1
A mixed bag, all bug fixes or similar small stuff.
1
Handful of bugfixes for rc2. None of these are particularly critical
2
or exciting.
2
3
3
thanks
4
-- PMM
4
-- PMM
5
5
6
The following changes since commit 45a150aa2b3492acf6691c7bdbeb25a8545d8345:
6
7
7
The following changes since commit 19eb2d4e736dc895f31fbd6b520e514f10cc08e0:
8
Merge remote-tracking branch 'remotes/ericb/tags/pull-bitmaps-2020-08-03' into staging (2020-08-03 15:13:49 +0100)
8
9
Merge remote-tracking branch 'remotes/thibault/tags/samuel-thibault' into staging (2019-05-07 10:43:32 +0100)
10
9
11
are available in the Git repository at:
10
are available in the Git repository at:
12
11
13
https://git.linaro.org/people/pmaydell/qemu-arm.git tags/pull-target-arm-20190507
12
https://git.linaro.org/people/pmaydell/qemu-arm.git tags/pull-target-arm-20200803
14
13
15
for you to fetch changes up to 63159601fb3e396b28da14cbb71e50ed3f5a0331:
14
for you to fetch changes up to 13557fd392890cbd985bceba7f717e01efd674b8:
16
15
17
target/arm: Stop using variable length array in dc_zva (2019-05-07 12:55:04 +0100)
16
hw/timer/imx_epit: Avoid assertion when CR.SWR is written (2020-08-03 17:56:11 +0100)
18
17
19
----------------------------------------------------------------
18
----------------------------------------------------------------
20
target-arm queue:
19
target-arm queue:
21
* Stop using variable length array in dc_zva
20
* hw/timer/imx_epit: Avoid assertion when CR.SWR is written
22
* Implement M-profile XPSR GE bits
21
* netduino2, netduinoplus2, microbit: set system_clock_scale so that
23
* Don't enable ARMV7M_EXCP_DEBUG from reset
22
SysTick running on the CPU clock works
24
* armv7m_nvic: NS BFAR and BFSR are RAZ/WI if BFHFNMINS == 0
23
* target/arm: Avoid maybe-uninitialized warning with gcc 4.9
25
* armv7m_nvic: Check subpriority in nvic_recompute_state_secure()
24
* target/arm: Fix AddPAC error indication
26
* fix various minor issues to allow building for Windows-on-ARM64
25
* Make AIRCR.SYSRESETREQ actually reset the system for the
27
* aspeed: Set SDRAM size
26
microbit, mps2-*, musca-*, netduino* boards
28
* Allow system registers for KVM guests to be changed by QEMU code
29
* raspi: Diagnose requests for too much RAM
30
* virt: Support firmware configuration with -blockdev
31
27
32
----------------------------------------------------------------
28
----------------------------------------------------------------
33
Cao Jiaxi (4):
29
Kaige Li (1):
34
QEMU_PACKED: Remove gcc_struct attribute in Windows non x86 targets
30
target/arm: Avoid maybe-uninitialized warning with gcc 4.9
35
qga: Fix mingw compilation warnings on enum conversion
36
util/cacheinfo: Use uint64_t on LLP64 model to satisfy Windows ARM64
37
osdep: Fix mingw compilation regarding stdio formats
38
31
39
Joel Stanley (1):
32
Peter Maydell (6):
40
arm: aspeed: Set SDRAM size
33
hw/arm/netduino2, netduinoplus2: Set system_clock_scale
34
include/hw/irq.h: New function qemu_irq_is_connected()
35
hw/intc/armv7m_nvic: Provide default "reset the system" behaviour for SYSRESETREQ
36
msf2-soc, stellaris: Don't wire up SYSRESETREQ
37
hw/arm/nrf51_soc: Set system_clock_scale
38
hw/timer/imx_epit: Avoid assertion when CR.SWR is written
41
39
42
Markus Armbruster (3):
40
Richard Henderson (1):
43
pc: Rearrange pc_system_firmware_init()'s legacy -drive loop
41
target/arm: Fix AddPAC error indication
44
pflash_cfi01: New pflash_cfi01_legacy_drive()
45
hw/arm/virt: Support firmware configuration with -blockdev
46
42
47
Peter Maydell (7):
43
include/hw/arm/armv7m.h | 4 +++-
48
hw/arm/raspi: Diagnose requests for too much RAM
44
include/hw/irq.h | 18 ++++++++++++++++++
49
arm: Allow system registers for KVM guests to be changed by QEMU code
45
hw/arm/msf2-soc.c | 11 -----------
50
hw/arm/armv7m_nvic: Check subpriority in nvic_recompute_state_secure()
46
hw/arm/netduino2.c | 10 ++++++++++
51
hw/intc/armv7m_nvic: NS BFAR and BFSR are RAZ/WI if BFHFNMINS == 0
47
hw/arm/netduinoplus2.c | 10 ++++++++++
52
hw/intc/armv7m_nvic: Don't enable ARMV7M_EXCP_DEBUG from reset
48
hw/arm/nrf51_soc.c | 5 +++++
53
target/arm: Implement XPSR GE bits
49
hw/arm/stellaris.c | 12 ------------
54
target/arm: Stop using variable length array in dc_zva
50
hw/intc/armv7m_nvic.c | 17 ++++++++++++++++-
51
hw/timer/imx_epit.c | 13 ++++++++++---
52
target/arm/pauth_helper.c | 6 +++++-
53
target/arm/translate-a64.c | 2 +-
54
tests/tcg/aarch64/pauth-5.c | 33 +++++++++++++++++++++++++++++++++
55
tests/tcg/aarch64/Makefile.target | 2 +-
56
13 files changed, 112 insertions(+), 31 deletions(-)
57
create mode 100644 tests/tcg/aarch64/pauth-5.c
55
58
56
contrib/libvhost-user/libvhost-user.h | 2 +-
57
include/hw/arm/aspeed.h | 1 +
58
include/hw/arm/virt.h | 2 +
59
include/hw/block/flash.h | 1 +
60
include/qemu/compiler.h | 2 +-
61
include/qemu/osdep.h | 10 +-
62
scripts/cocci-macro-file.h | 7 +-
63
target/arm/cpu.h | 13 ++-
64
hw/arm/aspeed.c | 8 ++
65
hw/arm/raspi.c | 7 ++
66
hw/arm/virt.c | 202 ++++++++++++++++++++++------------
67
hw/block/pflash_cfi01.c | 28 +++++
68
hw/i386/pc_sysfw.c | 18 +--
69
hw/intc/armv7m_nvic.c | 40 ++++++-
70
qga/commands-win32.c | 2 +-
71
target/arm/helper.c | 47 +++++++-
72
target/arm/kvm.c | 8 ++
73
target/arm/kvm32.c | 20 +---
74
target/arm/kvm64.c | 2 +
75
target/arm/machine.c | 2 +-
76
util/cacheinfo.c | 2 +-
77
21 files changed, 294 insertions(+), 130 deletions(-)
78
diff view generated by jsdifflib
1
Currently the dc_zva helper function uses a variable length
1
The netduino2 and netduinoplus2 boards forgot to set the system_clock_scale
2
array. In fact we know (as the comment above remarks) that
2
global, which meant that if guest code used the systick timer in "use
3
the length of this array is bounded because the architecture
3
the processor clock" mode it would hang because time never advances.
4
limits the block size and QEMU limits the target page size.
5
Use a fixed array size and assert that we don't run off it.
6
4
5
Set the global to match the documented CPU clock speed of these boards.
6
Judging by the data sheet this is slightly simplistic because the
7
SoC allows configuration of the SYSCLK source and frequency via the
8
RCC (reset and clock control) module, but we don't model that.
9
10
Fixes: https://bugs.launchpad.net/qemu/+bug/1876187
7
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
11
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
8
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
12
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
9
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
13
Message-id: 20200727162617.26227-1-peter.maydell@linaro.org
10
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
11
Message-id: 20190503120448.13385-1-peter.maydell@linaro.org
12
---
14
---
13
target/arm/helper.c | 8 ++++++--
15
hw/arm/netduino2.c | 10 ++++++++++
14
1 file changed, 6 insertions(+), 2 deletions(-)
16
hw/arm/netduinoplus2.c | 10 ++++++++++
17
2 files changed, 20 insertions(+)
15
18
16
diff --git a/target/arm/helper.c b/target/arm/helper.c
19
diff --git a/hw/arm/netduino2.c b/hw/arm/netduino2.c
17
index XXXXXXX..XXXXXXX 100644
20
index XXXXXXX..XXXXXXX 100644
18
--- a/target/arm/helper.c
21
--- a/hw/arm/netduino2.c
19
+++ b/target/arm/helper.c
22
+++ b/hw/arm/netduino2.c
20
@@ -XXX,XX +XXX,XX @@
23
@@ -XXX,XX +XXX,XX @@
21
#include "qemu/osdep.h"
24
#include "hw/arm/stm32f205_soc.h"
22
+#include "qemu/units.h"
25
#include "hw/arm/boot.h"
23
#include "target/arm/idau.h"
26
24
#include "trace.h"
27
+/* Main SYSCLK frequency in Hz (120MHz) */
25
#include "cpu.h"
28
+#define SYSCLK_FRQ 120000000ULL
26
@@ -XXX,XX +XXX,XX @@ void HELPER(dc_zva)(CPUARMState *env, uint64_t vaddr_in)
27
* We know that in fact for any v8 CPU the page size is at least 4K
28
* and the block size must be 2K or less, but TARGET_PAGE_SIZE is only
29
* 1K as an artefact of legacy v5 subpage support being present in the
30
- * same QEMU executable.
31
+ * same QEMU executable. So in practice the hostaddr[] array has
32
+ * two entries, given the current setting of TARGET_PAGE_BITS_MIN.
33
*/
34
int maxidx = DIV_ROUND_UP(blocklen, TARGET_PAGE_SIZE);
35
- void *hostaddr[maxidx];
36
+ void *hostaddr[DIV_ROUND_UP(2 * KiB, 1 << TARGET_PAGE_BITS_MIN)];
37
int try, i;
38
unsigned mmu_idx = cpu_mmu_index(env, false);
39
TCGMemOpIdx oi = make_memop_idx(MO_UB, mmu_idx);
40
41
+ assert(maxidx <= ARRAY_SIZE(hostaddr));
42
+
29
+
43
for (try = 0; try < 2; try++) {
30
static void netduino2_init(MachineState *machine)
44
31
{
45
for (i = 0; i < maxidx; i++) {
32
DeviceState *dev;
33
34
+ /*
35
+ * TODO: ideally we would model the SoC RCC and let it handle
36
+ * system_clock_scale, including its ability to define different
37
+ * possible SYSCLK sources.
38
+ */
39
+ system_clock_scale = NANOSECONDS_PER_SECOND / SYSCLK_FRQ;
40
+
41
dev = qdev_new(TYPE_STM32F205_SOC);
42
qdev_prop_set_string(dev, "cpu-type", ARM_CPU_TYPE_NAME("cortex-m3"));
43
sysbus_realize_and_unref(SYS_BUS_DEVICE(dev), &error_fatal);
44
diff --git a/hw/arm/netduinoplus2.c b/hw/arm/netduinoplus2.c
45
index XXXXXXX..XXXXXXX 100644
46
--- a/hw/arm/netduinoplus2.c
47
+++ b/hw/arm/netduinoplus2.c
48
@@ -XXX,XX +XXX,XX @@
49
#include "hw/arm/stm32f405_soc.h"
50
#include "hw/arm/boot.h"
51
52
+/* Main SYSCLK frequency in Hz (168MHz) */
53
+#define SYSCLK_FRQ 168000000ULL
54
+
55
static void netduinoplus2_init(MachineState *machine)
56
{
57
DeviceState *dev;
58
59
+ /*
60
+ * TODO: ideally we would model the SoC RCC and let it handle
61
+ * system_clock_scale, including its ability to define different
62
+ * possible SYSCLK sources.
