[Qemu-devel] [PATCH v1] GLib sucks - Remove any connections between me and GLib
hat OS to use as a firewall (in a school of all places) and FreeBSD won apparently i.e. `ipfw`. I was a Linux user with an IBM ThinkPad and he was a macOS X user with a Titanium macbook. BTW Age is not really an asset unless you have a secure job as a gatekeeper somewhere. I guess here, in QEMU-land, performance is number of merged PRs, ideally nicely batched along with lots of testing, and not just regressions; also performance. And if you think this commit is staged, then you are wrong. Committed, but to where I do not know (my personal repository, hopefully not a mental institution). Also, apologies to my brother for mentioning him here, but when the person who does your AR/AP and payroll quits, and one is supposedly a charity case, then they can do quite a lot of damage. He is the high flighing wealthy expat with all the perks, including dental and I am looking at the Dole queue. Of course he has to work with morons. I'll only work with nice people, so life is hard. Our old website is still up and running, so if anyone wants firewalls, storage, security, compute; then my brother knows how to do that and I can do C++ apps for iOS, android, macOS, Windows, Linux and now bare metal (ideally something easier than using git or dealing with people which I suck at): - https://www.metaparadigm.com/ You can pull my changes from my personal repo here: - https://github.com/michaeljclark/riscv-qemu/tree/qemu-mjc Apologies for the noise! and sorry for bringing GLib into this dispute between me and Palmer. Cc: Palmer Dabbelt Cc: Alistair Francis Signed-off-by: Michael Clark --- MAINTAINERS | 1 - 1 file changed, 1 deletion(-) diff --git a/MAINTAINERS b/MAINTAINERS index af339b86db76..47cb3c14298e 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -259,7 +259,6 @@ F: include/hw/ppc/ F: disas/ppc.c RISC-V -M: Michael Clark M: Palmer Dabbelt M: Alistair Francis M: Sagar Karandikar -- 2.17.1
Re: [Qemu-devel] [PR RFC] RISC-V Patches for 3.1-rc2
On Wed, Nov 14, 2018 at 12:52 PM Palmer Dabbelt wrote: > The following changes since commit > cb968d275c145467c8b385a3618a207ec111eab1: > > Update version for v3.1.0-rc1 release (2018-11-13 18:16:14 +) > > are available in the Git repository at: > > git://github.com/riscv/riscv-qemu.git tags/riscv-for-master-3.1-rc2 > > for you to fetch changes up to 3502dc824a7b0218abb49f4350e80a49829748cf: > > RISC-V: Respect fences for user-only emulators (2018-11-13 15:12:15 > -0800) > > > RISC-V Patches for 3.1-rc2 > > This pull request contains four patches that aren't really related to > each other aside from all being bug fixes that I think should go in for > 3.1.0: > Here's I again. I guess its a somewhat arbitrary set of fixes. Although one could say the same about the set of fixes in the github repo. At least you don't have review feedback asking you split a patch up to make it easier to review, after its already been reviewed (what I would call "make work"). No objections from me. We just have to get these patches into the other tree which also has RISC-V bug fixes (but I don't have an opinion about which bugs are more or less critical). There will be some minor merge conflicts because the RISC-V tree has writable misa suppoer. I'm planning to rebase at 3.1.0, just so we don't trample on each other. It's probably going to be quite a lot of work to get the patches in that tree here because my patches tend to attract a lot of pedantic feedback. i.e. !! to normalize a scalar into a bool and ~ to broadcast bit 1 e.g. C bitwise logic that you find frequently in spike; has to be dumbed down into macro wrappers. Feedback. These changes to keep people happy are the reason why we broke things for folk like changing gp$ to __globalPointer$ (in GCC/binutils). Suddendly all earlier asm is broken because someone wants to give feedback about how RISC-V should be; someone who does not need to bear the burden of those changes. Change simply because we can make you change stuff to break things for RISC-V users because we think those changes are a good idea. Don't get me wrong. We get good feedback (in QEMU, mostly from RIchard Henderson), but its not always the case. VSPILL/VFILL and vaddsz sp is a good idea. I need to find that email of Richard's on the GCC list and forward it to Krste. I hope the feedback makes it to folk working on the Vector spec because as far as I can tell it is a good idea. Apologies for the cynicism. I'll get over it. * The second half of Alistair's memory leak patch set that I missed last > week. > * A fix to make fclass.d availiable only on RV64IFD systems (without > this it's availiable on RV32IFD systems, truncating the result). > * A fix to make sfence.vm availiable only in priv-1.9.1, and sfence.vma > only availiable in priv-1.10. > * A change to respect fences in user-mode emulators, which were > previously treated as NOPs. > > As usual, this builds and boot Linux for me. I don't think I have > anything else planned for 3.1.0, but I may be wrong as things are a bit > hectic this week. > > > Alistair Francis (1): > hw/riscv/virt: Free the test device tree node name > > Bastian Koppelmann (2): > target/riscv: Fix FCLASS_D being treated as RV64 only > target/riscv: Fix sfence.vm/a both available in any priv version > > Palmer Dabbelt (1): > RISC-V: Respect fences for user-only emulators > > hw/riscv/virt.c | 1 + > target/riscv/translate.c | 24 > 2 files changed, 17 insertions(+), 8 deletions(-) > > >
Re: [Qemu-devel] [PULL 4/4] RISC-V: Respect fences for user-only emulators
Nits. Please improve your commit message to make them more impersonal... On Wed, Nov 14, 2018 at 12:52 PM Palmer Dabbelt wrote: > Our current fence implementation ignores fences for the user-only > configurations. This is incorrect but unlikely to manifest: it requires > multi-threaded user-only code that takes advantage of the weakness in > the host's memory model and can be inlined by TCG. > The RISC-V fence implementation... > This patch simply treats fences the same way for all our emulators. > I've given it to testing as I don't want to construct a test that would > actually trigger the failure. > Testing has been limited to... ? Reproducer? > Our fence implementation has an additional deficiency where we map all > RISC-V fences to full fences. Now that we have a formal memory model > for RISC-V we can start to take advantage of the strength bits on our > fence instructions. This requires a bit more though, so I'm going to > split it out because the implementation is still correct without taking > advantage of these weaker fences. > The fence implementation... A formal memory model for RISC-V allows... > Thanks to Richard Henderson for pointing out both of the issues. > > Signed-off-by: Palmer Dabbelt > Reviewed-by: Alistair Francis > Reviewed-by: Richard Henderson Assuming you fix up the commit message. I am not sure if I am present in any of the commit messages I wrote, however perhaps thats just a matter style with respect to writing (or re-writing) history. Reviewed-by: Michael Clark --- > target/riscv/translate.c | 2 -- > 1 file changed, 2 deletions(-) > > diff --git a/target/riscv/translate.c b/target/riscv/translate.c > index f44eb9c41b48..312bf298b3c2 100644 > --- a/target/riscv/translate.c > +++ b/target/riscv/translate.c > @@ -1776,7 +1776,6 @@ static void decode_RV32_64G(CPURISCVState *env, > DisasContext *ctx) > GET_RM(ctx->opcode)); > break; > case OPC_RISC_FENCE: > -#ifndef CONFIG_USER_ONLY > if (ctx->opcode & 0x1000) { > /* FENCE_I is a no-op in QEMU, > * however we need to end the translation block */ > @@ -1787,7 +1786,6 @@ static void decode_RV32_64G(CPURISCVState *env, > DisasContext *ctx) > /* FENCE is a full memory barrier. */ > tcg_gen_mb(TCG_MO_ALL | TCG_BAR_SC); > } > -#endif > break; > case OPC_RISC_SYSTEM: > gen_system(env, ctx, MASK_OP_SYSTEM(ctx->opcode), rd, rs1, > -- > 2.18.1 > > >
Re: [Qemu-devel] [PULL] First RISC-V Patch Set for the 3.1 Soft Freeze
Any patches intended for the RISC-V port should go through the maintainer tree. I have been pretty clear that I would like to run regression tests. I do not wish to pull surprises in via master. I also do not agree with this random patch approach. It's all well and good for inactive maintainers to pop up now, but we were not getting our patches reviewed. The riscv-qemu tree is actively managed and we rebase and test at release time. I don't like the idea of busted stuff getting into the tree with patch churn correcting it. In any case we disagree regards process. Please don't break any of the branches in the riscv-qemu repository. I have scripts to tag our stable branches there (we got positive feedback from Karsten on the "series"). You are adding risk to the process. Fine. You are in charge. Don't break stuff. On Thu, Oct 18, 2018 at 1:01 PM Palmer Dabbelt wrote: > On Wed, 17 Oct 2018 16:32:10 PDT (-0700), ebl...@redhat.com wrote: > > On 10/17/18 4:54 PM, Palmer Dabbelt wrote: > >> The following changes since commit > 09558375a634e17cea6cfbfec883ac2376d2dc7f: > >> > >>Merge remote-tracking branch > 'remotes/pmaydell/tags/pull-target-arm-20181016-1' into staging (2018-10-16 > 17:42:56 +0100) > >> > >> are available in the Git repository at: > >> > >>git://github.com/riscv/riscv-qemu.git tags/riscv-for-master-3.1-sf0 > >> > >> for you to fetch changes up to 7c28f4da20e5585dce7d575691dac5392b7c6f78: > >> > >>RISC-V: Don't add NULL bootargs to device-tree (2018-10-17 13:02:30 > -0700) > >> > >> -------- > >> First RISC-V Patch Set for the 3.1 Soft Freeze > >> > > > >> > >> Michael Clark (5): > >>RISC-V: Allow setting and clearing multiple irqs > >>RISC-V: Move non-ops from op_helper to cpu_helper > >>RISC-V: Update CSR and interrupt definitions > >>RISC-V: Add missing free for plic_hart_config > >>RISC-V: Don't add NULL bootargs to device-tree > >> > > > > Isn't this just a subset of Alistair's pull request? > > https://lists.gnu.org/archive/html/qemu-devel/2018-10/msg02342.html > > Yes, but there's still on-going discussion about the PCIe patches so > they're > not really fully reviewed. I think part of the trouble here is that there > wasn't someone submitting regular QEMU pull requests so there was always a > rush > to get things in. I've volunteered to wrangle the branches and submit > weekly > pull requests (just like I do for Linux), so now there won't be any more > big > cliffs. > > We've got a lot of patches to filter through because things have been > backed up > for a bit, so I thought it'd be best to just go with something simple for > this > first week. Assuming everything gets sorted out for the PCIe patches > they'll > just go up next week -- I'm super excited for them as well :) > > > which included: > >> -------- > >> Alistair Francis (5): > >> hw/riscv/virt: Increase the number of interrupts > >> hw/riscv/virt: Connect the gpex PCIe > >> riscv: Enable VGA and PCIE_VGA > >> hw/riscv/sifive_u: Connect the Xilinx PCIe > >> hw/riscv/virt: Connect a VirtIO net PCIe device > >> > >> Michael Clark (5): > >> RISC-V: Allow setting and clearing multiple irqs > >> RISC-V: Move non-ops from op_helper to cpu_helper > >> RISC-V: Update CSR and interrupt definitions > >> RISC-V: Add missing free for plic_hart_config > >> RISC-V: Don't add NULL bootargs to device-tree >
Re: [Qemu-devel] [PATCH v1 0/5] Misc RISC-V patches
Hi All, On Thu, Oct 11, 2018 at 7:22 AM Palmer Dabbelt wrote: > On Wed, 10 Oct 2018 11:10:07 PDT (-0700), peter.mayd...@linaro.org wrote: > > On 10 October 2018 at 18:49, Palmer Dabbelt wrote: > >> we should really > >> get the ball rolling on our big patch backlog. > > > > Yes, please do. Softfreeze is not all that far away and I > > would strongly prefer not to get an enormous sized pull > > request at the last minute. The ideal pattern is that > > code changes come in at a steady rate across the whole > > of the 'open' part of the development cycle. > > Ya, sorry, we've been a bit out of it. If I understand correctly, the > soft > freeze is the 30th? If so it's really time to get started, and it looks > like > Michael is busy so I'll have to go figure this out. > Yes. I should think twice about the Signed-off-by: on my commits. I need to run a regression on this out-of-order subset. I currently only run tests on the top of the riscv-qemu tree in-order, and when I rebase against master. If the commits need any significant effort to rebase because they are taken in some random order then the testing will be invalidated. i.e. I haven't checked the dependencies for these commits. I am happy to review whoever posts the contents of the tree. I can test apply the PRs against the riscv-qemu tree and if they give us lumps, we'll reject, including my own changes (if rebased). Alastair, I suggest you confer with Debian and Fedora folk. Don't break the Linux distros... I'm petrified that we might break Debian. Palmer, I disagree with idea, I would like to maintain the soft-fork until we have the CI running our regression test suite (currently manual) Peter, I have to pull in your remote wholesale. I don't cherry-pick from your tree. I think this is truly dumb. This might serve the needs of some folk running Linux but we have emulation fidelity fixes for the RISC-V community as a whole. Alastair is the only person not submitting his patches via the (sub)maintainer tree. BTW Who is the RISC-V port maintainer? Puzzled. Here is the pull queue. But I'm not ready to make a PR until we have the CI running the regression. I certainly don't want rebases of random commits to the riscv-qemu tree coming in when we pull. - https://github.com/riscv/riscv-qemu/tree/qemu-for-upstream That said, they have sign-off. There are plenty of other "RISC-V" maintainers. Do what you think is wise. Most important thing here is the Debian builders and other RISC-V virtual machines in production. Having the Debian folk or some other helpful tester running the entire tree. Pulling it in one go means we don't have a bisection problem interspersed with a whole lot of other random patches. You may not have all of the interrupt related changes that require extensive parallel burn-in tests (GCC bootstrap). i.e. we do significantly more than "make check" when we pull changes into our tree. Thanks and Regards, Michael.
Re: [Qemu-devel] qemu-riscv64 seg fault
On Mon, Sep 3, 2018 at 8:16 PM, Pranith Kumar wrote: > On Mon, Sep 3, 2018 at 1:07 AM Michael Clark wrote: > > > > Thanks. I was just about to log an issue in the riscv-qemu issue tracker > on GitHub. > > > > I reproduced it on my side. The fact that it is causes QEMU user to > crash in translate.c is interesting. > > > > I ran your program with -d in_asm and it appears to crash in thread::join > > Interestingly, qemu-x86_64 crashes too. But running natively on x86 > works fine. There is something off somewhere... > Indeed. Another data-point. I tested it in both qemu-2.12 and qemu-3.0. It still crashed in translate.c in qemu-2.12, but on another line. There was quite a bit of restructuring of target/riscv/translate.c in qemu-3.0, which is why I tried qemu-2.12, however, it seems that is likely not the cause. This was the backtrace from riscv-qemu-2.12: Thread 1 "qemu-riscv64" received signal SIGSEGV, Segmentation fault. 0x555e1347 in gen_intermediate_code (cs=cs@entry=0x57905590, tb=tb@entry=0x558e1a00 ) at /home/mjc/src/sifive/riscv-qemu/target/riscv/translate.c:1934 1934 if (num_insns == max_insns && (tb->cflags & CF_LAST_IO)) { (gdb) print tb $1 = (TranslationBlock *) 0x558e1a00 (gdb) quit A debugging session is active. ~/quickht$ qemu-x86_64 ./bench -t 2 -u 2 > qemu: uncaught target signal 11 (Segmentation fault) - core dumped > Segmentation fault > > ~/quickht$ ./bench -t 2 -u 2 > Comparisons: 23922 > Buckets touched: 4933 > Operations: 2000 >
Re: [Qemu-devel] qemu-riscv64 seg fault
Thanks. I was just about to log an issue in the riscv-qemu issue tracker on GitHub. I reproduced it on my side. The fact that it is causes QEMU user to crash in translate.c is interesting. I ran your program with -d in_asm and it appears to crash in thread::join On Mon, Sep 3, 2018 at 7:58 PM, Pranith Kumar wrote: > On second looks, running the benchmark on a RISCV processor is also > giving a seg fault. So may be there is something wrong with the > benchmark... OTOH, x86 version runs fine... hmm > > Please ignore this report, I will try to investigate further. > > Thanks, > On Mon, Sep 3, 2018 at 12:45 AM Pranith Kumar > wrote: > > > > Hi Michael, > > > > qemu-riscv64 seg faults for me on a static binary. You can build the > > binary from here: https://github.com/pranith/quickht > > > > $ STATIC=1 RISCV=1 make > > > > $ qemu-riscv64 ./bench -t 1 -u 1 > > > > > > Thanks, > > -- > > Pranith > > > > -- > Pranith >
Re: [Qemu-devel] microblaze build failure due to definition clash on riscv
On Thu, Aug 2, 2018 at 3:57 PM, Philippe Mathieu-Daudé wrote: > Oops I just realized I forgot to Cc the QEMU list, doing it now. > > On 07/31/2018 07:40 AM, Edgar E. Iglesias wrote: > > On Mon, Jul 30, 2018 at 03:22:46PM -0300, Philippe Mathieu-Daudé wrote: > >> Hi, I'm getting this error while building microblaze on riscv host: > >> > >> ... > >> CC disas/microblaze.o > >> disas/microblaze.c:179:0: error: "REG_SP" redefined [-Werror] > >> > >> #define REG_SP 1 /* stack pointer */ > >> > >> In file included from /usr/include/signal.h:306:0, > >> from include/qemu/osdep.h:101, > >> from disas/microblaze.c:36: > >> /usr/include/sys/ucontext.h:36:0: note: this is the location of the > >> previous definition > >> # define REG_SP 2 > >> > >> cc1: all warnings being treated as errors > >> make: *** [rules.mak:69: disas/microblaze.o] Error 1 > >> > >> > >> Having /usr/include/sys/ucontext.h: > >> > >> 31 #ifdef __USE_MISC > >> 32 # define NGREG 32 > >> 33 > >> 34 # define REG_PC 0 > >> 35 # define REG_RA 1 > >> 36 # define REG_SP 2 > >> 37 # define REG_TP 4 > >> 38 # define REG_S0 8 > >> 39 # define REG_S1 9 > >> 40 # define REG_A0 10 > >> 41 # define REG_S2 18 > >> 42 # define REG_NARGS 8 > >> > >> Edgar is it OK to rename REG_* -> MB_REG_*? > > > > Yes, no problems with me! > > > > Allthough it doesn't seem to be a good idea to have such generic names > in riscv system header files. > > OK this might be a Linux problem indeed, maybe this file lacks some > #ifdef __riscv (previous to the __USE_MISC?). > > I'll wait for the riscv folks before to send a dumb rename patch. > It seems to be a generic Linux problem or perhaps even a microblaze problem in this case, where QEMU microblaze definitions just happen to clash with system definitions. It depends on which way around you look at the problem. x86_64 has REG_R8, REG_R9, REG_R10, etc so RISC-V is just following the Linux scheme. Indeed RISC-V in QEMU unlike microblaze has this in disas/riscv.c typedef enum { rv_ireg_zero, rv_ireg_ra, rv_ireg_sp, rv_ireg_gp, rv_ireg_tp, rv_ireg_t0, rv_ireg_t1, rv_ireg_t2, rv_ireg_s0, rv_ireg_s1, rv_ireg_a0, rv_ireg_a1, rv_ireg_a2, rv_ireg_a3, rv_ireg_a4, rv_ireg_a5, rv_ireg_a6, rv_ireg_a7, rv_ireg_s2, rv_ireg_s3, rv_ireg_s4, rv_ireg_s5, rv_ireg_s6, rv_ireg_s7, rv_ireg_s8, rv_ireg_s9, rv_ireg_s10, rv_ireg_s11, rv_ireg_t3, rv_ireg_t4, rv_ireg_t5, rv_ireg_t6, } rv_ireg; So from that perspective, one could suggest to change microblaze. It will be a lot harder to change glibc given it is already baked into a large number of binary images, and they are system headers. Prefixing the definitions inside the mircoblaze disassembler would cause the least havoc. Changing system headers in glibc will take a long time to propagate via the current distros.
Re: [Qemu-devel] [PATCH] RISC-V: Correct typo in RV32 perf counters
On Mon, 30 Jul 2018 at 10:46 PM, Peter Maydell wrote: > On 25 May 2018 at 14:17, Richard Henderson wrote: > > On 05/24/2018 11:24 PM, Michael Clark wrote: > >> This patch enables mhpmcounter3h through mhpmcounter31h on RV32. > >> Previously the RV32 h versions (high 32-bits of 64-bit counters) > >> of these counters would trap with an illegal instruction instead > >> of returning 0 as intended. > >> > >> Reported-by: Richard Henderson > >> Signed-off-by: Michael Clark > >> --- > >> target/riscv/op_helper.c | 2 +- > >> 1 file changed, 1 insertion(+), 1 deletion(-) > > > > Fixes: Coverity CID 1390849 > > Reviewed-by: Richard Henderson > > Ping -- Coverity is still complaining about this -- did this > patch get lost? Sort of. I assumed it would go into a trivial queue. Feel free to apply, however it’s going to create (another) rebase conflict against my ‘for-upstream’ queue as this code has gone away, hence it is not in my queue. The bug is fixed in my tree in an alternate way as we have overhauled the CSR system to support atomic read/modify/write CSRs, and the new code is table driven versus using jumbo switch statements. When we transcribed this code to the new CSR system the bug disappeared as a consequence of the nature of the new mechanism which matches the table listings in the RISC-V Privileged ISA manual which makes this type of bug much more obvious: https://github.com/riscv/riscv-qemu/blob/qemu-for-upstream/target/riscv/csr.c#L912-L919 My tree is up-to-date as of 3.0-rc2, and has been rebased against Alistair’s changes but I have a feeling we are still missing reviews. We also have additional CSR fixes (such as vectored interrupts) that depend on the context of the new CSR system and we don’t have enough bandwidth to maintain a backport to the old code in upstream QEMU. I guess it is problematic going via my tree as last time I looked we didn’t have enough reviews. I can include this change in my pull for 3.1, along with the patches in my tree that have Reviewed-by and fix the rebase conflicts for anything that depends on new context (hopefully not creating new bugs during that process). I haven’t reposted the code that has not been reviewed as I’ll do that when I have a chance to reorder and split my queue based on what has been reviewed and what hasn’t. I’ve deferred that because i’m still working on new code that will be shipping from the SiFive tree (for which i’m currently writing test cases for). Michael.
