fw_cfg_mem: add read memory region callback
PATCH 1/1 fw_cfg_mem: add read memory region callback
The write/read should be paired, this can avoid the NULL-deref while the guest reads the fw_cfg port.
--- hw/nvram/fw_cfg.c | 6 ++++++ 1 file changed, 6 insertions(+)
hw/nvram/fw_cfg.c +6 -0
@@ -434,6 +434,11 @@ static bool fw_cfg_data_mem_valid(void *opaque, hwaddr addr,return addr ==0 ; } +static uint64_t fw_cfg_ctl_mem_read (void *opaque,hwaddr addr,unsigned size) +{ +return 0 ; +} +static void fw_cfg_ctl_mem_write (void *opaque,hwaddr addr,uint64_t value,unsigned size) {
@@ -468,6 +473,7 @@ static bool fw_cfg_comb_valid(void *opaque, hwaddr addr, }static const MemoryRegionOps fw_cfg_ctl_mem_ops = { + .read = fw_cfg_ctl_mem_read, .write = fw_cfg_ctl_mem_write, .endianness =DEVICE_BIG_ENDIAN , .valid.accepts = fw_cfg_ctl_mem_valid,
Signature · 1 line
2.11.0
Hi
Quoted 5 lines · Li Qiang
The write/read should be paired, this can avoid the NULL-deref while the guest reads the
fw_cfgport.Signed-off-by Li Qiang <[email protected]>
Do you have a reproducer and/or a backtrace? memory_region_dispatch_write() checks if ops->write != NULL.
Quoted 32 lines
--- hw/nvram/fw_cfg.c | 6 ++++++ 1 file changed, 6 insertions(+)hw/nvram/fw_cfg.c +6
-1@@ -434,6 +434,11 @@ static bool fw_cfg_data_mem_valid(void *opaque, hwaddr addr,return addr ==0 ; } +static uint64_t fw_cfg_ctl_mem_read (void *opaque,hwaddr addr,unsigned size) +{ +return 0 ; +} +static void fw_cfg_ctl_mem_write (void *opaque,hwaddr addr,uint64_t value,unsigned size) {@@ -468,6 +473,7 @@ static bool fw_cfg_comb_valid(void *opaque, hwaddr addr, }static const MemoryRegionOps fw_cfg_ctl_mem_ops = { + .read = fw_cfg_ctl_mem_read, .write = fw_cfg_ctl_mem_write, .endianness =DEVICE_BIG_ENDIAN , .valid.accepts = fw_cfg_ctl_mem_valid, --2.11.0
Signature · 1 line
Marc-André Lureau
Hi,
Quoted 4 lines · Marc-André Lureau <[email protected]> 于2018年9月12日周三 下午3:16写道:
Hi
Quoted 1 line · Li Qiang
Quoted 8 lines
Quoted 4 lines
The write/read should be paired, this can avoid the NULL-deref while the guest reads the
fw_cfgport.Signed-off-by Li Qiang <[email protected]>Do you have a reproducer and/or a backtrace?
memory_region_dispatch_write()checks ifops->write!= NULL.
As far as I can see, the fw_cfg_mem is not used in x86 and used in arm. And from my impression, this will NULL read will be a issue. So I just omit the 'read' field in fw_cfg_comb_mem_ops(this is used for x86) to emulate the
issue.
When using gdb, I got the following backtrack.