63
+ */
64
+ system_clock_scale = NANOSECONDS_PER_SECOND / SYSCLK_FRQ;
65
+
66
dev = qdev_new(TYPE_STM32F405_SOC);
67
qdev_prop_set_string(dev, "cpu-type", ARM_CPU_TYPE_NAME("cortex-m4"));
68
sysbus_realize_and_unref(SYS_BUS_DEVICE(dev), &error_fatal);
46
--
69
--
47
2.20.1
70
2.20.1
48
71
49
72
diff view generated by jsdifflib
1
In the M-profile architecture, if the CPU implements the DSP extension
1
Mostly devices don't need to care whether one of their output
2
then the XPSR has GE bits, in the same way as the A-profile CPSR. When
2
qemu_irq lines is connected, because functions like qemu_set_irq()
3
we added DSP extension support we forgot to add support for reading
3
silently do nothing if there is nothing on the other end. However
4
and writing the GE bits, which are stored in env->GE. We did put in
4
sometimes a device might want to implement default behaviour for the
5
the code to add XPSR_GE to the mask of bits to update in the v7m_msr
5
case where the machine hasn't wired the line up to anywhere.
6
helper, but forgot it in v7m_mrs. We also must not allow the XPSR we
6
7
pull off the stack on exception return to set the nonexistent GE bits.
7
Provide a function qemu_irq_is_connected() that devices can use for
8
Correct these errors:
8
this purpose. (The test is trivial but encapsulating it in a
9
* read and write env->GE in xpsr_read() and xpsr_write()
9
function makes it easier to see where we're doing it in case we need
10
* only set GE bits on exception return if DSP present
10
to change the implementation later.)
11
* read GE bits for MRS if DSP present
12
11
13
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
12
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
14
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
13
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
15
Message-id: 20190430131439.25251-5-peter.maydell@linaro.org
14
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
15
Message-id: 20200728103744.6909-2-peter.maydell@linaro.org
16
---
16
---
17
target/arm/cpu.h | 4 ++++
17
include/hw/irq.h | 18 ++++++++++++++++++
18
target/arm/helper.c | 12 ++++++++++--
18
1 file changed, 18 insertions(+)
19
2 files changed, 14 insertions(+), 2 deletions(-)
20
19
21
diff --git a/target/arm/cpu.h b/target/arm/cpu.h
20
diff --git a/include/hw/irq.h b/include/hw/irq.h
22
index XXXXXXX..XXXXXXX 100644
21
index XXXXXXX..XXXXXXX 100644
23
--- a/target/arm/cpu.h
22
--- a/include/hw/irq.h
24
+++ b/target/arm/cpu.h
23
+++ b/include/hw/irq.h
25
@@ -XXX,XX +XXX,XX @@ static inline uint32_t xpsr_read(CPUARMState *env)
24
@@ -XXX,XX +XXX,XX @@ qemu_irq qemu_irq_split(qemu_irq irq1, qemu_irq irq2);
26
| (env->CF << 29) | ((env->VF & 0x80000000) >> 3) | (env->QF << 27)
25
on an existing vector of qemu_irq. */
27
| (env->thumb << 24) | ((env->condexec_bits & 3) << 25)
26
void qemu_irq_intercept_in(qemu_irq *gpio_in, qemu_irq_handler handler, int n);
28
| ((env->condexec_bits & 0xfc) << 8)
27
29
+ | (env->GE << 16)
28
+/**
30
| env->v7m.exception;
29
+ * qemu_irq_is_connected: Return true if IRQ line is wired up
31
}
30
+ *
32
31
+ * If a qemu_irq has a device on the other (receiving) end of it,
33
@@ -XXX,XX +XXX,XX @@ static inline void xpsr_write(CPUARMState *env, uint32_t val, uint32_t mask)
32
+ * return true; otherwise return false.
34
if (mask & XPSR_Q) {
33
+ *
35
env->QF = ((val & XPSR_Q) != 0);
34
+ * Usually device models don't need to care whether the machine model
36
}
35
+ * has wired up their outbound qemu_irq lines, because functions like
37
+ if (mask & XPSR_GE) {
36
+ * qemu_set_irq() silently do nothing if there is nothing on the other
38
+ env->GE = (val & XPSR_GE) >> 16;
37
+ * end of the line. However occasionally a device model will want to
39
+ }
38
+ * provide default behaviour if its output is left floating, and
40
if (mask & XPSR_T) {
39
+ * it can use this function to identify when that is the case.
41
env->thumb = ((val & XPSR_T) != 0);
40
+ */
42
}
41
+static inline bool qemu_irq_is_connected(qemu_irq irq)
43
diff --git a/target/arm/helper.c b/target/arm/helper.c
42
+{
44
index XXXXXXX..XXXXXXX 100644
43
+ return irq != NULL;
45
--- a/target/arm/helper.c
44
+}
46
+++ b/target/arm/helper.c
47
@@ -XXX,XX +XXX,XX @@ static void do_v7m_exception_exit(ARMCPU *cpu)
48
{
49
CPUARMState *env = &cpu->env;
50
uint32_t excret;
51
- uint32_t xpsr;
52
+ uint32_t xpsr, xpsr_mask;
53
bool ufault = false;
54
bool sfault = false;
55
bool return_to_sp_process;
56
@@ -XXX,XX +XXX,XX @@ static void do_v7m_exception_exit(ARMCPU *cpu)
57
}
58
*frame_sp_p = frameptr;
59
}
60
+
45
+
61
+ xpsr_mask = ~(XPSR_SPREALIGN | XPSR_SFPA);
46
#endif
62
+ if (!arm_feature(env, ARM_FEATURE_THUMB_DSP)) {
63
+ xpsr_mask &= ~XPSR_GE;
64
+ }
65
/* This xpsr_write() will invalidate frame_sp_p as it may switch stack */
66
- xpsr_write(env, xpsr, ~(XPSR_SPREALIGN | XPSR_SFPA));
67
+ xpsr_write(env, xpsr, xpsr_mask);
68
69
if (env->v7m.secure) {
70
bool sfpa = xpsr & XPSR_SFPA;
71
@@ -XXX,XX +XXX,XX @@ uint32_t HELPER(v7m_mrs)(CPUARMState *env, uint32_t reg)
72
}
73
if (!(reg & 4)) {
74
mask |= XPSR_NZCV | XPSR_Q; /* APSR */
75
+ if (arm_feature(env, ARM_FEATURE_THUMB_DSP)) {
76
+ mask |= XPSR_GE;
77
+ }
78
}
79
/* EPSR reads as zero */
80
return xpsr_read(env) & mask;
81
--
47
--
82
2.20.1
48
2.20.1
83
49
84
50
diff view generated by jsdifflib
1
The M-profile architecture specifies that the DebugMonitor exception
1
The NVIC provides an outbound qemu_irq "SYSRESETREQ" which it signals
2
should be initially disabled, not enabled. It should be controlled
2
when the guest sets the SYSRESETREQ bit in the AIRCR register. This
3
by the DEMCR register's MON_EN bit, but we don't implement that
3
matches the hardware design (where the CPU has a signal of this name
4
register yet (like most of the debug architecture for M-profile).
4
and it is up to the SoC to connect that up to an actual reset
5
mechanism), but in QEMU it mostly results in duplicated code in SoC
6
objects and bugs where SoC model implementors forget to wire up the
7
SYSRESETREQ line.
5
8
6
Note that BKPT instructions will still work, because they
9
Provide a default behaviour for the case where SYSRESETREQ is not
7
will be escalated to HardFault.
10
actually connected to anything: use qemu_system_reset_request() to
11
perform a system reset. This will allow us to remove the
12
implementations of SYSRESETREQ handling from the boards where that's
13
exactly what it does, and also fixes the bugs in the board models
14
which forgot to wire up the signal:
15
16
* microbit
17
* mps2-an385
18
* mps2-an505
19
* mps2-an511
20
* mps2-an521
21
* musca-a
22
* musca-b1
23
* netduino
24
* netduinoplus2
25
26
We still allow the board to wire up the signal if it needs to, in case
27
we need to model more complicated reset controller logic or to model
28
buggy SoC hardware which forgot to wire up the line itself. But
29
defaulting to "reset the system" is more often going to be correct
30
than defaulting to "do nothing".
8
31
9
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
32
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
33
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
11
Message-id: 20190430131439.25251-4-peter.maydell@linaro.org
34
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
35
Message-id: 20200728103744.6909-3-peter.maydell@linaro.org
12
---
36
---
13
hw/intc/armv7m_nvic.c | 4 +++-
37
include/hw/arm/armv7m.h | 4 +++-
14
1 file changed, 3 insertions(+), 1 deletion(-)
38
hw/intc/armv7m_nvic.c | 17 ++++++++++++++++-
39
2 files changed, 19 insertions(+), 2 deletions(-)
15
40
41
diff --git a/include/hw/arm/armv7m.h b/include/hw/arm/armv7m.h
42
index XXXXXXX..XXXXXXX 100644
43
--- a/include/hw/arm/armv7m.h
44
+++ b/include/hw/arm/armv7m.h
45
@@ -XXX,XX +XXX,XX @@ typedef struct {
46
47
/* ARMv7M container object.
48
* + Unnamed GPIO input lines: external IRQ lines for the NVIC
49
- * + Named GPIO output SYSRESETREQ: signalled for guest AIRCR.SYSRESETREQ
50
+ * + Named GPIO output SYSRESETREQ: signalled for guest AIRCR.SYSRESETREQ.
51
+ * If this GPIO is not wired up then the NVIC will default to performing
52
+ * a qemu_system_reset_request(SHUTDOWN_CAUSE_GUEST_RESET).
53
* + Property "cpu-type": CPU type to instantiate
54
* + Property "num-irq": number of external IRQ lines
55
* + Property "memory": MemoryRegion defining the physical address space
16
diff --git a/hw/intc/armv7m_nvic.c b/hw/intc/armv7m_nvic.c
56
diff --git a/hw/intc/armv7m_nvic.c b/hw/intc/armv7m_nvic.c
17
index XXXXXXX..XXXXXXX 100644
57
index XXXXXXX..XXXXXXX 100644
18
--- a/hw/intc/armv7m_nvic.c
58
--- a/hw/intc/armv7m_nvic.c
19
+++ b/hw/intc/armv7m_nvic.c
59
+++ b/hw/intc/armv7m_nvic.c
20
@@ -XXX,XX +XXX,XX @@ static void armv7m_nvic_reset(DeviceState *dev)
60
@@ -XXX,XX +XXX,XX @@
21
* the System Handler Control register
61
#include "hw/intc/armv7m_nvic.h"
22
*/
62
#include "hw/irq.h"
23
s->vectors[ARMV7M_EXCP_SVC].enabled = 1;
63
#include "hw/qdev-properties.h"
24
- s->vectors[ARMV7M_EXCP_DEBUG].enabled = 1;
64
+#include "sysemu/runstate.h"
25
s->vectors[ARMV7M_EXCP_PENDSV].enabled = 1;
65
#include "target/arm/cpu.h"
26
s->vectors[ARMV7M_EXCP_SYSTICK].enabled = 1;
66
#include "exec/exec-all.h"
27
67
#include "exec/memop.h"
28
+ /* DebugMonitor is enabled via DEMCR.MON_EN */
68
@@ -XXX,XX +XXX,XX @@ static const uint8_t nvic_id[] = {
29
+ s->vectors[ARMV7M_EXCP_DEBUG].enabled = 0;
69
0x00, 0xb0, 0x1b, 0x00, 0x0d, 0xe0, 0x05, 0xb1
70
};
71
72
+static void signal_sysresetreq(NVICState *s)
73
+{
74
+ if (qemu_irq_is_connected(s->sysresetreq)) {
75
+ qemu_irq_pulse(s->sysresetreq);
76
+ } else {
77
+ /*
78
+ * Default behaviour if the SoC doesn't need to wire up
79
+ * SYSRESETREQ (eg to a system reset controller of some kind):
80
+ * perform a system reset via the usual QEMU API.