Re: [Qemu-devel] [PATCH v1 1/1] configure: Add RISC-V host support
On Sat, Jul 28, 2018 at 11:49 AM, Alistair Francis wrote: > Allow QEMU to be built to run on a RISC-V host. > > QEMU does not yet have a RISC-V TCG or user mode target port, but > running other architectures on RISC-V using TCI does work. > There is this RISC-V TCG backend here: https://github.com/riscv/riscv-qemu/tree/wip-riscv-tcg-backend I realized I had not sent an update to the list. There is a working TCG backend for RISC-V: - I've so far tested it with user-mode translation only. - It does not support large guests e.g. 64-bit hosts on riscv32. - The softmmu implementation is there but needs testing. - Some things are missing, e.g. memory fences. big-endian support You might find some other relevant changes in the configure hunks in the patch. IIRC there were a few changes to configure, one of them was to get the riscv disassembler working. Signed-off-by: Alistair Francis > --- > configure | 18 +- > 1 file changed, 17 insertions(+), 1 deletion(-) > > diff --git a/configure b/configure > index 2a7796ea80..c3ff3ae146 100755 > --- a/configure > +++ b/configure > @@ -606,6 +606,16 @@ EOF >compile_object > } > > +check_define_value() { > +cat > $TMPC < +#if (($1) != ($2)) > +#error $1 != ($2) > +#endif > +int main(void) { return 0; } > +EOF > + compile_object > +} > + > check_include() { > cat > $TMPC < #include <$1> > @@ -704,6 +714,12 @@ elif check_define __arm__ ; then >cpu="arm" > elif check_define __aarch64__ ; then >cpu="aarch64" > +elif check_define __riscv ; then > + if check_define_value __riscv_xlen 64 ; then > +cpu="riscv64" > + else > +cpu="riscv32" > + fi > else >cpu=$(uname -m) > fi > @@ -712,7 +728,7 @@ ARCH= > # Normalise host CPU name and set ARCH. > # Note that this case should only have supported host CPUs, not guests. > case "$cpu" in > - ppc|ppc64|s390|s390x|sparc64|x32) > + ppc|ppc64|s390|s390x|sparc64|x32|riscv32|riscv64) > cpu="$cpu" > supported_cpu="yes" > eval "cross_cc_${cpu}=\$host_cc" > -- > 2.17.1 > >
Re: [Qemu-devel] [PATCH v2] riscv: remove define cpu_init()
On Fri, May 18, 2018 at 8:02 PM, Igor Mammedov wrote: > On Fri, 18 May 2018 14:10:24 +1200 > Michael Clark wrote: > > > On Wed, May 16, 2018 at 4:00 AM, Igor Mammedov > wrote: > > > > > cpu_init() was removed since 2.12, so drop the define that is now > unused. > > > > > > Signed-off-by: Igor Mammedov > > > Reviewed-by: Philippe Mathieu-Daudé > > > > > > > Reviewed-by: Michael Clark > Michael, > > could you merge it through RISCV queue? Yes indeed. Sorry I missed this email. I'm going to bring this into the riscv-qemu queue right away and move it to the head of the queue, now we have Alastair's stuff in master. Although I won't be making any PRs until after 3.0 is released. I'm currently re-basing and testing the RISC-V tree against Alastair's changes. There were quite a few conflicts so it is going to take a bit of testing. I also need to spend some time re-testing master. I've been AWOL for the last month and a half due to some work related to SiFive's new E20 and E21 cores which were recently announced. In the QEMU 3.1 or 3.2 time-frame we should have the new CLIC interrupt controller which supports multiple levels of interrupt pre-emption, interrupt tail chaining and new vectored interrupt modes. In the proposed CLIC spec (Core Level Interrupt Controller), pre-emption is mostly handled in software but there is a tiny bit of hardware support. We don't have any magic hardware state machines that save multiple registers at a time. The proposed CLIC sticks pretty closely to a set of RISC-V ideals with respect to how much is done in a micro-op in the ISA and avoidance of complex hardware state machines (e.g. no wiring of the ABI caller save registers into hardware). I'm currently working on test cases for the CLIC and hopefully SiFive will have tests for other changes we have in the RISC-V tree, so we have "proofs" accompanying each patch that makes any sufficiantly complex change. It of course will take a bit of time for us to accompany each of our changes with a test. SiFive do actually have resources devoted to formal verification but I have not yet been able to convince them to devote resources to formally verifying the RISC-V QEMU implementation. Anyway we do at least hope to have some tests for the RISC-V QEMU target in the near future... Phillipe, I noticed the recent cross complier infrastructure support so we could potentially bring in riscv-tests as a submodule in QEMU... this will help... Michael. > > --- > > > v2: > > > * refine commit message (Eric Blake ) > > > --- > > > target/riscv/cpu.h | 1 - > > > 1 file changed, 1 deletion(-) > > > > > > diff --git a/target/riscv/cpu.h b/target/riscv/cpu.h > > > index 34abc38..0562a02 100644 > > > --- a/target/riscv/cpu.h > > > +++ b/target/riscv/cpu.h > > > @@ -251,7 +251,6 @@ int riscv_cpu_handle_mmu_fault(CPUState *cpu, > vaddr > > > address, int size, > > > char *riscv_isa_string(RISCVCPU *cpu); > > > void riscv_cpu_list(FILE *f, fprintf_function cpu_fprintf); > > > > > > -#define cpu_init(cpu_model) cpu_generic_init(TYPE_RISCV_CPU, > cpu_model) > > > #define cpu_signal_handler cpu_riscv_signal_handler > > > #define cpu_list riscv_cpu_list > > > #define cpu_mmu_index riscv_cpu_mmu_index > > > -- > > > 2.7.4 > > > > > > > >
Re: [Qemu-devel] [PATCH v1 2/5] sifive_u: Fix crash when introspecting the device
On Wed, Jul 18, 2018 at 8:28 AM, Alistair Francis wrote: > Use the new object_initialize_child() and sysbus_init_child_obj() to > fix the issue. > > Signed-off-by: Alistair Francis > Reviewed-by: Michael Clark > --- > hw/riscv/sifive_u.c | 15 +++ > 1 file changed, 7 insertions(+), 8 deletions(-) > > diff --git a/hw/riscv/sifive_u.c b/hw/riscv/sifive_u.c > index 3a6ffeb437..59ae1ce24a 100644 > --- a/hw/riscv/sifive_u.c > +++ b/hw/riscv/sifive_u.c > @@ -244,9 +244,9 @@ static void riscv_sifive_u_init(MachineState *machine) > int i; > > /* Initialize SoC */ > -object_initialize(>soc, sizeof(s->soc), TYPE_RISCV_U_SOC); > -object_property_add_child(OBJECT(machine), "soc", OBJECT(>soc), > - _abort); > +object_initialize_child(OBJECT(machine), "soc", >soc, > +sizeof(s->soc), TYPE_RISCV_U_SOC, > +_abort, NULL); > object_property_set_bool(OBJECT(>soc), true, "realized", > _abort); > > @@ -303,16 +303,15 @@ static void riscv_sifive_u_soc_init(Object *obj) > { > SiFiveUSoCState *s = RISCV_U_SOC(obj); > > -object_initialize(>cpus, sizeof(s->cpus), TYPE_RISCV_HART_ARRAY); > -object_property_add_child(obj, "cpus", OBJECT(>cpus), > - _abort); > +object_initialize_child(obj, "cpus", >cpus, sizeof(s->cpus), > +TYPE_RISCV_HART_ARRAY, _abort, NULL); > object_property_set_str(OBJECT(>cpus), SIFIVE_U_CPU, "cpu-type", > _abort); > object_property_set_int(OBJECT(>cpus), smp_cpus, "num-harts", > _abort); > > -object_initialize(>gem, sizeof(s->gem), TYPE_CADENCE_GEM); > -qdev_set_parent_bus(DEVICE(>gem), sysbus_get_default()); > +sysbus_init_child_obj(obj, "gem", >gem, sizeof(s->gem), > + TYPE_CADENCE_GEM); > } > > static void riscv_sifive_u_soc_realize(DeviceState *dev, Error **errp) > -- > 2.17.1 > >
Re: [Qemu-devel] [PATCH v1 5/5] spike: Fix crash when introspecting the device
On Wed, Jul 18, 2018 at 8:28 AM, Alistair Francis wrote: > Use the new object_initialize_child() and sysbus_init_child_obj() to > fix the issue. > > Signed-off-by: Alistair Francis > Reviewed-by: Michael Clark > --- > hw/riscv/spike.c | 10 -- > 1 file changed, 4 insertions(+), 6 deletions(-) > > diff --git a/hw/riscv/spike.c b/hw/riscv/spike.c > index f94e2b6707..c8c056c50b 100644 > --- a/hw/riscv/spike.c > +++ b/hw/riscv/spike.c > @@ -171,9 +171,8 @@ static void spike_v1_10_0_board_init(MachineState > *machine) > int i; > > /* Initialize SOC */ > -object_initialize(>soc, sizeof(s->soc), TYPE_RISCV_HART_ARRAY); > -object_property_add_child(OBJECT(machine), "soc", OBJECT(>soc), > - _abort); > +object_initialize_child(OBJECT(machine), "soc", >soc, > sizeof(s->soc), > +TYPE_RISCV_HART_ARRAY, _abort, NULL); > object_property_set_str(OBJECT(>soc), SPIKE_V1_10_0_CPU, > "cpu-type", > _abort); > object_property_set_int(OBJECT(>soc), smp_cpus, "num-harts", > @@ -254,9 +253,8 @@ static void spike_v1_09_1_board_init(MachineState > *machine) > int i; > > /* Initialize SOC */ > -object_initialize(>soc, sizeof(s->soc), TYPE_RISCV_HART_ARRAY); > -object_property_add_child(OBJECT(machine), "soc", OBJECT(>soc), > - _abort); > +object_initialize_child(OBJECT(machine), "soc", >soc, > sizeof(s->soc), > +TYPE_RISCV_HART_ARRAY, _abort, NULL); > object_property_set_str(OBJECT(>soc), SPIKE_V1_09_1_CPU, > "cpu-type", > _abort); > object_property_set_int(OBJECT(>soc), smp_cpus, "num-harts", > -- > 2.17.1 > >
Re: [Qemu-devel] [PATCH v1 4/5] riscv_hart: Fix crash when introspecting the device
On Wed, Jul 18, 2018 at 8:28 AM, Alistair Francis wrote: > Use the new object_initialize_child() and sysbus_init_child_obj() to > fix the issue. > > Signed-off-by: Alistair Francis > Reviewed-by: Michael Clark > --- > hw/riscv/riscv_hart.c | 7 +++ > 1 file changed, 3 insertions(+), 4 deletions(-) > > diff --git a/hw/riscv/riscv_hart.c b/hw/riscv/riscv_hart.c > index 75ba7ed579..e34a26a0ef 100644 > --- a/hw/riscv/riscv_hart.c > +++ b/hw/riscv/riscv_hart.c > @@ -45,11 +45,10 @@ static void riscv_harts_realize(DeviceState *dev, > Error **errp) > s->harts = g_new0(RISCVCPU, s->num_harts); > > for (n = 0; n < s->num_harts; n++) { > - > -object_initialize(>harts[n], sizeof(RISCVCPU), s->cpu_type); > +object_initialize_child(OBJECT(s), "harts[*]", >harts[n], > +sizeof(RISCVCPU), s->cpu_type, > +_abort, NULL); > s->harts[n].env.mhartid = n; > -object_property_add_child(OBJECT(s), "harts[*]", > OBJECT(>harts[n]), > - _abort); > qemu_register_reset(riscv_harts_cpu_reset, >harts[n]); > object_property_set_bool(OBJECT(>harts[n]), true, > "realized", ); > -- > 2.17.1 > >
Re: [Qemu-devel] [PATCH v1 3/5] virt: Fix crash when introspecting the device
On Wed, Jul 18, 2018 at 8:28 AM, Alistair Francis wrote: > Use the new object_initialize_child() and sysbus_init_child_obj() to > fix the issue. > > Signed-off-by: Alistair Francis > Reviewed-by: Michael Clark > --- > hw/riscv/virt.c | 5 ++--- > 1 file changed, 2 insertions(+), 3 deletions(-) > > diff --git a/hw/riscv/virt.c b/hw/riscv/virt.c > index aeada2498d..248bbdffd3 100644 > --- a/hw/riscv/virt.c > +++ b/hw/riscv/virt.c > @@ -274,9 +274,8 @@ static void riscv_virt_board_init(MachineState > *machine) > void *fdt; > > /* Initialize SOC */ > -object_initialize(>soc, sizeof(s->soc), TYPE_RISCV_HART_ARRAY); > -object_property_add_child(OBJECT(machine), "soc", OBJECT(>soc), > - _abort); > +object_initialize_child(OBJECT(machine), "soc", >soc, > sizeof(s->soc), > +TYPE_RISCV_HART_ARRAY, _abort, NULL); > object_property_set_str(OBJECT(>soc), VIRT_CPU, "cpu-type", > _abort); > object_property_set_int(OBJECT(>soc), smp_cpus, "num-harts", > -- > 2.17.1 > >
Re: [Qemu-devel] [PATCH v1 1/5] sifive_e: Fix crash when introspecting the device
On Wed, Jul 18, 2018 at 8:27 AM, Alistair Francis wrote: > Use the new object_initialize_child() and sysbus_init_child_obj() to > fix the issue. > > Signed-off-by: Alistair Francis > Reviewed-by: Michael Clark > --- > hw/riscv/sifive_e.c | 12 ++-- > 1 file changed, 6 insertions(+), 6 deletions(-) > > diff --git a/hw/riscv/sifive_e.c b/hw/riscv/sifive_e.c > index 8a8dbe1c00..4577d72037 100644 > --- a/hw/riscv/sifive_e.c > +++ b/hw/riscv/sifive_e.c > @@ -105,9 +105,9 @@ static void riscv_sifive_e_init(MachineState *machine) > int i; > > /* Initialize SoC */ > -object_initialize(>soc, sizeof(s->soc), TYPE_RISCV_E_SOC); > -object_property_add_child(OBJECT(machine), "soc", OBJECT(>soc), > - _abort); > +object_initialize_child(OBJECT(machine), "soc", >soc, > +sizeof(s->soc), TYPE_RISCV_E_SOC, > +_abort, NULL); > object_property_set_bool(OBJECT(>soc), true, "realized", > _abort); > > @@ -139,9 +139,9 @@ static void riscv_sifive_e_soc_init(Object *obj) > { > SiFiveESoCState *s = RISCV_E_SOC(obj); > > -object_initialize(>cpus, sizeof(s->cpus), TYPE_RISCV_HART_ARRAY); > -object_property_add_child(obj, "cpus", OBJECT(>cpus), > - _abort); > +object_initialize_child(obj, "cpus", >cpus, > +sizeof(s->cpus), TYPE_RISCV_HART_ARRAY, > +_abort, NULL); > object_property_set_str(OBJECT(>cpus), SIFIVE_E_CPU, "cpu-type", > _abort); > object_property_set_int(OBJECT(>cpus), smp_cpus, "num-harts", > -- > 2.17.1 > >
Re: [Qemu-devel] [PATCH v2 6/6] riscv64-softmmu.mak: Build Virtio Block support
On Tue, 10 Jul 2018 at 12:29 PM, Alistair Francis wrote: > Add build time support for the VirtIO block device. This allows us to > attach a drive using the virtio-blk-device. I’m not sure what has changed in master, but VirtIO block and net for both softmmu-riscv32 and softmmu-riscv64 were previously building by default and working. Perhaps i’ve missed some patches that split up VirtIO support into more granular config options. In any case, we should enable both CONFIG_VIRTIO_NET and CONFIG_VIRTIO_BLK for both riscv32 and riscv64 as these were working in QEMU 2.12. There are folk using the “virt” machine for riscv32 Linux. Thanks, at least I know what to expect when I rebase my patch queue against master and of course make a PR for the reviewed patches... I’m not sure of the soft-freeze date but the bulk of the patches are the same ones that were posted about a month ago... I also have several patches that are not yet on the list... Vectored interrupts. Trap tracing. More spec wording clarifications regarding load reservations and interrupts: - https://github.com/riscv/riscv-qemu/commits/qemu-2.13-for-upstream Michael Signed-off-by: Alistair Francis > --- > default-configs/riscv64-softmmu.mak | 2 ++ > 1 file changed, 2 insertions(+) > > diff --git a/default-configs/riscv64-softmmu.mak > b/default-configs/riscv64-softmmu.mak > index 35e74bebe9..8790f3b115 100644 > --- a/default-configs/riscv64-softmmu.mak > +++ b/default-configs/riscv64-softmmu.mak > @@ -12,3 +12,5 @@ CONFIG_PCI_XILINX=y > > CONFIG_VGA=y > CONFIG_VGA_PCI=y > + > +CONFIG_VIRTIO_BLK=y > -- > 2.17.1 > >
Re: [Qemu-devel] [PULL v4 0/7] riscv-pull queue
On Tue, Jul 10, 2018 at 9:52 AM, Alistair Francis wrote: > On Mon, Jul 9, 2018 at 3:00 AM, Andreas Schwab wrote: > > What is the state of the sifive_u emulation? When I tried to boot a bbl > > with an included kernel I get these errors: > > > > qemu-system-riscv64: plic: invalid register write: 2090 > > qemu-system-riscv64: plic: invalid register write: 2094 > > qemu-system-riscv64: plic: invalid register write: 2098 > > qemu-system-riscv64: plic: invalid register write: 209c > > qemu-system-riscv64: plic: invalid register write: 20a0 > > qemu-system-riscv64: plic: invalid register write: 20a4 > > qemu-system-riscv64: plic: invalid register write: 20a8 > > qemu-system-riscv64: plic: invalid register write: 20ac > > qemu-system-riscv64: plic: invalid register write: 20b0 > > qemu-system-riscv64: plic: invalid register write: 20b4 > > I see those as well. I haven't investigated but I assume we are just > not completely modelling the PLIC. In saying that it should still > boot. Do you not see the kernel booting? It could be a PLIC bug or it could be a Linux interrupt controller driver bug. We can see from the memory map docs for the U54 whether these memory addresses are in bounds based on the number of configured interrupt sources. I'm not sure how many sources we have configured on sifive_u. Last time I booted Linux on sifive_u I did not see these errors. I'd need your kernel config and to know what tree and branch you are building from. I will be able to look when I get time... The PLIC however seems stable in the 'virt' board at least... Sorry I've been incommunicado for several weeks. I have been working on a CLIC model (Core Local Interrupt Controller) which replaces the CLINT and has a CLINT backwards compatibility mode. It is a Core Local vs the PLIC which is the Platform Level router. Here is the "draft" spec. It will be a candidate proposed to the RISC-V Fast Interrupts working group for potential standardisation, however in any case it will be available from SiFive so we may eventuall want to include our implementation in QEMU: - https://github.com/sifive/clic-spec/blob/master/clic.adoc When i'm done with modelling the first iteration of the CLIC I'll go through my pending patch queue and make a PR for the Reviewed patches. I'll also do some testing on master to make sure we have not regressed anything in RISC-V QEMU given QEMU 2.12 and the riscv-qemu trees are both stable. BTW - with respect to 'sifive_e' and 'sifive_u' SOC changes, we'll have to see how the model matches SiFive's plans for these virtual machines. We want to avoid a proliferation of boards, and as I've mentioned before we want to be able to model the HiFive1, HiFiveU and other SiFive E Series and U Series Coreplex configurations. There are many permutations from SiFive's SOC generator so our goal is to avoid hardcoding all of the different SOC combinations (hence the removal of model numbers in my initial review of your patches). How we achieve this I do not know. We obviously want to invest our time in something that is acceptable to upstream, while also meeting the goal of modelling SiFive's many hardware configurations, given these boards also model softcore IP such as the e31 arty and u54 on Xilinx VC707. i.e. they are SiFive models. I'm not too concerned with the SOC changes assuming we don't regress any function, as we can always evolve the code in the future to match configurations from SiFive's SOC generator. At present the models are like a union of a subset of the real hardware as we are still missing many emulation models for various parts of the SOCs. After i've finished up this CLIC work, I'll go back through the list of things we still need to model. Adding the Cadence GEM to the SiFive U Series is really nice, and so is adding Xilinx PCI to the SiFive U and GPEX to virt (as discussed, given virt is generic, we want to use GPEX there). Thanks, Michael. > > > Andreas. > > > > -- > > Andreas Schwab, SUSE Labs, sch...@suse.de > > GPG Key fingerprint = 0196 BAD8 1CE9 1970 F4BE 1748 E4D4 88E3 0EEA B9D7 > > "And now for something completely different." >
Re: [Qemu-devel] [PATCH v1 3/5] hw/riscv/virt: Connect the Xilinx PCIe
> On 23/06/2018, at 1:07 PM, Peter Maydell wrote: > > On 22 June 2018 at 20:30, Alistair Francis wrote: >> Connect the Xilinx PCIe device based on the device tree included in the >> HiFive Unleashed ROM. > > Did you consider using the 'gpex' generic PCIe controller here? Yes. Alastair and I talked about this yesterday and we agreed in principle to using ‘gpex’ on the virt machine and the Xilinx PCIe on the SiFive U (’sifive_u’) machine, as this reflects one of the IP configurations of SiFive’s Coreplex U series when run on FPGA. By changing this patch to add Xilinx PCIe to ‘sifive_u’, we can plug IO devices into the U series machine, and instead add gpex to RISC-V virt. i.e. vendor agnostic generic PCIe controller for virt. We would like ‘virt’ to be a potential vendor agnostic hardware target when we have kvm, so using gpex fits with this strategy (speaking from a RISC-V perspective not a SiFive perspective).
Re: [Qemu-devel] [PATCH v4 16/40] hw/riscv: Use the IEC binary prefix definitions
On Mon, Jun 11, 2018 at 1:14 PM, Philippe Mathieu-Daudé wrote: > It eases code review, unit is explicit. > > Patch generated using: > > $ git grep -E '(1024|2048|4096|8192|(<<|>>).?(10|20|30))' hw/ > include/hw/ > > and modified manually. > > Signed-off-by: Philippe Mathieu-Daudé > Reviewed-by: Michael Clark > --- > hw/riscv/virt.c | 3 ++- > 1 file changed, 2 insertions(+), 1 deletion(-) > > diff --git a/hw/riscv/virt.c b/hw/riscv/virt.c > index ad03113e0f..34d48993a2 100644 > --- a/hw/riscv/virt.c > +++ b/hw/riscv/virt.c > @@ -19,6 +19,7 @@ > */ > > #include "qemu/osdep.h" > +#include "qemu/units.h" > #include "qemu/log.h" > #include "qemu/error-report.h" > #include "qapi/error.h" > @@ -84,7 +85,7 @@ static hwaddr load_initrd(const char *filename, uint64_t > mem_size, > * halfway into RAM, and for boards with 256MB of RAM or more we put > * the initrd at 128MB. > */ > -*start = kernel_entry + MIN(mem_size / 2, 128 * 1024 * 1024); > +*start = kernel_entry + MIN(mem_size / 2, 128 * MiB); > > size = load_ramdisk(filename, *start, mem_size - *start); > if (size == -1) { > -- > 2.17.1 > >
Re: [Qemu-devel] [PATCH v2] elf: Add RISC-V PSABI ELF header defines
On Fri, May 25, 2018 at 7:53 PM, Laurent Vivier <laur...@vivier.eu> wrote: > Le 25/05/2018 à 09:22, Michael Clark a écrit : > > Refer to the RISC-V PSABI specification for details: > > > > - https://github.com/riscv/riscv-elf-psabi-doc/blob/master/riscv-elf.md > > > > Cc: Michael Tokarev <m...@tls.msk.ru> > > Cc: Laurent Vivier <laur...@vivier.eu> > > Cc: Richard Henderson <richard.hender...@linaro.org> > > Cc: Alistair Francis <alistair.fran...@wdc.com> > > Signed-off-by: Michael Clark <m...@sifive.com> > > --- > > include/elf.h | 8 > > 1 file changed, 8 insertions(+) > > > > diff --git a/include/elf.h b/include/elf.h > > index 934dbbd6b3ae..bd0493f43d19 100644 > > --- a/include/elf.h > > +++ b/include/elf.h > > @@ -1285,6 +1285,14 @@ typedef struct { > > #define R_IA64_DTPREL64LSB 0xb7/* @dtprel(sym + add), data8 LSB */ > > #define R_IA64_LTOFF_DTPREL220xba/* @ltoff(@dtprel(s+a)), > imm22 */ > > > > +/* RISC-V specific definitions. */ > > +#define EF_RISCV_RVC 0x0001 > > +#define EF_RISCV_FLOAT_ABI_SINGLE 0x0002 > > +#define EF_RISCV_FLOAT_ABI_DOUBLE 0x0004 > > +#define EF_RISCV_FLOAT_ABI_QUAD 0x0006 > > +#define EF_RISCV_RVE 0x0008 > > +#define EF_RISCV_TSO 0x0010 > > + > > typedef struct elf32_rel { > >Elf32_Addr r_offset; > >Elf32_Word r_info; > > > > from glibc (and binutils), I have: > > /* RISC-V ELF Flags */ > #define EF_RISCV_RVC0x0001 > #define EF_RISCV_FLOAT_ABI 0x0006 > #define EF_RISCV_FLOAT_ABI_SOFT 0x > #define EF_RISCV_FLOAT_ABI_SINGLE 0x0002 > #define EF_RISCV_FLOAT_ABI_DOUBLE 0x0004 > #define EF_RISCV_FLOAT_ABI_QUAD 0x0006 > > Perhaps you can also define EF_RISCV_FLOAT_ABI_SOFT and EF_RISCV_FLOAT_ABI? > Will do. I'll align them also. Anyway: > > Reviewed-by: Laurent Vivier <laur...@vivier.eu> > > Thanks for checking glibc and binutils. I'll update to add the extra definitions before I make a PR. Thanks! Michael.
[Qemu-devel] [PATCH v2] elf: Add RISC-V PSABI ELF header defines
Refer to the RISC-V PSABI specification for details: - https://github.com/riscv/riscv-elf-psabi-doc/blob/master/riscv-elf.md Cc: Michael Tokarev <m...@tls.msk.ru> Cc: Laurent Vivier <laur...@vivier.eu> Cc: Richard Henderson <richard.hender...@linaro.org> Cc: Alistair Francis <alistair.fran...@wdc.com> Signed-off-by: Michael Clark <m...@sifive.com> --- include/elf.h | 8 1 file changed, 8 insertions(+) diff --git a/include/elf.h b/include/elf.h index 934dbbd6b3ae..bd0493f43d19 100644 --- a/include/elf.h +++ b/include/elf.h @@ -1285,6 +1285,14 @@ typedef struct { #define R_IA64_DTPREL64LSB 0xb7/* @dtprel(sym + add), data8 LSB */ #define R_IA64_LTOFF_DTPREL22 0xba/* @ltoff(@dtprel(s+a)), imm22 */ +/* RISC-V specific definitions. */ +#define EF_RISCV_RVC 0x0001 +#define EF_RISCV_FLOAT_ABI_SINGLE 0x0002 +#define EF_RISCV_FLOAT_ABI_DOUBLE 0x0004 +#define EF_RISCV_FLOAT_ABI_QUAD 0x0006 +#define EF_RISCV_RVE 0x0008 +#define EF_RISCV_TSO 0x0010 + typedef struct elf32_rel { Elf32_Addr r_offset; Elf32_Word r_info; -- 2.7.0
Re: [Qemu-devel] [PATCH v1 27/30] elf: Add RISC-V PSABI ELF header defines
On Wed, May 23, 2018 at 6:44 PM, Laurent Vivier <laur...@vivier.eu> wrote: > Le 23/05/2018 à 02:15, Michael Clark a écrit : > > Refer to the RISC-V PSABI specification for details: > > > > - https://github.com/riscv/riscv-elf-psabi-doc/blob/master/riscv-elf.md > > > > Cc: Michael Tokarev <m...@tls.msk.ru> > > Cc: Laurent Vivier <laur...@vivier.eu> > > Cc: Richard Henderson <richard.hender...@linaro.org> > > Cc: Alistair Francis <alistair.fran...@wdc.com> > > Signed-off-by: Michael Clark <m...@sifive.com> > > --- > > include/elf.h | 8 > > 1 file changed, 8 insertions(+) > > > > diff --git a/include/elf.h b/include/elf.h > > index 934dbbd6b3ae..d363ba85a688 100644 > > --- a/include/elf.h > > +++ b/include/elf.h > > @@ -1285,6 +1285,14 @@ typedef struct { > > #define R_IA64_DTPREL64LSB 0xb7/* @dtprel(sym + add), data8 LSB */ > > #define R_IA64_LTOFF_DTPREL220xba/* @ltoff(@dtprel(s+a)), > imm22 */ > > > > +/* RISC-V specific definitions. */ > > +#define EF_RISCV_RVC 0x0001 > > +#define EF_RISCV_FLOAT_ABI_SINGLE 0x0002 > > +#define EF_RISCV_FLOAT_ABI_DOUBLE 0x0004 > > +#define EF_RISCV_FLOAT_ABI_QUAD (0x0006 >^ > Typo here -| > Thanks! My mistake. The original patch had only EF_RISCV_RVE, which is the define we need for a subsequent patch, however I decided to add the remaining flags from the spec and indeed there was a paren in the spec. I'll respin this as a separate patch.
[Qemu-devel] [PATCH] RISC-V: Correct typo in RV32 perf counters
This patch enables mhpmcounter3h through mhpmcounter31h on RV32. Previously the RV32 h versions (high 32-bits of 64-bit counters) of these counters would trap with an illegal instruction instead of returning 0 as intended. Reported-by: Richard Henderson <r...@twiddle.net> Signed-off-by: Michael Clark <m...@sifive.com> --- target/riscv/op_helper.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/target/riscv/op_helper.c b/target/riscv/op_helper.c index 3abf52453cfc..1f6dc9a85852 100644 --- a/target/riscv/op_helper.c +++ b/target/riscv/op_helper.c @@ -406,7 +406,7 @@ target_ulong csr_read_helper(CPURISCVState *env, target_ulong csrno) return 0; } #if defined(TARGET_RISCV32) -if (csrno >= CSR_MHPMCOUNTER3 && csrno <= CSR_MHPMCOUNTER31) { +if (csrno >= CSR_MHPMCOUNTER3H && csrno <= CSR_MHPMCOUNTER31H) { return 0; } #endif -- 2.7.0
Re: [Qemu-devel] [RISC-V] Coverity 1390849, Logically dead code
On Fri, May 25, 2018 at 9:54 AM, Richard Hendersonwrote: > In the latest Coverity scan, it reports > > 405if (csrno >= CSR_MHPMCOUNTER3 && csrno <= CSR_MHPMCOUNTER31) { > 406return 0; > 407} > 408#if defined(TARGET_RISCV32) > 409if (csrno >= CSR_MHPMCOUNTER3 && csrno <= CSR_MHPMCOUNTER31) { > > CID 1390849 (#1 of 1): Logically dead code (DEADCODE) > dead_error_line: Execution cannot reach this statement: return 0U;. > > 410return 0; > 411} > 412#endif > > > I believe the condition at op_helper.c:409 should be testing > CSR_MHPMCOUNTER3H > and CSR_MHPMCOUNTER31H. Must run now, otherwise I'd also send the trivial > patch. Thanks! Yes indeed they should be the H versions. I will send the trivial patch...
