Since commit 4047368938 "accel/tcg: Introduce tlb_set_page_full" we
have been seeing this assert
../accel/tcg/cputlb.c:1294: tlb_set_page_with_attrs: Assertion `is_power_of_2(size)' failed.
When running Tock on the OpenTitan machine.
The issue is that pmp_get_tlb_size() would return a TLB size that wasn't
a power of 2. The size was also smaller then TARGET_PAGE_SIZE.
This patch ensures that any TLB size less then TARGET_PAGE_SIZE is
rounded down to 1 to ensure it's a valid size.
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: LIU Zhiwei<zhiwei_liu@linux.alibaba.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20221012011449.506928-1-alistair.francis@opensource.wdc.com
Message-Id: <20221012011449.506928-1-alistair.francis@opensource.wdc.com>
For a TOR entry to match, the stard address must be lower than the end
address. Normally this is always the case, but correct code might still
run into the following scenario:
Initial state:
pmpaddr3 = 0x2000 pmp3cfg = OFF
pmpaddr4 = 0x3000 pmp4cfg = TOR
Execution:
1. write 0x40ff to pmpaddr3
2. write 0x32ff to pmpaddr4
3. set pmp3cfg to NAPOT with a read-modify-write on pmpcfg0
4. set pmp4cfg to NAPOT with a read-modify-write on pmpcfg1
When (2) is emulated, a call to pmp_update_rule() creates a negative
range for pmp4 as pmp4cfg is still set to TOR. And when (3) is emulated,
a call to tlb_flush() is performed, causing pmp_get_tlb_size() to return
a very creatively large TLB size for pmp4. This, in turn, may result in
accesses to non-existent/unitialized memory regions and a fault, so that
(4) ends up never being executed.
This is in m-mode with MPRV unset, meaning that unlocked PMP entries
should have no effect. Therefore such a behavior based on PMP content
is very unexpected.
Make sure no negative PMP range can be created, whether explicitly by
the emulated code or implicitly like the above.
Signed-off-by: Nicolas Pitre <nico@fluxnic.net>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-Id: <3oq0sqs1-67o0-145-5n1s-453o118804q@syhkavp.arg>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
There is an overflow with the current code where a pmpaddr value of
0x1fffffff is decoded as sa=0 and ea=0 whereas it should be sa=0 and
ea=0xffffffff.
Fix that by simplifying the computation. There is in fact no need for
ctz64() nor special case for -1 to achieve proper results.
Signed-off-by: Nicolas Pitre <nico@fluxnic.net>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-Id: <rq81o86n-17ps-92no-p65o-79o88476266@syhkavp.arg>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: LIU Zhiwei <zhiwei_liu@c-sky.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-id: 20220120122050.41546-2-zhiwei_liu@c-sky.com
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
%s/CSP/CSR
%s/thie/the
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-id: 20210627115716.3552-1-bmeng.cn@gmail.com
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Although we construct epmp_operation in such a way that it can only be
between 0 and 15 Coverity complains that we don't handle the other
possible cases. To fix Coverity and make it easier for humans to read
add a default case to the switch statement that calls
g_assert_not_reached().
Fixes: CID 1453108
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: LIU Zhiwei <zhiwei_liu@c-sky.com>
Message-id: ec5f225928eec448278c82fcb1f6805ee61dde82.1621550996.git.alistair.francis@wdc.com
This commit adds support for ePMP v0.9.1.
The ePMP spec can be found in:
https://docs.google.com/document/d/1Mh_aiHYxemL0umN3GTTw8vsbmzHZ_nxZXgjgOUzbvc8
Signed-off-by: Hongzheng-Li <Ethan.Lee.QNL@gmail.com>
Signed-off-by: Hou Weiying <weiying_hou@outlook.com>
Signed-off-by: Myriad-Dreamin <camiyoru@gmail.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Message-id: fef23b885f9649a4d54e7c98b168bdec5d297bb1.1618812899.git.alistair.francis@wdc.com
[ Changes by AF:
- Rebase on master
- Update to latest spec
- Use a switch case to handle ePMP MML permissions
- Fix a few bugs
]
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
The RISC-V spec says:
if PMP entry i is locked and pmpicfg.A is set to TOR, writes to
pmpaddri-1 are ignored.
The current QEMU code ignores accesses to pmpaddri-1 and pmpcfgi-1 which
is incorrect.
Update the pmp_is_locked() function to not check the supporting fields
and instead enforce the lock functionality in the pmpaddr write operation.