81
+ */
82
+ qemu_system_reset_request(SHUTDOWN_CAUSE_GUEST_RESET);
83
+ }
84
+}
30
+
85
+
31
resetprio = arm_feature(&s->cpu->env, ARM_FEATURE_V8) ? -4 : -3;
86
static int nvic_pending_prio(NVICState *s)
32
s->vectors[ARMV7M_EXCP_RESET].prio = resetprio;
87
{
33
s->vectors[ARMV7M_EXCP_NMI].prio = -2;
88
/* return the group priority of the current pending interrupt,
89
@@ -XXX,XX +XXX,XX @@ static void nvic_writel(NVICState *s, uint32_t offset, uint32_t value,
90
if (value & R_V7M_AIRCR_SYSRESETREQ_MASK) {
91
if (attrs.secure ||
92
!(cpu->env.v7m.aircr & R_V7M_AIRCR_SYSRESETREQS_MASK)) {
93
- qemu_irq_pulse(s->sysresetreq);
94
+ signal_sysresetreq(s);
95
}
96
}
97
if (value & R_V7M_AIRCR_VECTCLRACTIVE_MASK) {
34
--
98
--
35
2.20.1
99
2.20.1
36
100
37
101
diff view generated by jsdifflib
1
At the moment the Arm implementations of kvm_arch_{get,put}_registers()
1
The MSF2 SoC model and the Stellaris board code both wire
2
don't support having QEMU change the values of system registers
2
SYSRESETREQ up to a function that just invokes
3
(aka coprocessor registers for AArch32). This is because although
3
qemu_system_reset_request(SHUTDOWN_CAUSE_GUEST_RESET);
4
kvm_arch_get_registers() calls write_list_to_cpustate() to
4
This is now the default action that the NVIC does if the line is
5
update the CPU state struct fields (so QEMU code can read the
5
not connected, so we can delete the handling code.
6
values in the usual way), kvm_arch_put_registers() does not
7
call write_cpustate_to_list(), meaning that any changes to
8
the CPU state struct fields will not be passed back to KVM.
9
10
The rationale for this design is documented in a comment in the
11
AArch32 kvm_arch_put_registers() -- writing the values in the
12
cpregs list into the CPU state struct is "lossy" because the
13
write of a register might not succeed, and so if we blindly
14
copy the CPU state values back again we will incorrectly
15
change register values for the guest. The assumption was that
16
no QEMU code would need to write to the registers.
17
18
However, when we implemented debug support for KVM guests, we
19
broke that assumption: the code to handle "set the guest up
20
to take a breakpoint exception" does so by updating various
21
guest registers including ESR_EL1.
22
23
Support this by making kvm_arch_put_registers() synchronize
24
CPU state back into the list. We sync only those registers
25
where the initial write succeeds, which should be sufficient.
26
27
This commit is the same as commit 823e1b3818f9b10b824ddc which we
28
had to revert in commit 942f99c825fc94c8b1a4, except that the bug
29
which was preventing EDK2 guest firmware running has been fixed:
30
kvm_arm_reset_vcpu() now calls write_list_to_cpustate().
31
6
32
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
33
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
8
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
34
Tested-by: Eric Auger <eric.auger@redhat.com>
9
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
10
Message-id: 20200728103744.6909-4-peter.maydell@linaro.org
35
---
11
---
36
target/arm/cpu.h | 9 ++++++++-
12
hw/arm/msf2-soc.c | 11 -----------
37
target/arm/helper.c | 27 +++++++++++++++++++++++++--
13
hw/arm/stellaris.c | 12 ------------
38
target/arm/kvm.c | 8 ++++++++
14
2 files changed, 23 deletions(-)
39
target/arm/kvm32.c | 20 ++------------------
40
target/arm/kvm64.c | 2 ++
41
target/arm/machine.c | 2 +-
42
6 files changed, 46 insertions(+), 22 deletions(-)
43
15
44
diff --git a/target/arm/cpu.h b/target/arm/cpu.h
16
diff --git a/hw/arm/msf2-soc.c b/hw/arm/msf2-soc.c
45
index XXXXXXX..XXXXXXX 100644
17
index XXXXXXX..XXXXXXX 100644
46
--- a/target/arm/cpu.h
18
--- a/hw/arm/msf2-soc.c
47
+++ b/target/arm/cpu.h
19
+++ b/hw/arm/msf2-soc.c
48
@@ -XXX,XX +XXX,XX @@ bool write_list_to_cpustate(ARMCPU *cpu);
20
@@ -XXX,XX +XXX,XX @@
49
/**
21
#include "hw/irq.h"
50
* write_cpustate_to_list:
22
#include "hw/arm/msf2-soc.h"
51
* @cpu: ARMCPU
23
#include "hw/misc/unimp.h"
52
+ * @kvm_sync: true if this is for syncing back to KVM
24
-#include "sysemu/runstate.h"
53
*
25
#include "sysemu/sysemu.h"
54
* For each register listed in the ARMCPU cpreg_indexes list, write
26
55
* its value from the ARMCPUState structure into the cpreg_values list.
27
#define MSF2_TIMER_BASE 0x40004000
56
* This is used to copy info from TCG's working data structures into
28
@@ -XXX,XX +XXX,XX @@ static const int spi_irq[MSF2_NUM_SPIS] = { 2, 3 };
57
* KVM or for outbound migration.
29
static const int uart_irq[MSF2_NUM_UARTS] = { 10, 11 };
58
*
30
static const int timer_irq[MSF2_NUM_TIMERS] = { 14, 15 };
59
+ * @kvm_sync is true if we are doing this in order to sync the
31
60
+ * register state back to KVM. In this case we will only update
32
-static void do_sys_reset(void *opaque, int n, int level)
61
+ * values in the list if the previous list->cpustate sync actually
33
-{
62
+ * successfully wrote the CPU state. Otherwise we will keep the value
34
- if (level) {
63
+ * that is in the list.
35
- qemu_system_reset_request(SHUTDOWN_CAUSE_GUEST_RESET);
64
+ *
36
- }
65
* Returns: true if all register values were read correctly,
37
-}
66
* false if some register was unknown or could not be read.
38
-
67
* Note that we do not stop early on failure -- we will attempt
39
static void m2sxxx_soc_initfn(Object *obj)
68
* reading all registers in the list.
40
{
69
*/
41
MSF2State *s = MSF2_SOC(obj);
70
-bool write_cpustate_to_list(ARMCPU *cpu);
42
@@ -XXX,XX +XXX,XX @@ static void m2sxxx_soc_realize(DeviceState *dev_soc, Error **errp)
71
+bool write_cpustate_to_list(ARMCPU *cpu, bool kvm_sync);
43
return;
72
44
}
73
#define ARM_CPUID_TI915T 0x54029152
45
74
#define ARM_CPUID_TI925T 0x54029252
46
- qdev_connect_gpio_out_named(DEVICE(&s->armv7m.nvic), "SYSRESETREQ", 0,
75
diff --git a/target/arm/helper.c b/target/arm/helper.c
47
- qemu_allocate_irq(&do_sys_reset, NULL, 0));
48
-
49
system_clock_scale = NANOSECONDS_PER_SECOND / s->m3clk;
50
51
for (i = 0; i < MSF2_NUM_UARTS; i++) {
52
diff --git a/hw/arm/stellaris.c b/hw/arm/stellaris.c
76
index XXXXXXX..XXXXXXX 100644
53
index XXXXXXX..XXXXXXX 100644
77
--- a/target/arm/helper.c
54
--- a/hw/arm/stellaris.c
78
+++ b/target/arm/helper.c
55
+++ b/hw/arm/stellaris.c
79
@@ -XXX,XX +XXX,XX @@ static bool raw_accessors_invalid(const ARMCPRegInfo *ri)
56
@@ -XXX,XX +XXX,XX @@
80
return true;
57
#include "hw/boards.h"
58
#include "qemu/log.h"
59
#include "exec/address-spaces.h"
60
-#include "sysemu/runstate.h"
61
#include "sysemu/sysemu.h"
62
#include "hw/arm/armv7m.h"
63
#include "hw/char/pl011.h"
64
@@ -XXX,XX +XXX,XX @@ static void stellaris_adc_init(Object *obj)
65
qdev_init_gpio_in(dev, stellaris_adc_trigger, 1);
81
}
66
}
82
67
83
-bool write_cpustate_to_list(ARMCPU *cpu)
68
-static
84
+bool write_cpustate_to_list(ARMCPU *cpu, bool kvm_sync)
69
-void do_sys_reset(void *opaque, int n, int level)
85
{
70
-{
86
/* Write the coprocessor state from cpu->env to the (index,value) list. */
71
- if (level) {
87
int i;
72
- qemu_system_reset_request(SHUTDOWN_CAUSE_GUEST_RESET);
88
@@ -XXX,XX +XXX,XX @@ bool write_cpustate_to_list(ARMCPU *cpu)
73
- }
89
for (i = 0; i < cpu->cpreg_array_len; i++) {
74
-}
90
uint32_t regidx = kvm_to_cpreg_id(cpu->cpreg_indexes[i]);
75
-
91
const ARMCPRegInfo *ri;
76
/* Board init. */
92
+ uint64_t newval;
77
static stellaris_board_info stellaris_boards[] = {
93
78
{ "LM3S811EVB",
94
ri = get_arm_cp_reginfo(cpu->cp_regs, regidx);
79
@@ -XXX,XX +XXX,XX @@ static void stellaris_init(MachineState *ms, stellaris_board_info *board)
95
if (!ri) {
80
/* This will exit with an error if the user passed us a bad cpu_type */
96
@@ -XXX,XX +XXX,XX @@ bool write_cpustate_to_list(ARMCPU *cpu)
81
sysbus_realize_and_unref(SYS_BUS_DEVICE(nvic), &error_fatal);
97
if (ri->type & ARM_CP_NO_RAW) {
82
98
continue;
83
- qdev_connect_gpio_out_named(nvic, "SYSRESETREQ", 0,
99
}
84
- qemu_allocate_irq(&do_sys_reset, NULL, 0));
100
- cpu->cpreg_values[i] = read_raw_cp_reg(&cpu->env, ri);
85
-
101
+
86
if (board->dc1 & (1 << 16)) {
102
+ newval = read_raw_cp_reg(&cpu->env, ri);
87
dev = sysbus_create_varargs(TYPE_STELLARIS_ADC, 0x40038000,
103
+ if (kvm_sync) {
88
qdev_get_gpio_in(nvic, 14),
104
+ /*
105
+ * Only sync if the previous list->cpustate sync succeeded.
106
+ * Rather than tracking the success/failure state for every
107
+ * item in the list, we just recheck "does the raw write we must
108
+ * have made in write_list_to_cpustate() read back OK" here.
109
+ */
110
+ uint64_t oldval = cpu->cpreg_values[i];
111
+
112
+ if (oldval == newval) {
113
+ continue;
114
+ }
115
+
116
+ write_raw_cp_reg(&cpu->env, ri, oldval);
117
+ if (read_raw_cp_reg(&cpu->env, ri) != oldval) {
118
+ continue;
119
+ }
120
+
121
+ write_raw_cp_reg(&cpu->env, ri, newval);
122
+ }
123
+ cpu->cpreg_values[i] = newval;
124
}
125
return ok;
126
}
127
diff --git a/target/arm/kvm.c b/target/arm/kvm.c
128
index XXXXXXX..XXXXXXX 100644
129
--- a/target/arm/kvm.c
130
+++ b/target/arm/kvm.c
131
@@ -XXX,XX +XXX,XX @@ void kvm_arm_reset_vcpu(ARMCPU *cpu)
132
fprintf(stderr, "write_kvmstate_to_list failed\n");
133
abort();
134
}
135
+ /*
136
+ * Sync the reset values also into the CPUState. This is necessary
137
+ * because the next thing we do will be a kvm_arch_put_registers()
138
+ * which will update the list values from the CPUState before copying
139
+ * the list values back to KVM. It's OK to ignore failure returns here
140
+ * for the same reason we do so in kvm_arch_get_registers().
