaction commits, hence both locks are taken +* while writing to as->current_map (with the BQL taken outside). +*/ +static QemuMutex flat_view_mutex; + static QTAILQ_HEAD(memory_listeners, MemoryListener) memory_listeners =QTAILQ_HEAD_INITIALIZER (memory_listeners); static QTAILQ_HEAD(, AddressSpace) address_spaces =QTAILQ_HEAD_INITIALIZER (address_spaces); +static void memory_init (void ) +{ +qemu_mutex_init (&flat_view_mutex); +} +typedef struct AddrRange AddrRange ;/*
@@ -222,6 +236,7 @@ struct FlatRange {* order. */ struct FlatView { +unsigned ref;FlatRange *ranges;unsigned nr;unsigned nr_allocated;
@@ -243,6 +258,7 @@ static bool flatrange_equal(FlatRange *a, FlatRange *b)static void flatview_init (FlatView *view) { + view->ref =1 ; view->ranges =NULL ; view->nr =0 ; view->nr_allocated =0 ;
@@ -276,6 +292,18 @@ static void flatview_destroy(FlatView *view)g_free (view); } +static void flatview_ref (FlatView *view) +{ +__sync_fetch_and_add (&view->ref,1 ); +} + +static void flatview_unref (FlatView *view) +{ +if (__sync_fetch_and_sub (&view->ref,1 ) ==1 ) { +flatview_destroy (view); + } +} +static bool can_merge (FlatRange *r1,FlatRange *r2) {return int128_eq (addrrange_end (r1->addr), r2->addr.start)
@@ -728,16 +756,38 @@ static void address_space_update_topology_pass(AddressSpace *as, } +static FlatView *address_space_get_flatview (AddressSpace *as) +{ +FlatView *view; + +qemu_mutex_lock (&flat_view_mutex); + view = as->current_map; +flatview_ref (view); +qemu_mutex_unlock (&flat_view_mutex); +return view; +} +static void address_space_update_topology (AddressSpace *as) { -FlatView *old_view = as->current_map; +FlatView *old_view =address_space_get_flatview (as);FlatView *new_view =generate_memory_topology (as->root);address_space_update_topology_pass (as, old_view, new_view, false);address_space_update_topology_pass (as, old_view, new_view, true); +qemu_mutex_lock (&flat_view_mutex); +flatview_unref (as->current_map); as->current_map = new_view; -flatview_destroy (old_view); +qemu_mutex_unlock (&flat_view_mutex); + +/* Note that all the old MemoryRegions are still alive up to this +* point. This relieves most MemoryListeners from the need to +* ref/unref the MemoryRegions they get---unless they use them +* outside the iothread mutex, in which case precise reference +* counting is necessary. +*/ +flatview_unref (old_view); +address_space_update_ioeventfds (as); }
@@ -1173,12 +1223,13 @@ void memory_region_sync_dirty_bitmap(MemoryRegion *mr)FlatRange *fr; QTAILQ_FOREACH(as, &address_spaces, address_spaces_link) { -FlatView *view = as->current_map; +FlatView *view =address_space_get_flatview (as); FOR_EACH_FLAT_RANGE(fr, view) {if (fr->mr == mr) {MEMORY_LISTENER_UPDATE_REGION (fr, as, Forward, log_sync); } } +flatview_unref (view); } }
@@ -1230,7 +1281,7 @@ static void memory_region_update_coalesced_range_as(MemoryRegion *mr, AddressSpaAddrRange tmp;MemoryRegionSection section; - view = as->current_map; + view =address_space_get_flatview (as); FOR_EACH_FLAT_RANGE(fr, view) {if (fr->mr == mr) { section = (MemoryRegionSection) {
@@ -1256,6 +1307,7 @@ static void memory_region_update_coalesced_range_as(MemoryRegion *mr, AddressSpa } } } +flatview_unref (view); }static void memory_region_update_coalesced_range (MemoryRegion *mr)
@@ -1543,7 +1595,7 @@ MemoryRegionSection memory_region_find(MemoryRegion *mr, as =memory_region_to_address_space (root); range =addrrange_make (int128_make64 (addr),int128_make64 (size)); - view = as->current_map; + view =address_space_get_flatview (as); fr =flatview_lookup (view, range);if (!fr) {return ret;
@@ -1564,6 +1616,7 @@ MemoryRegionSection memory_region_find(MemoryRegion *mr, ret.readonly = fr->readonly;memory_region_ref (ret.mr); +flatview_unref (view);return ret; }