Re: [Qemu-devel] [PATCH v1 02/30] RISC-V: Improve page table walker spec compliance
Hi Phil, Alastair, Richard, et al... Apologies if I'm slow to respond to your other review emails. I'm going to go through all of them and address each of them one by one. Currently, I want to either get some MMU tests from the verification team or separately write tests for the newly added comments in the page walker. 1). +/* Invalid PTE */ 2). +/* Inner PTE, continue walking */ 3). +/* Reserved leaf PTE flags: PTE_W */ 4). +/* Reserved leaf PTE flags: PTE_W + PTE_X */ 5). +/* User PTE flags when not U mode and mstatus.SUM is not set, 6). +/* Supervisor PTE flags when not S mode */ 7). +/* Misasligned PPN */ 8). +/* Read access check failed */ 9). +/* Write access check failed */ 10). +/* Fetch access check failed */ 11). /* if necessary, set accessed and dirty bits. */ That's why I might be slow to respond to the other emails but I will get to them in due course. SiFive has internal verification tests which require RTL test harness. Currently, the open source MMU tests need expanding. It might be better that we expand the latter as it will be helpful for other implementations to test their page walkers (hardware and emulators alike): - https://github.com/riscv/riscv-tests/tree/master/isa/rv64si/ I can then compare master vs this patch. I think master may fail Misasligned PPN. I may just test critical function such as U-mode can't access U=0. That's pretty easy as linux-kernel has a shared kernel address space and riscv-linux doesn't yet have ASLR so I can just try to deref a kernel address in userspace. In fact we still need to implement kernel memory protection for riscv-linux i.e. .rodata +R, .text +RX and .data +RW (W^X) before we add ASLR. The linux kernel port is indeed quite new but this patch does not regress any user-facing functionality. One of the goals with this patch is to make the logical clauses and comments match the English text in the RISC-V Privileged ISA v1.10 with respect to the MMU. Michael. On Wed, May 23, 2018 at 12:14 PM, Michael Clark <m...@sifive.com> wrote: > - Inline PTE_TABLE check for better readability > - Change access checks from ternary operator to if > - Improve readibility of User page U mode and SUM test > - Disallow non U mode from fetching from User pages > - Add reserved PTE flag check: W or W|X > - Add misaligned PPN check > - Set READ protection for PTE X flag and mstatus.mxr > - Use memory_region_is_ram in pte update > > Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> > Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> > Cc: Palmer Dabbelt <pal...@sifive.com> > Cc: Alistair Francis <alistair.fran...@wdc.com> > Signed-off-by: Michael Clark <m...@sifive.com> > Reviewed-by: Alistair Francis <alistair.fran...@wdc.com> > --- > target/riscv/cpu_bits.h | 2 -- > target/riscv/helper.c | 64 ++ > --- > 2 files changed, 45 insertions(+), 21 deletions(-) > > diff --git a/target/riscv/cpu_bits.h b/target/riscv/cpu_bits.h > index 64aa097181fa..12b4757088f4 100644 > --- a/target/riscv/cpu_bits.h > +++ b/target/riscv/cpu_bits.h > @@ -407,5 +407,3 @@ > #define PTE_SOFT 0x300 /* Reserved for Software */ > > #define PTE_PPN_SHIFT 10 > - > -#define PTE_TABLE(PTE) (((PTE) & (PTE_V | PTE_R | PTE_W | PTE_X)) == > PTE_V) > diff --git a/target/riscv/helper.c b/target/riscv/helper.c > index 95889f23b94d..3b57e1360549 100644 > --- a/target/riscv/helper.c > +++ b/target/riscv/helper.c > @@ -185,16 +185,39 @@ restart: > #endif > target_ulong ppn = pte >> PTE_PPN_SHIFT; > > -if (PTE_TABLE(pte)) { /* next level of page table */ > +if (!(pte & PTE_V)) { > +/* Invalid PTE */ > +return TRANSLATE_FAIL; > +} else if (!(pte & (PTE_R | PTE_W | PTE_X))) { > +/* Inner PTE, continue walking */ > base = ppn << PGSHIFT; > -} else if ((pte & PTE_U) ? (mode == PRV_S) && !sum : !(mode == > PRV_S)) { > -break; > -} else if (!(pte & PTE_V) || (!(pte & PTE_R) && (pte & PTE_W))) { > -break; > -} else if (access_type == MMU_INST_FETCH ? !(pte & PTE_X) : > - access_type == MMU_DATA_LOAD ? !(pte & PTE_R) && > - !(mxr && (pte & PTE_X)) : !((pte & PTE_R) && (pte & > PTE_W))) { > -break; > +} else if ((pte & (PTE_R | PTE_W | PTE_X)) == PTE_W) { > +/* Reserved leaf PTE flags: PTE_W */ > +return TRANSLATE_FAIL; > +} else if ((pte & (PTE_R | PTE_W | PTE_X)) == (PTE_W | PTE_X))
[Qemu-devel] [PATCH v1 30/30] RISC-V: Support separate firmware and kernel payload
Support for separate firmware and kernel payload is added by updating BBL to read optional preloaded kernel address attributes from device-tree using a similar mechanism to that used to pass init ramdisk addresses to linux kernel. chosen { riscv,kernel-start = <0x 0x8020>; riscv,kernel-end = <0x 0x80590634>; }; These attributes are added by QEMU and read by BBL when combining -bios and -kernel options. e.g. $ qemu-system-riscv64 -machine virt -bios bbl -kernel vmlinux With this change, bbl can be compiled without --with-payload and the dummy payload alignment is altered to make the memory footprint of the firmware-only bbl smaller. The dummy payload message is updated to indicate the alternative load method. This load method could also be supported by a first stage boot loader that reads seperate firmware and kernel from SPI flash. The main advantage of this new mechanism is that it eases kernel development by avoiding the riscv-pk packaging step after kernel builds, makes building per repository artefacts for CI simpler, and mimics bootloaders on other platforms that can load a kernel image file directly. Ultimately BBL should use an SPI driver to load the kernel image however this mechanism supports use cases such such as QEMU's -bios, -kernel and -initrd options following examples from other platforms that pass kernel entry to firmware via device-tree. The board is also changed to use the firmware address from the loaded firmware or combined firmware+kernel. This is normally equal to the DRAM base address of 0x8000_, however now it is possible to boot firmware at different load addresses because the reset code jumps to the actual firmware entry address. Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Alistair Francis <alistair.fran...@wdc.com> Signed-off-by: Michael Clark <m...@sifive.com> --- hw/riscv/Makefile.objs | 1 + hw/riscv/boot.c | 172 hw/riscv/virt.c | 67 +++ include/hw/riscv/boot.h | 30 + 4 files changed, 213 insertions(+), 57 deletions(-) create mode 100644 hw/riscv/boot.c create mode 100644 include/hw/riscv/boot.h diff --git a/hw/riscv/Makefile.objs b/hw/riscv/Makefile.objs index 1dde01d39dcc..d36b004ab0f9 100644 --- a/hw/riscv/Makefile.objs +++ b/hw/riscv/Makefile.objs @@ -1,3 +1,4 @@ +obj-y += boot.o obj-y += riscv_htif.o obj-y += riscv_hart.o obj-y += sifive_e.o diff --git a/hw/riscv/boot.c b/hw/riscv/boot.c new file mode 100644 index ..cf4e5d594638 --- /dev/null +++ b/hw/riscv/boot.c @@ -0,0 +1,172 @@ +/* + * QEMU RISCV firmware and kernel loader + * + * Copyright (c) 2017-2018 SiFive, Inc. + * + * Holds the state of a heterogenous array of RISC-V harts + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2 or later, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program. If not, see <http://www.gnu.org/licenses/>. + */ + +#include "qemu/osdep.h" +#include "qemu/log.h" +#include "qemu/error-report.h" +#include "hw/loader.h" +#include "hw/boards.h" +#include "sysemu/device_tree.h" +#include "elf.h" +#include "hw/riscv/boot.h" + +#define RISCV_BOOT_DEBUG 0 + +#define boot_debug(fs, ...) \ +if (RISCV_BOOT_DEBUG) { \ +fprintf(stderr, "boot: %s: "fs, __func__, ##__VA_ARGS__); \ +} + +static uint64_t kernel_offset; + +static uint64_t kernel_translate(void *opaque, uint64_t addr) +{ +/* mask kernel virtual address and offset by load address */ +if (kernel_offset) { +return (addr & 0x7fff) + kernel_offset; +} else { +return addr; +} +} + +hwaddr riscv_load_firmware(const char *filename) +{ +uint64_t firmware_entry, firmware_start, firmware_end; + +if (load_elf(filename, NULL, NULL, + _entry, _start, _end, + 0, EM_RISCV, 1, 0) < 0) { +error_report("riscv_boot: could not load firmware '%s'", filename); +exit(1); +} + +/* align kernel load address to the megapage after the firmware */ +#if defined(TARGET_RISCV32) +kernel_offset = (firmware_end + 0x3f) & ~0x3f; +#else +kernel_offset = (firmware_end + 0x1f) & ~0x1f; +#endif + +boot_debug("entry=0x" TARGET_FMT_plx " start=0x" TARGET_FMT_plx " " + "end=0x" TARGET_FMT_plx " kernel_offset
[Qemu-devel] [PATCH v1 29/30] RISC-V: Don't add NULL bootargs to device-tree
--- hw/riscv/sifive_u.c | 4 +++- hw/riscv/spike.c| 6 -- hw/riscv/virt.c | 4 +++- 3 files changed, 10 insertions(+), 4 deletions(-) diff --git a/hw/riscv/sifive_u.c b/hw/riscv/sifive_u.c index 326b0f434cff..02721d43c474 100644 --- a/hw/riscv/sifive_u.c +++ b/hw/riscv/sifive_u.c @@ -206,7 +206,9 @@ static void create_fdt(SiFiveUState *s, const struct MemmapEntry *memmap, qemu_fdt_add_subnode(fdt, "/chosen"); qemu_fdt_setprop_string(fdt, "/chosen", "stdout-path", nodename); -qemu_fdt_setprop_string(fdt, "/chosen", "bootargs", cmdline); +if (cmdline) { +qemu_fdt_setprop_string(fdt, "/chosen", "bootargs", cmdline); +} g_free(nodename); } diff --git a/hw/riscv/spike.c b/hw/riscv/spike.c index f94e2b670799..5dd7d28aed4e 100644 --- a/hw/riscv/spike.c +++ b/hw/riscv/spike.c @@ -156,8 +156,10 @@ static void create_fdt(SpikeState *s, const struct MemmapEntry *memmap, g_free(cells); g_free(nodename); -qemu_fdt_add_subnode(fdt, "/chosen"); -qemu_fdt_setprop_string(fdt, "/chosen", "bootargs", cmdline); +if (cmdline) { +qemu_fdt_add_subnode(fdt, "/chosen"); +qemu_fdt_setprop_string(fdt, "/chosen", "bootargs", cmdline); +} } static void spike_v1_10_0_board_init(MachineState *machine) diff --git a/hw/riscv/virt.c b/hw/riscv/virt.c index 321fa6e8122a..c889aa3cd269 100644 --- a/hw/riscv/virt.c +++ b/hw/riscv/virt.c @@ -253,7 +253,9 @@ static void *create_fdt(RISCVVirtState *s, const struct MemmapEntry *memmap, qemu_fdt_add_subnode(fdt, "/chosen"); qemu_fdt_setprop_string(fdt, "/chosen", "stdout-path", nodename); -qemu_fdt_setprop_string(fdt, "/chosen", "bootargs", cmdline); +if (cmdline) { +qemu_fdt_setprop_string(fdt, "/chosen", "bootargs", cmdline); +} g_free(nodename); return fdt; -- 2.7.0
[Qemu-devel] [PATCH v1 28/30] RISC-V: linux-user support for RVE ABI
From: Kito Cheng <kito.ch...@gmail.com> This change checks elf_flags for EF_RISCV_RVE and if present uses the RVE linux syscall ABI which uses t0 for the syscall number instead of a7. Warn and exit if a non-RVE ABI binary is run on a cpu with the RVE extension as it is incompatible. Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Cc: Alistair Francis <alistair.fran...@wdc.com> Co-authored-by: Kito Cheng <kito.ch...@gmail.com> Co-authored-by: Michael Clark <m...@sifive.com> Signed-off-by: Michael Clark <m...@sifive.com> --- linux-user/riscv/cpu_loop.c | 14 +- target/riscv/cpu.h | 4 target/riscv/cpu_user.h | 3 ++- 3 files changed, 19 insertions(+), 2 deletions(-) diff --git a/linux-user/riscv/cpu_loop.c b/linux-user/riscv/cpu_loop.c index f137d39d7e82..5f6a941c6c19 100644 --- a/linux-user/riscv/cpu_loop.c +++ b/linux-user/riscv/cpu_loop.c @@ -20,6 +20,7 @@ #include "qemu/osdep.h" #include "qemu.h" #include "cpu_loop-common.h" +#include "elf.h" void cpu_loop(CPURISCVState *env) { @@ -53,7 +54,8 @@ void cpu_loop(CPURISCVState *env) ret = 0; } else { ret = do_syscall(env, - env->gpr[xA7], + env->gpr[(env->elf_flags & EF_RISCV_RVE) +? xT0 : xA7], env->gpr[xA0], env->gpr[xA1], env->gpr[xA2], @@ -113,6 +115,16 @@ void cpu_loop(CPURISCVState *env) void target_cpu_copy_regs(CPUArchState *env, struct target_pt_regs *regs) { +CPUState *cpu = ENV_GET_CPU(env); +TaskState *ts = cpu->opaque; +struct image_info *info = ts->info; + env->pc = regs->sepc; env->gpr[xSP] = regs->sp; +env->elf_flags = info->elf_flags; + +if ((env->misa & RVE) && !(env->elf_flags & EF_RISCV_RVE)) { +fprintf(stderr, "Incompatible ELF: RVE cpu requires RVE ABI binary\n"); +exit(EXIT_FAILURE); +} } diff --git a/target/riscv/cpu.h b/target/riscv/cpu.h index 830a9d476dce..0823461ae9f2 100644 --- a/target/riscv/cpu.h +++ b/target/riscv/cpu.h @@ -123,6 +123,10 @@ struct CPURISCVState { uint32_t features; +#ifdef CONFIG_USER_ONLY +uint32_t elf_flags; +#endif + #ifndef CONFIG_USER_ONLY target_ulong priv; target_ulong resetvec; diff --git a/target/riscv/cpu_user.h b/target/riscv/cpu_user.h index c2199610abff..52d380aa98c2 100644 --- a/target/riscv/cpu_user.h +++ b/target/riscv/cpu_user.h @@ -10,4 +10,5 @@ #define xA4 14 #define xA5 15 #define xA6 16 -#define xA7 17 /* syscall number goes here */ +#define xA7 17 /* syscall number for RVI ABI */ +#define xT0 5 /* syscall number for RVE ABI */ -- 2.7.0
[Qemu-devel] [PATCH v1 26/30] RISC-V: Remove unnecessary disassembler constraints
Remove machine generated constraints that are not referenced by the pseudo-instruction constraints. Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Cc: Alistair Francis <alistair.fran...@wdc.com> Signed-off-by: Michael Clark <m...@sifive.com> --- disas/riscv.c | 138 -- 1 file changed, 138 deletions(-) diff --git a/disas/riscv.c b/disas/riscv.c index 7fd1019623ee..27546dd7902c 100644 --- a/disas/riscv.c +++ b/disas/riscv.c @@ -87,33 +87,10 @@ typedef enum { typedef enum { rvc_end, -rvc_simm_6, -rvc_imm_6, -rvc_imm_7, -rvc_imm_8, -rvc_imm_9, -rvc_imm_10, -rvc_imm_12, -rvc_imm_18, -rvc_imm_nz, -rvc_imm_x2, -rvc_imm_x4, -rvc_imm_x8, -rvc_imm_x16, -rvc_rd_b3, -rvc_rs1_b3, -rvc_rs2_b3, -rvc_rd_eq_rs1, rvc_rd_eq_ra, -rvc_rd_eq_sp, rvc_rd_eq_x0, -rvc_rs1_eq_sp, rvc_rs1_eq_x0, rvc_rs2_eq_x0, -rvc_rd_ne_x0_x2, -rvc_rd_ne_x0, -rvc_rs1_ne_x0, -rvc_rs2_ne_x0, rvc_rs2_eq_rs1, rvc_rs1_eq_ra, rvc_imm_eq_zero, @@ -2522,111 +2499,16 @@ static bool check_constraints(rv_decode *dec, const rvc_constraint *c) uint8_t rd = dec->rd, rs1 = dec->rs1, rs2 = dec->rs2; while (*c != rvc_end) { switch (*c) { -case rvc_simm_6: -if (!(imm >= -32 && imm < 32)) { -return false; -} -break; -case rvc_imm_6: -if (!(imm <= 63)) { -return false; -} -break; -case rvc_imm_7: -if (!(imm <= 127)) { -return false; -} -break; -case rvc_imm_8: -if (!(imm <= 255)) { -return false; -} -break; -case rvc_imm_9: -if (!(imm <= 511)) { -return false; -} -break; -case rvc_imm_10: -if (!(imm <= 1023)) { -return false; -} -break; -case rvc_imm_12: -if (!(imm <= 4095)) { -return false; -} -break; -case rvc_imm_18: -if (!(imm <= 262143)) { -return false; -} -break; -case rvc_imm_nz: -if (!(imm != 0)) { -return false; -} -break; -case rvc_imm_x2: -if (!((imm & 0b1) == 0)) { -return false; -} -break; -case rvc_imm_x4: -if (!((imm & 0b11) == 0)) { -return false; -} -break; -case rvc_imm_x8: -if (!((imm & 0b111) == 0)) { -return false; -} -break; -case rvc_imm_x16: -if (!((imm & 0b) == 0)) { -return false; -} -break; -case rvc_rd_b3: -if (!(rd >= 8 && rd <= 15)) { -return false; -} -break; -case rvc_rs1_b3: -if (!(rs1 >= 8 && rs1 <= 15)) { -return false; -} -break; -case rvc_rs2_b3: -if (!(rs2 >= 8 && rs2 <= 15)) { -return false; -} -break; -case rvc_rd_eq_rs1: -if (!(rd == rs1)) { -return false; -} -break; case rvc_rd_eq_ra: if (!(rd == 1)) { return false; } break; -case rvc_rd_eq_sp: -if (!(rd == 2)) { -return false; -} -break; case rvc_rd_eq_x0: if (!(rd == 0)) { return false; } break; -case rvc_rs1_eq_sp: -if (!(rs1 == 2)) { -return false; -} -break; case rvc_rs1_eq_x0: if (!(rs1 == 0)) { return false; @@ -2637,26 +2519,6 @@ static bool check_constraints(rv_decode *dec, const rvc_constraint *c) return false; } break; -case rvc_rd_ne_x0_x2: -if (!(rd != 0 && rd != 2)) { -return false; -} -break; -case rvc_rd_ne_x0: -if (!(rd != 0)) { -return false; -} -break; -case rvc_rs1_ne_x0: -if (!(rs1 != 0)) { -return false; -} -break; -case rvc_rs2_ne_x0: -if (!(rs2 != 0)) { -re
[Qemu-devel] [PATCH v1 22/30] RISC-V: Add misa runtime write support
This patch adds support for writing misa. misa is validated based on rules in the ISA specification. 'E' is mutually exclusive with all other extensions. 'D' depends on 'F' so 'D' bit is dropped if 'F' is not present. A conservative approach to consistency is taken by flushing the translation cache on misa writes. misa_mask is added to the CPU struct to store the original set of extensions. Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Cc: Alistair Francis <alistair.fran...@wdc.com> Signed-off-by: Michael Clark <m...@sifive.com> --- target/riscv/cpu.c | 2 +- target/riscv/cpu.h | 4 +++- target/riscv/cpu_bits.h | 11 +++ target/riscv/csr.c | 52 - 4 files changed, 66 insertions(+), 3 deletions(-) diff --git a/target/riscv/cpu.c b/target/riscv/cpu.c index 9f47f2072c56..d59e1c0c146d 100644 --- a/target/riscv/cpu.c +++ b/target/riscv/cpu.c @@ -86,7 +86,7 @@ typedef struct RISCVCPUInfo { static void set_misa(CPURISCVState *env, target_ulong misa) { -env->misa = misa; +env->misa_mask = env->misa = misa; } static void set_versions(CPURISCVState *env, int user_ver, int priv_ver) diff --git a/target/riscv/cpu.h b/target/riscv/cpu.h index ae0e3f6a544d..830a9d476dce 100644 --- a/target/riscv/cpu.h +++ b/target/riscv/cpu.h @@ -86,7 +86,8 @@ so a cpu features bitfield is required, likewise for optional PMP support */ enum { RISCV_FEATURE_MMU, -RISCV_FEATURE_PMP +RISCV_FEATURE_PMP, +RISCV_FEATURE_MISA_RW }; #define USER_VERSION_2_02_0 0x00020200 @@ -118,6 +119,7 @@ struct CPURISCVState { target_ulong user_ver; target_ulong priv_ver; target_ulong misa; +target_ulong misa_mask; uint32_t features; diff --git a/target/riscv/cpu_bits.h b/target/riscv/cpu_bits.h index 878de6233846..55b45a40ca3f 100644 --- a/target/riscv/cpu_bits.h +++ b/target/riscv/cpu_bits.h @@ -311,10 +311,21 @@ #define MSTATUS32_SD0x8000 #define MSTATUS64_SD0x8000ULL +#define MISA32_MXL 0xC000 +#define MISA64_MXL 0xC000ULL + +#define MXL_RV321 +#define MXL_RV642 +#define MXL_RV128 3 + #if defined(TARGET_RISCV32) #define MSTATUS_SD MSTATUS32_SD +#define MISA_MXL MISA32_MXL +#define MXL_VAL MXL_RV32 #elif defined(TARGET_RISCV64) #define MSTATUS_SD MSTATUS64_SD +#define MISA_MXL MISA64_MXL +#define MXL_VAL MXL_RV64 #endif /* sstatus CSR bits */ diff --git a/target/riscv/csr.c b/target/riscv/csr.c index 9bbe81a110a5..0be973002ee2 100644 --- a/target/riscv/csr.c +++ b/target/riscv/csr.c @@ -341,6 +341,56 @@ static int read_misa(CPURISCVState *env, int csrno, target_ulong *val) return 0; } +static int write_misa(CPURISCVState *env, int csrno, target_ulong val) +{ +if (!riscv_feature(env, RISCV_FEATURE_MISA_RW)) { +/* drop write to misa */ +return 0; +} + +/* 'I' or 'E' must be present */ +if (!(val & (RVI | RVE))) { +/* it not, drop write to misa */ +return 0; +} + +/* 'E' excludes all other extensions */ +if (val & RVE) { +/* when we support 'E' we can do "val = RVE;" however + * for now we just drop writes if 'E' is present */ +return 0; +} + +/* Mask extensions that are not supported by this hart */ +val &= env->misa_mask; + +/* Mask extensions that are not supported by QEMU */ +val &= (RVI | RVE | RVM | RVA | RVF | RVD | RVC | RVS | RVU); + +/* 'D' depends on 'F', so clear 'D' if 'F' is not present */ +if ((val & RVD) && !(val & RVF)) { +val &= ~RVD; +} + +/* Suppress 'C' if next instruction is not aligned + TODO: this should check next_pc */ +if ((val & RVC) && (GETPC() & ~3) != 0) { +val &= ~RVC; +} + +/* misa.MXL writes are not supported by QEMU */ +val = (env->misa & MISA_MXL) | (val & ~MISA_MXL); + +/* flush translation cache */ +if (val != env->misa) { +tb_flush(CPU(riscv_env_get_cpu(env))); +} + +env->misa = val; + +return 0; +} + static int read_medeleg(CPURISCVState *env, int csrno, target_ulong *val) { *val = env->medeleg; @@ -821,7 +871,7 @@ static riscv_csr_operations csr_ops[CSR_TABLE_SIZE] = { /* Machine Trap Setup */ [CSR_MSTATUS] = { any, read_mstatus, write_mstatus }, -[CSR_MISA] ={ any, read_misa }, +[CSR_MISA] ={ any, read_misa,write_misa}, [CSR_MIDELEG] = { any, read_mideleg, write_mideleg }, [CSR_MEDELEG] = { any, read_medeleg, write_medeleg }, [CSR_MIE] = { any, read_mie, write_mie }, -- 2.7.0
[Qemu-devel] [PATCH v1 20/30] RISC-V: Add misa to DisasContext