141
+ */
142
+ write_list_to_cpustate(cpu);
143
}
144
145
/*
146
diff --git a/target/arm/kvm32.c b/target/arm/kvm32.c
147
index XXXXXXX..XXXXXXX 100644
148
--- a/target/arm/kvm32.c
149
+++ b/target/arm/kvm32.c
150
@@ -XXX,XX +XXX,XX @@ int kvm_arch_put_registers(CPUState *cs, int level)
151
return ret;
152
}
153
154
- /* Note that we do not call write_cpustate_to_list()
155
- * here, so we are only writing the tuple list back to
156
- * KVM. This is safe because nothing can change the
157
- * CPUARMState cp15 fields (in particular gdb accesses cannot)
158
- * and so there are no changes to sync. In fact syncing would
159
- * be wrong at this point: for a constant register where TCG and
160
- * KVM disagree about its value, the preceding write_list_to_cpustate()
161
- * would not have had any effect on the CPUARMState value (since the
162
- * register is read-only), and a write_cpustate_to_list() here would
163
- * then try to write the TCG value back into KVM -- this would either
164
- * fail or incorrectly change the value the guest sees.
165
- *
166
- * If we ever want to allow the user to modify cp15 registers via
167
- * the gdb stub, we would need to be more clever here (for instance
168
- * tracking the set of registers kvm_arch_get_registers() successfully
169
- * managed to update the CPUARMState with, and only allowing those
170
- * to be written back up into the kernel).
171
- */
172
+ write_cpustate_to_list(cpu, true);
173
+
174
if (!write_list_to_kvmstate(cpu, level)) {
175
return EINVAL;
176
}
177
diff --git a/target/arm/kvm64.c b/target/arm/kvm64.c
178
index XXXXXXX..XXXXXXX 100644
179
--- a/target/arm/kvm64.c
180
+++ b/target/arm/kvm64.c
181
@@ -XXX,XX +XXX,XX @@ int kvm_arch_put_registers(CPUState *cs, int level)
182
return ret;
183
}
184
185
+ write_cpustate_to_list(cpu, true);
186
+
187
if (!write_list_to_kvmstate(cpu, level)) {
188
return EINVAL;
189
}
190
diff --git a/target/arm/machine.c b/target/arm/machine.c
191
index XXXXXXX..XXXXXXX 100644
192
--- a/target/arm/machine.c
193
+++ b/target/arm/machine.c
194
@@ -XXX,XX +XXX,XX @@ static int cpu_pre_save(void *opaque)
195
abort();
196
}
197
} else {
198
- if (!write_cpustate_to_list(cpu)) {
199
+ if (!write_cpustate_to_list(cpu, false)) {
200
/* This should never fail. */
201
abort();
202
}
203
--
89
--
204
2.20.1
90
2.20.1
205
91
206
92
diff view generated by jsdifflib
1
From: Markus Armbruster <armbru@redhat.com>
1
From: Richard Henderson <richard.henderson@linaro.org>
2
2
3
The loop does two things: map legacy -drive to properties, and collect
3
The definition of top_bit used in this function is one higher
4
all the backends for use after the loop. The next patch will factor
4
than that used in the Arm ARM psuedo-code, which put the error
5
out the former for reuse in hw/arm/virt.c. To make that easier,
5
indication at top_bit - 1 at the wrong place, which meant that
6
rearrange the loop so it does the first thing first, and the second
6
it wasn't visible to Auth.
7
thing second.
8
7
9
Signed-off-by: Markus Armbruster <armbru@redhat.com>
8
Fixing the definition of top_bit requires more changes, because
10
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
9
its most common use is for the count of bits in top_bit:bot_bit,
11
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
10
which would then need to be computed as top_bit - bot_bit + 1.
12
Message-id: 20190416091348.26075-2-armbru@redhat.com
11
12
For now, prefer the minimal fix to the error indication alone.
13
14
Fixes: 63ff0ca94cb
15
Reported-by: Derrick McKee <derrick.mckee@gmail.com>
16
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
17
Message-id: 20200728195706.11087-1-richard.henderson@linaro.org
18
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
19
[PMM: added comment about the divergence from the pseudocode]
13
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
20
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
14
---
21
---
15
hw/i386/pc_sysfw.c | 24 +++++++++++-------------
22
target/arm/pauth_helper.c | 6 +++++-
16
1 file changed, 11 insertions(+), 13 deletions(-)
23
tests/tcg/aarch64/pauth-5.c | 33 +++++++++++++++++++++++++++++++
24
tests/tcg/aarch64/Makefile.target | 2 +-
25
3 files changed, 39 insertions(+), 2 deletions(-)
26
create mode 100644 tests/tcg/aarch64/pauth-5.c
17
27
18
diff --git a/hw/i386/pc_sysfw.c b/hw/i386/pc_sysfw.c
28
diff --git a/target/arm/pauth_helper.c b/target/arm/pauth_helper.c
19
index XXXXXXX..XXXXXXX 100644
29
index XXXXXXX..XXXXXXX 100644
20
--- a/hw/i386/pc_sysfw.c
30
--- a/target/arm/pauth_helper.c
21
+++ b/hw/i386/pc_sysfw.c
31
+++ b/target/arm/pauth_helper.c
22
@@ -XXX,XX +XXX,XX @@ void pc_system_firmware_init(PCMachineState *pcms,
32
@@ -XXX,XX +XXX,XX @@ static uint64_t pauth_addpac(CPUARMState *env, uint64_t ptr, uint64_t modifier,
23
33
*/
24
/* Map legacy -drive if=pflash to machine properties */
34
test = sextract64(ptr, bot_bit, top_bit - bot_bit);
25
for (i = 0; i < ARRAY_SIZE(pcms->flash); i++) {
35
if (test != 0 && test != -1) {
26
- pflash_blk[i] = pflash_cfi01_get_blk(pcms->flash[i]);
36
- pac ^= MAKE_64BIT_MASK(top_bit - 1, 1);
27
pflash_drv = drive_get(IF_PFLASH, 0, i);
37
+ /*
28
- if (!pflash_drv) {
38
+ * Note that our top_bit is one greater than the pseudocode's
29
- continue;
39
+ * version, hence "- 2" here.
30
+ if (pflash_drv) {
40
+ */
31
+ loc_push_none(&loc);
41
+ pac ^= MAKE_64BIT_MASK(top_bit - 2, 1);
32
+ qemu_opts_loc_restore(pflash_drv->opts);
33
+ if (pflash_cfi01_get_blk(pcms->flash[i])) {
34
+ error_report("clashes with -machine");
35
+ exit(1);
36
+ }
37
+ qdev_prop_set_drive(DEVICE(pcms->flash[i]), "drive",
38
+ blk_by_legacy_dinfo(pflash_drv), &error_fatal);
39
+ loc_pop(&loc);
40
}
41
- loc_push_none(&loc);
42
- qemu_opts_loc_restore(pflash_drv->opts);
43
- if (pflash_blk[i]) {
44
- error_report("clashes with -machine");
45
- exit(1);
46
- }
47
- pflash_blk[i] = blk_by_legacy_dinfo(pflash_drv);
48
- qdev_prop_set_drive(DEVICE(pcms->flash[i]),
49
- "drive", pflash_blk[i], &error_fatal);
50
- loc_pop(&loc);
51
+ pflash_blk[i] = pflash_cfi01_get_blk(pcms->flash[i]);
52
}
42
}
53
43
54
/* Reject gaps */
44
/*
45
diff --git a/tests/tcg/aarch64/pauth-5.c b/tests/tcg/aarch64/pauth-5.c
46
new file mode 100644
47
index XXXXXXX..XXXXXXX
48
--- /dev/null
49
+++ b/tests/tcg/aarch64/pauth-5.c
50
@@ -XXX,XX +XXX,XX @@
51
+#include <assert.h>
52
+
53
+static int x;
54
+
55
+int main()
56
+{
57
+ int *p0 = &x, *p1, *p2, *p3;
58
+ unsigned long salt = 0;
59
+
60
+ /*
61
+ * With TBI enabled and a 48-bit VA, there are 7 bits of auth, and so
62
+ * a 1/128 chance of auth = pac(ptr,key,salt) producing zero.
63
+ * Find a salt that creates auth != 0.
64
+ */
65
+ do {
66
+ salt++;
67
+ asm("pacda %0, %1" : "=r"(p1) : "r"(salt), "0"(p0));
68
+ } while (p0 == p1);
69
+
70
+ /*
71
+ * This pac must fail, because the input pointer bears an encryption,
72
+ * and so is not properly extended within bits [55:47]. This will
73
+ * toggle bit 54 in the output...
74
+ */
75
+ asm("pacda %0, %1" : "=r"(p2) : "r"(salt), "0"(p1));
76
+
77
+ /* ... so that the aut must fail, setting bit 53 in the output ... */
78
+ asm("autda %0, %1" : "=r"(p3) : "r"(salt), "0"(p2));
79
+
80
+ /* ... which means this equality must not hold. */
81
+ assert(p3 != p0);
82
+ return 0;
83
+}
84
diff --git a/tests/tcg/aarch64/Makefile.target b/tests/tcg/aarch64/Makefile.target
85
index XXXXXXX..XXXXXXX 100644
86
--- a/tests/tcg/aarch64/Makefile.target
87
+++ b/tests/tcg/aarch64/Makefile.target
88
@@ -XXX,XX +XXX,XX @@ run-fcvt: fcvt
89
90
# Pauth Tests
91
ifneq ($(DOCKER_IMAGE)$(CROSS_CC_HAS_ARMV8_3),)
92
-AARCH64_TESTS += pauth-1 pauth-2 pauth-4
93
+AARCH64_TESTS += pauth-1 pauth-2 pauth-4 pauth-5
94
pauth-%: CFLAGS += -march=armv8.3-a
95
run-pauth-%: QEMU_OPTS += -cpu max
96
run-plugin-pauth-%: QEMU_OPTS += -cpu max
55
--
97
--
56
2.20.1
98
2.20.1
57
99
58
100
diff view generated by jsdifflib
Deleted patch
1
From: Markus Armbruster <armbru@redhat.com>
2
1
3
Factored out of pc_system_firmware_init() so the next commit can reuse
4
it in hw/arm/virt.c.
5
6
Signed-off-by: Markus Armbruster <armbru@redhat.com>
7
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
8
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
9
Message-id: 20190416091348.26075-3-armbru@redhat.com
10
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
11
---
12
include/hw/block/flash.h | 1 +
13
hw/block/pflash_cfi01.c | 28 ++++++++++++++++++++++++++++
14
hw/i386/pc_sysfw.c | 16 ++--------------
15
3 files changed, 31 insertions(+), 14 deletions(-)
16
17
diff --git a/include/hw/block/flash.h b/include/hw/block/flash.h
18
index XXXXXXX..XXXXXXX 100644
19
--- a/include/hw/block/flash.h
20
+++ b/include/hw/block/flash.h
21
@@ -XXX,XX +XXX,XX @@ PFlashCFI01 *pflash_cfi01_register(hwaddr base,
22
int be);
23
BlockBackend *pflash_cfi01_get_blk(PFlashCFI01 *fl);
24
MemoryRegion *pflash_cfi01_get_memory(PFlashCFI01 *fl);
25
+void pflash_cfi01_legacy_drive(PFlashCFI01 *dev, DriveInfo *dinfo);
26
27
/* pflash_cfi02.c */
28
29
diff --git a/hw/block/pflash_cfi01.c b/hw/block/pflash_cfi01.c
30
index XXXXXXX..XXXXXXX 100644
31
--- a/hw/block/pflash_cfi01.c
32
+++ b/hw/block/pflash_cfi01.c
33
@@ -XXX,XX +XXX,XX @@
34
#include "qapi/error.h"
35
#include "qemu/timer.h"
36
#include "qemu/bitops.h"
37
+#include "qemu/error-report.h"
38
#include "qemu/host-utils.h"
39
#include "qemu/log.h"
40
+#include "qemu/option.h"
41
#include "hw/sysbus.h"
42
+#include "sysemu/blockdev.h"
43
#include "sysemu/sysemu.h"
44
#include "trace.h"
45
46
@@ -XXX,XX +XXX,XX @@ MemoryRegion *pflash_cfi01_get_memory(PFlashCFI01 *fl)
47
return &fl->mem;
48
}
49
50
+/*
51
+ * Handle -drive if=pflash for machines that use properties.