Thread 5 "qemu-system-x86" received signal SIGSEGV, Segmentation fault. [Switching to Thread 0x7fffca15b700 (LWP 7637)] 0x0000000000000000 in ?? ()
(gdb) bt
#0 0x0000000000000000 in ?? ()
#1 0x0000555555872492 in memory_region_oldmmio_read_accessor (mr=0x5555567fa870, addr=1, value=0x7fffca158510, size=1, shift=0, mask=255, attrs=...) at /home/liqiang02/qemu-devel/qemu/memory.c:409 #2 0x000055555586f2dd in access_with_adjusted_size (addr=addr@entry=1, value=value@entry=0x7fffca158510, size=size@entry=1, access_size_min=<optimized out>, access_size_max=<optimized out>,
access_fn=0x555555872440 <memory_region_oldmmio_read_accessor>,
mr=0x5555567fa870, attrs=...) at /home/liqiang02/qemu-devel/qemu/memory.c:593
#3 0x0000555555873e90 in memory_region_dispatch_read1 (attrs=..., size=1, pval=0x7fffca158510, addr=1, mr=0x5555567fa870) at /home/liqiang02/qemu-devel/qemu/memory.c:1404
#4 memory_region_dispatch_read (mr=mr@entry=0x5555567fa870, addr=1, pval=pval@entry=0x7fffca158510, size=1, attrs=attrs@entry=...) at /home/liqiang02/qemu-devel/qemu/memory.c:1423
#5 0x0000555555821e42 in flatview_read_continue (fv=fv@entry=0x7fffbc03f370, addr=addr@entry=1297, attrs=..., buf=<optimized out>, buf@entry=0x7ffff7fee000 "", len=len@entry=1, addr1=<optimized out>, l=<optimized out>, mr=0x5555567fa870)
at /home/liqiang02/qemu-devel/qemu/exec.c:3293
#6 0x0000555555822006 in flatview_read (fv=0x7fffbc03f370, addr=1297, attrs=..., buf=0x7ffff7fee000 "", len=1) at /home/liqiang02/qemu-devel/qemu/exec.c:3331
#7 0x000055555582211f in address_space_read_full (as=<optimized out>, addr=addr@entry=1297, attrs=..., buf=buf@entry=0x7ffff7fee000 "", len=len@entry=1) at /home/liqiang02/qemu-devel/qemu/exec.c:3344 #8 0x000055555582225a in address_space_rw (as=<optimized out>, addr=addr@entry=1297, attrs=..., attrs@entry=..., buf=buf@entry=0x7ffff7fee000 "", len=len@entry=1, is_write=is_write@entry=false)
at /home/liqiang02/qemu-devel/qemu/exec.c:3374
#9 0x0000555555886239 in kvm_handle_io (count=1, size=1, direction=<optimized out>, data=<optimized out>, attrs=..., port=1297) at /home/liqiang02/qemu-devel/qemu/accel/kvm/kvm-all.c:1731
#10 kvm_cpu_exec (cpu=cpu@entry=0x5555566e9990) at /home/liqiang02/qemu-devel/qemu/accel/kvm/kvm-all.c:1971
#11 0x000055555585d3de in qemu_kvm_cpu_thread_fn (arg=0x5555566e9990) at /home/liqiang02/qemu-devel/qemu/cpus.c:1257
#12 0x00007fffdbd58494 in start_thread () from /lib/x86_64-linux-gnu/libpthread.so.0
#13 0x00007fffdba9aacf in clone () from /lib/x86_64-linux-gnu/libc.so.6
So I send out this path.
But this time when I not use gdb, there is no segmentation.
So this may not a issue.
If who has a arm environment, he can try this, the PoC is just easy. just read the 0x510 port with word.
Thanks,
Li Qiang
Quoted 41 lines
hw/nvram/fw_cfg.c static bool fw_cfg_data_mem_valid(void *opaque,
Quoted 10 lines
--- hw/nvram/fw_cfg.c | 6 ++++++ 1 file changed, 6 insertions(+)hw/nvram/fw_cfg.c +0
-0@@ -434,6 +434,11 @@ static bool fw_cfg_data_mem_valid(void *opaque,hwaddr addr,
hw/nvram/fw_cfg.c static bool fw_cfg_data_mem_valid(void *opaque, hwaddr addr,
Quoted 4 lines
hw/nvram/fw_cfg.c +1
-0return addr ==0 ; } +static uint64_t fw_cfg_ctl_mem_read(void *opaque, hwaddr addr, unsignedsize)
hw/nvram/fw_cfg.c static bool fw_cfg_comb_valid(void *opaque, hwaddr
Quoted 8 lines
hw/nvram/fw_cfg.c +4
-0+{ +return 0 ; +} +static void fw_cfg_ctl_mem_write (void *opaque,hwaddr addr,uint64_t value,unsigned size) {@@ -468,6 +473,7 @@ static bool fw_cfg_comb_valid(void *opaque, hwaddraddr,
Quoted 11 lines
hw/nvram/fw_cfg.c +1
-1}static const MemoryRegionOps fw_cfg_ctl_mem_ops = { + .read = fw_cfg_ctl_mem_read, .write = fw_cfg_ctl_mem_write, .endianness =DEVICE_BIG_ENDIAN , .valid.accepts = fw_cfg_ctl_mem_valid, --2.11.0
--
Marc-André Lureau
Quoted 71 lines · Li Qiang
Hi,
Quoted 4 lines · Marc-André Lureau <[email protected]> 于2018年9月12日周三 下午3:16写道:
Hi
Quoted 1 line · Li Qiang
Quoted 8 lines
Quoted 4 lines
The write/read should be paired, this can avoid the NULL-deref while the guest reads the
fw_cfgport.Signed-off-by Li Qiang <[email protected]>Do you have a reproducer and/or a backtrace?