gen methods should access state from DisasContext. Add misa field to the DisasContext struct and remove CPURISCVState argument from all gen methods. Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Cc: Alistair Francis <alistair.fran...@wdc.com> Cc: Emilio G. Cota <c...@braap.org> Signed-off-by: Michael Clark <m...@sifive.com> Reviewed-by: Richard Henderson <richard.hender...@linaro.org> --- target/riscv/translate.c | 78 ++-- 1 file changed, 42 insertions(+), 36 deletions(-) diff --git a/target/riscv/translate.c b/target/riscv/translate.c index a980611eb611..fd21b133a5a4 100644 --- a/target/riscv/translate.c +++ b/target/riscv/translate.c @@ -45,6 +45,7 @@ typedef struct DisasContext { target_ulong pc_succ_insn; uint32_t opcode; uint32_t mstatus_fs; +uint32_t misa; uint32_t mem_idx; /* Remember the rounding mode encoded in the previous fp instruction, which we have already installed into env->fp_status. Or -1 for @@ -74,6 +75,11 @@ static const int tcg_memop_lookup[8] = { #define CASE_OP_32_64(X) case X #endif +static inline bool has_ext(DisasContext *ctx, uint32_t ext) +{ +return ctx->misa & ext; +} + static void generate_exception(DisasContext *ctx, int excp) { tcg_gen_movi_tl(cpu_pc, ctx->base.pc_next); @@ -505,14 +511,13 @@ static void gen_arith_imm(DisasContext *ctx, uint32_t opc, int rd, tcg_temp_free(source1); } -static void gen_jal(CPURISCVState *env, DisasContext *ctx, int rd, -target_ulong imm) +static void gen_jal(DisasContext *ctx, int rd, target_ulong imm) { target_ulong next_pc; /* check misaligned: */ next_pc = ctx->base.pc_next + imm; -if (!riscv_has_ext(env, RVC)) { +if (!has_ext(ctx, RVC)) { if ((next_pc & 0x3) != 0) { gen_exception_inst_addr_mis(ctx); return; @@ -526,8 +531,8 @@ static void gen_jal(CPURISCVState *env, DisasContext *ctx, int rd, ctx->base.is_jmp = DISAS_NORETURN; } -static void gen_jalr(CPURISCVState *env, DisasContext *ctx, uint32_t opc, - int rd, int rs1, target_long imm) +static void gen_jalr(DisasContext *ctx, uint32_t opc, int rd, int rs1, + target_long imm) { /* no chaining with JALR */ TCGLabel *misaligned = NULL; @@ -539,7 +544,7 @@ static void gen_jalr(CPURISCVState *env, DisasContext *ctx, uint32_t opc, tcg_gen_addi_tl(cpu_pc, cpu_pc, imm); tcg_gen_andi_tl(cpu_pc, cpu_pc, (target_ulong)-2); -if (!riscv_has_ext(env, RVC)) { +if (!has_ext(ctx, RVC)) { misaligned = gen_new_label(); tcg_gen_andi_tl(t0, cpu_pc, 0x2); tcg_gen_brcondi_tl(TCG_COND_NE, t0, 0x0, misaligned); @@ -564,8 +569,8 @@ static void gen_jalr(CPURISCVState *env, DisasContext *ctx, uint32_t opc, tcg_temp_free(t0); } -static void gen_branch(CPURISCVState *env, DisasContext *ctx, uint32_t opc, - int rs1, int rs2, target_long bimm) +static void gen_branch(DisasContext *ctx, uint32_t opc, int rs1, int rs2, + target_long bimm) { TCGLabel *l = gen_new_label(); TCGv source1, source2; @@ -602,7 +607,7 @@ static void gen_branch(CPURISCVState *env, DisasContext *ctx, uint32_t opc, gen_goto_tb(ctx, 1, ctx->pc_succ_insn); gen_set_label(l); /* branch taken */ -if (!riscv_has_ext(env, RVC) && ((ctx->base.pc_next + bimm) & 0x3)) { +if (!has_ext(ctx, RVC) && ((ctx->base.pc_next + bimm) & 0x3)) { /* misaligned */ gen_exception_inst_addr_mis(ctx); } else { @@ -1311,8 +1316,8 @@ static void gen_fp_arith(DisasContext *ctx, uint32_t opc, int rd, } } -static void gen_system(CPURISCVState *env, DisasContext *ctx, uint32_t opc, - int rd, int rs1, int csr) +static void gen_system(DisasContext *ctx, uint32_t opc, int rd, int rs1, + int csr) { TCGv source1, csr_store, dest, rs1_pass, imm_rs1; source1 = tcg_temp_new(); @@ -1354,7 +1359,7 @@ static void gen_system(CPURISCVState *env, DisasContext *ctx, uint32_t opc, gen_exception_illegal(ctx); break; case 0x102: /* SRET */ -if (riscv_has_ext(env, RVS)) { +if (has_ext(ctx, RVS)) { gen_helper_sret(cpu_pc, cpu_env, cpu_pc); tcg_gen_exit_tb(0); /* no chaining */ ctx->base.is_jmp = DISAS_NORETURN; @@ -1495,7 +1500,7 @@ static void decode_RV32_64C0(DisasContext *ctx) } } -static void decode_RV32_64C1(CPURISCVState *env, DisasContext *ctx) +static void decode_RV32_64C1(DisasContext *ctx) { uint8_t funct3 = extract32(ctx->opcode, 13, 3); uint8_t rd_rs1 = GET_C_RS1
[Qemu-devel] [PATCH v1 11/30] RISC-V: Split out mstatus_fs from tb_flags
From: Richard Henderson <richard.hender...@linaro.org> Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Alistair Francis <alistair.fran...@wdc.com> Cc: Richard Henderson <richard.hender...@linaro.org> Signed-off-by: Michael Clark <m...@sifive.com> Reviewed-by: Michael Clark <m...@sifive.com> --- target/riscv/cpu.h | 6 +++--- target/riscv/translate.c | 10 +- 2 files changed, 8 insertions(+), 8 deletions(-) diff --git a/target/riscv/cpu.h b/target/riscv/cpu.h index 3a3d91447736..242a8fcbe180 100644 --- a/target/riscv/cpu.h +++ b/target/riscv/cpu.h @@ -276,8 +276,8 @@ void QEMU_NORETURN do_raise_exception_err(CPURISCVState *env, target_ulong cpu_riscv_get_fflags(CPURISCVState *env); void cpu_riscv_set_fflags(CPURISCVState *env, target_ulong); -#define TB_FLAGS_MMU_MASK 3 -#define TB_FLAGS_FP_ENABLE MSTATUS_FS +#define TB_FLAGS_MMU_MASK 3 +#define TB_FLAGS_MSTATUS_FS MSTATUS_FS static inline void cpu_get_tb_cpu_state(CPURISCVState *env, target_ulong *pc, target_ulong *cs_base, uint32_t *flags) @@ -285,7 +285,7 @@ static inline void cpu_get_tb_cpu_state(CPURISCVState *env, target_ulong *pc, *pc = env->pc; *cs_base = 0; #ifdef CONFIG_USER_ONLY -*flags = TB_FLAGS_FP_ENABLE; +*flags = TB_FLAGS_MSTATUS_FS; #else *flags = cpu_mmu_index(env, 0) | (env->mstatus & MSTATUS_FS); #endif diff --git a/target/riscv/translate.c b/target/riscv/translate.c index ee2bbc55b051..466b9551cbd9 100644 --- a/target/riscv/translate.c +++ b/target/riscv/translate.c @@ -44,7 +44,7 @@ typedef struct DisasContext { /* pc_succ_insn points to the instruction following base.pc_next */ target_ulong pc_succ_insn; uint32_t opcode; -uint32_t flags; +uint32_t mstatus_fs; uint32_t mem_idx; /* Remember the rounding mode encoded in the previous fp instruction, which we have already installed into env->fp_status. Or -1 for @@ -656,7 +656,7 @@ static void gen_fp_load(DisasContext *ctx, uint32_t opc, int rd, { TCGv t0; -if (!(ctx->flags & TB_FLAGS_FP_ENABLE)) { +if (ctx->mstatus_fs == 0) { gen_exception_illegal(ctx); return; } @@ -686,7 +686,7 @@ static void gen_fp_store(DisasContext *ctx, uint32_t opc, int rs1, { TCGv t0; -if (!(ctx->flags & TB_FLAGS_FP_ENABLE)) { +if (ctx->mstatus_fs == 0) { gen_exception_illegal(ctx); return; } @@ -945,7 +945,7 @@ static void gen_fp_arith(DisasContext *ctx, uint32_t opc, int rd, { TCGv t0 = NULL; -if (!(ctx->flags & TB_FLAGS_FP_ENABLE)) { +if (ctx->mstatus_fs == 0) { goto do_illegal; } @@ -1810,8 +1810,8 @@ static void riscv_tr_init_disas_context(DisasContextBase *dcbase, CPUState *cs) DisasContext *ctx = container_of(dcbase, DisasContext, base); ctx->pc_succ_insn = ctx->base.pc_first; -ctx->flags = ctx->base.tb->flags; ctx->mem_idx = ctx->base.tb->flags & TB_FLAGS_MMU_MASK; +ctx->mstatus_fs = ctx->base.tb->flags & TB_FLAGS_MSTATUS_FS; ctx->frm = -1; /* unknown rounding mode */ } -- 2.7.0
[Qemu-devel] [PATCH v1 24/30] RISC-V: Fix PLIC pending bitfield reads
The address calculation for the pending bitfield had a copy paste bug. This bug went unnoticed because the Linux PLIC driver does not read the pending bitfield, rather it reads pending interrupt numbers from the claim register and writes acknowledgements back to the claim register. Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Cc: Alistair Francis <alistair.fran...@wdc.com> Reported-by: Vincent Siles <vincent.si...@ens-lyon.org> Signed-off-by: Michael Clark <m...@sifive.com> --- hw/riscv/sifive_plic.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/hw/riscv/sifive_plic.c b/hw/riscv/sifive_plic.c index 28e28d932f7c..b81d29faff99 100644 --- a/hw/riscv/sifive_plic.c +++ b/hw/riscv/sifive_plic.c @@ -215,7 +215,7 @@ static uint64_t sifive_plic_read(void *opaque, hwaddr addr, unsigned size) } else if (addr >= plic->pending_base && /* 1 bit per source */ addr < plic->pending_base + (plic->num_sources >> 3)) { -uint32_t word = (addr - plic->priority_base) >> 2; +uint32_t word = (addr - plic->pending_base) >> 2; if (RISCV_DEBUG_PLIC) { qemu_log("plic: read pending: word=%d value=%d\n", word, plic->pending[word]); -- 2.7.0
[Qemu-devel] [PATCH v1 27/30] elf: Add RISC-V PSABI ELF header defines
Refer to the RISC-V PSABI specification for details: - https://github.com/riscv/riscv-elf-psabi-doc/blob/master/riscv-elf.md Cc: Michael Tokarev <m...@tls.msk.ru> Cc: Laurent Vivier <laur...@vivier.eu> Cc: Richard Henderson <richard.hender...@linaro.org> Cc: Alistair Francis <alistair.fran...@wdc.com> Signed-off-by: Michael Clark <m...@sifive.com> --- include/elf.h | 8 1 file changed, 8 insertions(+) diff --git a/include/elf.h b/include/elf.h index 934dbbd6b3ae..d363ba85a688 100644 --- a/include/elf.h +++ b/include/elf.h @@ -1285,6 +1285,14 @@ typedef struct { #define R_IA64_DTPREL64LSB 0xb7/* @dtprel(sym + add), data8 LSB */ #define R_IA64_LTOFF_DTPREL22 0xba/* @ltoff(@dtprel(s+a)), imm22 */ +/* RISC-V specific definitions. */ +#define EF_RISCV_RVC 0x0001 +#define EF_RISCV_FLOAT_ABI_SINGLE 0x0002 +#define EF_RISCV_FLOAT_ABI_DOUBLE 0x0004 +#define EF_RISCV_FLOAT_ABI_QUAD (0x0006 +#define EF_RISCV_RVE 0x0008 +#define EF_RISCV_TSO 0x0010 + typedef struct elf32_rel { Elf32_Addr r_offset; Elf32_Word r_info; -- 2.7.0
[Qemu-devel] [PATCH v1 10/30] RISC-V: Implement existential predicates for CSRs
CSR predicate functions are added to the CSR table. mstatus.FS and counter enable checks are moved to predicate functions and two new predicates are added to check misa.S for s* CSRs and a new PMP CPU feature for pmp* CSRs. Processors that don't implement S-mode will trap on access to s* CSRs and processors that don't implement PMP will trap on accesses to pmp* CSRs. PMP checks are disabled in riscv_cpu_handle_mmu_fault when the PMP CPU feature is not present. Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Alistair Francis <alistair.fran...@wdc.com> Signed-off-by: Michael Clark <m...@sifive.com> --- target/riscv/cpu.c| 6 ++ target/riscv/cpu.h| 5 +- target/riscv/cpu_helper.c | 3 +- target/riscv/csr.c| 172 ++ 4 files changed, 107 insertions(+), 79 deletions(-) diff --git a/target/riscv/cpu.c b/target/riscv/cpu.c index d630e8fd6c62..9f47f2072c56 100644 --- a/target/riscv/cpu.c +++ b/target/riscv/cpu.c @@ -124,6 +124,7 @@ static void rv32gcsu_priv1_09_1_cpu_init(Object *obj) set_versions(env, USER_VERSION_2_02_0, PRIV_VERSION_1_09_1); set_resetvec(env, DEFAULT_RSTVEC); set_feature(env, RISCV_FEATURE_MMU); +set_feature(env, RISCV_FEATURE_PMP); } static void rv32gcsu_priv1_10_0_cpu_init(Object *obj) @@ -133,6 +134,7 @@ static void rv32gcsu_priv1_10_0_cpu_init(Object *obj) set_versions(env, USER_VERSION_2_02_0, PRIV_VERSION_1_10_0); set_resetvec(env, DEFAULT_RSTVEC); set_feature(env, RISCV_FEATURE_MMU); +set_feature(env, RISCV_FEATURE_PMP); } static void rv32imacu_nommu_cpu_init(Object *obj) @@ -141,6 +143,7 @@ static void rv32imacu_nommu_cpu_init(Object *obj) set_misa(env, RV32 | RVI | RVM | RVA | RVC | RVU); set_versions(env, USER_VERSION_2_02_0, PRIV_VERSION_1_10_0); set_resetvec(env, DEFAULT_RSTVEC); +set_feature(env, RISCV_FEATURE_PMP); } #elif defined(TARGET_RISCV64) @@ -152,6 +155,7 @@ static void rv64gcsu_priv1_09_1_cpu_init(Object *obj) set_versions(env, USER_VERSION_2_02_0, PRIV_VERSION_1_09_1); set_resetvec(env, DEFAULT_RSTVEC); set_feature(env, RISCV_FEATURE_MMU); +set_feature(env, RISCV_FEATURE_PMP); } static void rv64gcsu_priv1_10_0_cpu_init(Object *obj) @@ -161,6 +165,7 @@ static void rv64gcsu_priv1_10_0_cpu_init(Object *obj) set_versions(env, USER_VERSION_2_02_0, PRIV_VERSION_1_10_0); set_resetvec(env, DEFAULT_RSTVEC); set_feature(env, RISCV_FEATURE_MMU); +set_feature(env, RISCV_FEATURE_PMP); } static void rv64imacu_nommu_cpu_init(Object *obj) @@ -169,6 +174,7 @@ static void rv64imacu_nommu_cpu_init(Object *obj) set_misa(env, RV64 | RVI | RVM | RVA | RVC | RVU); set_versions(env, USER_VERSION_2_02_0, PRIV_VERSION_1_10_0); set_resetvec(env, DEFAULT_RSTVEC); +set_feature(env, RISCV_FEATURE_PMP); } #endif diff --git a/target/riscv/cpu.h b/target/riscv/cpu.h index 02c60c45631b..3a3d91447736 100644 --- a/target/riscv/cpu.h +++ b/target/riscv/cpu.h @@ -83,9 +83,10 @@ /* S extension denotes that Supervisor mode exists, however it is possible to have a core that support S mode but does not have an MMU and there is currently no bit in misa to indicate whether an MMU exists or not - so a cpu features bitfield is required */ + so a cpu features bitfield is required, likewise for optional PMP support */ enum { -RISCV_FEATURE_MMU +RISCV_FEATURE_MMU, +RISCV_FEATURE_PMP }; #define USER_VERSION_2_02_0 0x00020200 diff --git a/target/riscv/cpu_helper.c b/target/riscv/cpu_helper.c index 1f523861b7cb..bc15e19022cc 100644 --- a/target/riscv/cpu_helper.c +++ b/target/riscv/cpu_helper.c @@ -404,7 +404,8 @@ int riscv_cpu_handle_mmu_fault(CPUState *cs, vaddr address, int size, qemu_log_mask(CPU_LOG_MMU, "%s address=%" VADDR_PRIx " ret %d physical " TARGET_FMT_plx " prot %d\n", __func__, address, ret, pa, prot); -if (!pmp_hart_has_privs(env, pa, TARGET_PAGE_SIZE, 1 << rw)) { +if (riscv_feature(env, RISCV_FEATURE_PMP) && +!pmp_hart_has_privs(env, pa, TARGET_PAGE_SIZE, 1 << rw)) { ret = TRANSLATE_FAIL; } if (ret == TRANSLATE_SUCCESS) { diff --git a/target/riscv/csr.c b/target/riscv/csr.c index 631a5ff9f7d8..0e6c0c365154 100644 --- a/target/riscv/csr.c +++ b/target/riscv/csr.c @@ -26,6 +26,7 @@ /* Control and Status Register function table forward declaration */ +typedef int (*riscv_csr_predicate_fn)(CPURISCVState *env, int csrno); typedef int (*riscv_csr_read_fn)(CPURISCVState *env, int csrno, target_ulong *ret_value); typedef int (*riscv_csr_write_fn)(CPURISCVState *env, int csrno, @@ -34,6 +35,7 @@ typedef int (*riscv_csr_op_fn)(CPURISCVState *env, int csrno, target_ulong *ret_value, target_ulong new_value, tar
[Qemu-devel] [PATCH v1 23/30] RISC-V: Fix CLINT timecmp low 32-bit writes
A missing shift made updates to the low order bits of timecmp erroneously copy the old low order bits into the high order bits of the 64-bit timecmp register. Add the missing shift and rename timecmp local variables to timecmp_hi and timecmp_lo. This bug didn't show up as the low order bits are usually written first followed by the high order bits meaning the high order bits contained an invalid value between the timecmp_lo and timecmp_hi update. Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Cc: Alistair Francis <alistair.fran...@wdc.com> Co-Authored-by: Johannes Haring <johannes.har...@gmx.net> Signed-off-by: Michael Clark <m...@sifive.com> --- hw/riscv/sifive_clint.c | 8 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/hw/riscv/sifive_clint.c b/hw/riscv/sifive_clint.c index 0d2fd52487e6..d4c159e93736 100644 --- a/hw/riscv/sifive_clint.c +++ b/hw/riscv/sifive_clint.c @@ -146,15 +146,15 @@ static void sifive_clint_write(void *opaque, hwaddr addr, uint64_t value, error_report("clint: invalid timecmp hartid: %zu", hartid); } else if ((addr & 0x7) == 0) { /* timecmp_lo */ -uint64_t timecmp = env->timecmp; +uint64_t timecmp_hi = env->timecmp >> 32; sifive_clint_write_timecmp(RISCV_CPU(cpu), -timecmp << 32 | (value & 0x)); +timecmp_hi << 32 | (value & 0x)); return; } else if ((addr & 0x7) == 4) { /* timecmp_hi */ -uint64_t timecmp = env->timecmp; +uint64_t timecmp_lo = env->timecmp; sifive_clint_write_timecmp(RISCV_CPU(cpu), -value << 32 | (timecmp & 0x)); +value << 32 | (timecmp_lo & 0x)); } else { error_report("clint: invalid timecmp write: %08x", (uint32_t)addr); } -- 2.7.0
[Qemu-devel] [PATCH v1 25/30] RISC-V: Enable second UART on sifive_e and sifive_u
Previously the second UARTs on the sifive_e and sifive_u machines where disabled due to check-qtest-riscv32 and check-qtest-riscv64 failures. Recent changes in the QEMU core serial code have resolved these failures so the second UARTs can be instantiated. Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Cc: Alistair Francis <alistair.fran...@wdc.com> Signed-off-by: Michael Clark <m...@sifive.com> --- hw/riscv/sifive_e.c | 4 ++-- hw/riscv/sifive_u.c | 4 ++-- 2 files changed, 4 insertions(+), 4 deletions(-) diff --git a/hw/riscv/sifive_e.c b/hw/riscv/sifive_e.c index e4ecb7aa4bb6..159209199537 100644 --- a/hw/riscv/sifive_e.c +++ b/hw/riscv/sifive_e.c @@ -155,8 +155,8 @@ static void riscv_sifive_e_init(MachineState *machine) memmap[SIFIVE_E_QSPI0].base, memmap[SIFIVE_E_QSPI0].size); sifive_mmio_emulate(sys_mem, "riscv.sifive.e.pwm0", memmap[SIFIVE_E_PWM0].base, memmap[SIFIVE_E_PWM0].size); -/* sifive_uart_create(sys_mem, memmap[SIFIVE_E_UART1].base, -serial_hd(1), SIFIVE_PLIC(s->plic)->irqs[SIFIVE_E_UART1_IRQ]); */ +sifive_uart_create(sys_mem, memmap[SIFIVE_E_UART1].base, +serial_hd(1), SIFIVE_PLIC(s->plic)->irqs[SIFIVE_E_UART1_IRQ]); sifive_mmio_emulate(sys_mem, "riscv.sifive.e.qspi1", memmap[SIFIVE_E_QSPI1].base, memmap[SIFIVE_E_QSPI1].size); sifive_mmio_emulate(sys_mem, "riscv.sifive.e.pwm1", diff --git a/hw/riscv/sifive_u.c b/hw/riscv/sifive_u.c index c05dcbba955e..326b0f434cff 100644 --- a/hw/riscv/sifive_u.c +++ b/hw/riscv/sifive_u.c @@ -299,8 +299,8 @@ static void riscv_sifive_u_init(MachineState *machine) memmap[SIFIVE_U_PLIC].size); sifive_uart_create(system_memory, memmap[SIFIVE_U_UART0].base, serial_hd(0), SIFIVE_PLIC(s->plic)->irqs[SIFIVE_U_UART0_IRQ]); -/* sifive_uart_create(system_memory, memmap[SIFIVE_U_UART1].base, -serial_hd(1), SIFIVE_PLIC(s->plic)->irqs[SIFIVE_U_UART1_IRQ]); */ +sifive_uart_create(system_memory, memmap[SIFIVE_U_UART1].base, +serial_hd(1), SIFIVE_PLIC(s->plic)->irqs[SIFIVE_U_UART1_IRQ]); sifive_clint_create(memmap[SIFIVE_U_CLINT].base, memmap[SIFIVE_U_CLINT].size, smp_cpus, SIFIVE_SIP_BASE, SIFIVE_TIMECMP_BASE, SIFIVE_TIME_BASE); -- 2.7.0
[Qemu-devel] [PATCH v1 09/30] RISC-V: Implement atomic mip/sip CSR updates
Use the new CSR read/modify/write interface to implement atomic updates to mip/sip. Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Alistair Francis <alistair.fran...@wdc.com> Signed-off-by: Michael Clark <m...@sifive.com> --- target/riscv/csr.c | 56 +++--- 1 file changed, 28 insertions(+), 28 deletions(-) diff --git a/target/riscv/csr.c b/target/riscv/csr.c index e08f3523d854..631a5ff9f7d8 100644 --- a/target/riscv/csr.c +++ b/target/riscv/csr.c @@ -493,25 +493,31 @@ static int write_mbadaddr(CPURISCVState *env, int csrno, target_ulong val) return 0; } -static int read_mip(CPURISCVState *env, int csrno, target_ulong *val) -{ -*val = atomic_read(>mip); -return 0; -} - -static int write_mip(CPURISCVState *env, int csrno, target_ulong val) +static int rmw_mip(CPURISCVState *env, int csrno, target_ulong *ret_value, + target_ulong new_value, target_ulong write_mask) { RISCVCPU *cpu = riscv_env_get_cpu(env); +target_ulong mask = write_mask & delegable_ints; +uint32_t old_mip; + +/* We can't allow the supervisor to control SEIP as this would allow the + * supervisor to clear a pending external interrupt which will result in + * lost a interrupt in the case a PLIC is attached. The SEIP bit must be + * hardware controlled when a PLIC is attached. This should be an option + * for CPUs with software-delegated Supervisor External Interrupts. */ +mask &= ~MIP_SEIP; + +if (mask) { +qemu_mutex_lock_iothread(); +old_mip = riscv_cpu_update_mip(cpu, mask, (new_value & mask)); +qemu_mutex_unlock_iothread(); +} else { +old_mip = atomic_read(>mip); +} -/* - * csrs, csrc on mip.SEIP is not decomposable into separate read and - * write steps, so a different implementation is needed - */ - -qemu_mutex_lock_iothread(); -riscv_cpu_update_mip(cpu, MIP_SSIP | MIP_STIP, - (val & (MIP_SSIP | MIP_STIP))); -qemu_mutex_unlock_iothread(); +if (ret_value) { +*ret_value = old_mip; +} return 0; } @@ -631,17 +637,11 @@ static int write_sbadaddr(CPURISCVState *env, int csrno, target_ulong val) return 0; } -static int read_sip(CPURISCVState *env, int csrno, target_ulong *val) -{ -*val = atomic_read(>mip) & env->mideleg; -return 0; -} - -static int write_sip(CPURISCVState *env, int csrno, target_ulong val) +static int rmw_sip(CPURISCVState *env, int csrno, target_ulong *ret_value, + target_ulong new_value, target_ulong write_mask) { -target_ulong newval = (atomic_read(>mip) & ~env->mideleg) - | (val & env->mideleg); -return write_mip(env, CSR_MIP, newval); +return rmw_mip(env, CSR_MSTATUS, ret_value, new_value, + write_mask & env->mideleg); } /* Supervisor Protection and Translation */ @@ -823,7 +823,7 @@ static const riscv_csr_operations csr_ops[0xfff] = { [CSR_MEPC] ={ read_mepc,write_mepc}, [CSR_MCAUSE] = { read_mcause, write_mcause }, [CSR_MBADADDR] ={ read_mbadaddr,write_mbadaddr}, -[CSR_MIP] = { read_mip, write_mip }, +[CSR_MIP] = { NULL, NULL, rmw_mip }, /* Supervisor Trap Setup */ [CSR_SSTATUS] = { read_sstatus, write_sstatus }, @@ -836,7 +836,7 @@ static const riscv_csr_operations csr_ops[0xfff] = { [CSR_SEPC] ={ read_sepc,write_sepc}, [CSR_SCAUSE] = { read_scause, write_scause }, [CSR_SBADADDR] ={ read_sbadaddr,write_sbadaddr}, -[CSR_SIP] = { read_sip, write_sip }, +[CSR_SIP] = { NULL, NULL, rmw_sip }, /* Supervisor Protection and Translation */ [CSR_SATP] ={ read_satp,write_satp}, -- 2.7.0
[Qemu-devel] [PATCH v1 21/30] RISC-V: Add misa.MAFD checks to translate