52
+ * If @dinfo is null, do nothing.
53
+ * Else if @fl's property "drive" is already set, fatal error.
54
+ * Else set it to the BlockBackend with @dinfo.
55
+ */
56
+void pflash_cfi01_legacy_drive(PFlashCFI01 *fl, DriveInfo *dinfo)
57
+{
58
+ Location loc;
59
+
60
+ if (!dinfo) {
61
+ return;
62
+ }
63
+
64
+ loc_push_none(&loc);
65
+ qemu_opts_loc_restore(dinfo->opts);
66
+ if (fl->blk) {
67
+ error_report("clashes with -machine");
68
+ exit(1);
69
+ }
70
+ qdev_prop_set_drive(DEVICE(fl), "drive",
71
+ blk_by_legacy_dinfo(dinfo), &error_fatal);
72
+ loc_pop(&loc);
73
+}
74
+
75
static void postload_update_cb(void *opaque, int running, RunState state)
76
{
77
PFlashCFI01 *pfl = opaque;
78
diff --git a/hw/i386/pc_sysfw.c b/hw/i386/pc_sysfw.c
79
index XXXXXXX..XXXXXXX 100644
80
--- a/hw/i386/pc_sysfw.c
81
+++ b/hw/i386/pc_sysfw.c
82
@@ -XXX,XX +XXX,XX @@ void pc_system_firmware_init(PCMachineState *pcms,
83
{
84
PCMachineClass *pcmc = PC_MACHINE_GET_CLASS(pcms);
85
int i;
86
- DriveInfo *pflash_drv;
87
BlockBackend *pflash_blk[ARRAY_SIZE(pcms->flash)];
88
- Location loc;
89
90
if (!pcmc->pci_enabled) {
91
old_pc_system_rom_init(rom_memory, true);
92
@@ -XXX,XX +XXX,XX @@ void pc_system_firmware_init(PCMachineState *pcms,
93
94
/* Map legacy -drive if=pflash to machine properties */
95
for (i = 0; i < ARRAY_SIZE(pcms->flash); i++) {
96
- pflash_drv = drive_get(IF_PFLASH, 0, i);
97
- if (pflash_drv) {
98
- loc_push_none(&loc);
99
- qemu_opts_loc_restore(pflash_drv->opts);
100
- if (pflash_cfi01_get_blk(pcms->flash[i])) {
101
- error_report("clashes with -machine");
102
- exit(1);
103
- }
104
- qdev_prop_set_drive(DEVICE(pcms->flash[i]), "drive",
105
- blk_by_legacy_dinfo(pflash_drv), &error_fatal);
106
- loc_pop(&loc);
107
- }
108
+ pflash_cfi01_legacy_drive(pcms->flash[i],
109
+ drive_get(IF_PFLASH, 0, i));
110
pflash_blk[i] = pflash_cfi01_get_blk(pcms->flash[i]);
111
}
112
113
--
114
2.20.1
115
116
diff view generated by jsdifflib
1
From: Cao Jiaxi <driver1998@foxmail.com>
1
From: Kaige Li <likaige@loongson.cn>
2
2
3
Windows ARM64 uses LLP64 model, which breaks current assumptions.
3
GCC version 4.9.4 isn't clever enough to figure out that all
4
execution paths in disas_ldst() that use 'fn' will have initialized
5
it first, and so it warns:
4
6
5
Signed-off-by: Cao Jiaxi <driver1998@foxmail.com>
7
/home/LiKaige/qemu/target/arm/translate-a64.c: In function ‘disas_ldst’:
6
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
8
/home/LiKaige/qemu/target/arm/translate-a64.c:3392:5: error: ‘fn’ may be used uninitialized in this function [-Werror=maybe-uninitialized]
7
Reviewed-by: Thomas Huth <thuth@redhat.com>
9
fn(cpu_reg(s, rt), clean_addr, tcg_rs, get_mem_index(s),
8
Message-id: 20190503003707.10185-1-driver1998@foxmail.com
10
^
9
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
11
/home/LiKaige/qemu/target/arm/translate-a64.c:3318:22: note: ‘fn’ was declared here
10
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
12
AtomicThreeOpFn *fn;
13
^
14
15
Make it happy by initializing the variable to NULL.
16
17
Signed-off-by: Kaige Li <likaige@loongson.cn>
18
Message-id: 1596110248-7366-2-git-send-email-likaige@loongson.cn
19
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
20
[PMM: Clean up commit message and note which gcc version this was]
11
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
21
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
12
---
22
---
13
util/cacheinfo.c | 2 +-
23
target/arm/translate-a64.c | 2 +-
14
1 file changed, 1 insertion(+), 1 deletion(-)
24
1 file changed, 1 insertion(+), 1 deletion(-)
15
25
16
diff --git a/util/cacheinfo.c b/util/cacheinfo.c
26
diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c
17
index XXXXXXX..XXXXXXX 100644
27
index XXXXXXX..XXXXXXX 100644
18
--- a/util/cacheinfo.c
28
--- a/target/arm/translate-a64.c
19
+++ b/util/cacheinfo.c
29
+++ b/target/arm/translate-a64.c
20
@@ -XXX,XX +XXX,XX @@ static void sys_cache_info(int *isize, int *dsize)
30
@@ -XXX,XX +XXX,XX @@ static void disas_ldst_atomic(DisasContext *s, uint32_t insn,
21
static void arch_cache_info(int *isize, int *dsize)
31
bool r = extract32(insn, 22, 1);
22
{
32
bool a = extract32(insn, 23, 1);
23
if (*isize == 0 || *dsize == 0) {
33
TCGv_i64 tcg_rs, clean_addr;
24
- unsigned long ctr;
34
- AtomicThreeOpFn *fn;
25
+ uint64_t ctr;
35
+ AtomicThreeOpFn *fn = NULL;
26
36
27
/* The real cache geometry is in CCSIDR_EL1/CLIDR_EL1/CSSELR_EL1,
37
if (is_vector || !dc_isar_feature(aa64_atomics, s)) {
28
but (at least under Linux) these are marked protected by the
38
unallocated_encoding(s);
29
--
39
--
30
2.20.1
40
2.20.1
31
41
32
42
diff view generated by jsdifflib
1
The non-secure versions of the BFAR and BFSR registers are
1
The nrf51 SoC model wasn't setting the system_clock_scale
2
supposed to be RAZ/WI if AICR.BFHFNMINS == 0; we were
2
global.which meant that if guest code used the systick timer in "use
3
incorrectly allowing NS code to access the real values.
3
the processor clock" mode it would hang because time never advances.
4
5
Set the global to match the documented CPU clock speed for this SoC.
6
7
This SoC in fact doesn't have a SysTick timer (which is the only thing
8
currently that cares about the system_clock_scale), because it's
9
a configurable option in the Cortex-M0. However our Cortex-M0 and
10
thus our nrf51 and our micro:bit board do provide a SysTick, so
11
we ought to provide a functional one rather than a broken one.
4
12
5
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
13
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
6
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
14
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
7
Message-id: 20190430131439.25251-3-peter.maydell@linaro.org
15
Message-id: 20200727193458.31250-1-peter.maydell@linaro.org
8
---
16
---
9
hw/intc/armv7m_nvic.c | 27 ++++++++++++++++++++++++---
17
hw/arm/nrf51_soc.c | 5 +++++
10
1 file changed, 24 insertions(+), 3 deletions(-)
18
1 file changed, 5 insertions(+)
11
19
12
diff --git a/hw/intc/armv7m_nvic.c b/hw/intc/armv7m_nvic.c
20
diff --git a/hw/arm/nrf51_soc.c b/hw/arm/nrf51_soc.c
13
index XXXXXXX..XXXXXXX 100644
21
index XXXXXXX..XXXXXXX 100644
14
--- a/hw/intc/armv7m_nvic.c
22
--- a/hw/arm/nrf51_soc.c
15
+++ b/hw/intc/armv7m_nvic.c
23
+++ b/hw/arm/nrf51_soc.c
16
@@ -XXX,XX +XXX,XX @@ static uint32_t nvic_readl(NVICState *s, uint32_t offset, MemTxAttrs attrs)
24
@@ -XXX,XX +XXX,XX @@
17
if (!arm_feature(&cpu->env, ARM_FEATURE_M_MAIN)) {
25
18
goto bad_offset;
26
#define BASE_TO_IRQ(base) ((base >> 12) & 0x1F)
19
}
27
20
+ if (!attrs.secure &&
28
+/* HCLK (the main CPU clock) on this SoC is always 16MHz */
21
+ !(s->cpu->env.v7m.aircr & R_V7M_AIRCR_BFHFNMINS_MASK)) {
29
+#define HCLK_FRQ 16000000
22
+ return 0;
30
+
23
+ }
31
static uint64_t clock_read(void *opaque, hwaddr addr, unsigned int size)
24
return cpu->env.v7m.bfar;
32
{
25
case 0xd3c: /* Aux Fault Status. */
33
qemu_log_mask(LOG_UNIMP, "%s: 0x%" HWADDR_PRIx " [%u]\n",
26
/* TODO: Implement fault status registers. */
34
@@ -XXX,XX +XXX,XX @@ static void nrf51_soc_realize(DeviceState *dev_soc, Error **errp)
27
@@ -XXX,XX +XXX,XX @@ static void nvic_writel(NVICState *s, uint32_t offset, uint32_t value,
28
if (!arm_feature(&cpu->env, ARM_FEATURE_M_MAIN)) {
29
goto bad_offset;
30
}
31
+ if (!attrs.secure &&
32
+ !(s->cpu->env.v7m.aircr & R_V7M_AIRCR_BFHFNMINS_MASK)) {
33
+ return;
34
+ }
35
cpu->env.v7m.bfar = value;
36
return;
35
return;
37
case 0xd3c: /* Aux Fault Status. */
36
}
38
@@ -XXX,XX +XXX,XX @@ static MemTxResult nvic_sysreg_read(void *opaque, hwaddr addr,
37
39
val = 0;
38
+ system_clock_scale = NANOSECONDS_PER_SECOND / HCLK_FRQ;
40
break;
41
};
42
- /* The BFSR bits [15:8] are shared between security states
43
- * and we store them in the NS copy
44
+ /*
45
+ * The BFSR bits [15:8] are shared between security states
46
+ * and we store them in the NS copy. They are RAZ/WI for
47
+ * NS code if AIRCR.BFHFNMINS is 0.
48
*/
49
val = s->cpu->env.v7m.cfsr[attrs.secure];
50
- val |= s->cpu->env.v7m.cfsr[M_REG_NS] & R_V7M_CFSR_BFSR_MASK;
51
+ if (!attrs.secure &&
52
+ !(s->cpu->env.v7m.aircr & R_V7M_AIRCR_BFHFNMINS_MASK)) {
53
+ val &= ~R_V7M_CFSR_BFSR_MASK;
54
+ } else {
55
+ val |= s->cpu->env.v7m.cfsr[M_REG_NS] & R_V7M_CFSR_BFSR_MASK;
56
+ }
57
val = extract32(val, (offset - 0xd28) * 8, size * 8);
58
break;
59
case 0xfe0 ... 0xfff: /* ID. */
60
@@ -XXX,XX +XXX,XX @@ static MemTxResult nvic_sysreg_write(void *opaque, hwaddr addr,
61
*/
62
value <<= ((offset - 0xd28) * 8);
63
64
+ if (!attrs.secure &&
65
+ !(s->cpu->env.v7m.aircr & R_V7M_AIRCR_BFHFNMINS_MASK)) {
66
+ /* BFSR bits are RAZ/WI for NS if BFHFNMINS is set */
67
+ value &= ~R_V7M_CFSR_BFSR_MASK;
68
+ }
69
+
39
+
70
s->cpu->env.v7m.cfsr[attrs.secure] &= ~value;
40
object_property_set_link(OBJECT(&s->cpu), "memory", OBJECT(&s->container),
71
if (attrs.secure) {
41
&error_abort);
72
/* The BFSR bits [15:8] are shared between security states
42
if (!sysbus_realize(SYS_BUS_DEVICE(&s->cpu), errp)) {
73
--
43
--
74
2.20.1
44
2.20.1
75
45
76
46
diff view generated by jsdifflib
1
From: Markus Armbruster <armbru@redhat.com>
1
The imx_epit device has a software-controllable reset triggered by
2
setting the SWR bit in the CR register. An error in commit cc2722ec83ad9
3
means that we will end up assert()ing if the guest does this, because
4
the code in imx_epit_write() starts ptimer transactions, and then
5
imx_epit_reset() also starts ptimer transactions, triggering
6
"ptimer_transaction_begin: Assertion `!s->in_transaction' failed".