memory_region_dispatch_write()checks ifops->write!= NULL.As far as I can see, the
fw_cfg_memis not used in x86 and used in arm. And from my impression, this will NULL read will be a issue. So I just omit the 'read' field infw_cfg_comb_mem_ops(this is used for x86)to emulate the
issue.When using gdb, I got the following backtrack.
Thread 5 "qemu-system-x86" received signal SIGSEGV, Segmentation fault. [Switching to Thread
0x7fffca15b700(LWP 7637)]0x0000000000000000in ?? ()
(gdb) bt
#00x0000000000000000in ?? ()
#10x0000555555872492inmemory_region_oldmmio_read_accessor(mr=0x5555567fa870, addr=1, value=0x7fffca158510, size=1, shift=0, mask=255, attrs=...) at/home/liqiang02/qemu-devel/qemu/memory.c:409#20x000055555586f2ddinaccess_with_adjusted_size(addr=addr@entry=1, value=value@entry=0x7fffca158510, size=size@entry=1, access_size_min=<optimized out>, access_size_max=<optimized out>,access_fn=0x555555872440 <memory_region_oldmmio_read_accessor>,mr=0x5555567fa870, attrs=...) at
/home/liqiang02/qemu-devel/qemu/memory.c:593
#30x0000555555873e90inmemory_region_dispatch_read1(attrs=..., size=1, pval=0x7fffca158510, addr=1, mr=0x5555567fa870) at/home/liqiang02/qemu-devel/qemu/memory.c:1404
#4memory_region_dispatch_read(mr=mr@entry=0x5555567fa870, addr=1, pval=pval@entry=0x7fffca158510, size=1, attrs=attrs@entry=...) at/home/liqiang02/qemu-devel/qemu/memory.c:1423
#50x0000555555821e42inflatview_read_continue(fv=fv@entry=0x7fffbc03f370, addr=addr@entry=1297, attrs=..., buf=<optimized out>, buf@entry=0x7ffff7fee000 "", len=len@entry=1, addr1=<optimized out>, l=<optimized out>, mr=0x5555567fa870)at /home/liqiang02/qemu-devel/qemu/exec.c:3293#6
0x0000555555822006inflatview_read(fv=0x7fffbc03f370, addr=1297, attrs=..., buf=0x7ffff7fee000 "", len=1) at/home/liqiang02/qemu-devel/qemu/exec.c:3331
#70x000055555582211finaddress_space_read_full(as=<optimized out>, addr=addr@entry=1297, attrs=..., buf=buf@entry=0x7ffff7fee000 "", len=len@entry=1) at/home/liqiang02/qemu-devel/qemu/exec.c:3344#80x000055555582225ainaddress_space_rw(as=<optimized out>, addr=addr@entry=1297, attrs=..., attrs@entry=..., buf=buf@entry=0x7ffff7fee000 "", len=len@entry=1, is_write=is_write@entry=false)at /home/liqiang02/qemu-devel/qemu/exec.c:3374#9
0x0000555555886239inkvm_handle_io(count=1, size=1, direction=<optimized out>, data=<optimized out>, attrs=..., port=1297) at/home/liqiang02/qemu-devel/qemu/accel/kvm/kvm-all.c:1731
#10kvm_cpu_exec(cpu=cpu@entry=0x5555566e9990) at/home/liqiang02/qemu-devel/qemu/accel/kvm/kvm-all.c:1971
#110x000055555585d3deinqemu_kvm_cpu_thread_fn(arg=0x5555566e9990) at/home/liqiang02/qemu-devel/qemu/cpus.c:1257
#120x00007fffdbd58494instart_thread() from/lib/x86_64-linux-gnu/libpthread.so.0
#130x00007fffdba9aacfin clone () from/lib/x86_64-linux-gnu/libc.so.6
Quoted 8 lines
So I send out this path.
But this time when I not use gdb, there is no segmentation.