Add misa checks for M, A, F and D extensions and if they are not present generate illegal instructions. This improves emulation accurary for harts with a limited set of extensions. Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Cc: Alistair Francis <alistair.fran...@wdc.com> Cc: Emilio G. Cota <c...@braap.org> Signed-off-by: Michael Clark <m...@sifive.com> --- target/riscv/translate.c | 158 +++ 1 file changed, 158 insertions(+) diff --git a/target/riscv/translate.c b/target/riscv/translate.c index fd21b133a5a4..e488101ff56d 100644 --- a/target/riscv/translate.c +++ b/target/riscv/translate.c @@ -290,24 +290,42 @@ static void gen_arith(DisasContext *ctx, uint32_t opc, int rd, int rs1, tcg_gen_and_tl(source1, source1, source2); break; CASE_OP_32_64(OPC_RISC_MUL): +if (!has_ext(ctx, RVM)) { +goto do_illegal; +} tcg_gen_mul_tl(source1, source1, source2); break; case OPC_RISC_MULH: +if (!has_ext(ctx, RVM)) { +goto do_illegal; +} tcg_gen_muls2_tl(source2, source1, source1, source2); break; case OPC_RISC_MULHSU: +if (!has_ext(ctx, RVM)) { +goto do_illegal; +} gen_mulhsu(source1, source1, source2); break; case OPC_RISC_MULHU: +if (!has_ext(ctx, RVM)) { +goto do_illegal; +} tcg_gen_mulu2_tl(source2, source1, source1, source2); break; #if defined(TARGET_RISCV64) case OPC_RISC_DIVW: +if (!has_ext(ctx, RVM)) { +goto do_illegal; +} tcg_gen_ext32s_tl(source1, source1); tcg_gen_ext32s_tl(source2, source2); /* fall through to DIV */ #endif case OPC_RISC_DIV: +if (!has_ext(ctx, RVM)) { +goto do_illegal; +} /* Handle by altering args to tcg_gen_div to produce req'd results: * For overflow: want source1 in source1 and 1 in source2 * For div by zero: want -1 in source1 and 1 in source2 -> -1 result */ @@ -339,11 +357,17 @@ static void gen_arith(DisasContext *ctx, uint32_t opc, int rd, int rs1, break; #if defined(TARGET_RISCV64) case OPC_RISC_DIVUW: +if (!has_ext(ctx, RVM)) { +goto do_illegal; +} tcg_gen_ext32u_tl(source1, source1); tcg_gen_ext32u_tl(source2, source2); /* fall through to DIVU */ #endif case OPC_RISC_DIVU: +if (!has_ext(ctx, RVM)) { +goto do_illegal; +} cond1 = tcg_temp_new(); zeroreg = tcg_const_tl(0); resultopt1 = tcg_temp_new(); @@ -363,11 +387,17 @@ static void gen_arith(DisasContext *ctx, uint32_t opc, int rd, int rs1, break; #if defined(TARGET_RISCV64) case OPC_RISC_REMW: +if (!has_ext(ctx, RVM)) { +goto do_illegal; +} tcg_gen_ext32s_tl(source1, source1); tcg_gen_ext32s_tl(source2, source2); /* fall through to REM */ #endif case OPC_RISC_REM: +if (!has_ext(ctx, RVM)) { +goto do_illegal; +} cond1 = tcg_temp_new(); cond2 = tcg_temp_new(); zeroreg = tcg_const_tl(0); @@ -395,11 +425,17 @@ static void gen_arith(DisasContext *ctx, uint32_t opc, int rd, int rs1, break; #if defined(TARGET_RISCV64) case OPC_RISC_REMUW: +if (!has_ext(ctx, RVM)) { +goto do_illegal; +} tcg_gen_ext32u_tl(source1, source1); tcg_gen_ext32u_tl(source2, source2); /* fall through to REMU */ #endif case OPC_RISC_REMU: +if (!has_ext(ctx, RVM)) { +goto do_illegal; +} cond1 = tcg_temp_new(); zeroreg = tcg_const_tl(0); resultopt1 = tcg_temp_new(); @@ -417,6 +453,7 @@ static void gen_arith(DisasContext *ctx, uint32_t opc, int rd, int rs1, tcg_temp_free(zeroreg); tcg_temp_free(resultopt1); break; +do_illegal: default: gen_exception_illegal(ctx); return; @@ -697,13 +734,20 @@ static void gen_fp_load(DisasContext *ctx, uint32_t opc, int rd, switch (opc) { case OPC_RISC_FLW: +if (!has_ext(ctx, RVF)) { +goto do_illegal; +} tcg_gen_qemu_ld_i64(cpu_fpr[rd], t0, ctx->mem_idx, MO_TEUL); /* RISC-V requires NaN-boxing of narrower width floating point values */ tcg_gen_ori_i64(cpu_fpr[rd], cpu_fpr[rd], 0xULL); break; case OPC_RISC_FLD: +if (!has_ext(ctx, RVD)) { +goto do_illegal; +} tcg_gen_qemu_ld_i64(cpu_fpr[rd], t0, ctx->mem_idx, MO_TEQ); break; +do_illegal: default: gen_exception_illegal(ctx); break; @@ -729,11 +77
[Qemu-devel] [PATCH v1 18/30] RISC-V: Add missing free for plic_hart_config
Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Cc: Alistair Francis <alistair.fran...@wdc.com> Signed-off-by: Michael Clark <m...@sifive.com> --- hw/riscv/virt.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/hw/riscv/virt.c b/hw/riscv/virt.c index ad03113e0f72..321fa6e8122a 100644 --- a/hw/riscv/virt.c +++ b/hw/riscv/virt.c @@ -385,6 +385,8 @@ static void riscv_virt_board_init(MachineState *machine) serial_mm_init(system_memory, memmap[VIRT_UART0].base, 0, SIFIVE_PLIC(s->plic)->irqs[UART0_IRQ], 399193, serial_hd(0), DEVICE_LITTLE_ENDIAN); + +g_free(plic_hart_config); } static void riscv_virt_board_machine_init(MachineClass *mc) -- 2.7.0
[Qemu-devel] [PATCH v1 15/30] RISC-V: Add hartid and \n to interrupt logging
Add carriage return that was erroneously removed when converting to qemu_log. Change hard coded core number to the actual hartid. Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Alistair Francis <alistair.fran...@wdc.com> Signed-off-by: Michael Clark <m...@sifive.com> --- target/riscv/cpu_helper.c | 18 ++ 1 file changed, 10 insertions(+), 8 deletions(-) diff --git a/target/riscv/cpu_helper.c b/target/riscv/cpu_helper.c index bc15e19022cc..69592c037042 100644 --- a/target/riscv/cpu_helper.c +++ b/target/riscv/cpu_helper.c @@ -446,11 +446,13 @@ void riscv_cpu_do_interrupt(CPUState *cs) if (RISCV_DEBUG_INTERRUPT) { int log_cause = cs->exception_index & RISCV_EXCP_INT_MASK; if (cs->exception_index & RISCV_EXCP_INT_FLAG) { -qemu_log_mask(LOG_TRACE, "core 0: trap %s, epc 0x" TARGET_FMT_lx, -riscv_intr_names[log_cause], env->pc); +qemu_log_mask(LOG_TRACE, "core " +TARGET_FMT_ld ": trap %s, epc 0x" TARGET_FMT_lx "\n", +env->mhartid, riscv_intr_names[log_cause], env->pc); } else { -qemu_log_mask(LOG_TRACE, "core 0: intr %s, epc 0x" TARGET_FMT_lx, -riscv_excp_names[log_cause], env->pc); +qemu_log_mask(LOG_TRACE, "core " +TARGET_FMT_ld ": intr %s, epc 0x" TARGET_FMT_lx "\n", +env->mhartid, riscv_excp_names[log_cause], env->pc); } } @@ -512,8 +514,8 @@ void riscv_cpu_do_interrupt(CPUState *cs) if (hasbadaddr) { if (RISCV_DEBUG_INTERRUPT) { -qemu_log_mask(LOG_TRACE, "core " TARGET_FMT_ld -": badaddr 0x" TARGET_FMT_lx, env->mhartid, env->badaddr); +qemu_log_mask(LOG_TRACE, "core " TARGET_FMT_ld ": badaddr 0x" +TARGET_FMT_lx "\n", env->mhartid, env->badaddr); } env->sbadaddr = env->badaddr; } else { @@ -537,8 +539,8 @@ void riscv_cpu_do_interrupt(CPUState *cs) if (hasbadaddr) { if (RISCV_DEBUG_INTERRUPT) { -qemu_log_mask(LOG_TRACE, "core " TARGET_FMT_ld -": badaddr 0x" TARGET_FMT_lx, env->mhartid, env->badaddr); +qemu_log_mask(LOG_TRACE, "core " TARGET_FMT_ld ": badaddr 0x" +TARGET_FMT_lx "\n", env->mhartid, env->badaddr); } env->mbadaddr = env->badaddr; } else { -- 2.7.0
[Qemu-devel] [PATCH v1 06/30] RISC-V: Move non-ops from op_helper to cpu_helper
This patch makes op_helper.c contain only instruction operation helpers used by translate.c and moves any unrelated cpu helpers into cpu_helper.c. No logic is changed by this patch. Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Alistair Francis <alistair.fran...@wdc.com> Signed-off-by: Michael Clark <m...@sifive.com> Reviewed-by: Alistair Francis <alistair.fran...@wdc.com> --- target/riscv/Makefile.objs | 2 +- target/riscv/{helper.c => cpu_helper.c} | 35 - target/riscv/op_helper.c| 34 3 files changed, 35 insertions(+), 36 deletions(-) rename target/riscv/{helper.c => cpu_helper.c} (95%) diff --git a/target/riscv/Makefile.objs b/target/riscv/Makefile.objs index abd0a7cde333..fcc5d34c1f2e 100644 --- a/target/riscv/Makefile.objs +++ b/target/riscv/Makefile.objs @@ -1 +1 @@ -obj-y += translate.o op_helper.o helper.o cpu.o fpu_helper.o gdbstub.o pmp.o +obj-y += translate.o op_helper.o cpu_helper.o cpu.o fpu_helper.o gdbstub.o pmp.o diff --git a/target/riscv/helper.c b/target/riscv/cpu_helper.c similarity index 95% rename from target/riscv/helper.c rename to target/riscv/cpu_helper.c index 47d116e9c13f..6c886e99055a 100644 --- a/target/riscv/helper.c +++ b/target/riscv/cpu_helper.c @@ -1,5 +1,5 @@ /* - * RISC-V emulation helpers for qemu. + * RISC-V CPU helpers for qemu. * * Copyright (c) 2016-2017 Sagar Karandikar, sag...@eecs.berkeley.edu * Copyright (c) 2017-2018 SiFive, Inc. @@ -72,6 +72,39 @@ bool riscv_cpu_exec_interrupt(CPUState *cs, int interrupt_request) #if !defined(CONFIG_USER_ONLY) +/* iothread_mutex must be held */ +uint32_t riscv_cpu_update_mip(RISCVCPU *cpu, uint32_t mask, uint32_t value) +{ +CPURISCVState *env = >env; +uint32_t old, new, cmp = atomic_read(>mip); + +do { +old = cmp; +new = (old & ~mask) | (value & mask); +cmp = atomic_cmpxchg(>mip, old, new); +} while (old != cmp); + +if (new && !old) { +cpu_interrupt(CPU(cpu), CPU_INTERRUPT_HARD); +} else if (!new && old) { +cpu_reset_interrupt(CPU(cpu), CPU_INTERRUPT_HARD); +} + +return old; +} + +void riscv_set_mode(CPURISCVState *env, target_ulong newpriv) +{ +if (newpriv > PRV_M) { +g_assert_not_reached(); +} +if (newpriv == PRV_H) { +newpriv = PRV_U; +} +/* tlb_flush is unnecessary as mode is contained in mmu_idx */ +env->priv = newpriv; +} + /* get_physical_address - get the physical address for this virtual address * * Do a page table walk to obtain the physical address corresponding to a diff --git a/target/riscv/op_helper.c b/target/riscv/op_helper.c index 5a02795bf931..2b9dd9da6486 100644 --- a/target/riscv/op_helper.c +++ b/target/riscv/op_helper.c @@ -652,39 +652,6 @@ target_ulong helper_csrrc(CPURISCVState *env, target_ulong src, #ifndef CONFIG_USER_ONLY -/* iothread_mutex must be held */ -uint32_t riscv_cpu_update_mip(RISCVCPU *cpu, uint32_t mask, uint32_t value) -{ -CPURISCVState *env = >env; -uint32_t old, new, cmp = atomic_read(>mip); - -do { -old = cmp; -new = (old & ~mask) | (value & mask); -cmp = atomic_cmpxchg(>mip, old, new); -} while (old != cmp); - -if (new && !old) { -cpu_interrupt(CPU(cpu), CPU_INTERRUPT_HARD); -} else if (!new && old) { -cpu_reset_interrupt(CPU(cpu), CPU_INTERRUPT_HARD); -} - -return old; -} - -void riscv_set_mode(CPURISCVState *env, target_ulong newpriv) -{ -if (newpriv > PRV_M) { -g_assert_not_reached(); -} -if (newpriv == PRV_H) { -newpriv = PRV_U; -} -/* tlb_flush is unnecessary as mode is contained in mmu_idx */ -env->priv = newpriv; -} - target_ulong helper_sret(CPURISCVState *env, target_ulong cpu_pc_deb) { if (!(env->priv >= PRV_S)) { @@ -735,7 +702,6 @@ target_ulong helper_mret(CPURISCVState *env, target_ulong cpu_pc_deb) return retpc; } - void helper_wfi(CPURISCVState *env) { CPUState *cs = CPU(riscv_env_get_cpu(env)); -- 2.7.0
[Qemu-devel] [PATCH v1 17/30] RISC-V: Replace __builtin_popcount with ctpop8 in PLIC
The mode variable only uses the lower 4-bits (M,H,S,U) so replace the GCC specific __builtin_popcount with ctpop8. Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Cc: Alistair Francis <alistair.fran...@wdc.com> Signed-off-by: Michael Clark <m...@sifive.com> --- hw/riscv/sifive_plic.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/hw/riscv/sifive_plic.c b/hw/riscv/sifive_plic.c index b267ff88902d..dc6f4924e282 100644 --- a/hw/riscv/sifive_plic.c +++ b/hw/riscv/sifive_plic.c @@ -383,7 +383,7 @@ static void parse_hart_config(SiFivePLICState *plic) p = plic->hart_config; while ((c = *p++)) { if (c == ',') { -addrid += __builtin_popcount(modes); +addrid += ctpop8(modes); modes = 0; hartid++; } else { @@ -397,7 +397,7 @@ static void parse_hart_config(SiFivePLICState *plic) } } if (modes) { -addrid += __builtin_popcount(modes); +addrid += ctpop8(modes); } hartid++; -- 2.7.0
[Qemu-devel] [PATCH v1 14/30] RISC-V: Add public API for the CSR dispatch table
This allows hardware and/or derived cpu instances to override or implement new CSR operations. Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Alistair Francis <alistair.fran...@wdc.com> Signed-off-by: Michael Clark <m...@sifive.com> --- target/riscv/cpu.h | 18 ++ target/riscv/csr.c | 35 ++- 2 files changed, 36 insertions(+), 17 deletions(-) diff --git a/target/riscv/cpu.h b/target/riscv/cpu.h index 242a8fcbe180..1ade90d23bbc 100644 --- a/target/riscv/cpu.h +++ b/target/riscv/cpu.h @@ -307,6 +307,24 @@ static inline target_ulong csr_read_helper(CPURISCVState *env, int csrno) return val; } +typedef int (*riscv_csr_predicate_fn)(CPURISCVState *env, int csrno); +typedef int (*riscv_csr_read_fn)(CPURISCVState *env, int csrno, +target_ulong *ret_value); +typedef int (*riscv_csr_write_fn)(CPURISCVState *env, int csrno, +target_ulong new_value); +typedef int (*riscv_csr_op_fn)(CPURISCVState *env, int csrno, +target_ulong *ret_value, target_ulong new_value, target_ulong write_mask); + +typedef struct { +riscv_csr_predicate_fn predicate; +riscv_csr_read_fn read; +riscv_csr_write_fn write; +riscv_csr_op_fn op; +} riscv_csr_operations; + +void riscv_get_csr_ops(int csrno, riscv_csr_operations *ops); +void riscv_set_csr_ops(int csrno, riscv_csr_operations *ops); + #include "exec/cpu-all.h" #endif /* RISCV_CPU_H */ diff --git a/target/riscv/csr.c b/target/riscv/csr.c index 509215327243..0f886e04b130 100644 --- a/target/riscv/csr.c +++ b/target/riscv/csr.c @@ -23,28 +23,29 @@ #include "qemu/main-loop.h" #include "exec/exec-all.h" +/* CSR function table */ -/* Control and Status Register function table forward declaration */ +static riscv_csr_operations csr_ops[]; -typedef int (*riscv_csr_predicate_fn)(CPURISCVState *env, int csrno); -typedef int (*riscv_csr_read_fn)(CPURISCVState *env, int csrno, -target_ulong *ret_value); -typedef int (*riscv_csr_write_fn)(CPURISCVState *env, int csrno, -target_ulong new_value); -typedef int (*riscv_csr_op_fn)(CPURISCVState *env, int csrno, -target_ulong *ret_value, target_ulong new_value, target_ulong write_mask); +/* CSR function table constants */ -typedef struct { -riscv_csr_predicate_fn predicate; -riscv_csr_read_fn read; -riscv_csr_write_fn write; -riscv_csr_op_fn op; -} riscv_csr_operations; +enum { +CSR_TABLE_SIZE = 0xfff +}; + +/* CSR function table public API */ -static const riscv_csr_operations csr_ops[]; +void riscv_get_csr_ops(int csrno, riscv_csr_operations *ops) +{ +*ops = csr_ops[csrno & CSR_TABLE_SIZE]; +} +void riscv_set_csr_ops(int csrno, riscv_csr_operations *ops) +{ +csr_ops[csrno & CSR_TABLE_SIZE] = *ops; +} -/* Predicates */ +/* CSR function table predicates (private) */ static int fs(CPURISCVState *env, int csrno) { @@ -784,7 +785,7 @@ int riscv_csrrw(CPURISCVState *env, int csrno, target_ulong *ret_value, /* Control and Status Register function table */ -static const riscv_csr_operations csr_ops[0xfff] = { +static riscv_csr_operations csr_ops[CSR_TABLE_SIZE] = { /* User Floating-Point CSRs */ [CSR_FFLAGS] = { fs, read_fflags, write_fflags }, [CSR_FRM] = { fs, read_frm, write_frm }, -- 2.7.0
[Qemu-devel] [PATCH v1 05/30] RISC-V: Allow setting and clearing multiple irqs
Change the API of riscv_set_local_interrupt to take a write mask and value to allow setting and clearing of multiple local interrupts atomically in a single call. Rename the new function to riscv_cpu_update_mip. Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Alistair Francis <alistair.fran...@wdc.com> Signed-off-by: Michael Clark <m...@sifive.com> --- hw/riscv/sifive_clint.c | 8 hw/riscv/sifive_plic.c | 4 ++-- target/riscv/cpu.h | 22 +- target/riscv/op_helper.c | 24 +++- 4 files changed, 34 insertions(+), 24 deletions(-) diff --git a/hw/riscv/sifive_clint.c b/hw/riscv/sifive_clint.c index 7cc606e06546..0d2fd52487e6 100644 --- a/hw/riscv/sifive_clint.c +++ b/hw/riscv/sifive_clint.c @@ -47,12 +47,12 @@ static void sifive_clint_write_timecmp(RISCVCPU *cpu, uint64_t value) if (cpu->env.timecmp <= rtc_r) { /* if we're setting an MTIMECMP value in the "past", immediately raise the timer interrupt */ -riscv_set_local_interrupt(cpu, MIP_MTIP, 1); +riscv_cpu_update_mip(cpu, MIP_MTIP, BOOL_TO_MASK(1)); return; } /* otherwise, set up the future timer interrupt */ -riscv_set_local_interrupt(cpu, MIP_MTIP, 0); +riscv_cpu_update_mip(cpu, MIP_MTIP, BOOL_TO_MASK(0)); diff = cpu->env.timecmp - rtc_r; /* back to ns (note args switched in muldiv64) */ next = qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL) + @@ -67,7 +67,7 @@ static void sifive_clint_write_timecmp(RISCVCPU *cpu, uint64_t value) static void sifive_clint_timer_cb(void *opaque) { RISCVCPU *cpu = opaque; -riscv_set_local_interrupt(cpu, MIP_MTIP, 1); +riscv_cpu_update_mip(cpu, MIP_MTIP, BOOL_TO_MASK(1)); } /* CPU wants to read rtc or timecmp register */ @@ -132,7 +132,7 @@ static void sifive_clint_write(void *opaque, hwaddr addr, uint64_t value, if (!env) { error_report("clint: invalid timecmp hartid: %zu", hartid); } else if ((addr & 0x3) == 0) { -riscv_set_local_interrupt(RISCV_CPU(cpu), MIP_MSIP, value != 0); +riscv_cpu_update_mip(RISCV_CPU(cpu), MIP_MSIP, BOOL_TO_MASK(value)); } else { error_report("clint: invalid sip write: %08x", (uint32_t)addr); } diff --git a/hw/riscv/sifive_plic.c b/hw/riscv/sifive_plic.c index 1af23c76e603..b267ff88902d 100644 --- a/hw/riscv/sifive_plic.c +++ b/hw/riscv/sifive_plic.c @@ -142,10 +142,10 @@ static void sifive_plic_update(SiFivePLICState *plic) int level = sifive_plic_irqs_pending(plic, addrid); switch (mode) { case PLICMode_M: -riscv_set_local_interrupt(RISCV_CPU(cpu), MIP_MEIP, level); +riscv_cpu_update_mip(RISCV_CPU(cpu), MIP_MEIP, BOOL_TO_MASK(level)); break; case PLICMode_S: -riscv_set_local_interrupt(RISCV_CPU(cpu), MIP_SEIP, level); +riscv_cpu_update_mip(RISCV_CPU(cpu), MIP_SEIP, BOOL_TO_MASK(level)); break; default: break; diff --git a/target/riscv/cpu.h b/target/riscv/cpu.h index e0608e6d5f08..c5d485769cde 100644 --- a/target/riscv/cpu.h +++ b/target/riscv/cpu.h @@ -126,13 +126,18 @@ struct CPURISCVState { target_ulong mhartid; target_ulong mstatus; + /* * CAUTION! Unlike the rest of this struct, mip is accessed asynchonously - * by I/O threads and other vCPUs, so hold the iothread mutex before - * operating on it. CPU_INTERRUPT_HARD should be in effect iff this is - * non-zero. Use riscv_cpu_set_local_interrupt. + * by I/O threads. It should be read with atomic_read. It should be updated + * using riscv_cpu_update_mip with the iothread mutex held. The iothread + * mutex must be held because mip must be consistent with the CPU inturrept + * state. riscv_cpu_update_mip calls cpu_interrupt or cpu_reset_interrupt + * wuth the invariant that CPU_INTERRUPT_HARD is set iff mip is non-zero. + * mip is 32-bits to allow atomic_read on 32-bit hosts. */ -uint32_t mip;/* allow atomic_read for >= 32-bit hosts */ +uint32_t mip; + target_ulong mie; target_ulong mideleg; @@ -247,7 +252,6 @@ void riscv_cpu_do_unaligned_access(CPUState *cs, vaddr addr, uintptr_t retaddr); int riscv_cpu_handle_mmu_fault(CPUState *cpu, vaddr address, int size, int rw, int mmu_idx); - char *riscv_isa_string(RISCVCPU *cpu); void riscv_cpu_list(FILE *f, fprintf_function cpu_fprintf); @@ -256,6 +260,10 @@ void riscv_cpu_list(FILE *f, fprintf_function cpu_fprintf); #define cpu_list riscv_cpu_list #define cpu_mmu_index riscv_cpu_mmu_index +#ifndef CONFIG_USER_ONLY +uint32_t riscv_cpu_update_mip(RISCVCPU *cpu, uint32_t mask, uint32_t
[Qemu-devel] [PATCH v1 16/30] RISC-V: Use riscv prefix consistently on cpu helpers
* Add riscv prefix to raise_exception function * Add riscv prefix to CSR read/write functions * Add riscv prefix to signal handler function * Add riscv prefix to get fflags function * Remove redundant declaration of riscv_cpu_init and rename cpu_riscv_init to riscv_cpu_init * rename riscv_set_mode to riscv_cpu_set_mode Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Alistair Francis <alistair.fran...@wdc.com> Signed-off-by: Michael Clark <m...@sifive.com> --- linux-user/riscv/signal.c | 4 ++-- target/riscv/cpu.h| 21 ++--- target/riscv/cpu_helper.c | 10 +- target/riscv/csr.c| 8 target/riscv/fpu_helper.c | 6 +++--- target/riscv/op_helper.c | 28 ++-- 6 files changed, 38 insertions(+), 39 deletions(-) diff --git a/linux-user/riscv/signal.c b/linux-user/riscv/signal.c index ef599e319a10..f03bc7212b49 100644 --- a/linux-user/riscv/signal.c +++ b/linux-user/riscv/signal.c @@ -84,7 +84,7 @@ static void setup_sigcontext(struct target_sigcontext *sc, CPURISCVState *env) __put_user(env->fpr[i], >fpr[i]); } -uint32_t fcsr = csr_read_helper(env, CSR_FCSR); /*riscv_get_fcsr(env);*/ +uint32_t fcsr = riscv_csr_read(env, CSR_FCSR); __put_user(fcsr, >fcsr); } @@ -160,7 +160,7 @@ static void restore_sigcontext(CPURISCVState *env, struct target_sigcontext *sc) uint32_t fcsr; __get_user(fcsr, >fcsr); -csr_write_helper(env, fcsr, CSR_FCSR); +riscv_csr_write(env, CSR_FCSR, fcsr); } static void restore_ucontext(CPURISCVState *env, struct target_ucontext *uc) diff --git a/target/riscv/cpu.h b/target/riscv/cpu.h index 1ade90d23bbc..d6bb3136db18 100644 --- a/target/riscv/cpu.h +++ b/target/riscv/cpu.h @@ -257,7 +257,7 @@ char *riscv_isa_string(RISCVCPU *cpu); void riscv_cpu_list(FILE *f, fprintf_function cpu_fprintf); #define cpu_init(cpu_model) cpu_generic_init(TYPE_RISCV_CPU, cpu_model) -#define cpu_signal_handler cpu_riscv_signal_handler +#define cpu_signal_handler riscv_cpu_signal_handler #define cpu_list riscv_cpu_list #define cpu_mmu_index riscv_cpu_mmu_index @@ -265,16 +265,15 @@ void riscv_cpu_list(FILE *f, fprintf_function cpu_fprintf); uint32_t riscv_cpu_update_mip(RISCVCPU *cpu, uint32_t mask, uint32_t value); #define BOOL_TO_MASK(x) (-!!(x)) /* helper for riscv_cpu_update_mip value */ #endif -void riscv_set_mode(CPURISCVState *env, target_ulong newpriv); +void riscv_cpu_set_mode(CPURISCVState *env, target_ulong newpriv); void riscv_translate_init(void); -RISCVCPU *cpu_riscv_init(const char *cpu_model); -int cpu_riscv_signal_handler(int host_signum, void *pinfo, void *puc); -void QEMU_NORETURN do_raise_exception_err(CPURISCVState *env, - uint32_t exception, uintptr_t pc); +int riscv_cpu_signal_handler(int host_signum, void *pinfo, void *puc); +void QEMU_NORETURN riscv_raise_exception(CPURISCVState *env, + uint32_t exception, uintptr_t pc); -target_ulong cpu_riscv_get_fflags(CPURISCVState *env); -void cpu_riscv_set_fflags(CPURISCVState *env, target_ulong); +target_ulong riscv_cpu_get_fflags(CPURISCVState *env); +void riscv_cpu_set_fflags(CPURISCVState *env, target_ulong); #define TB_FLAGS_MMU_MASK 3 #define TB_FLAGS_MSTATUS_FS MSTATUS_FS @@ -294,13 +293,13 @@ static inline void cpu_get_tb_cpu_state(CPURISCVState *env, target_ulong *pc, int riscv_csrrw(CPURISCVState *env, int csrno, target_ulong *ret_value, target_ulong new_value, target_ulong write_mask); -static inline void csr_write_helper(CPURISCVState *env, target_ulong val, -int csrno) +static inline void riscv_csr_write(CPURISCVState *env, int csrno, + target_ulong val) { riscv_csrrw(env, csrno, NULL, val, -1); } -static inline target_ulong csr_read_helper(CPURISCVState *env, int csrno) +static inline target_ulong riscv_csr_read(CPURISCVState *env, int csrno) { target_ulong val = 0; riscv_csrrw(env, csrno, , 0, 0); diff --git a/target/riscv/cpu_helper.c b/target/riscv/cpu_helper.c index 69592c037042..b4bbf7a9fa0a 100644 --- a/target/riscv/cpu_helper.c +++ b/target/riscv/cpu_helper.c @@ -93,7 +93,7 @@ uint32_t riscv_cpu_update_mip(RISCVCPU *cpu, uint32_t mask, uint32_t value) return old; } -void riscv_set_mode(CPURISCVState *env, target_ulong newpriv) +void riscv_cpu_set_mode(CPURISCVState *env, target_ulong newpriv) { if (newpriv > PRV_M) { g_assert_not_reached(); @@ -366,7 +366,7 @@ void riscv_cpu_do_unaligned_access(CPUState *cs, vaddr addr, g_assert_not_reached(); } env->badaddr = addr; -do_raise_exception_err(env, cs->exception_index, retaddr); +riscv_raise_exception(env, cs->exception_index, retaddr);