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Message-id: 2831241458163f445a89bd59c59990247265b0c6.1618812899.git.alistair.francis@wdc.com
If PMP permission of any address has been changed by updating PMP entry,
flush all TLB pages to prevent from getting old permission.
Signed-off-by: Jim Shu <cwshu@andestech.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-id: 1613916082-19528-4-git-send-email-cwshu@andestech.com
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Currently, PMP permission checking of TLB page is bypassed if TLB hits
Fix it by propagating PMP permission to TLB page permission.
PMP permission checking also use MMU-style API to change TLB permission
and size.
Signed-off-by: Jim Shu <cwshu@andestech.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-id: 1613916082-19528-2-git-send-email-cwshu@andestech.com
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
As per the privilege specification, any access from S/U mode should fail
if no pmp region is configured.
Signed-off-by: Atish Patra <atish.patra@wdc.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-id: 20201223192553.332508-1-atish.patra@wdc.com
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
In the case of supporting PMP feature, add PMP state description
to vmstate_riscv_cpu.
'vmstate_pmp_addr' and 'num_rules' could be regenerated by
pmp_update_rule(). But there exists the problem of updating
num_rules repeatedly in pmp_update_rule(). So here extracts
pmp_update_rule_addr() and pmp_update_rule_nums() to update
'vmstate_pmp_addr' and 'num_rules' respectively.
Signed-off-by: Yifei Jiang <jiangyifei@huawei.com>
Signed-off-by: Yipeng Yin <yinyipeng1@huawei.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-id: 20201026115530.304-4-jiangyifei@huawei.com
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
The minimum granularity of PMP is 4 bytes, it is small than 4KB page
size, therefore, the pmp checking would be ignored if its range doesn't
start from the alignment of one page. This patch detects the pmp entries
and sets the small page size to TLB if there is a PMP entry which cover
the page size.
Signed-off-by: Zong Li <zong.li@sifive.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-Id: <6b0bf48662ef26ab4c15381a08e78a74ebd7ca79.1595924470.git.zong.li@sifive.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
First, sizeof(target_ulong) equals to 4 on riscv32, so this change
does not change the function on riscv32. Second, sizeof(target_ulong)
equals to 8 on riscv64, and 'reg_index * 8 + i' is not a legal
pmp_index (we will explain later), which should be 'reg_index * 4 + i'.
If the parameter reg_index equals to 2 (means that we will change the
value of pmpcfg2, or the second pmpcfg on riscv64), then
pmpcfg_csr_write(env, 2, val) will map write tasks to
pmp_write_cfg(env, 2 * 8 + [0...7], val). However, no cfg csr is indexed
by value 16 or 23 on riscv64, so we consider it as a bug.
We are looking for constant (e.g., define a new constant named
RISCV_WORD_SIZE) in QEMU to help others understand code better,
but none was found. A possible good explanation of this literal is it is
the minimum word length on riscv is 4 bytes (32 bit).
Signed-off-by: Hongzheng-Li <Ethan.Lee.QNL@gmail.com>
Signed-off-by: Hou Weiying <weiying_hou@outlook.com>
Signed-off-by: Myriad-Dreamin <camiyoru@gmail.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-Id: <SG2PR02MB263420036254AC8841F66CE393460@SG2PR02MB2634.apcprd02.prod.outlook.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
The end address calculation for NA4 mode is wrong because the address
used isn't shifted.
It doesn't watch 4 bytes but a huge range because the end address
calculation is wrong.
The solution is to use the shifted address calculated for start address
variable.
Modifications are tested on Zephyr OS userspace test suite which works
for other RISC-V boards (E31 and E34 core).
Signed-off-by: Alexandre Mergnat <amergnat@baylibre.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-id: 20200706084550.24117-1-amergnat@baylibre.com
Message-Id: <20200706084550.24117-1-amergnat@baylibre.com>
[ Changes by AF:
- Improve the commit title and message
]
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
riscv_cpu_tlb_fill() uses the `size` parameter to check PMP violation
using pmp_hart_has_privs().
However, if the size is unknown (=0), the ending address will be
`addr - 1` as it is `addr + size - 1` in `pmp_hart_has_privs()`.
This always causes a false PMP violation on the starting address of the
range, as `addr - 1` is not in the range.
In order to fix, we just assume that all bytes from addr to the end of
the page will be accessed if the size is unknown.
Signed-off-by: Dayeol Lee <dayeol@berkeley.edu>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
Use the always-compiled trace events, remove the now unused
RISCV_DEBUG_PMP definition.
Note pmpaddr_csr_read() could previously do out-of-bound accesses
passing addr_index >= MAX_RISCV_PMPS.