2
7
3
The ARM virt machines put firmware in flash memory. To configure it,
8
The cleanest way to avoid this double-transaction is to move the
4
you use -drive if=pflash,unit=0,... and optionally -drive
9
start-transaction for the CR write handling down below the check of
5
if=pflash,unit=1,...
10
the SWR bit.
6
11
7
Why two -drive? This permits setting up one part of the flash memory
12
Fixes: https://bugs.launchpad.net/qemu/+bug/1880424
8
read-only, and the other part read/write. It also makes upgrading
13
Fixes: cc2722ec83ad944505fe
9
firmware on the host easier. Below the hood, we get two separate
14
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10
flash devices, because we were too lazy to improve our flash device
15
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
11
models to support sector protection.
16
Message-id: 20200727154550.3409-1-peter.maydell@linaro.org
17
---
18
hw/timer/imx_epit.c | 13 ++++++++++---
19
1 file changed, 10 insertions(+), 3 deletions(-)
12
20
13
The problem at hand is to do the same with -blockdev somehow, as one
21
diff --git a/hw/timer/imx_epit.c b/hw/timer/imx_epit.c
14
more step towards deprecating -drive.
15
16
We recently solved this problem for x86 PC machines, in commit
17
ebc29e1beab. See the commit message for design rationale.
18
19
This commit solves it for ARM virt basically the same way: new machine
20
properties pflash0, pflash1 forward to the onboard flash devices'
21
properties. Requires creating the onboard devices in the
22
.instance_init() method virt_instance_init(). The existing code to
23
pick up drives defined with -drive if=pflash is replaced by code to
24
desugar into the machine properties.
25
26
There are a few behavioral differences, though:
27
28
* The flash devices are always present (x86: only present if
29
configured)
30
31
* Flash base addresses and sizes are fixed (x86: sizes depend on
32
images, mapped back to back below a fixed address)
33
34
* -bios configures contents of first pflash (x86: -bios configures ROM
35
contents)
36
37
* -bios is rejected when first pflash is also configured with -machine
38
pflash0=... (x86: bios is silently ignored then)
39
40
* -machine pflash1=... does not require -machine pflash0=... (x86: it
41
does).
42
43
The actual code is a bit simpler than for x86 mostly due to the first
44
two differences.
45
46
Before the patch, all the action is in create_flash(), called from the
47
machine's .init() method machvirt_init():
48
49
main()
50
machine_run_board_init()
51
machvirt_init()
52
create_flash()
53
create_one_flash() for flash[0]
54
create
55
configure
56
includes obeying -drive if=pflash,unit=0
57
realize
58
map
59
fall back to -bios
60
create_one_flash() for flash[1]
61
create
62
configure
63
includes obeying -drive if=pflash,unit=1
64
realize
65
map
66
update FDT
67
68
To make the machine properties work, we need to move device creation
69
to its .instance_init() method virt_instance_init().
70
71
Another complication is machvirt_init()'s computation of
72
@firmware_loaded: it predicts what create_flash() will do. Instead of
73
predicting what create_flash()'s replacement virt_firmware_init() will
74
do, I decided to have virt_firmware_init() return what it did.
75
Requires calling it a bit earlier.
76
77
Resulting call tree:
78
79
main()
80
current_machine = object_new()
81
...
82
virt_instance_init()
83
virt_flash_create()
84
virt_flash_create1() for flash[0]
85
create
86
configure: set defaults
87
become child of machine [NEW]
88
add machine prop pflash0 as alias for drive [NEW]
89
virt_flash_create1() for flash[1]
90
create
91
configure: set defaults
92
become child of machine [NEW]
93
add machine prop pflash1 as alias for drive [NEW]
94
for all machine props from the command line: machine_set_property()
95
...
96
property_set_alias() for machine props pflash0, pflash1
97
...
98
set_drive() for cfi.pflash01 prop drive
99
this is how -machine pflash0=... etc set
100
machine_run_board_init(current_machine);
101
virt_firmware_init()
102
pflash_cfi01_legacy_drive()
103
legacy -drive if=pflash,unit=0 and =1 [NEW]
104
virt_flash_map()
105
virt_flash_map1() for flash[0]
106
configure: num-blocks
107
realize
108
map
109
virt_flash_map1() for flash[1]
110
configure: num-blocks
111
realize
112
map
113
fall back to -bios
114
virt_flash_fdt()
115
update FDT
116
117
You have László to thank for making me explain this in detail.
118
119
Signed-off-by: Markus Armbruster <armbru@redhat.com>
120
Acked-by: Laszlo Ersek <lersek@redhat.com>
121
Message-id: 20190416091348.26075-4-armbru@redhat.com
122
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
123
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
124
---
125
include/hw/arm/virt.h | 2 +
126
hw/arm/virt.c | 202 +++++++++++++++++++++++++++---------------
127
2 files changed, 132 insertions(+), 72 deletions(-)
128
129
diff --git a/include/hw/arm/virt.h b/include/hw/arm/virt.h
130
index XXXXXXX..XXXXXXX 100644
22
index XXXXXXX..XXXXXXX 100644
131
--- a/include/hw/arm/virt.h
23
--- a/hw/timer/imx_epit.c
132
+++ b/include/hw/arm/virt.h
24
+++ b/hw/timer/imx_epit.c
133
@@ -XXX,XX +XXX,XX @@
25
@@ -XXX,XX +XXX,XX @@ static void imx_epit_write(void *opaque, hwaddr offset, uint64_t value,
134
#include "qemu/notify.h"
26
135
#include "hw/boards.h"
27
switch (offset >> 2) {
136
#include "hw/arm/arm.h"
28
case 0: /* CR */
137
+#include "hw/block/flash.h"
29
- ptimer_transaction_begin(s->timer_cmp);
138
#include "sysemu/kvm.h"
30
- ptimer_transaction_begin(s->timer_reload);
139
#include "hw/intc/arm_gicv3_common.h"
31
140
32
oldcr = s->cr;
141
@@ -XXX,XX +XXX,XX @@ typedef struct {
33
s->cr = value & 0x03ffffff;
142
Notifier machine_done;
34
if (s->cr & CR_SWR) {
143
DeviceState *platform_bus_dev;
35
/* handle the reset */
144
FWCfgState *fw_cfg;
36
imx_epit_reset(DEVICE(s));
145
+ PFlashCFI01 *flash[2];
37
- } else {
146
bool secure;
38
+ /*
147
bool highmem;
39
+ * TODO: could we 'break' here? following operations appear
148
bool highmem_ecam;
40
+ * to duplicate the work imx_epit_reset() already did.
149
diff --git a/hw/arm/virt.c b/hw/arm/virt.c
41
+ */
150
index XXXXXXX..XXXXXXX 100644
151
--- a/hw/arm/virt.c
152
+++ b/hw/arm/virt.c
153
@@ -XXX,XX +XXX,XX @@
154
155
#include "qemu/osdep.h"
156
#include "qemu/units.h"
157
+#include "qemu/option.h"
158
#include "qapi/error.h"
159
#include "hw/sysbus.h"
160
#include "hw/arm/arm.h"
161
@@ -XXX,XX +XXX,XX @@ static void create_virtio_devices(const VirtMachineState *vms, qemu_irq *pic)
162
}
163
}
164
165
-static void create_one_flash(const char *name, hwaddr flashbase,
166
- hwaddr flashsize, const char *file,
167
- MemoryRegion *sysmem)
168
+#define VIRT_FLASH_SECTOR_SIZE (256 * KiB)
169
+
170
+static PFlashCFI01 *virt_flash_create1(VirtMachineState *vms,
171
+ const char *name,
172
+ const char *alias_prop_name)
173
{
174
- /* Create and map a single flash device. We use the same
175
- * parameters as the flash devices on the Versatile Express board.
176
+ /*
177
+ * Create a single flash device. We use the same parameters as
178
+ * the flash devices on the Versatile Express board.
179
*/
180
- DriveInfo *dinfo = drive_get_next(IF_PFLASH);
181
DeviceState *dev = qdev_create(NULL, TYPE_PFLASH_CFI01);
182
- SysBusDevice *sbd = SYS_BUS_DEVICE(dev);
183
- const uint64_t sectorlength = 256 * 1024;
184
185
- if (dinfo) {
186
- qdev_prop_set_drive(dev, "drive", blk_by_legacy_dinfo(dinfo),
187
- &error_abort);
188
- }
189
-
190
- qdev_prop_set_uint32(dev, "num-blocks", flashsize / sectorlength);
191
- qdev_prop_set_uint64(dev, "sector-length", sectorlength);
192
+ qdev_prop_set_uint64(dev, "sector-length", VIRT_FLASH_SECTOR_SIZE);
193
qdev_prop_set_uint8(dev, "width", 4);
194
qdev_prop_set_uint8(dev, "device-width", 2);
195
qdev_prop_set_bit(dev, "big-endian", false);
196
@@ -XXX,XX +XXX,XX @@ static void create_one_flash(const char *name, hwaddr flashbase,
197
qdev_prop_set_uint16(dev, "id2", 0x00);
198
qdev_prop_set_uint16(dev, "id3", 0x00);
199
qdev_prop_set_string(dev, "name", name);
200
- qdev_init_nofail(dev);
201
-
202
- memory_region_add_subregion(sysmem, flashbase,
203
- sysbus_mmio_get_region(SYS_BUS_DEVICE(dev), 0));
204
-
205
- if (file) {
206
- char *fn;
207
- int image_size;
208
-
209
- if (drive_get(IF_PFLASH, 0, 0)) {
210
- error_report("The contents of the first flash device may be "
211
- "specified with -bios or with -drive if=pflash... "
212
- "but you cannot use both options at once");
213
- exit(1);
214
- }
215
- fn = qemu_find_file(QEMU_FILE_TYPE_BIOS, file);
216
- if (!fn) {
217
- error_report("Could not find ROM image '%s'", file);
218
- exit(1);
219
- }
220
- image_size = load_image_mr(fn, sysbus_mmio_get_region(sbd, 0));
221
- g_free(fn);
222
- if (image_size < 0) {
223
- error_report("Could not load ROM image '%s'", file);
224
- exit(1);
225
- }
226
- }
227
+ object_property_add_child(OBJECT(vms), name, OBJECT(dev),
228
+ &error_abort);
229
+ object_property_add_alias(OBJECT(vms), alias_prop_name,
230
+ OBJECT(dev), "drive", &error_abort);
231
+ return PFLASH_CFI01(dev);
232
}
233
234
-static void create_flash(const VirtMachineState *vms,
235
- MemoryRegion *sysmem,
236
- MemoryRegion *secure_sysmem)
237
+static void virt_flash_create(VirtMachineState *vms)
238
{
239
- /* Create two flash devices to fill the VIRT_FLASH space in the memmap.
240
- * Any file passed via -bios goes in the first of these.
241
+ vms->flash[0] = virt_flash_create1(vms, "virt.flash0", "pflash0");
242
+ vms->flash[1] = virt_flash_create1(vms, "virt.flash1", "pflash1");
243
+}
244
+
245
+static void virt_flash_map1(PFlashCFI01 *flash,
246
+ hwaddr base, hwaddr size,
247
+ MemoryRegion *sysmem)
248
+{
249
+ DeviceState *dev = DEVICE(flash);
250
+
251
+ assert(size % VIRT_FLASH_SECTOR_SIZE == 0);
252
+ assert(size / VIRT_FLASH_SECTOR_SIZE <= UINT32_MAX);
253
+ qdev_prop_set_uint32(dev, "num-blocks", size / VIRT_FLASH_SECTOR_SIZE);
254
+ qdev_init_nofail(dev);
255
+
256
+ memory_region_add_subregion(sysmem, base,
257
+ sysbus_mmio_get_region(SYS_BUS_DEVICE(dev),
258
+ 0));
259
+}
260
+
261
+static void virt_flash_map(VirtMachineState *vms,
262
+ MemoryRegion *sysmem,
263
+ MemoryRegion *secure_sysmem)
264
+{
265
+ /*
266
+ * Map two flash devices to fill the VIRT_FLASH space in the memmap.