[Qemu-devel] [PATCH v1 12/30] RISC-V: Mark mstatus.fs dirty
From: Richard Henderson <richard.hender...@linaro.org> Modifed from Richard Henderson's patch [1] to integrate with the new control and status register implementation. [1] https://lists.nongnu.org/archive/html/qemu-devel/2018-03/msg07034.html Note: the f* CSRs already mark mstatus.FS dirty using env->mstatus |= mstatus.FS so the bug in the first spin of this patch has been fixed in a prior commit. Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Alistair Francis <alistair.fran...@wdc.com> Cc: Richard Henderson <richard.hender...@linaro.org> Signed-off-by: Michael Clark <m...@sifive.com> Reviewed-by: Michael Clark <m...@sifive.com> Co-authored-by: Richard Henderson <richard.hender...@linaro.org> Co-authored-by: Michael Clark <m...@sifive.com> --- target/riscv/csr.c | 12 target/riscv/translate.c | 40 +++- 2 files changed, 39 insertions(+), 13 deletions(-) diff --git a/target/riscv/csr.c b/target/riscv/csr.c index 0e6c0c365154..b4452388ff02 100644 --- a/target/riscv/csr.c +++ b/target/riscv/csr.c @@ -325,18 +325,6 @@ static int write_mstatus(CPURISCVState *env, int csrno, target_ulong val) mstatus = (mstatus & ~mask) | (val & mask); -/* Note: this is a workaround for an issue where mstatus.FS - does not report dirty after floating point operations - that modify floating point state. This workaround is - technically compliant with the RISC-V Privileged - specification as it is legal to return only off, or dirty. - at the expense of extra floating point save/restore. */ - -/* FP is always dirty or off */ -if (mstatus & MSTATUS_FS) { -mstatus |= MSTATUS_FS; -} - int dirty = ((mstatus & MSTATUS_FS) == MSTATUS_FS) | ((mstatus & MSTATUS_XS) == MSTATUS_XS); mstatus = set_field(mstatus, MSTATUS_SD, dirty); diff --git a/target/riscv/translate.c b/target/riscv/translate.c index 466b9551cbd9..a980611eb611 100644 --- a/target/riscv/translate.c +++ b/target/riscv/translate.c @@ -651,6 +651,31 @@ static void gen_store(DisasContext *ctx, uint32_t opc, int rs1, int rs2, tcg_temp_free(dat); } +#ifndef CONFIG_USER_ONLY +/* The states of mstatus_fs are: + * 0 = disabled, 1 = initial, 2 = clean, 3 = dirty + * We will have already diagnosed disabled state, + * and need to turn initial/clean into dirty. + */ +static void mark_fs_dirty(DisasContext *ctx) +{ +TCGv tmp; +if (ctx->mstatus_fs == MSTATUS_FS) { +return; +} +/* Remember the state change for the rest of the TB. */ +ctx->mstatus_fs = MSTATUS_FS; + +tmp = tcg_temp_new(); +tcg_gen_ld_tl(tmp, cpu_env, offsetof(CPURISCVState, mstatus)); +tcg_gen_ori_tl(tmp, tmp, MSTATUS_FS); +tcg_gen_st_tl(tmp, cpu_env, offsetof(CPURISCVState, mstatus)); +tcg_temp_free(tmp); +} +#else +static inline void mark_fs_dirty(DisasContext *ctx) { } +#endif + static void gen_fp_load(DisasContext *ctx, uint32_t opc, int rd, int rs1, target_long imm) { @@ -679,6 +704,8 @@ static void gen_fp_load(DisasContext *ctx, uint32_t opc, int rd, break; } tcg_temp_free(t0); + +mark_fs_dirty(ctx); } static void gen_fp_store(DisasContext *ctx, uint32_t opc, int rs1, @@ -944,6 +971,7 @@ static void gen_fp_arith(DisasContext *ctx, uint32_t opc, int rd, int rs1, int rs2, int rm) { TCGv t0 = NULL; +bool fp_output = true; if (ctx->mstatus_fs == 0) { goto do_illegal; @@ -1006,6 +1034,7 @@ static void gen_fp_arith(DisasContext *ctx, uint32_t opc, int rd, } gen_set_gpr(rd, t0); tcg_temp_free(t0); +fp_output = false; break; case OPC_RISC_FCVT_W_S: @@ -1035,6 +1064,7 @@ static void gen_fp_arith(DisasContext *ctx, uint32_t opc, int rd, } gen_set_gpr(rd, t0); tcg_temp_free(t0); +fp_output = false; break; case OPC_RISC_FCVT_S_W: @@ -1085,6 +1115,7 @@ static void gen_fp_arith(DisasContext *ctx, uint32_t opc, int rd, } gen_set_gpr(rd, t0); tcg_temp_free(t0); +fp_output = false; break; case OPC_RISC_FMV_S_X: @@ -1177,6 +1208,7 @@ static void gen_fp_arith(DisasContext *ctx, uint32_t opc, int rd, } gen_set_gpr(rd, t0); tcg_temp_free(t0); +fp_output = false; break; case OPC_RISC_FCVT_W_D: @@ -1206,6 +1238,7 @@ static void gen_fp_arith(DisasContext *ctx, uint32_t opc, int rd, } gen_set_gpr(rd, t0); tcg_temp_free(t0); +fp_output = false; break; case OPC_RISC_FCVT_D_W: @@ -1253,6 +1286,7 @@ static void gen_fp_arith(DisasContext *ctx, uint32_t opc, int rd, default: goto do_
[Qemu-devel] [PATCH v1 03/30] RISC-V: Use atomic_cmpxchg to update PLIC bitmaps
The PLIC previously used a mutex to protect against concurrent access to the claimed and pending bitfields. Instead of using a mutex, we update the bitfields using atomic_cmpxchg. Rename sifive_plic_num_irqs_pending to sifive_plic_irqs_pending and add an early out if any interrupts are pending as the count of pending interrupts is not used. Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Alistair Francis <alistair.fran...@wdc.com> Signed-off-by: Michael Clark <m...@sifive.com> Reviewed-by: Richard Henderson <richard.hender...@linaro.org> --- hw/riscv/sifive_plic.c | 49 +++--- include/hw/riscv/sifive_plic.h | 1 - 2 files changed, 22 insertions(+), 28 deletions(-) diff --git a/hw/riscv/sifive_plic.c b/hw/riscv/sifive_plic.c index 874de2ebaf77..1af23c76e603 100644 --- a/hw/riscv/sifive_plic.c +++ b/hw/riscv/sifive_plic.c @@ -81,36 +81,32 @@ static void sifive_plic_print_state(SiFivePLICState *plic) } } -static -void sifive_plic_set_pending(SiFivePLICState *plic, int irq, bool pending) +static uint32_t atomic_set_masked(uint32_t *a, uint32_t mask, uint32_t value) { -qemu_mutex_lock(>lock); -uint32_t word = irq >> 5; -if (pending) { -plic->pending[word] |= (1 << (irq & 31)); -} else { -plic->pending[word] &= ~(1 << (irq & 31)); -} -qemu_mutex_unlock(>lock); +uint32_t old, new, cmp = atomic_read(a); + +do { +old = cmp; +new = (old & ~mask) | (value & mask); +cmp = atomic_cmpxchg(a, old, new); +} while (old != cmp); + +return old; } -static -void sifive_plic_set_claimed(SiFivePLICState *plic, int irq, bool claimed) +static void sifive_plic_set_pending(SiFivePLICState *plic, int irq, bool level) { -qemu_mutex_lock(>lock); -uint32_t word = irq >> 5; -if (claimed) { -plic->claimed[word] |= (1 << (irq & 31)); -} else { -plic->claimed[word] &= ~(1 << (irq & 31)); -} -qemu_mutex_unlock(>lock); +atomic_set_masked(>pending[irq >> 5], 1 << (irq & 31), -!!level); } -static -int sifive_plic_num_irqs_pending(SiFivePLICState *plic, uint32_t addrid) +static void sifive_plic_set_claimed(SiFivePLICState *plic, int irq, bool level) { -int i, j, count = 0; +atomic_set_masked(>claimed[irq >> 5], 1 << (irq & 31), -!!level); +} + +static int sifive_plic_irqs_pending(SiFivePLICState *plic, uint32_t addrid) +{ +int i, j; for (i = 0; i < plic->bitfield_words; i++) { uint32_t pending_enabled_not_claimed = (plic->pending[i] & ~plic->claimed[i]) & @@ -123,11 +119,11 @@ int sifive_plic_num_irqs_pending(SiFivePLICState *plic, uint32_t addrid) uint32_t prio = plic->source_priority[irq]; int enabled = pending_enabled_not_claimed & (1 << j); if (enabled && prio > plic->target_priority[addrid]) { -count++; +return 1; } } } -return count; +return 0; } static void sifive_plic_update(SiFivePLICState *plic) @@ -143,7 +139,7 @@ static void sifive_plic_update(SiFivePLICState *plic) if (!env) { continue; } -int level = sifive_plic_num_irqs_pending(plic, addrid) > 0; +int level = sifive_plic_irqs_pending(plic, addrid); switch (mode) { case PLICMode_M: riscv_set_local_interrupt(RISCV_CPU(cpu), MIP_MEIP, level); @@ -440,7 +436,6 @@ static void sifive_plic_realize(DeviceState *dev, Error **errp) memory_region_init_io(>mmio, OBJECT(dev), _plic_ops, plic, TYPE_SIFIVE_PLIC, plic->aperture_size); parse_hart_config(plic); -qemu_mutex_init(>lock); plic->bitfield_words = (plic->num_sources + 31) >> 5; plic->source_priority = g_new0(uint32_t, plic->num_sources); plic->target_priority = g_new(uint32_t, plic->num_addrs); diff --git a/include/hw/riscv/sifive_plic.h b/include/hw/riscv/sifive_plic.h index 11a5a98df1f9..ff09a288261e 100644 --- a/include/hw/riscv/sifive_plic.h +++ b/include/hw/riscv/sifive_plic.h @@ -55,7 +55,6 @@ typedef struct SiFivePLICState { uint32_t *pending; uint32_t *claimed; uint32_t *enable; -QemuMutex lock; qemu_irq *irqs; /* config */ -- 2.7.0
[Qemu-devel] [PATCH v1 19/30] RISC-V: Allow interrupt controllers to claim interrupts
We can't allow the supervisor to control SEIP as this would allow the supervisor to clear a pending external interrupt which will result in lost a interrupt in the case a PLIC is attached. The SEIP bit must be hardware controlled when a PLIC is attached. This logic was previously hard-coded so SEIP was always masked even if no PLIC was attached. This patch adds riscv_cpu_claim_interrupts so that the PLIC can register control of SEIP. In the case of models without a PLIC (spike), the SEIP bit remains software controlled. This interface allows for hardware control of supervisor timer and software interrupts by other interrupt controller models. Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Cc: Alistair Francis <alistair.fran...@wdc.com> Signed-off-by: Michael Clark <m...@sifive.com> --- hw/riscv/sifive_plic.c| 13 + target/riscv/cpu.h| 2 ++ target/riscv/cpu_helper.c | 11 +++ target/riscv/csr.c| 12 4 files changed, 30 insertions(+), 8 deletions(-) diff --git a/hw/riscv/sifive_plic.c b/hw/riscv/sifive_plic.c index dc6f4924e282..28e28d932f7c 100644 --- a/hw/riscv/sifive_plic.c +++ b/hw/riscv/sifive_plic.c @@ -23,6 +23,7 @@ #include "qemu/error-report.h" #include "hw/sysbus.h" #include "target/riscv/cpu.h" +#include "sysemu/sysemu.h" #include "hw/riscv/sifive_plic.h" #define RISCV_DEBUG_PLIC 0 @@ -447,6 +448,18 @@ static void sifive_plic_realize(DeviceState *dev, Error **errp) for (i = 0; i <= plic->num_sources; i++) { plic->irqs[i] = qemu_allocate_irq(sifive_plic_irq_request, plic, i); } + +/* We can't allow the supervisor to control SEIP as this would allow the + * supervisor to clear a pending external interrupt which will result in + * lost a interrupt in the case a PLIC is attached. The SEIP bit must be + * hardware controlled when a PLIC is attached. */ +for (i = 0; i < smp_cpus; i++) { +RISCVCPU *cpu = RISCV_CPU(qemu_get_cpu(i)); +if (riscv_cpu_claim_interrupts(cpu, MIP_SEIP) < 0) { +error_report("sifive_plic_realize: SEIP already claimed"); +exit(1); +} +} } static void sifive_plic_class_init(ObjectClass *klass, void *data) diff --git a/target/riscv/cpu.h b/target/riscv/cpu.h index d6bb3136db18..ae0e3f6a544d 100644 --- a/target/riscv/cpu.h +++ b/target/riscv/cpu.h @@ -138,6 +138,7 @@ struct CPURISCVState { * mip is 32-bits to allow atomic_read on 32-bit hosts. */ uint32_t mip; +uint32_t miclaim; target_ulong mie; target_ulong mideleg; @@ -262,6 +263,7 @@ void riscv_cpu_list(FILE *f, fprintf_function cpu_fprintf); #define cpu_mmu_index riscv_cpu_mmu_index #ifndef CONFIG_USER_ONLY +int riscv_cpu_claim_interrupts(RISCVCPU *cpu, uint32_t interrupts); uint32_t riscv_cpu_update_mip(RISCVCPU *cpu, uint32_t mask, uint32_t value); #define BOOL_TO_MASK(x) (-!!(x)) /* helper for riscv_cpu_update_mip value */ #endif diff --git a/target/riscv/cpu_helper.c b/target/riscv/cpu_helper.c index b4bbf7a9fa0a..7c9f6c46c75a 100644 --- a/target/riscv/cpu_helper.c +++ b/target/riscv/cpu_helper.c @@ -72,6 +72,17 @@ bool riscv_cpu_exec_interrupt(CPUState *cs, int interrupt_request) #if !defined(CONFIG_USER_ONLY) +int riscv_cpu_claim_interrupts(RISCVCPU *cpu, uint32_t interrupts) +{ +CPURISCVState *env = >env; +if (env->miclaim & interrupts) { +return -1; +} else { +env->miclaim |= interrupts; +return 0; +} +} + /* iothread_mutex must be held */ uint32_t riscv_cpu_update_mip(RISCVCPU *cpu, uint32_t mask, uint32_t value) { diff --git a/target/riscv/csr.c b/target/riscv/csr.c index 45e33d876034..9bbe81a110a5 100644 --- a/target/riscv/csr.c +++ b/target/riscv/csr.c @@ -502,15 +502,11 @@ static int rmw_mip(CPURISCVState *env, int csrno, target_ulong *ret_value, target_ulong new_value, target_ulong write_mask) { RISCVCPU *cpu = riscv_env_get_cpu(env); -target_ulong mask = write_mask & delegable_ints; -uint32_t old_mip; -/* We can't allow the supervisor to control SEIP as this would allow the - * supervisor to clear a pending external interrupt which will result in - * lost a interrupt in the case a PLIC is attached. The SEIP bit must be - * hardware controlled when a PLIC is attached. This should be an option - * for CPUs with software-delegated Supervisor External Interrupts. */ -mask &= ~MIP_SEIP; +/* Allow software control of delegable interrupts not claimed by hardware */ +target_ulong mask = write_mask & delegable_ints & ~env->miclaim; + +uint32_t old_mip; if (mask) { qemu_mutex_lock_iothread(); -- 2.7.0
[Qemu-devel] [PATCH v1 13/30] RISC-V: Implement mstatus.TSR/TW/TVM
This adds the necessary minimum to support S-mode virtualization for priv ISA >= v1.10 Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Alistair Francis <alistair.fran...@wdc.com> Cc: Matthew Suozzo <msuo...@google.com> Signed-off-by: Michael Clark <m...@sifive.com> Co-authored-by: Matthew Suozzo <msuo...@google.com> Co-authored-by: Michael Clark <m...@sifive.com> --- target/riscv/csr.c | 17 + target/riscv/op_helper.c | 25 + 2 files changed, 34 insertions(+), 8 deletions(-) diff --git a/target/riscv/csr.c b/target/riscv/csr.c index b4452388ff02..509215327243 100644 --- a/target/riscv/csr.c +++ b/target/riscv/csr.c @@ -313,7 +313,8 @@ static int write_mstatus(CPURISCVState *env, int csrno, target_ulong val) } mask = MSTATUS_SIE | MSTATUS_SPIE | MSTATUS_MIE | MSTATUS_MPIE | MSTATUS_SPP | MSTATUS_FS | MSTATUS_MPRV | MSTATUS_SUM | -MSTATUS_MPP | MSTATUS_MXR; +MSTATUS_MPP | MSTATUS_MXR | MSTATUS_TVM | MSTATUS_TSR | +MSTATUS_TW; } /* silenty discard mstatus.mpp writes for unsupported modes */ @@ -654,7 +655,11 @@ static int read_satp(CPURISCVState *env, int csrno, target_ulong *val) if (!riscv_feature(env, RISCV_FEATURE_MMU)) { *val = 0; } else if (env->priv_ver >= PRIV_VERSION_1_10_0) { -*val = env->satp; +if (env->priv == PRV_S && get_field(env->mstatus, MSTATUS_TVM)) { +return -1; +} else { +*val = env->satp; +} } else { *val = env->sptbr; } @@ -675,8 +680,12 @@ static int write_satp(CPURISCVState *env, int csrno, target_ulong val) validate_vm(env, get_field(val, SATP_MODE)) && ((val ^ env->satp) & (SATP_MODE | SATP_ASID | SATP_PPN))) { -tlb_flush(CPU(riscv_env_get_cpu(env))); -env->satp = val; +if (env->priv == PRV_S && get_field(env->mstatus, MSTATUS_TVM)) { +return -1; +} else { +tlb_flush(CPU(riscv_env_get_cpu(env))); +env->satp = val; +} } return 0; } diff --git a/target/riscv/op_helper.c b/target/riscv/op_helper.c index 81bd1a77ea90..77c79ba36e0b 100644 --- a/target/riscv/op_helper.c +++ b/target/riscv/op_helper.c @@ -82,6 +82,11 @@ target_ulong helper_sret(CPURISCVState *env, target_ulong cpu_pc_deb) do_raise_exception_err(env, RISCV_EXCP_INST_ADDR_MIS, GETPC()); } +if (env->priv_ver >= PRIV_VERSION_1_10_0 && +get_field(env->mstatus, MSTATUS_TSR)) { +do_raise_exception_err(env, RISCV_EXCP_ILLEGAL_INST, GETPC()); +} + target_ulong mstatus = env->mstatus; target_ulong prev_priv = get_field(mstatus, MSTATUS_SPP); mstatus = set_field(mstatus, @@ -125,16 +130,28 @@ void helper_wfi(CPURISCVState *env) { CPUState *cs = CPU(riscv_env_get_cpu(env)); -cs->halted = 1; -cs->exception_index = EXCP_HLT; -cpu_loop_exit(cs); +if (env->priv == PRV_S && +env->priv_ver >= PRIV_VERSION_1_10_0 && +get_field(env->mstatus, MSTATUS_TW)) { +do_raise_exception_err(env, RISCV_EXCP_ILLEGAL_INST, GETPC()); +} else { +cs->halted = 1; +cs->exception_index = EXCP_HLT; +cpu_loop_exit(cs); +} } void helper_tlb_flush(CPURISCVState *env) { RISCVCPU *cpu = riscv_env_get_cpu(env); CPUState *cs = CPU(cpu); -tlb_flush(cs); +if (env->priv == PRV_S && +env->priv_ver >= PRIV_VERSION_1_10_0 && +get_field(env->mstatus, MSTATUS_TVM)) { +do_raise_exception_err(env, RISCV_EXCP_ILLEGAL_INST, GETPC()); +} else { +tlb_flush(cs); +} } #endif /* !CONFIG_USER_ONLY */ -- 2.7.0
[Qemu-devel] [PATCH v1 07/30] RISC-V: Update CSR and interrupt definitions
* Add user-mode CSR defininitions. * Reorder CSR definitions to match the specification. * Change H mode interrupt comment to 'reserved'. * Remove unused X_COP interrupt. * Add user-mode and core-level interrupts. * Remove erroneous until comemnts on machine mode interrupts. * Move together paging mode and page table bit definitions. * Move together interrupt and exception cause definitions. Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Alistair Francis <alistair.fran...@wdc.com> Signed-off-by: Michael Clark <m...@sifive.com> Reviewed-by: Alistair Francis <alistair.fran...@wdc.com> --- target/riscv/cpu_bits.h | 692 +-- target/riscv/op_helper.c | 4 +- 2 files changed, 376 insertions(+), 320 deletions(-) diff --git a/target/riscv/cpu_bits.h b/target/riscv/cpu_bits.h index 12b4757088f4..878de6233846 100644 --- a/target/riscv/cpu_bits.h +++ b/target/riscv/cpu_bits.h @@ -6,242 +6,283 @@ (((target_ulong)(val) * ((mask) & ~((mask) << 1))) & \ (target_ulong)(mask))) -#define PGSHIFT 12 - -#define FSR_RD_SHIFT 5 -#define FSR_RD (0x7 << FSR_RD_SHIFT) - -#define FPEXC_NX 0x01 -#define FPEXC_UF 0x02 -#define FPEXC_OF 0x04 -#define FPEXC_DZ 0x08 -#define FPEXC_NV 0x10 - -#define FSR_AEXC_SHIFT 0 -#define FSR_NVA (FPEXC_NV << FSR_AEXC_SHIFT) -#define FSR_OFA (FPEXC_OF << FSR_AEXC_SHIFT) -#define FSR_UFA (FPEXC_UF << FSR_AEXC_SHIFT) -#define FSR_DZA (FPEXC_DZ << FSR_AEXC_SHIFT) -#define FSR_NXA (FPEXC_NX << FSR_AEXC_SHIFT) -#define FSR_AEXC (FSR_NVA | FSR_OFA | FSR_UFA | FSR_DZA | FSR_NXA) - -/* CSR numbers */ -#define CSR_FFLAGS 0x1 -#define CSR_FRM 0x2 -#define CSR_FCSR 0x3 -#define CSR_CYCLE 0xc00 -#define CSR_TIME 0xc01 -#define CSR_INSTRET 0xc02 -#define CSR_HPMCOUNTER3 0xc03 -#define CSR_HPMCOUNTER4 0xc04 -#define CSR_HPMCOUNTER5 0xc05 -#define CSR_HPMCOUNTER6 0xc06 -#define CSR_HPMCOUNTER7 0xc07 -#define CSR_HPMCOUNTER8 0xc08 -#define CSR_HPMCOUNTER9 0xc09 -#define CSR_HPMCOUNTER10 0xc0a -#define CSR_HPMCOUNTER11 0xc0b -#define CSR_HPMCOUNTER12 0xc0c -#define CSR_HPMCOUNTER13 0xc0d -#define CSR_HPMCOUNTER14 0xc0e -#define CSR_HPMCOUNTER15 0xc0f -#define CSR_HPMCOUNTER16 0xc10 -#define CSR_HPMCOUNTER17 0xc11 -#define CSR_HPMCOUNTER18 0xc12 -#define CSR_HPMCOUNTER19 0xc13 -#define CSR_HPMCOUNTER20 0xc14 -#define CSR_HPMCOUNTER21 0xc15 -#define CSR_HPMCOUNTER22 0xc16 -#define CSR_HPMCOUNTER23 0xc17 -#define CSR_HPMCOUNTER24 0xc18 -#define CSR_HPMCOUNTER25 0xc19 -#define CSR_HPMCOUNTER26 0xc1a -#define CSR_HPMCOUNTER27 0xc1b -#define CSR_HPMCOUNTER28 0xc1c -#define CSR_HPMCOUNTER29 0xc1d -#define CSR_HPMCOUNTER30 0xc1e -#define CSR_HPMCOUNTER31 0xc1f -#define CSR_SSTATUS 0x100 -#define CSR_SIE 0x104 -#define CSR_STVEC 0x105 -#define CSR_SCOUNTEREN 0x106 -#define CSR_SSCRATCH 0x140 -#define CSR_SEPC 0x141 -#define CSR_SCAUSE 0x142 -#define CSR_SBADADDR 0x143 -#define CSR_SIP 0x144 -#define CSR_SPTBR 0x180 -#define CSR_SATP 0x180 -#define CSR_MSTATUS 0x300 -#define CSR_MISA 0x301 -#define CSR_MEDELEG 0x302 -#define CSR_MIDELEG 0x303 -#define CSR_MIE 0x304 -#define CSR_MTVEC 0x305 -#define CSR_MCOUNTEREN 0x306 -#define CSR_MSCRATCH 0x340 -#define CSR_MEPC 0x341 -#define CSR_MCAUSE 0x342 -#define CSR_MBADADDR 0x343 -#define CSR_MIP 0x344 -#define CSR_PMPCFG0 0x3a0 -#define CSR_PMPCFG1 0x3a1 -#define CSR_PMPCFG2 0x3a2 -#define CSR_PMPCFG3 0x3a3 -#define CSR_PMPADDR0 0x3b0 -#define CSR_PMPADDR1 0x3b1 -#define CSR_PMPADDR2 0x3b2 -#define CSR_PMPADDR3 0x3b3 -#define CSR_PMPADDR4 0x3b4 -#define CSR_PMPADDR5 0x3b5 -#define CSR_PMPADDR6 0x3b6 -#define CSR_PMPADDR7 0x3b7 -#define CSR_PMPADDR8 0x3b8 -#define CSR_PMPADDR9 0x3b9 -#define CSR_PMPADDR10 0x3ba -#define CSR_PMPADDR11 0x3bb -#define CSR_PMPADDR12 0x3bc -#define CSR_PMPADDR13 0x3bd -#define CSR_PMPADDR14 0x3be -#define CSR_PMPADDR15 0x3bf -#define CSR_TSELECT 0x7a0 -#define CSR_TDATA1 0x7a1 -#define CSR_TDATA2 0x7a2 -#define CSR_TDATA3 0x7a3 -#define CSR_DCSR 0x7b0 -#define CSR_DPC 0x7b1 -#define CSR_DSCRATCH 0x7b2 -#define CSR_MCYCLE 0xb00 -#define CSR_MINSTRET 0xb02 -#define CSR_MHPMCOUNTER3 0xb03 -#define CSR_MHPMCOUNTER4 0xb04 -#define CSR_MHPMCOUNTER5 0xb05 -#define CSR_MHPMCOUNTER6 0xb06 -#define CSR_MHPMCOUNTER7 0xb07 -#define CSR_MHPMCOUNTER8 0xb08 -#define CSR_MHPMCOUNTER9 0xb09 -#define CSR_MHPMCOUNTER10 0xb0a -#define CSR_MHPMCOUNTER11 0xb0b -#define CSR_MHPMCOUNTER12 0xb0c -#define CSR_MHPMCOUNTER13 0xb0d -#define CSR_MHPMCOUNTER14 0xb0e -#define CSR_MHPMCOUNTER15 0xb0f -#define CSR_MHPMCOUNTER16 0xb10 -#define CSR_MHPMCOUNTER17 0xb11 -#define CSR_MHPMCOUNTER18 0xb12 -#define CSR_MHPMCOUNTER19 0xb13 -#define CSR_MHPMCOUNTER20 0xb14 -#define CSR_MHPMCOUNTER21 0xb15 -#define CSR_MHPMCOUNTER22 0xb16 -#define CSR_MHPMCOUNTER23 0xb17 -#define CSR_MHPMCOUNTER24 0xb18 -#define CSR