@@ -1572,10 +1625,11 @@ void address_space_sync_dirty_bitmap(AddressSpace *as)FlatView *view;FlatRange *fr; - view = as->current_map; + view =address_space_get_flatview (as); FOR_EACH_FLAT_RANGE(fr, view) {MEMORY_LISTENER_UPDATE_REGION (fr, as, Forward, log_sync); } +flatview_unref (view); }void memory_global_dirty_log_start (void )
@@ -1607,7 +1661,7 @@ static void listener_add_address_space(MemoryListener *listener, } } - view = as->current_map; + view =address_space_get_flatview (as); FOR_EACH_FLAT_RANGE(fr, view) {MemoryRegionSection section = { .mr = fr->mr,
@@ -1621,6 +1675,7 @@ static void listener_add_address_space(MemoryListener *listener, listener->region_add (listener, §ion); } } +flatview_unref (view); }void memory_listener_register (MemoryListener *listener,AddressSpace *filter)
@@ -1654,6 +1709,10 @@ void memory_listener_unregister(MemoryListener *listener)void address_space_init (AddressSpace *as,MemoryRegion *root,const char *name) { +if (QTAILQ_EMPTY (&address_spaces)) { +memory_init (); + } +memory_region_transaction_begin (); as->root = root; as->current_map =g_new (FlatView,1 );
@@ -1675,7 +1734,7 @@ void address_space_destroy(AddressSpace *as)memory_region_transaction_commit ();QTAILQ_REMOVE (&address_spaces, as, address_spaces_link);address_space_destroy_dispatch (as); -flatview_destroy (as->current_map); +flatview_unref (as->current_map);g_free (as->name);g_free (as->current_map);g_free (as->ioeventfds);
Signature · 1 line
1.7.1
Re: [Qemu-devel] [PATCH 40/40] memory: add reference counting to FlatView
Quoted 10 lines · Paolo Bonzini
With this change, a FlatView can be used even after a concurrent update has replaced it. Because we do not have RCU, we use a mutex to protect the small critical sections that read/write the
as->current_mappointer. Accesses to the FlatView can be done outside the mutex.If a MemoryRegion will be used after the FlatView is unref-ed (or after a MemoryListener callback is returned), a reference has to be added to that MemoryRegion. For example,
memory_region_findadds a reference to the MemoryRegion that it returns.
For my understanding: Every lookup, e.g. triggered by address_space_rw, will briefly reference the FlatView of that address space and drop that reference again after referencing the target memory region.
Provided that is true: If we run that lookup on an address space that happens to be modified in parallel, the lookup may actually cause the final deref and, thus, release of the FlatView - even if the target memory region was totally unrelated to the ongoing change. That could make a hot-path pay the price of an action a slow path caused. Not really a beautiful concept. Even if the FlatView finalization is a simple free() (that is the minimum), we would pull the memory allocator into code paths that might try hard to keep a safe distance for the sake of predictability.
Jan
Signature · 2 lines
Siemens AG, Corporate Technology, CT RTC ITP SDP-DE Corporate Competence Center Embedded Linux
Re: [Qemu-devel] [PATCH 40/40] memory: add reference counting to FlatView
Quoted 16 lines · Jan Kiszka
Quoted 10 lines · Paolo Bonzini
With this change, a FlatView can be used even after a concurrent update has replaced it. Because we do not have RCU, we use a mutex to protect the small critical sections that read/write the
as->current_mappointer. Accesses to the FlatView can be done outside the mutex.If a MemoryRegion will be used after the FlatView is unref-ed (or after a MemoryListener callback is returned), a reference has to be added to that MemoryRegion. For example,
memory_region_findadds a reference to the MemoryRegion that it returns.For my understanding: Every lookup, e.g. triggered by
address_space_rw, will briefly reference the FlatView of that address space and drop that reference again after referencing the target memory region.