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
The RISC-V Physical Memory Protection is restricted to privileged
modes. Restrict its compilation to QEMU system builds.
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
The current implementation returns 1 (PMP check success) if the address is in
range even if the PMP entry is off. This is a bug.
For example, if there is a PMP check in S-Mode which is in range, but its PMP
entry is off, this will succeed, which it should not.
The patch fixes this bug by only checking the PMP permissions if the address is
in range and its corresponding PMP entry it not off. Otherwise, it will keep
the ret = -1 which will be checked and handled correctly at the end of the
function.
Signed-off-by: Hesham Almatary <Hesham.Almatary@cl.cam.ac.uk>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
The current PMP check function checks for env->priv which is not the effective
memory privilege mode.
For example, mstatus.MPRV could be set while executing in M-Mode, and in that
case the privilege mode for the PMP check should be S-Mode rather than M-Mode
(in env->priv) if mstatus.MPP == PRV_S.
This patch passes the effective memory privilege mode to the PMP check.
Functions that call the PMP check should pass the correct memory privilege mode
after reading mstatus' MPRV/MPP or hstatus.SPRV (if Hypervisor mode exists).
Suggested-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Hesham Almatary <Hesham.Almatary@cl.cam.ac.uk>
Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
A wrong address is passed to `pmp_is_in_range` while checking if a
memory access is within a PMP range.
Since the ending address of the pmp range (i.e., pmp_state.addr[i].ea)
is set to the last address in the range (i.e., pmp base + pmp size - 1),
memory accesses containg the last address in the range will always fail.
For example, assume that a PMP range is 4KB from 0x87654000 such that
the last address within the range is 0x87654fff.
1-byte access to 0x87654fff should be considered to be fully inside the
PMP range.
However the access now fails and complains partial inclusion because
pmp_is_in_range(env, i, addr + size) returns 0 whereas
pmp_is_in_range(env, i, addr) returns 1.
Signed-off-by: Dayeol Lee <dayeol@berkeley.edu>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Michael Clark <mjc@sifive.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
No header includes qemu-common.h after this commit, as prescribed by
qemu-common.h's file comment.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20190523143508.25387-5-armbru@redhat.com>
[Rebased with conflicts resolved automatically, except for
include/hw/arm/xlnx-zynqmp.h hw/arm/nrf51_soc.c hw/arm/msf2-soc.c
block/qcow2-refcount.c block/qcow2-cluster.c block/qcow2-cache.c
target/arm/cpu.h target/lm32/cpu.h target/m68k/cpu.h target/mips/cpu.h
target/moxie/cpu.h target/nios2/cpu.h target/openrisc/cpu.h
target/riscv/cpu.h target/tilegx/cpu.h target/tricore/cpu.h
target/unicore32/cpu.h target/xtensa/cpu.h; bsd-user/main.c and
net/tap-bsd.c fixed up]
Currently, start and end address of a PMP region are not decoded
correctly by pmp_decode_napot().
Let's say we have a 128KB PMP region with base address as 0x80000000.
Now, the PMPADDRx CSR value for this region will be 0x20003fff.
The current pmp_decode_napot() implementation will decode PMPADDRx
CSR as t1=14, base=0x100000000, and range=0x1ffff whereas it should
have decoded PMPADDRx CSR as t1=14, base=0x80000000, and range=0x1fff.
This patch fixes the base value decoding in pmp_decode_napot() when
PMPADDRx CSR is not -1 (i.e. 0xffffffffffffffff).
Signed-off-by: Anup Patel <anup@brainfault.org>
Signed-off-by: Anup Patel <anup.patel@wdc.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
pmp_read_cfg() returns 8-bit value, which is combined together to form a single pmpcfg CSR.
The default promotion rules will result in an integer here ("i*8" is integer, which
flows through) resulting in a 32-bit signed value on most hosts.
That's bogus on RV64I, with the high bits of the CSR being wrong.
Signed-off-by: Dayeol Lee <dayeol@berkeley.edu>
Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
Implements the physical memory protection extension as specified in
Privileged ISA Version 1.10.
PMP (Physical Memory Protection) is as-of-yet unused and needs testing.
The SiFive verification team have PMP test cases that will be run.
Nothing currently depends on PMP support. It would be preferable to keep
the code in-tree for folk that are interested in RISC-V PMP support.
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Daire McNamara <daire.mcnamara@emdalo.com>
Signed-off-by: Ivan Griffin <ivan.griffin@emdalo.com>
Signed-off-by: Michael Clark <mjc@sifive.com>