267
* sysmem is the system memory space. secure_sysmem is the secure view
268
* of the system, and the first flash device should be made visible only
269
* there. The second flash device is visible to both secure and nonsecure.
270
@@ -XXX,XX +XXX,XX @@ static void create_flash(const VirtMachineState *vms,
271
*/
272
hwaddr flashsize = vms->memmap[VIRT_FLASH].size / 2;
273
hwaddr flashbase = vms->memmap[VIRT_FLASH].base;
274
- char *nodename;
275
276
- create_one_flash("virt.flash0", flashbase, flashsize,
277
- bios_name, secure_sysmem);
278
- create_one_flash("virt.flash1", flashbase + flashsize, flashsize,
279
- NULL, sysmem);
280
+ virt_flash_map1(vms->flash[0], flashbase, flashsize,
281
+ secure_sysmem);
282
+ virt_flash_map1(vms->flash[1], flashbase + flashsize, flashsize,
283
+ sysmem);
284
+}
285
+
286
+static void virt_flash_fdt(VirtMachineState *vms,
287
+ MemoryRegion *sysmem,
288
+ MemoryRegion *secure_sysmem)
289
+{
290
+ hwaddr flashsize = vms->memmap[VIRT_FLASH].size / 2;
291
+ hwaddr flashbase = vms->memmap[VIRT_FLASH].base;
292
+ char *nodename;
293
294
if (sysmem == secure_sysmem) {
295
/* Report both flash devices as a single node in the DT */
296
@@ -XXX,XX +XXX,XX @@ static void create_flash(const VirtMachineState *vms,
297
qemu_fdt_setprop_cell(vms->fdt, nodename, "bank-width", 4);
298
g_free(nodename);
299
} else {
300
- /* Report the devices as separate nodes so we can mark one as
301
+ /*
302
+ * Report the devices as separate nodes so we can mark one as
303
* only visible to the secure world.
304
*/
305
nodename = g_strdup_printf("/secflash@%" PRIx64, flashbase);
306
@@ -XXX,XX +XXX,XX @@ static void create_flash(const VirtMachineState *vms,
307
}
308
}
309
310
+static bool virt_firmware_init(VirtMachineState *vms,
311
+ MemoryRegion *sysmem,
312
+ MemoryRegion *secure_sysmem)
313
+{
314
+ int i;
315
+ BlockBackend *pflash_blk0;
316
+
317
+ /* Map legacy -drive if=pflash to machine properties */
318
+ for (i = 0; i < ARRAY_SIZE(vms->flash); i++) {
319
+ pflash_cfi01_legacy_drive(vms->flash[i],
320
+ drive_get(IF_PFLASH, 0, i));
321
+ }
322
+
323
+ virt_flash_map(vms, sysmem, secure_sysmem);
324
+
325
+ pflash_blk0 = pflash_cfi01_get_blk(vms->flash[0]);
326
+
327
+ if (bios_name) {
328
+ char *fname;
329
+ MemoryRegion *mr;
330
+ int image_size;
331
+
332
+ if (pflash_blk0) {
333
+ error_report("The contents of the first flash device may be "
334
+ "specified with -bios or with -drive if=pflash... "
335
+ "but you cannot use both options at once");
336
+ exit(1);
337
+ }
42
+ }
338
+
43
+
339
+ /* Fall back to -bios */
44
+ ptimer_transaction_begin(s->timer_cmp);
45
+ ptimer_transaction_begin(s->timer_reload);
340
+
46
+
341
+ fname = qemu_find_file(QEMU_FILE_TYPE_BIOS, bios_name);
47
+ if (!(s->cr & CR_SWR)) {
342
+ if (!fname) {
48
imx_epit_set_freq(s);
343
+ error_report("Could not find ROM image '%s'", bios_name);
49
}
344
+ exit(1);
50
345
+ }
346
+ mr = sysbus_mmio_get_region(SYS_BUS_DEVICE(vms->flash[0]), 0);
347
+ image_size = load_image_mr(fname, mr);
348
+ g_free(fname);
349
+ if (image_size < 0) {
350
+ error_report("Could not load ROM image '%s'", bios_name);
351
+ exit(1);
352
+ }
353
+ }
354
+
355
+ return pflash_blk0 || bios_name;
356
+}
357
+
358
static FWCfgState *create_fw_cfg(const VirtMachineState *vms, AddressSpace *as)
359
{
360
hwaddr base = vms->memmap[VIRT_FW_CFG].base;
361
@@ -XXX,XX +XXX,XX @@ static void machvirt_init(MachineState *machine)
362
MemoryRegion *secure_sysmem = NULL;
363
int n, virt_max_cpus;
364
MemoryRegion *ram = g_new(MemoryRegion, 1);
365
- bool firmware_loaded = bios_name || drive_get(IF_PFLASH, 0, 0);
366
+ bool firmware_loaded;
367
bool aarch64 = true;
368
369
/*
370
@@ -XXX,XX +XXX,XX @@ static void machvirt_init(MachineState *machine)
371
exit(1);
372
}
373
374
+ if (vms->secure) {
375
+ if (kvm_enabled()) {
376
+ error_report("mach-virt: KVM does not support Security extensions");
377
+ exit(1);
378
+ }
379
+
380
+ /*
381
+ * The Secure view of the world is the same as the NonSecure,
382
+ * but with a few extra devices. Create it as a container region
383
+ * containing the system memory at low priority; any secure-only
384
+ * devices go in at higher priority and take precedence.
385
+ */
386
+ secure_sysmem = g_new(MemoryRegion, 1);
387
+ memory_region_init(secure_sysmem, OBJECT(machine), "secure-memory",
388
+ UINT64_MAX);
389
+ memory_region_add_subregion_overlap(secure_sysmem, 0, sysmem, -1);
390
+ }
391
+
392
+ firmware_loaded = virt_firmware_init(vms, sysmem,
393
+ secure_sysmem ?: sysmem);
394
+
395
/* If we have an EL3 boot ROM then the assumption is that it will
396
* implement PSCI itself, so disable QEMU's internal implementation
397
* so it doesn't get in the way. Instead of starting secondary
398
@@ -XXX,XX +XXX,XX @@ static void machvirt_init(MachineState *machine)
399
exit(1);
400
}
401
402
- if (vms->secure) {
403
- if (kvm_enabled()) {
404
- error_report("mach-virt: KVM does not support Security extensions");
405
- exit(1);
406
- }
407
-
408
- /* The Secure view of the world is the same as the NonSecure,
409
- * but with a few extra devices. Create it as a container region
410
- * containing the system memory at low priority; any secure-only
411
- * devices go in at higher priority and take precedence.
412
- */
413
- secure_sysmem = g_new(MemoryRegion, 1);
414
- memory_region_init(secure_sysmem, OBJECT(machine), "secure-memory",
415
- UINT64_MAX);
416
- memory_region_add_subregion_overlap(secure_sysmem, 0, sysmem, -1);
417
- }
418
-
419
create_fdt(vms);
420
421
possible_cpus = mc->possible_cpu_arch_ids(machine);
422
@@ -XXX,XX +XXX,XX @@ static void machvirt_init(MachineState *machine)
423
&machine->device_memory->mr);
424
}
425
426
- create_flash(vms, sysmem, secure_sysmem ? secure_sysmem : sysmem);
427
+ virt_flash_fdt(vms, sysmem, secure_sysmem);
428
429
create_gic(vms, pic);
430
431
@@ -XXX,XX +XXX,XX @@ static void virt_instance_init(Object *obj)
432
NULL);
433
434
vms->irqmap = a15irqmap;
435
+
436
+ virt_flash_create(vms);
437
}
438
439
static const TypeInfo virt_machine_info = {
440
--
51
--
441
2.20.1
52
2.20.1
442
53
443
54
diff view generated by jsdifflib
Deleted patch
1
The Raspberry Pi boards have a physical memory map which does
2
not allow for more than 1GB of RAM. Currently if the user tries
3
to ask for more then we fail in a confusing way:
4
1
5
$ qemu-system-aarch64 --machine raspi3 -m 8G
6
Unexpected error in visit_type_uintN() at qapi/qapi-visit-core.c:164:
7
qemu-system-aarch64: Parameter 'vcram-base' expects uint32_t
8
Aborted (core dumped)
9
10
Catch this earlier and diagnose it with a more friendly message:
11
$ qemu-system-aarch64 --machine raspi3 -m 8G
12
qemu-system-aarch64: Requested ram size is too large for this machine: maximum is 1GB
13
14
Fixes: https://bugs.launchpad.net/qemu/+bug/1794187
15
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
16
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
17
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
18
Reviewed-by: Wainer dos Santos Moschetta <wainersm@redhat.com>
19
---
20
hw/arm/raspi.c | 7 +++++++
21
1 file changed, 7 insertions(+)
22
23
diff --git a/hw/arm/raspi.c b/hw/arm/raspi.c
24
index XXXXXXX..XXXXXXX 100644
25
--- a/hw/arm/raspi.c
26
+++ b/hw/arm/raspi.c
27
@@ -XXX,XX +XXX,XX @@
28
*/
29
30
#include "qemu/osdep.h"
31
+#include "qemu/units.h"
32
#include "qapi/error.h"
33
#include "qemu-common.h"
34
#include "cpu.h"
35
@@ -XXX,XX +XXX,XX @@ static void raspi_init(MachineState *machine, int version)
36
BusState *bus;
37
DeviceState *carddev;
38
39
+ if (machine->ram_size > 1 * GiB) {
40
+ error_report("Requested ram size is too large for this machine: "
41
+ "maximum is 1GB");
42
+ exit(1);
43
+ }
44
+
45
object_initialize(&s->soc, sizeof(s->soc),
46
version == 3 ? TYPE_BCM2837 : TYPE_BCM2836);
47
object_property_add_child(OBJECT(machine), "soc", OBJECT(&s->soc),
48
--
49
2.20.1
50
51
diff view generated by jsdifflib
Deleted patch
1
From: Joel Stanley <joel@jms.id.au>
2
1
3
We currently use Qemu's default of 128MB. As we know how much ram each
4
machine ships with, make it easier on users by setting a default.
5
6
It can still be overridden with -m on the command line.