So this may not a issue.
If who has a arm environment, he can try this, the PoC is just easy. just read the
0x510port with word.
I don't understand your description of how you managed to trigger this SIGSEGV.
FWIW, looking at the codebase, there's a good number of static MemoryRegionOps structures for which the "read_with_attrs" and "read" members are default-initialized to NULL. It seems unlikely they are all wrong.
- exec.c: notdirty_mem_ops, readonly_mem_ops - hw/misc/debugexit.c: debug_exit_ops - hw/misc/hyperv_testdev.c: synic_test_sint_ops - hw/misc/pc-testdev.c: test_irq_ops, test_flush_ops - hw/pci-host/designware.c: designware_pci_host_msi_ops
- hw/rdma/vmw/pvrdma_main.c: uar_ops
- hw/sparc64/sun4u.c: power_mem_ops
Laszlo
Hi Laszlo,
Quoted 85 lines · Laszlo Ersek <[email protected]> 于2018年9月12日周三 下午6:36写道:
Quoted 36 lines · Li Qiang
Hi,
Quoted 4 lines · Marc-André Lureau <[email protected]> 于2018年9月12日周三 下午3:16写道:
Hi
Quoted 1 line · Li Qiang
Quoted 8 lines
Quoted 4 lines
The write/read should be paired, this can avoid the NULL-deref while the guest reads the
fw_cfgport.Signed-off-by Li Qiang <[email protected]>Do you have a reproducer and/or a backtrace?
memory_region_dispatch_write()checks ifops->write!= NULL.As far as I can see, the
fw_cfg_memis not used in x86 and used in arm. And from my impression, this will NULL read will be a issue. So I just omit the 'read' field infw_cfg_comb_mem_ops(this is used for x86)to emulate the
issue.When using gdb, I got the following backtrack.
Thread 5 "qemu-system-x86" received signal SIGSEGV, Segmentation fault. [Switching to Thread
0x7fffca15b700(LWP 7637)]0x0000000000000000in ?? ()
(gdb) bt
#00x0000000000000000in ?? ()
#10x0000555555872492inmemory_region_oldmmio_read_accessor(mr=0x5555567fa870, addr=1, value=0x7fffca158510, size=1, shift=0, mask=255, attrs=...) at/home/liqiang02/qemu-devel/qemu/memory.c:409#20x000055555586f2ddinaccess_with_adjusted_size(addr=addr@entry=1, value=value@entry=0x7fffca158510, size=size@entry=1, access_size_min=<optimized out>, access_size_max=<optimized out>,Quoted 4 lines
access_fn=0x555555872440 <memory_region_oldmmio_read_accessor>,mr=0x5555567fa870, attrs=...) at
/home/liqiang02/qemu-devel/qemu/memory.c:593
#30x0000555555873e90inmemory_region_dispatch_read1(attrs=...,size=1,
Quoted 6 lines
pval=0x7fffca158510, addr=1, mr=0x5555567fa870) at
/home/liqiang02/qemu-devel/qemu/memory.c:1404
#4memory_region_dispatch_read(mr=mr@entry=0x5555567fa870, addr=1, pval=pval@entry=0x7fffca158510, size=1, attrs=attrs@entry=...) at/home/liqiang02/qemu-devel/qemu/memory.c:1423
#50x0000555555821e42inflatview_read_continue(fv=fv@entry=0x7fffbc03f370,
Quoted 1 line
addr=addr@entry=1297, attrs=..., buf=<optimized out>, buf@entry
=0x7ffff7fee000
Quoted 11 lines
"", len=len@entry=1, addr1=<optimized out>, l=<optimized out>, mr=0x5555567fa870)
at /home/liqiang02/qemu-devel/qemu/exec.c:3293#6
0x0000555555822006inflatview_read(fv=0x7fffbc03f370, addr=1297, attrs=..., buf=0x7ffff7fee000 "", len=1) at/home/liqiang02/qemu-devel/qemu/exec.c:3331
#70x000055555582211finaddress_space_read_full(as=<optimized out>, addr=addr@entry=1297, attrs=..., buf=buf@entry=0x7ffff7fee000 "", len=len@entry=1) at/home/liqiang02/qemu-devel/qemu/exec.c:3344#80x000055555582225ainaddress_space_rw(as=<optimized out>, addr=addr@entry=1297, attrs=..., attrs@entry=..., buf=buf@entry=0x7ffff7fee000
Quoted 13 lines
"", len=len@entry=1, is_write=is_write@entry=false)
at /home/liqiang02/qemu-devel/qemu/exec.c:3374#9
0x0000555555886239inkvm_handle_io(count=1, size=1, direction=<optimized out>, data=<optimized out>, attrs=..., port=1297) at/home/liqiang02/qemu-devel/qemu/accel/kvm/kvm-all.c:1731
#10kvm_cpu_exec(cpu=cpu@entry=0x5555566e9990) at/home/liqiang02/qemu-devel/qemu/accel/kvm/kvm-all.c:1971
#110x000055555585d3deinqemu_kvm_cpu_thread_fn(arg=0x5555566e9990) at/home/liqiang02/qemu-devel/qemu/cpus.c:1257
#120x00007fffdbd58494instart_thread() from/lib/x86_64-linux-gnu/libpthread.so.0
#130x00007fffdba9aacfin clone () from/lib/x86_64-linux-gnu/libc.so.6Quoted 8 lines
So I send out this path.