[Qemu-devel] [PATCH v1 01/30] RISC-V: Update address bits to support sv39 and sv48
Cc: Sagar KarandikarCc: Bastian Koppelmann Cc: Palmer Dabbelt Cc: Alistair Francis Signed-off-by: Palmer Dabbelt --- target/riscv/cpu.h | 8 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/target/riscv/cpu.h b/target/riscv/cpu.h index 34abc383e3d4..e0608e6d5f08 100644 --- a/target/riscv/cpu.h +++ b/target/riscv/cpu.h @@ -24,12 +24,12 @@ #define TARGET_PAGE_BITS 12 /* 4 KiB Pages */ #if defined(TARGET_RISCV64) #define TARGET_LONG_BITS 64 -#define TARGET_PHYS_ADDR_SPACE_BITS 50 -#define TARGET_VIRT_ADDR_SPACE_BITS 39 +#define TARGET_PHYS_ADDR_SPACE_BITS 56 /* 44-bit PPN */ +#define TARGET_VIRT_ADDR_SPACE_BITS 48 /* sv48 */ #elif defined(TARGET_RISCV32) #define TARGET_LONG_BITS 32 -#define TARGET_PHYS_ADDR_SPACE_BITS 34 -#define TARGET_VIRT_ADDR_SPACE_BITS 32 +#define TARGET_PHYS_ADDR_SPACE_BITS 34 /* 22-bit PPN */ +#define TARGET_VIRT_ADDR_SPACE_BITS 32 /* sv32 */ #endif #define TCG_GUEST_DEFAULT_MO 0 -- 2.7.0
[Qemu-devel] [PATCH v1 08/30] RISC-V: Implement modular CSR helper interface
Previous CSR code uses csr_read_helper and csr_write_helper to update CSR registers however this interface prevents atomic read/modify/write CSR operations; in addition there is no trap-free method to access to CSRs due to the monolithic CSR functions call longjmp. The current iCSR interface is not safe to be called by target/riscv/gdbstub.c as privilege checks or missing CSRs may call longjmp to generate exceptions. It needs to indicate existence so traps can be generated in the CSR instruction helpers. This commit moves CSR access from the monolithic switch statements in target/riscv/op_helper.c into modular read/write functions in target/riscv/csr.c using a new function pointer table for dispatch (which can later be used to allow CPUs to hook up model specific CSRs). A read/modify/write interface is added to support atomic CSR operations and a non-trapping interface is added to allow exception-free access to CSRs by the debugger. The CSR functions and CSR dispatch table are ordered to match The RISC-V Instruction Set Manual, Volume II: Privileged Architecture Version 1.10, 2.2 CSR Listing. Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Alistair Francis <alistair.fran...@wdc.com> Signed-off-by: Michael Clark <m...@sifive.com> --- target/riscv/Makefile.objs | 2 +- target/riscv/cpu.h | 18 +- target/riscv/cpu_helper.c | 4 +- target/riscv/csr.c | 857 + target/riscv/gdbstub.c | 10 +- target/riscv/op_helper.c | 611 +--- 6 files changed, 898 insertions(+), 604 deletions(-) create mode 100644 target/riscv/csr.c diff --git a/target/riscv/Makefile.objs b/target/riscv/Makefile.objs index fcc5d34c1f2e..4072abe3e45c 100644 --- a/target/riscv/Makefile.objs +++ b/target/riscv/Makefile.objs @@ -1 +1 @@ -obj-y += translate.o op_helper.o cpu_helper.o cpu.o fpu_helper.o gdbstub.o pmp.o +obj-y += translate.o op_helper.o cpu_helper.o cpu.o csr.o fpu_helper.o gdbstub.o pmp.o diff --git a/target/riscv/cpu.h b/target/riscv/cpu.h index c5d485769cde..02c60c45631b 100644 --- a/target/riscv/cpu.h +++ b/target/riscv/cpu.h @@ -290,9 +290,21 @@ static inline void cpu_get_tb_cpu_state(CPURISCVState *env, target_ulong *pc, #endif } -void csr_write_helper(CPURISCVState *env, target_ulong val_to_write, -target_ulong csrno); -target_ulong csr_read_helper(CPURISCVState *env, target_ulong csrno); +int riscv_csrrw(CPURISCVState *env, int csrno, target_ulong *ret_value, +target_ulong new_value, target_ulong write_mask); + +static inline void csr_write_helper(CPURISCVState *env, target_ulong val, +int csrno) +{ +riscv_csrrw(env, csrno, NULL, val, -1); +} + +static inline target_ulong csr_read_helper(CPURISCVState *env, int csrno) +{ +target_ulong val = 0; +riscv_csrrw(env, csrno, , 0, 0); +return val; +} #include "exec/cpu-all.h" diff --git a/target/riscv/cpu_helper.c b/target/riscv/cpu_helper.c index 6c886e99055a..1f523861b7cb 100644 --- a/target/riscv/cpu_helper.c +++ b/target/riscv/cpu_helper.c @@ -526,7 +526,7 @@ void riscv_cpu_do_interrupt(CPUState *cs) get_field(s, MSTATUS_SIE) : get_field(s, MSTATUS_UIE << env->priv)); s = set_field(s, MSTATUS_SPP, env->priv); s = set_field(s, MSTATUS_SIE, 0); -csr_write_helper(env, s, CSR_MSTATUS); +env->mstatus = s; riscv_set_mode(env, PRV_S); } else { /* No need to check MTVEC for misaligned - lower 2 bits cannot be set */ @@ -551,7 +551,7 @@ void riscv_cpu_do_interrupt(CPUState *cs) get_field(s, MSTATUS_MIE) : get_field(s, MSTATUS_UIE << env->priv)); s = set_field(s, MSTATUS_MPP, env->priv); s = set_field(s, MSTATUS_MIE, 0); -csr_write_helper(env, s, CSR_MSTATUS); +env->mstatus = s; riscv_set_mode(env, PRV_M); } /* TODO yield load reservation */ diff --git a/target/riscv/csr.c b/target/riscv/csr.c new file mode 100644 index ..e08f3523d854 --- /dev/null +++ b/target/riscv/csr.c @@ -0,0 +1,857 @@ +/* + * RISC-V Control and Status Registers. + * + * Copyright (c) 2016-2017 Sagar Karandikar, sag...@eecs.berkeley.edu + * Copyright (c) 2017-2018 SiFive, Inc. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2 or later, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public
[Qemu-devel] [PATCH v1 04/30] RISC-V: Simplify riscv_cpu_local_irqs_pending
This commit is intended to improve readability. There is no change to the logic. Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Alistair Francis <alistair.fran...@wdc.com> Signed-off-by: Michael Clark <m...@sifive.com> Reviewed-by: Alistair Francis <alistair.fran...@wdc.com> --- target/riscv/helper.c | 34 -- 1 file changed, 12 insertions(+), 22 deletions(-) diff --git a/target/riscv/helper.c b/target/riscv/helper.c index 3b57e1360549..47d116e9c13f 100644 --- a/target/riscv/helper.c +++ b/target/riscv/helper.c @@ -35,28 +35,18 @@ int riscv_cpu_mmu_index(CPURISCVState *env, bool ifetch) } #ifndef CONFIG_USER_ONLY -/* - * Return RISC-V IRQ number if an interrupt should be taken, else -1. - * Used in cpu-exec.c - * - * Adapted from Spike's processor_t::take_interrupt() - */ -static int riscv_cpu_hw_interrupts_pending(CPURISCVState *env) +static int riscv_cpu_local_irq_pending(CPURISCVState *env) { -target_ulong pending_interrupts = atomic_read(>mip) & env->mie; - -target_ulong mie = get_field(env->mstatus, MSTATUS_MIE); -target_ulong m_enabled = env->priv < PRV_M || (env->priv == PRV_M && mie); -target_ulong enabled_interrupts = pending_interrupts & - ~env->mideleg & -m_enabled; - -target_ulong sie = get_field(env->mstatus, MSTATUS_SIE); -target_ulong s_enabled = env->priv < PRV_S || (env->priv == PRV_S && sie); -enabled_interrupts |= pending_interrupts & env->mideleg & - -s_enabled; - -if (enabled_interrupts) { -return ctz64(enabled_interrupts); /* since non-zero */ +target_ulong mstatus_mie = get_field(env->mstatus, MSTATUS_MIE); +target_ulong mstatus_sie = get_field(env->mstatus, MSTATUS_SIE); +target_ulong pending = atomic_read(>mip) & env->mie; +target_ulong mie = env->priv < PRV_M || (env->priv == PRV_M && mstatus_mie); +target_ulong sie = env->priv < PRV_S || (env->priv == PRV_S && mstatus_sie); +target_ulong irqs = (pending & ~env->mideleg & -mie) | +(pending & env->mideleg & -sie); + +if (irqs) { +return ctz64(irqs); /* since non-zero */ } else { return EXCP_NONE; /* indicates no pending interrupt */ } @@ -69,7 +59,7 @@ bool riscv_cpu_exec_interrupt(CPUState *cs, int interrupt_request) if (interrupt_request & CPU_INTERRUPT_HARD) { RISCVCPU *cpu = RISCV_CPU(cs); CPURISCVState *env = >env; -int interruptno = riscv_cpu_hw_interrupts_pending(env); +int interruptno = riscv_cpu_local_irq_pending(env); if (interruptno >= 0) { cs->exception_index = RISCV_EXCP_INT_FLAG | interruptno; riscv_cpu_do_interrupt(cs); -- 2.7.0
[Qemu-devel] [PATCH v1 02/30] RISC-V: Improve page table walker spec compliance
- Inline PTE_TABLE check for better readability - Change access checks from ternary operator to if - Improve readibility of User page U mode and SUM test - Disallow non U mode from fetching from User pages - Add reserved PTE flag check: W or W|X - Add misaligned PPN check - Set READ protection for PTE X flag and mstatus.mxr - Use memory_region_is_ram in pte update Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Alistair Francis <alistair.fran...@wdc.com> Signed-off-by: Michael Clark <m...@sifive.com> Reviewed-by: Alistair Francis <alistair.fran...@wdc.com> --- target/riscv/cpu_bits.h | 2 -- target/riscv/helper.c | 64 ++--- 2 files changed, 45 insertions(+), 21 deletions(-) diff --git a/target/riscv/cpu_bits.h b/target/riscv/cpu_bits.h index 64aa097181fa..12b4757088f4 100644 --- a/target/riscv/cpu_bits.h +++ b/target/riscv/cpu_bits.h @@ -407,5 +407,3 @@ #define PTE_SOFT 0x300 /* Reserved for Software */ #define PTE_PPN_SHIFT 10 - -#define PTE_TABLE(PTE) (((PTE) & (PTE_V | PTE_R | PTE_W | PTE_X)) == PTE_V) diff --git a/target/riscv/helper.c b/target/riscv/helper.c index 95889f23b94d..3b57e1360549 100644 --- a/target/riscv/helper.c +++ b/target/riscv/helper.c @@ -185,16 +185,39 @@ restart: #endif target_ulong ppn = pte >> PTE_PPN_SHIFT; -if (PTE_TABLE(pte)) { /* next level of page table */ +if (!(pte & PTE_V)) { +/* Invalid PTE */ +return TRANSLATE_FAIL; +} else if (!(pte & (PTE_R | PTE_W | PTE_X))) { +/* Inner PTE, continue walking */ base = ppn << PGSHIFT; -} else if ((pte & PTE_U) ? (mode == PRV_S) && !sum : !(mode == PRV_S)) { -break; -} else if (!(pte & PTE_V) || (!(pte & PTE_R) && (pte & PTE_W))) { -break; -} else if (access_type == MMU_INST_FETCH ? !(pte & PTE_X) : - access_type == MMU_DATA_LOAD ? !(pte & PTE_R) && - !(mxr && (pte & PTE_X)) : !((pte & PTE_R) && (pte & PTE_W))) { -break; +} else if ((pte & (PTE_R | PTE_W | PTE_X)) == PTE_W) { +/* Reserved leaf PTE flags: PTE_W */ +return TRANSLATE_FAIL; +} else if ((pte & (PTE_R | PTE_W | PTE_X)) == (PTE_W | PTE_X)) { +/* Reserved leaf PTE flags: PTE_W + PTE_X */ +return TRANSLATE_FAIL; +} else if ((pte & PTE_U) && ((mode != PRV_U) && + (!sum || access_type == MMU_INST_FETCH))) { +/* User PTE flags when not U mode and mstatus.SUM is not set, + or the access type is an instruction fetch */ +return TRANSLATE_FAIL; +} else if (!(pte & PTE_U) && (mode != PRV_S)) { +/* Supervisor PTE flags when not S mode */ +return TRANSLATE_FAIL; +} else if (ppn & ((1ULL << ptshift) - 1)) { +/* Misasligned PPN */ +return TRANSLATE_FAIL; +} else if (access_type == MMU_DATA_LOAD && !((pte & PTE_R) || + ((pte & PTE_X) && mxr))) { +/* Read access check failed */ +return TRANSLATE_FAIL; +} else if (access_type == MMU_DATA_STORE && !(pte & PTE_W)) { +/* Write access check failed */ +return TRANSLATE_FAIL; +} else if (access_type == MMU_INST_FETCH && !(pte & PTE_X)) { +/* Fetch access check failed */ +return TRANSLATE_FAIL; } else { /* if necessary, set accessed and dirty bits. */ target_ulong updated_pte = pte | PTE_A | @@ -202,16 +225,19 @@ restart: /* Page table updates need to be atomic with MTTCG enabled */ if (updated_pte != pte) { -/* if accessed or dirty bits need updating, and the PTE is - * in RAM, then we do so atomically with a compare and swap. - * if the PTE is in IO space, then it can't be updated. - * if the PTE changed, then we must re-walk the page table - as the PTE is no longer valid */ +/* + * - if accessed or dirty bits need updating, and the PTE is + * in RAM, then we do so atomically with a compare and swap. + * - if the PTE is in IO space or ROM, then it can't be updated + * and we return TRANSLATE_FAIL. + * - if the PTE changed by the time we went to update it, then + * it is no longer valid and we must re-walk the page table. + */ MemoryRegion *mr; hwadd
[Qemu-devel] [PATCH v1 00/30] QEMU 2.13 RISC-V updates
This is a series of bug fixes, specification conformance fixes and CPU feature modularily updates to allow more precise modelling of the SiFive U Series CPUs (multi-core application processors with MMU, Supervisor and User modes) and SiFive E Series CPUs (embedded microcontroller cores without MMU or Supervisor mode). This series adds several updates to CPU features such as writable 'misa', minimal hypervisor support for mstatus.TSR/TW/TVM, a new modular interface for truly atomic CSRs, bug fixes and several new features such as separate firmware and kernel payload. These changes are available in the RISC-V GitHub repo: - https://github.com/riscv/riscv-qemu/tree/qemu-2.13-for-upstream Summary of changes * Support separate firmware and kernel payload * Can use -bios bbl -kernel vmlinux on 'virt' machine * Add infrastructure to support runtime 'misa' writes * Implemented TSR, TW and TVM for privileged ISA v1.10 * Update floating-point to correctly mark mstatus.FS dirty. * Made PMP (Physical Memory Protection) an optional CPU feature. * Disabled access to s* CSRs on cores without misa.S set. * Added CSR feature predicates to improve CPU emulation support and to allow for easier CPU model differentiation. * SiFive U series application processors (MMU, S-mode, U-mode) * SiFive E series embedded microcontrollers (no MMU, U-mode) * Add non-trapping interface to CSRs so that gdbstub.c can accesses CSRs without longjmp being called. * Implements an interface for atomic CSR accesses and convert accesses to 'mip' and 'sip' to the atomic interface: * The previous implementation using separate methods for csr_read_helper and csr_write_helper was incompatible with atomic CSR accesses. The previous implementation used monolithic switch statements and was not modular. * Add public API so that CPUs can implement custom CSRs. * Replaces locks with atomic lock-free updates for interrupt * Reduce idle Linux SMP CPU usage by up to 35%. * Increases CPU performance under load by up to 15%. * Improved specification conformance of the page table walker * Change access checks from ternary operator to if statements. * Checks for misaligned PPNs. * Disallow M-mode or S-mode from fetching from User pages. * Adds reserved PTE flag check: W or W|X. * Set READ flag for PTE X flag if mstatus.mxr is in effect. * Improves page walker comments and code readability . Testing Coverage * Linux Fedora SMP mstatus.FS scheduler test: pass * Linux Fedora SMP MTTCG tests (~22 hr GCC bootstrap): pass * spike_v1.9.1 bbl/linux-4.6.2 board test: pass * spike_v1.10 bbl/linux-4.14 board test: pass * virt bbl/linux-4.16-rc2 board test: pass * sifive_e board test (HiFive1 binaries): pass * sifive_u board test (HiFive Unleashed): pass * riscv-tests: pass * checkpatch: pass Kito Cheng (1): RISC-V: linux-user support for RVE ABI Michael Clark (27): RISC-V: Update address bits to support sv39 and sv48 RISC-V: Improve page table walker spec compliance RISC-V: Use atomic_cmpxchg to update PLIC bitmaps RISC-V: Simplify riscv_cpu_local_irqs_pending RISC-V: Allow setting and clearing multiple irqs RISC-V: Move non-ops from op_helper to cpu_helper RISC-V: Update CSR and interrupt definitions RISC-V: Implement modular CSR helper interface RISC-V: Implement atomic mip/sip CSR updates RISC-V: Implement existential predicates for CSRs RISC-V: Implement mstatus.TSR/TW/TVM RISC-V: Add public API for the CSR dispatch table RISC-V: Add hartid and \n to interrupt logging RISC-V: Use riscv prefix consistently on cpu helpers RISC-V: Replace __builtin_popcount with ctpop8 in PLIC RISC-V: Add missing free for plic_hart_config RISC-V: Allow interrupt controllers to claim interrupts RISC-V: Add misa to DisasContext RISC-V: Add misa.MAFD checks to translate RISC-V: Add misa runtime write support RISC-V: Fix CLINT timecmp low 32-bit writes RISC-V: Fix PLIC pending bitfield reads RISC-V: Enable second UART on sifive_e and sifive_u RISC-V: Remove unnecessary disassembler constraints elf: Add RISC-V PSABI ELF header defines RISC-V: Don't add NULL bootargs to device-tree RISC-V: Support separate firmware and kernel payload Richard Henderson (2): RISC-V: Split out mstatus_fs from tb_flags RISC-V: Mark mstatus.fs dirty disas/riscv.c | 138 - hw/riscv/Makefile.objs | 1 + hw/riscv/boot.c | 172 ++ hw/riscv/sifive_clint.c | 16 +- hw/riscv/sifive_e.c | 4 +- hw/riscv/sifive_plic.c | 72 +-- hw/riscv/sifive_u.c | 8 +- hw/riscv/spike.c| 6 +- hw/riscv/virt.c | 73 +-- include/elf.h | 8 + include/hw/riscv/boot.h | 30 ++ include/hw/riscv/sifive_plic.h | 1 - linux-user/riscv/cpu_loop.c | 14 +- linux-user
Re: [Qemu-devel] [PATCH] RISC-V: make it possible to alter default reset vector
On Tue, May 8, 2018 at 9:08 AM, Antony Pavlov <antonynpav...@gmail.com> wrote: > The RISC-V Instruction Set Manual, Volume II: > Privileged Architecture, Version 1.10 states > that upon reset the pc is set to > an implementation-defined reset vector > (see chapter 3.3 Reset). > > This patch makes it possible to alter default > reset vector by setting "rstvec" property > for TYPE_RISCV_HART_ARRAY. > This one needs some thought. We have already made it possible for a CPU class to override the reset vector, with consideration of this exact use case. The idea with the current approach is that you instantiate your specific CPU model (target/riscv) in the hardware machine model (hw/riscv) and the reset vector is a property of the CPU model you are using on your machine, which is how it is now. RISCVHartArray needs some work. First it is in 'hw/riscv' not 'target/riscv'. Secondly RISCVHartArray is commented as "heterogenous" (nevertheless the current implementation is "homogeneous" so apologies if this mislead you). The intention is that RISCVHartArray can construct a heterogenous array of different cpu models with some topology. The current shortcoming of homogeneity is in the current constructor which has only one model property. The SiFive U54-MC for example has 5 cores, a 'e51' no-MMU monitor core and 4 'u54' application cores e.g. "e51,u54,u54,u54,u54". The reset vector should be a property of each CPU, given they can be different models with different reset vectors. RISCVHartArray is a work in progress. Background on whay RISCVHartArray exists, as a placeholder that will be expanded to add support for configuration of "heterogeneous" core complexes. There is some periphery that needs to reflect on the CPU array properties. The SiFivePLIC memory layout is dependent on the cores in the " heterogenous" core complex. We haven't yet wired the PLIC to RISCVHartArray to get topoology information so it currently uses a mode list property "M,MS,MS,MS,MS". The idea is that it can eventuall reflect on topology configuration from the RISCVHartArray. Example: the 'e51' coreplex in the U54-MC does not support S-mode but the 'u54' coreplex application cores do, and the interrupt controller memory map is dependent on the topology. Our idea was to re-use RISCVHartArray logic for instantiating "heterogenous" core complexes as part of an SOC using some configuration e.g. an array of cpu models. It's likely that individial CPUs will have different modes, different reset vectors, different extensions, etc, etc. Q1. Which CPU are you actually modelling? Q2. Can you achieve your goal by defining your own CPU? I believe yes. Q3. What object should have the property of a reset vector? I believe the answer to Q3 is a specific cpu model, which is how we have it, not a heterogenous array where some CPUs may indeed have different reset vectors. Apologies if RISCVHartArray gave you the idea the harts where homogeneous. I believe moving a per cpu property onto the array is IMHO not the right thing to do. This leads into an RFC that I need to write on modelling dynamically reconfigurable hardware models that can be produced by SiFive's core generator. We want an interface that is kinder to the user than complex command line options... or ones deemed inappropriate such as inferring toplogy from a device-tree passed with -dtb e.g. an SOC class that instantiates its cpu cores and hardware blocks as defined in the device-tree at the given memory addresses with the given interrupt routing, etc (we understand that this may work for some simple configurations, but not for more complex configurations). The reset vector is a good example of a property that is not available in device-tree. In any case the RFC on configuration models for dynamically reconfigurable hardware will be another email, however this serves as context for it. i.e. where a property should be located. In fact we should fix the RISCVHartArray constructor or possible move the class altogether until we have a good model for constructing topology without hardcoding it in SOC structures, which for SiFive's use of QEMU, would be a combinatorial explosion, given the combinations of cores, extensions and blocks that can be generated by their core generator, and that we wish to model in QEMU or some derivative, if we have to maintain reconfigurable hardware support in a SiFive tree. I'll leave the RFC proper for another email. This is just an abstract. BTW - there are plently of others you can get to accept this patch ;-) See the 'Cc. Signed-off-by: Antony Pavlov <antonynpav...@gmail.com> > Cc: Michael Clark <m...@sifive.com> > Cc: Palmer Dabbelt <pal...@sifive.com> > Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> > Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> > Cc: Peter Crosthwaite <crosthwaite.pe.