Quoted 9 lines
Provided that is true: If we run that lookup on an address space that happens to be modified in parallel, the lookup may actually cause the final deref and, thus, release of the FlatView - even if the target memory region was totally unrelated to the ongoing change. That could make a hot-path pay the price of an action a slow path caused. Not really a beautiful concept. Even if the FlatView finalization is a simple
free()(that is the minimum), we would pull the memory allocator into code paths that might try hard to keep a safe distance for the sake of predictability.
A mitigation could be to off-load the finalization to a bottom-half e.g., ie. let the main iothread do the dirty work. Similar to the RCU handler off-loading of the kernel. There is just a theoretical risk that we pile up finalization requests and run out of memory. But all that is pretty complicated and not really elegant. Better avoid the ref/unref over hot-paths for objects of such broad scope like the FlatView.
Jan
Signature · 2 lines
Siemens AG, Corporate Technology, CT RTC ITP SDP-DE Corporate Competence Center Embedded Linux
Re: [Qemu-devel] [PATCH 40/40] memory: add reference counting to FlatView
In reply to Jan KiszkaIl 07/05/2013 20:00, Jan Kiszka ha scritto:
Quoted 16 lines · Paolo Bonzini
Quoted 10 lines
With this change, a FlatView can be used even after a concurrent update has replaced it. Because we do not have RCU, we use a mutex to protect the small critical sections that read/write the
as->current_mappointer. Accesses to the FlatView can be done outside the mutex.If a MemoryRegion will be used after the FlatView is unref-ed (or after a MemoryListener callback is returned), a reference has to be added to that MemoryRegion. For example,
memory_region_findadds a reference to the MemoryRegion that it returns.For my understanding: Every lookup, e.g. triggered by
address_space_rw, will briefly reference the FlatView of that address space and drop that reference again after referencing the target memory region.
Quoted 9 lines
Provided that is true: If we run that lookup on an address space that happens to be modified in parallel, the lookup may actually cause the final deref and, thus, release of the FlatView - even if the target memory region was totally unrelated to the ongoing change. That could make a hot-path pay the price of an action a slow path caused. Not really a beautiful concept. Even if the FlatView finalization is a simple
free()(that is the minimum), we would pull the memory allocator into code paths that might try hard to keep a safe distance for the sake of predictability.
All this is correct. But I hope to get RCU in 1.6, otherwise we can use a bottom half. In any case, this is obviously no worse than current code that requires the BQL (hence the lookup would serialize against the free).
Paolo
Re: [Qemu-devel] [PATCH 40/40] memory: add reference counting to FlatView
Quoted 27 lines · Paolo Bonzini
Il 07/05/2013 20:00, Jan Kiszka ha scritto:
Quoted 16 lines · Paolo Bonzini
Quoted 10 lines
With this change, a FlatView can be used even after a concurrent update has replaced it. Because we do not have RCU, we use a mutex to protect the small critical sections that read/write the
as->current_mappointer. Accesses to the FlatView can be done outside the mutex.If a MemoryRegion will be used after the FlatView is unref-ed (or after a MemoryListener callback is returned), a reference has to be added to that MemoryRegion. For example,
memory_region_findadds a reference to the MemoryRegion that it returns.For my understanding: Every lookup, e.g. triggered by
address_space_rw, will briefly reference the FlatView of that address space and drop that reference again after referencing the target memory region.Quoted 9 lines
Provided that is true: If we run that lookup on an address space that happens to be modified in parallel, the lookup may actually cause the final deref and, thus, release of the FlatView - even if the target memory region was totally unrelated to the ongoing change. That could make a hot-path pay the price of an action a slow path caused. Not really a beautiful concept. Even if the FlatView finalization is a simple
free()(that is the minimum), we would pull the memory allocator into code paths that might try hard to keep a safe distance for the sake of predictability.
Quoted 4 lines
All this is correct. But I hope to get RCU in 1.6, otherwise we can use a bottom half. In any case, this is obviously no worse than current code that requires the BQL (hence the lookup would serialize against the free).
Sure, not a regression. But I'd like to avoid that we are moving in the potentially wrong direction /wrt final goal. And I'm very interested in having increments that can be used in RT demonstration scenarios without having to hack too much on the code.
Jan
Signature · 2 lines
Siemens AG, Corporate Technology, CT RTC ITP SDP-DE Corporate Competence Center Embedded Linux