But this time when I not use gdb, there is no segmentation.
So this may not a issue.
If who has a arm environment, he can try this, the PoC is just easy. just read the
0x510port with word.
Quoted 9 lines
I don't understand your description of how you managed to trigger this SIGSEGV.
FWIW, looking at the codebase, there's a good number of static MemoryRegionOps structures for which the "
read_with_attrs" and "read" members are default-initialized to NULL. It seems unlikely they are all wrong.
I uses the debugexit.
QEMU command: gdb --args qemu-system-x86_64 -m 2048 -hda /home/liqiang02/ubuntu1801.img -enable-kvm -vnc :100 -device isa-debug-exit
guest: inw(0x501)
We can get the following backtrack.
Starting program: /usr/local/bin/qemu-system-x86_64 -m 2048 -hda /home/liqiang02/ubuntu1801.img -enable-kvm -vnc :100 -device isa-debug-exit [Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1". [New Thread 0x7fffcde12700 (LWP 52824)]
[New Thread 0x7fffcd41f700 (LWP 52825)]
WARNING: Image format was not specified for '/home/liqiang02/ubuntu1801.img' and probing guessed raw.
Automatically detecting the format is dangerous for raw images,
write operations on block 0 will be restricted.
Specify the 'raw' format explicitly to remove the restrictions.
[New Thread 0x7fffcca3c700 (LWP 52826)]
[New Thread 0x7fff463ff700 (LWP 52828)]
[Thread 0x7fffcd41f700 (LWP 52825) exited]
[New Thread 0x7fffcd41f700 (LWP 52857)]
[Thread 0x7fffcd41f700 (LWP 52857) exited]
[New Thread 0x7fffcd41f700 (LWP 52935)]
Thread 4 "qemu-system-x86" received signal SIGSEGV, Segmentation fault. [Switching to Thread 0x7fffcca3c700 (LWP 52826)] 0x0000000000000000 in ?? ()
(gdb) bt
#0 0x0000000000000000 in ?? ()
#1 0x00005555557c1f81 in memory_region_oldmmio_read_accessor (mr=0x5555577b32b0, addr=0, value=0x7fffcca39568, size=2, shift=0, mask=65535, attrs=...) at /home/liqiang02/qemu_max_cpu/qemu-2.8/memory.c:409 #2 0x00005555557c2697 in access_with_adjusted_size (addr=0, value=0x7fffcca39568, size=2, access_size_min=1, access_size_max=4, access=0x5555557c1f31 <memory_region_oldmmio_read_accessor>, mr=0x5555577b32b0, attrs=...)