7
8
Signed-off-by: Joel Stanley <joel@jms.id.au>
9
Reviewed-by: Andrew Jeffery <andrew@aj.id.au>
10
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
11
Message-id: 20190503022958.1394-1-joel@jms.id.au
12
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
13
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
14
---
15
include/hw/arm/aspeed.h | 1 +
16
hw/arm/aspeed.c | 8 ++++++++
17
2 files changed, 9 insertions(+)
18
19
diff --git a/include/hw/arm/aspeed.h b/include/hw/arm/aspeed.h
20
index XXXXXXX..XXXXXXX 100644
21
--- a/include/hw/arm/aspeed.h
22
+++ b/include/hw/arm/aspeed.h
23
@@ -XXX,XX +XXX,XX @@ typedef struct AspeedBoardConfig {
24
const char *spi_model;
25
uint32_t num_cs;
26
void (*i2c_init)(AspeedBoardState *bmc);
27
+ uint32_t ram;
28
} AspeedBoardConfig;
29
30
#define TYPE_ASPEED_MACHINE MACHINE_TYPE_NAME("aspeed")
31
diff --git a/hw/arm/aspeed.c b/hw/arm/aspeed.c
32
index XXXXXXX..XXXXXXX 100644
33
--- a/hw/arm/aspeed.c
34
+++ b/hw/arm/aspeed.c
35
@@ -XXX,XX +XXX,XX @@
36
#include "sysemu/block-backend.h"
37
#include "hw/loader.h"
38
#include "qemu/error-report.h"
39
+#include "qemu/units.h"
40
41
static struct arm_boot_info aspeed_board_binfo = {
42
.board_id = -1, /* device-tree-only board */
43
@@ -XXX,XX +XXX,XX @@ static void aspeed_machine_class_init(ObjectClass *oc, void *data)
44
mc->no_floppy = 1;
45
mc->no_cdrom = 1;
46
mc->no_parallel = 1;
47
+ if (board->ram) {
48
+ mc->default_ram_size = board->ram;
49
+ }
50
amc->board = board;
51
}
52
53
@@ -XXX,XX +XXX,XX @@ static const AspeedBoardConfig aspeed_boards[] = {
54
.spi_model = "mx25l25635e",
55
.num_cs = 1,
56
.i2c_init = palmetto_bmc_i2c_init,
57
+ .ram = 256 * MiB,
58
}, {
59
.name = MACHINE_TYPE_NAME("ast2500-evb"),
60
.desc = "Aspeed AST2500 EVB (ARM1176)",
61
@@ -XXX,XX +XXX,XX @@ static const AspeedBoardConfig aspeed_boards[] = {
62
.spi_model = "mx25l25635e",
63
.num_cs = 1,
64
.i2c_init = ast2500_evb_i2c_init,
65
+ .ram = 512 * MiB,
66
}, {
67
.name = MACHINE_TYPE_NAME("romulus-bmc"),
68
.desc = "OpenPOWER Romulus BMC (ARM1176)",
69
@@ -XXX,XX +XXX,XX @@ static const AspeedBoardConfig aspeed_boards[] = {
70
.spi_model = "mx66l1g45g",
71
.num_cs = 2,
72
.i2c_init = romulus_bmc_i2c_init,
73
+ .ram = 512 * MiB,
74
}, {
75
.name = MACHINE_TYPE_NAME("witherspoon-bmc"),
76
.desc = "OpenPOWER Witherspoon BMC (ARM1176)",
77
@@ -XXX,XX +XXX,XX @@ static const AspeedBoardConfig aspeed_boards[] = {
78
.spi_model = "mx66l1g45g",
79
.num_cs = 2,
80
.i2c_init = witherspoon_bmc_i2c_init,
81
+ .ram = 512 * MiB,
82
},
83
};
84
85
--
86
2.20.1
87
88
diff view generated by jsdifflib
Deleted patch
1
From: Cao Jiaxi <driver1998@foxmail.com>
2
1
3
gcc_struct is for x86 only, and it generates an warning on ARM64 Clang/MinGW targets.
4
5
Signed-off-by: Cao Jiaxi <driver1998@foxmail.com>
6
Reviewed-by: Thomas Huth <thuth@redhat.com>
7
Message-id: 20190503003618.10089-1-driver1998@foxmail.com
8
[PMM: dropped the slirp change as slirp is now a submodule]
9
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10
---
11
contrib/libvhost-user/libvhost-user.h | 2 +-
12
include/qemu/compiler.h | 2 +-
13
scripts/cocci-macro-file.h | 7 ++++++-
14
3 files changed, 8 insertions(+), 3 deletions(-)
15
16
diff --git a/contrib/libvhost-user/libvhost-user.h b/contrib/libvhost-user/libvhost-user.h
17
index XXXXXXX..XXXXXXX 100644
18
--- a/contrib/libvhost-user/libvhost-user.h
19
+++ b/contrib/libvhost-user/libvhost-user.h
20
@@ -XXX,XX +XXX,XX @@ typedef struct VhostUserInflight {
21
uint16_t queue_size;
22
} VhostUserInflight;
23
24
-#if defined(_WIN32)
25
+#if defined(_WIN32) && (defined(__x86_64__) || defined(__i386__))
26
# define VU_PACKED __attribute__((gcc_struct, packed))
27
#else
28
# define VU_PACKED __attribute__((packed))
29
diff --git a/include/qemu/compiler.h b/include/qemu/compiler.h
30
index XXXXXXX..XXXXXXX 100644
31
--- a/include/qemu/compiler.h
32
+++ b/include/qemu/compiler.h
33
@@ -XXX,XX +XXX,XX @@
34
35
#define QEMU_SENTINEL __attribute__((sentinel))
36
37
-#if defined(_WIN32)
38
+#if defined(_WIN32) && (defined(__x86_64__) || defined(__i386__))
39
# define QEMU_PACKED __attribute__((gcc_struct, packed))
40
#else
41
# define QEMU_PACKED __attribute__((packed))
42
diff --git a/scripts/cocci-macro-file.h b/scripts/cocci-macro-file.h
43
index XXXXXXX..XXXXXXX 100644
44
--- a/scripts/cocci-macro-file.h
45
+++ b/scripts/cocci-macro-file.h
46
@@ -XXX,XX +XXX,XX @@
47
#define QEMU_NORETURN __attribute__ ((__noreturn__))
48
#define QEMU_WARN_UNUSED_RESULT __attribute__((warn_unused_result))
49
#define QEMU_SENTINEL __attribute__((sentinel))
50
-#define QEMU_PACKED __attribute__((gcc_struct, packed))
51
+
52
+#if defined(_WIN32) && (defined(__x86_64__) || defined(__i386__))
53
+# define QEMU_PACKED __attribute__((gcc_struct, packed))
54
+#else
55
+# define QEMU_PACKED __attribute__((packed))
56
+#endif
57
58
#define cat(x,y) x ## y
59
#define cat2(x,y) cat(x,y)
60
--
61
2.20.1
62
63
diff view generated by jsdifflib
Deleted patch
1
From: Cao Jiaxi <driver1998@foxmail.com>
2
1
3
The win2qemu[] is supposed to be the conversion table to convert between
4
STORAGE_BUS_TYPE in Windows SDK and GuestDiskBusType in qga.
5
6
But it was incorrectly written that it forces to set a GuestDiskBusType
7
value to STORAGE_BUS_TYPE, which generates an enum conversion warning in clang.
8
9
Suggested-by: Eric Blake <eblake@redhat.com>
10
Signed-off-by: Cao Jiaxi <driver1998@foxmail.com>
11
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
12
Reviewed-by: Thomas Huth <thuth@redhat.com>
13
Message-id: 20190503003650.10137-1-driver1998@foxmail.com
14
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
15
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
16
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
17
---
18
qga/commands-win32.c | 2 +-
19
1 file changed, 1 insertion(+), 1 deletion(-)
20
21
diff --git a/qga/commands-win32.c b/qga/commands-win32.c
22
index XXXXXXX..XXXXXXX 100644
23
--- a/qga/commands-win32.c
24
+++ b/qga/commands-win32.c
25
@@ -XXX,XX +XXX,XX @@ void qmp_guest_file_flush(int64_t handle, Error **errp)
26
27
#ifdef CONFIG_QGA_NTDDSCSI
28
29
-static STORAGE_BUS_TYPE win2qemu[] = {
30
+static GuestDiskBusType win2qemu[] = {
31
[BusTypeUnknown] = GUEST_DISK_BUS_TYPE_UNKNOWN,
32
[BusTypeScsi] = GUEST_DISK_BUS_TYPE_SCSI,
33
[BusTypeAtapi] = GUEST_DISK_BUS_TYPE_IDE,
34
--
35
2.20.1
36
37
diff view generated by jsdifflib
Deleted patch
1
From: Cao Jiaxi <driver1998@foxmail.com>
2
1
3
I encountered the following compilation error on mingw:
4
5
/mnt/d/qemu/include/qemu/osdep.h:97:9: error: '__USE_MINGW_ANSI_STDIO' macro redefined [-Werror,-Wmacro-redefined]
6
#define __USE_MINGW_ANSI_STDIO 1
7
^
8
/mnt/d/llvm-mingw/aarch64-w64-mingw32/include/_mingw.h:433:9: note: previous definition is here
9
#define __USE_MINGW_ANSI_STDIO 0 /* was not defined so it should be 0 */
10
11
It turns out that __USE_MINGW_ANSI_STDIO must be set before any
12
system headers are included, not just before stdio.h.
13
14
Signed-off-by: Cao Jiaxi <driver1998@foxmail.com>
15
Reviewed-by: Thomas Huth <thuth@redhat.com>
16
Reviewed-by: Stefan Weil <sw@weilnetz.de>
17
Message-id: 20190503003719.10233-1-driver1998@foxmail.com
18
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
19
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
20
---
21
include/qemu/osdep.h | 10 +++++-----
22
1 file changed, 5 insertions(+), 5 deletions(-)
23
24
diff --git a/include/qemu/osdep.h b/include/qemu/osdep.h
25
index XXXXXXX..XXXXXXX 100644
26
--- a/include/qemu/osdep.h
27
+++ b/include/qemu/osdep.h
28
@@ -XXX,XX +XXX,XX @@ extern int daemon(int, int);
29
#endif
30
#endif
31
32
+/* enable C99/POSIX format strings (needs mingw32-runtime 3.15 or later) */
33
+#ifdef __MINGW32__
34
+#define __USE_MINGW_ANSI_STDIO 1
35
+#endif
36
+
37
#include <stdarg.h>
38
#include <stddef.h>
39
#include <stdbool.h>
40
#include <stdint.h>
41
#include <sys/types.h>
42
#include <stdlib.h>
43
-
44
-/* enable C99/POSIX format strings (needs mingw32-runtime 3.15 or later) */
45
-#ifdef __MINGW32__
46
-#define __USE_MINGW_ANSI_STDIO 1
47
-#endif
48
#include <stdio.h>
49
50
#include <string.h>
51
--
52
2.20.1
53
54
diff view generated by jsdifflib
Deleted patch
1
Rule R_CQRV says that if two pending interrupts have the same
2
group priority then ties are broken by looking at the subpriority.
3
We had a comment describing this but had forgotten to actually
4
implement the subpriority comparison. Correct the omission.
5
1
6
(The further tie break rules of "lowest exception number" and
7
"secure before non-secure" are handled implicitly by the order
8
in which we iterate through the exceptions in the loops.)
9
10
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
11
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
12
Message-id: 20190430131439.25251-2-peter.maydell@linaro.org
13
---
14
hw/intc/armv7m_nvic.c | 9 +++++++--
15
1 file changed, 7 insertions(+), 2 deletions(-)
16
17
diff --git a/hw/intc/armv7m_nvic.c b/hw/intc/armv7m_nvic.c
18
index XXXXXXX..XXXXXXX 100644
19
--- a/hw/intc/armv7m_nvic.c
20
+++ b/hw/intc/armv7m_nvic.c
21
@@ -XXX,XX +XXX,XX @@ static void nvic_recompute_state_secure(NVICState *s)
22
int active_prio = NVIC_NOEXC_PRIO;
23
int pend_irq = 0;
24
bool pending_is_s_banked = false;
25
+ int pend_subprio = 0;
26
27
/* R_CQRV: precedence is by:
28
* - lowest group priority; if both the same then
29
@@ -XXX,XX +XXX,XX @@ static void nvic_recompute_state_secure(NVICState *s)
30
for (i = 1; i < s->num_irq; i++) {
31
for (bank = M_REG_S; bank >= M_REG_NS; bank--) {
32
VecInfo *vec;
33
- int prio;
34
+ int prio, subprio;
35
bool targets_secure;
36
37
if (bank == M_REG_S) {
38
@@ -XXX,XX +XXX,XX @@ static void nvic_recompute_state_secure(NVICState *s)
39
}
40
41
prio = exc_group_prio(s, vec->prio, targets_secure);
42
- if (vec->enabled && vec->pending && prio < pend_prio) {
43
+ subprio = vec->prio & ~nvic_gprio_mask(s, targets_secure);
44
+ if (vec->enabled && vec->pending &&
45
+ ((prio < pend_prio) ||
46
+ (prio == pend_prio && prio >= 0 && subprio < pend_subprio))) {
47
pend_prio = prio;
48
+ pend_subprio = subprio;
49
pend_irq = i;
50
pending_is_s_banked = (bank == M_REG_S);
51
}
52
--
53
2.20.1
54
55
diff view generated by jsdifflib