at /home/liqiang02/qemu_max_cpu/qemu-2.8/memory.c:592
#3 0x00005555557c4aa2 in memory_region_dispatch_read1 (mr=0x5555577b32b0, addr=0, pval=0x7fffcca39568, size=2, attrs=...) at /home/liqiang02/qemu_max_cpu/qemu-2.8/memory.c:1254
#4 0x00005555557c4b26 in memory_region_dispatch_read (mr=0x5555577b32b0, addr=0, pval=0x7fffcca39568, size=2, attrs=...) at /home/liqiang02/qemu_max_cpu/qemu-2.8/memory.c:1273
#5 0x00005555557708ba in address_space_read_continue (as=0x5555562cc760 <address_space_io>, addr=1281, attrs=..., buf=0x7ffff7fee000 "", len=2, addr1=0, l=2, mr=0x5555577b32b0) at /home/liqiang02/qemu_max_cpu/qemu-2.8/exec.c:2692
#6 0x0000555555770a53 in address_space_read_full (as=0x5555562cc760 <address_space_io>, addr=1281, attrs=..., buf=0x7ffff7fee000 "", len=2) at /home/liqiang02/qemu_max_cpu/qemu-2.8/exec.c:2743
#7 0x0000555555770b15 in address_space_read (len=2, buf=0x7ffff7fee000 "", attrs=..., addr=1281, as=0x5555562cc760 <address_space_io>) at /home/liqiang02/qemu_max_cpu/qemu-2.8/include/exec/memory.h:1527 #8 address_space_rw (as=0x5555562cc760 <address_space_io>, addr=1281, attrs=..., buf=0x7ffff7fee000 "", len=2, is_write=false) at /home/liqiang02/qemu_max_cpu/qemu-2.8/exec.c:2757
#9 0x00005555557bea59 in kvm_handle_io (port=1281, attrs=..., data=0x7ffff7fee000, direction=0, size=2, count=1) at /home/liqiang02/qemu_max_cpu/qemu-2.8/kvm-all.c:1800
#10 0x00005555557bef5f in kvm_cpu_exec (cpu=0x5555568279c0) at /home/liqiang02/qemu_max_cpu/qemu-2.8/kvm-all.c:1958
#11 0x00005555557a5d5f in qemu_kvm_cpu_thread_fn (arg=0x5555568279c0) at /home/liqiang02/qemu_max_cpu/qemu-2.8/cpus.c:998
#12 0x00007fffdbf89494 in start_thread () from /lib/x86_64-linux-gnu/libpthread.so.0
#13 0x00007fffdbccbacf in clone () from /lib/x86_64-linux-gnu/libc.so.6 (gdb)
Seems the root cause is in 'memory_region_oldmmio_read_accessor '. Anyway I have sent out another patch to address issue.
Thanks,
Li Qiang
Quoted 10 lines
- exec.c: notdirty_mem_ops, readonly_mem_ops - hw/misc/debugexit.c: debug_exit_ops - hw/misc/hyperv_testdev.c: synic_test_sint_ops - hw/misc/pc-testdev.c: test_irq_ops, test_flush_ops - hw/pci-host/designware.c: designware_pci_host_msi_ops-
hw/rdma/vmw/pvrdma_main.c:uar_ops
-hw/sparc64/sun4u.c:power_mem_opsLaszlo
Quoted 23 lines · Li Qiang
Quoted 11 lines
Quoted 4 lines
Quoted 4 lines
Do you have a reproducer and/or a backtrace?
memory_region_dispatch_write()checks ifops->write!= NULL.FWIW, looking at the codebase, there's a good number of static MemoryRegionOps structures for which the "
read_with_attrs" and "read" members are default-initialized to NULL. It seems unlikely they are all wrong.I uses the debugexit.
QEMU command: gdb --args qemu-system-x86_64 -m 2048 -hda
/home/liqiang02/ubuntu1801.img-enable-kvm -vnc :100 -device isa-debug-exitguest: inw(0x501)
We can get the following backtrack.
Starting program:
/usr/local/bin/qemu-system-x86_64-m 2048 -hda/home/liqiang02/ubuntu1801.img-enable-kvm -vnc :100 -device isa-debug-exit
Quoted 8 lines
Thread 4 "qemu-system-x86" received signal SIGSEGV, Segmentation fault. [Switching to Thread
0x7fffcca3c700(LWP 52826)]0x0000000000000000in ?? ()
(gdb) bt
#00x0000000000000000in ?? ()
#10x00005555557c1f81inmemory_region_oldmmio_read_accessor(mr=0x5555577b32b0, addr=0, value=0x7fffcca39568, size=2, shift=0, mask=65535, attrs=...) at/home/liqiang02/qemu_max_cpu/qemu-2.8/memory.c:409
Doesn't that mean we should fix memory_region_oldmmio_read_accessor() to deal with a NULL callback, rather than hacking up a large number of devices to supply a no-op callback?
Signature · 3 lines
Eric Blake, Principal Software Engineer Red Hat, Inc. +1-919-301-3266 Virtualization: qemu.org | libvirt.org