Re: [Qemu-devel] [PATCH 4/9] target/riscv: Introduce cpu_riscv_get_fcsr
On Fri, May 11, 2018 at 3:52 PM, Richard Henderson < richard.hender...@linaro.org> wrote: > Cc: Michael Clark <m...@sifive.com> > Cc: Palmer Dabbelt <pal...@sifive.com> > Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> > Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> > Signed-off-by: Richard Henderson <richard.hender...@linaro.org> > I'm not against this change but it conflicts with changes in the riscv repo. I should post my patch queue to the list... We have made a somewhat medium sized change and have unraveled two monolithic switch statements out of csr_read_helper switch and csr_write_helper into clearly decomposed functions for modifying control and status registers, along with an interface to allow CPUs to hook custom control and status registers. This was done to support atomic read/modify/write CSRs which was not possible to achieve with the current helpers which separately called via the csr_read_helper followed by csr_write_helper. Given the only way to modify CSRs was via the switch statements, we needed to move them out to provide a mechanism for CSRs that wish to be truly atomic. e.g. 'mip'. The CSR functions are defined in The RISC-V Instruction Set Manual Volume I: User-Level ISA Document Version 2.2 as "atomic" instructions: - CSRRW (Atomic Read/Write CSR) - CSRRS (Atomic Read and Set Bits in CSR) - CSRRC (Atomic Read and Clear Bits in CSR) We have thus changed QEMU to allow truly atomic CSR implementations. The new implementation replaces the compiler doing compare/branch vs jump table switch codegen for a sparse CSR address space with a single array of function pointers. i.e. load, indirect jump. Along with this change we have also renamed functions in target/riscv to use riscv_ prefix and added a public interface to hook custom CSRs. The CSR changes will allow out of tree code to hook custom CSRs without needing to change target/riscv code. - riscv_cpu_ won over cpu_riscv_ given the number of functions conforming with the former riscv_ prefix and the desire for consistency in target/riscv In the riscv tree we now have riscv_csr_read(env, CSR_FCSR) and riscv_csr_write(env, CSR_FCSR, fcsr) as the method to read and write the composite. There is also a user in linux-user/riscv/signal.c that should probably use the new interface. We could change linux-user/riscv/signal.c to use your new interface however your interface only provides a read method and no write method, so the write interface remains in the (current) big CSR switch statement, leaving an inconsitency between the encapsulation of read and write. We currently have the new fcsr read and write encapsulated in static functions read_fcsr and write_fcsr in a new csr module (which should perhaps be called csr_helper.c). See: - https://github.com/riscv/riscv-qemu/commits/qemu-2.13-for-upstream - https://github.com/riscv/riscv-qemu/commit/0783ce5ea580552b1f8e2f16a3e3cc1af19db69b - https://github.com/riscv/riscv-qemu/commit/fa17549fbc726e83a3c163b1534c7465147c6718 > --- > target/riscv/cpu.h| 1 + > target/riscv/fpu_helper.c | 6 ++ > target/riscv/op_helper.c | 3 +-- > 3 files changed, 8 insertions(+), 2 deletions(-) > > diff --git a/target/riscv/cpu.h b/target/riscv/cpu.h > index 34abc383e3..f2bc243b95 100644 > --- a/target/riscv/cpu.h > +++ b/target/riscv/cpu.h > @@ -265,6 +265,7 @@ void QEMU_NORETURN do_raise_exception_err(CPURISCVState > *env, >uint32_t exception, uintptr_t > pc); > > target_ulong cpu_riscv_get_fflags(CPURISCVState *env); > +target_ulong cpu_riscv_get_fcsr(CPURISCVState *env); > void cpu_riscv_set_fflags(CPURISCVState *env, target_ulong); > > #define TB_FLAGS_MMU_MASK 3 > diff --git a/target/riscv/fpu_helper.c b/target/riscv/fpu_helper.c > index abbadead5c..41c7352115 100644 > --- a/target/riscv/fpu_helper.c > +++ b/target/riscv/fpu_helper.c > @@ -37,6 +37,12 @@ target_ulong cpu_riscv_get_fflags(CPURISCVState *env) > return hard; > } > > +target_ulong cpu_riscv_get_fcsr(CPURISCVState *env) > +{ > +return (cpu_riscv_get_fflags(env) << FSR_AEXC_SHIFT) > + | (env->frm << FSR_RD_SHIFT); > +} > + > void cpu_riscv_set_fflags(CPURISCVState *env, target_ulong hard) > { > int soft = 0; > diff --git a/target/riscv/op_helper.c b/target/riscv/op_helper.c > index 3abf52453c..fd2d8c0a9d 100644 > --- a/target/riscv/op_helper.c > +++ b/target/riscv/op_helper.c > @@ -423,8 +423,7 @@ target_ulong csr_read_helper(CPURISCVState *env, > target_ulong csrno) > return env->frm; > case CSR_FCSR: > validate_mstatus_fs(env, GETPC()); > -return (cpu_riscv_get_fflags(env) << FSR_AEXC_SHIFT) > -| (env->frm << FSR_RD_SHIFT); > +return cpu_riscv_get_fcsr(env); > /* rdtime/rdtimeh is trapped and emulated by bbl in system mode */ > #ifdef CONFIG_USER_ONLY > case CSR_TIME: > -- > 2.17.0 > >
Re: [Qemu-devel] [PATCH 5/9] target/riscv: Honor CPU_DUMP_FPU
On Sun, May 13, 2018 at 12:52 PM, Philippe Mathieu-Daudé <f4...@amsat.org> wrote: > On 05/11/2018 12:52 AM, Richard Henderson wrote: > > Cc: Michael Clark <m...@sifive.com> > > Cc: Palmer Dabbelt <pal...@sifive.com> > > Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> > > Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> > > Signed-off-by: Richard Henderson <richard.hender...@linaro.org> > > Reviewed-by: Philippe Mathieu-Daudé <f4...@amsat.org> Reviewed-by: Michael Clark <m...@sifive.com> > > --- > > target/riscv/cpu.c | 16 +++- > > 1 file changed, 11 insertions(+), 5 deletions(-) > > > > diff --git a/target/riscv/cpu.c b/target/riscv/cpu.c > > index 4e5a56d4e3..4612f324c9 100644 > > --- a/target/riscv/cpu.c > > +++ b/target/riscv/cpu.c > > @@ -199,6 +199,10 @@ static void riscv_cpu_dump_state(CPUState *cs, FILE > *f, > > int i; > > > > cpu_fprintf(f, " %s " TARGET_FMT_lx "\n", "pc ", env->pc); > > +if (flags & CPU_DUMP_FPU) { > > +cpu_fprintf(f, " %s " TARGET_FMT_lx "\n", "fcsr", > > +cpu_riscv_get_fcsr(env)); > > +} > > #ifndef CONFIG_USER_ONLY > > cpu_fprintf(f, " %s " TARGET_FMT_lx "\n", "mhartid ", env->mhartid); > > cpu_fprintf(f, " %s " TARGET_FMT_lx "\n", "mstatus ", env->mstatus); > > @@ -219,11 +223,13 @@ static void riscv_cpu_dump_state(CPUState *cs, > FILE *f, > > cpu_fprintf(f, "\n"); > > } > > } > > -for (i = 0; i < 32; i++) { > > -cpu_fprintf(f, " %s %016" PRIx64, > > -riscv_fpr_regnames[i], env->fpr[i]); > > -if ((i & 3) == 3) { > > -cpu_fprintf(f, "\n"); > > +if (flags & CPU_DUMP_FPU) { > > +for (i = 0; i < 32; i++) { > > +cpu_fprintf(f, " %s %016" PRIx64, > > +riscv_fpr_regnames[i], env->fpr[i]); > > +if ((i & 3) == 3) { > > +cpu_fprintf(f, "\n"); > > +} > > } > > } > > } > > >
Re: [Qemu-devel] [PATCH v3 4/7] hw/riscv/sifive_u: Set the soc device tree node as a simple-bus
On Tue, May 15, 2018 at 12:07 PM, Alistair Francis <alistair.fran...@wdc.com > wrote: > To allow Linux to ennumerate devices on the /soc/ node set it as a > "simple-bus". > > Signed-off-by: Alistair Francis <alistair.fran...@wdc.com> > Reviewed-by: Michael Clark <m...@sifive.com> > --- > hw/riscv/sifive_u.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/hw/riscv/sifive_u.c b/hw/riscv/sifive_u.c > index 3883d7ff9c..f438a72c27 100644 > --- a/hw/riscv/sifive_u.c > +++ b/hw/riscv/sifive_u.c > @@ -97,7 +97,7 @@ static void create_fdt(SiFiveUState *s, const struct > MemmapEntry *memmap, > > qemu_fdt_add_subnode(fdt, "/soc"); > qemu_fdt_setprop(fdt, "/soc", "ranges", NULL, 0); > -qemu_fdt_setprop_string(fdt, "/soc", "compatible", > "ucbbar,spike-bare-soc"); > +qemu_fdt_setprop_string(fdt, "/soc", "compatible", "simple-bus"); > qemu_fdt_setprop_cell(fdt, "/soc", "#size-cells", 0x2); > qemu_fdt_setprop_cell(fdt, "/soc", "#address-cells", 0x2); > > -- > 2.17.0 > >
Re: [Qemu-devel] [PATCH v3 2/7] hw/riscv/sifive_e: Create a SiFive E SoC object
On Tue, May 15, 2018 at 12:07 PM, Alistair Francis <alistair.fran...@wdc.com > wrote: > Signed-off-by: Alistair Francis <alistair.fran...@wdc.com> > Reviewed-by: Michael Clark <m...@sifive.com> > --- > hw/riscv/sifive_e.c | 97 +++-- > include/hw/riscv/sifive_e.h | 16 +- > 2 files changed, 86 insertions(+), 27 deletions(-) > > diff --git a/hw/riscv/sifive_e.c b/hw/riscv/sifive_e.c > index e4ecb7aa4b..384b456540 100644 > --- a/hw/riscv/sifive_e.c > +++ b/hw/riscv/sifive_e.c > @@ -102,18 +102,12 @@ static void riscv_sifive_e_init(MachineState > *machine) > SiFiveEState *s = g_new0(SiFiveEState, 1); > MemoryRegion *sys_mem = get_system_memory(); > MemoryRegion *main_mem = g_new(MemoryRegion, 1); > -MemoryRegion *mask_rom = g_new(MemoryRegion, 1); > -MemoryRegion *xip_mem = g_new(MemoryRegion, 1); > int i; > > -/* Initialize SOC */ > -object_initialize(>soc, sizeof(s->soc), TYPE_RISCV_HART_ARRAY); > +/* Initialize SoC */ > +object_initialize(>soc, sizeof(s->soc), TYPE_RISCV_E_SOC); > object_property_add_child(OBJECT(machine), "soc", OBJECT(>soc), >_abort); > -object_property_set_str(OBJECT(>soc), SIFIVE_E_CPU, "cpu-type", > -_abort); > -object_property_set_int(OBJECT(>soc), smp_cpus, "num-harts", > -_abort); > object_property_set_bool(OBJECT(>soc), true, "realized", > _abort); > > @@ -123,11 +117,57 @@ static void riscv_sifive_e_init(MachineState > *machine) > memory_region_add_subregion(sys_mem, > memmap[SIFIVE_E_DTIM].base, main_mem); > > +/* Mask ROM reset vector */ > +uint32_t reset_vec[2] = { > +0x204002b7,/* 0x1000: lui t0,0x20400 */ > +0x00028067,/* 0x1004: jr t0 */ > +}; > + > +/* copy in the reset vector in little_endian byte order */ > +for (i = 0; i < sizeof(reset_vec) >> 2; i++) { > +reset_vec[i] = cpu_to_le32(reset_vec[i]); > +} > +rom_add_blob_fixed_as("mrom.reset", reset_vec, sizeof(reset_vec), > + memmap[SIFIVE_E_MROM].base, > _space_memory); > + > +if (machine->kernel_filename) { > +load_kernel(machine->kernel_filename); > +} > +} > + > +static void riscv_sifive_e_soc_init(Object *obj) > +{ > +const struct MemmapEntry *memmap = sifive_e_memmap; > + > +SiFiveESoCState *s = RISCV_E_SOC(obj); > +MemoryRegion *sys_mem = get_system_memory(); > +MemoryRegion *mask_rom = g_new(MemoryRegion, 1); > + > +object_initialize(>cpus, sizeof(s->cpus), TYPE_RISCV_HART_ARRAY); > +object_property_add_child(obj, "cpus", OBJECT(>cpus), > + _abort); > +object_property_set_str(OBJECT(>cpus), SIFIVE_E_CPU, "cpu-type", > +_abort); > +object_property_set_int(OBJECT(>cpus), smp_cpus, "num-harts", > +_abort); > + > /* Mask ROM */ > memory_region_init_rom(mask_rom, NULL, "riscv.sifive.e.mrom", > memmap[SIFIVE_E_MROM].size, _fatal); > memory_region_add_subregion(sys_mem, > memmap[SIFIVE_E_MROM].base, mask_rom); > +} > + > +static void riscv_sifive_e_soc_realize(DeviceState *dev, Error **errp) > +{ > +const struct MemmapEntry *memmap = sifive_e_memmap; > + > +SiFiveESoCState *s = RISCV_E_SOC(dev); > +MemoryRegion *sys_mem = get_system_memory(); > +MemoryRegion *xip_mem = g_new(MemoryRegion, 1); > + > +object_property_set_bool(OBJECT(>cpus), true, "realized", > +_abort); > > /* MMIO */ > s->plic = sifive_plic_create(memmap[SIFIVE_E_PLIC].base, > @@ -171,23 +211,6 @@ static void riscv_sifive_e_init(MachineState > *machine) > memmap[SIFIVE_E_XIP].size, _fatal); > memory_region_set_readonly(xip_mem, true); > memory_region_add_subregion(sys_mem, memmap[SIFIVE_E_XIP].base, > xip_mem); > - > -/* Mask ROM reset vector */ > -uint32_t reset_vec[2] = { > -0x204002b7,/* 0x1000: lui t0,0x20400 */ > -0x00028067,/* 0x1004: jr t0 */ > -}; > - > -/* copy in the reset vector in little_endian byte order */ > -for (i = 0; i < sizeof(reset_vec) >> 2; i++) { > -reset_vec[i] = cpu_to_le32(reset_vec[i]); > -} > -rom_add_blob_fixed_as("mrom.reset", reset_vec, sizeof
Re: [Qemu-devel] [PATCH v2] riscv: remove define cpu_init()
On Wed, May 16, 2018 at 4:00 AM, Igor Mammedov <imamm...@redhat.com> wrote: > cpu_init() was removed since 2.12, so drop the define that is now unused. > > Signed-off-by: Igor Mammedov <imamm...@redhat.com> > Reviewed-by: Philippe Mathieu-Daudé <f4...@amsat.org> > Reviewed-by: Michael Clark <m...@sifive.com> --- > v2: > * refine commit message (Eric Blake <ebl...@redhat.com>) > --- > target/riscv/cpu.h | 1 - > 1 file changed, 1 deletion(-) > > diff --git a/target/riscv/cpu.h b/target/riscv/cpu.h > index 34abc38..0562a02 100644 > --- a/target/riscv/cpu.h > +++ b/target/riscv/cpu.h > @@ -251,7 +251,6 @@ int riscv_cpu_handle_mmu_fault(CPUState *cpu, vaddr > address, int size, > char *riscv_isa_string(RISCVCPU *cpu); > void riscv_cpu_list(FILE *f, fprintf_function cpu_fprintf); > > -#define cpu_init(cpu_model) cpu_generic_init(TYPE_RISCV_CPU, cpu_model) > #define cpu_signal_handler cpu_riscv_signal_handler > #define cpu_list riscv_cpu_list > #define cpu_mmu_index riscv_cpu_mmu_index > -- > 2.7.4 > >
Re: [Qemu-devel] [PATCH v2 20/27] target/riscv: Remove floatX_maybe_silence_nan from conversions
On Sat, May 12, 2018 at 12:43 PM, Richard Henderson < richard.hender...@linaro.org> wrote: > This is now handled properly by the generic softfloat code. > > Cc: Michael Clark <m...@sifive.com> > Cc: Palmer Dabbelt <pal...@sifive.com> > Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> > Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> > Signed-off-by: Richard Henderson <richard.hender...@linaro.org> > Reviewed-by: Michael Clark <m...@sifive.com> --- > target/riscv/fpu_helper.c | 6 ++ > 1 file changed, 2 insertions(+), 4 deletions(-) > > diff --git a/target/riscv/fpu_helper.c b/target/riscv/fpu_helper.c > index abbadead5c..fdb87d8d82 100644 > --- a/target/riscv/fpu_helper.c > +++ b/target/riscv/fpu_helper.c > @@ -279,14 +279,12 @@ uint64_t helper_fmax_d(CPURISCVState *env, uint64_t > frs1, uint64_t frs2) > > uint64_t helper_fcvt_s_d(CPURISCVState *env, uint64_t rs1) > { > -rs1 = float64_to_float32(rs1, >fp_status); > -return float32_maybe_silence_nan(rs1, >fp_status); > +return float64_to_float32(rs1, >fp_status); > } > > uint64_t helper_fcvt_d_s(CPURISCVState *env, uint64_t rs1) > { > -rs1 = float32_to_float64(rs1, >fp_status); > -return float64_maybe_silence_nan(rs1, >fp_status); > +return float32_to_float64(rs1, >fp_status); > } > > uint64_t helper_fsqrt_d(CPURISCVState *env, uint64_t frs1) > -- > 2.17.0 > >
Re: [Qemu-devel] [PATCH v2 5/7] hw/riscv/sifive_u: Set the interrupt controler number of interrupts
On Sat, May 12, 2018 at 11:28 AM, Alistair Francis <alistair.fran...@wdc.com > wrote: > Set the interrupt-controller ndev to the correct number taken from the > HiFive Unleashed board. > > Signed-off-by: Alistair Francis <alistair.fran...@wdc.com> > If you look at hw/riscv/virt.c we have removed hardcoding a few more constants using in the device tree. e.g. we allocate and resolve phandles vs hardcoding them. We can alwauys make a follow up commits to move some of these magic numbers into constants in the headers, preferably with enum vs #define. Reviewed-by: Michael Clark <m...@sifive.com> --- > hw/riscv/sifive_u.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/hw/riscv/sifive_u.c b/hw/riscv/sifive_u.c > index 859f43c6f9..50389cdc90 100644 > --- a/hw/riscv/sifive_u.c > +++ b/hw/riscv/sifive_u.c > @@ -187,7 +187,7 @@ static void create_fdt(SiFiveUState *s, const struct > MemmapEntry *memmap, > 0x0, memmap[SIFIVE_U_PLIC].size); > qemu_fdt_setprop_string(fdt, nodename, "reg-names", "control"); > qemu_fdt_setprop_cell(fdt, nodename, "riscv,max-priority", 7); > -qemu_fdt_setprop_cell(fdt, nodename, "riscv,ndev", 4); > +qemu_fdt_setprop_cell(fdt, nodename, "riscv,ndev", 0x35); > qemu_fdt_setprop_cells(fdt, nodename, "phandle", 2); > qemu_fdt_setprop_cells(fdt, nodename, "linux,phandle", 2); > plic_phandle = qemu_fdt_get_phandle(fdt, nodename); > -- > 2.17.0 > >
Re: [Qemu-devel] [PATCH v2 7/7] hw/riscv/sifive_u: Connect the Cadence GEM Ethernet device
On Sat, May 12, 2018 at 11:28 AM, Alistair Francis <alistair.fran...@wdc.com > wrote: > Connect the Cadence GEM ethernet device. This also requires us to > expose the plic interrupt lines. > > Signed-off-by: Alistair Francis <alistair.fran...@wdc.com> > Reviewed-by: Michael Clark <m...@sifive.com> > --- > default-configs/riscv32-softmmu.mak | 1 + > default-configs/riscv64-softmmu.mak | 1 + > hw/riscv/sifive_u.c | 50 + > include/hw/riscv/sifive_u.h | 9 -- > 4 files changed, 59 insertions(+), 2 deletions(-) > > diff --git a/default-configs/riscv32-softmmu.mak > b/default-configs/riscv32-softmmu.mak > index f9e742120c..9a1c42e8b2 100644 > --- a/default-configs/riscv32-softmmu.mak > +++ b/default-configs/riscv32-softmmu.mak > @@ -2,3 +2,4 @@ > > CONFIG_SERIAL=y > CONFIG_VIRTIO=y > +CONFIG_CADENCE=y > diff --git a/default-configs/riscv64-softmmu.mak > b/default-configs/riscv64-softmmu.mak > index f9e742120c..9a1c42e8b2 100644 > --- a/default-configs/riscv64-softmmu.mak > +++ b/default-configs/riscv64-softmmu.mak > @@ -2,3 +2,4 @@ > > CONFIG_SERIAL=y > CONFIG_VIRTIO=y > +CONFIG_CADENCE=y > diff --git a/hw/riscv/sifive_u.c b/hw/riscv/sifive_u.c > index 540d53bf2f..fff36bd04d 100644 > --- a/hw/riscv/sifive_u.c > +++ b/hw/riscv/sifive_u.c > @@ -60,8 +60,11 @@ static const struct MemmapEntry { > [SIFIVE_U_UART0] ={ 0x10013000, 0x1000 }, > [SIFIVE_U_UART1] ={ 0x10023000, 0x1000 }, > [SIFIVE_U_DRAM] = { 0x8000,0x0 }, > +[SIFIVE_U_GEM] = { 0x100900FC, 0x2000 }, > }; > > +#define GEM_REVISION0x10070109 > + > static uint64_t load_kernel(const char *kernel_filename) > { > uint64_t kernel_entry, kernel_high; > @@ -194,6 +197,27 @@ static void create_fdt(SiFiveUState *s, const struct > MemmapEntry *memmap, > g_free(cells); > g_free(nodename); > > +nodename = g_strdup_printf("/soc/ethernet@%lx", > +(long)memmap[SIFIVE_U_GEM].base); > +qemu_fdt_add_subnode(fdt, nodename); > +qemu_fdt_setprop_string(fdt, nodename, "compatible", "cdns,macb"); > +qemu_fdt_setprop_cells(fdt, nodename, "reg", > +0x0, memmap[SIFIVE_U_GEM].base, > +0x0, memmap[SIFIVE_U_GEM].size); > +qemu_fdt_setprop_string(fdt, nodename, "reg-names", "control"); > +qemu_fdt_setprop_string(fdt, nodename, "phy-mode", "gmii"); > +qemu_fdt_setprop_cells(fdt, nodename, "interrupt-parent", > plic_phandle); > +qemu_fdt_setprop_cells(fdt, nodename, "interrupts", SIFIVE_U_GEM_IRQ); > +qemu_fdt_setprop_cells(fdt, nodename, "#address-cells", 1); > +qemu_fdt_setprop_cells(fdt, nodename, "#size-cells", 0); > +g_free(nodename); > + > +nodename = g_strdup_printf("/soc/ethernet@%lx/ethernet-phy@0", > +(long)memmap[SIFIVE_U_GEM].base); > +qemu_fdt_add_subnode(fdt, nodename); > +qemu_fdt_setprop_cells(fdt, nodename, "reg", 0x0); > +g_free(nodename); > + > nodename = g_strdup_printf("/soc/uart@%lx", > (long)memmap[SIFIVE_U_UART0].base); > qemu_fdt_add_subnode(fdt, nodename); > @@ -296,6 +320,9 @@ static void riscv_sifive_u54_init(Object *obj) > memmap[SIFIVE_U_MROM].size, _fatal); > memory_region_add_subregion(system_memory, > memmap[SIFIVE_U_MROM].base, > mask_rom); > + > +object_initialize(>gem, sizeof(s->gem), TYPE_CADENCE_GEM); > +qdev_set_parent_bus(DEVICE(>gem), sysbus_get_default()); > } > > static void riscv_sifive_u54_realize(DeviceState *dev, Error **errp) > @@ -303,6 +330,10 @@ static void riscv_sifive_u54_realize(DeviceState > *dev, Error **errp) > SiFiveU54State *s = RISCV_U54_SOC(dev); > const struct MemmapEntry *memmap = sifive_u_memmap; > MemoryRegion *system_memory = get_system_memory(); > +qemu_irq plic_gpios[SIFIVE_U_PLIC_NUM_SOURCES]; > +int i; > +Error *err = NULL; > +NICInfo *nd = _table[0]; > > object_property_set_bool(OBJECT(>cpus), true, "realized", > _abort); > @@ -327,6 +358,25 @@ static void riscv_sifive_u54_realize(DeviceState > *dev, Error **errp) > sifive_clint_create(memmap[SIFIVE_U_CLINT].base, > memmap[SIFIVE_U_CLINT].size, smp_cpus, > SIFIVE_SIP_BASE, SIFIVE_TIMECMP_BASE, SIFIVE_TIME_BASE); > + > +for (i = 0; i < SIFIVE_U_PLIC_NUM_SOURCES; i++) { > +plic_gpios[i] = qdev_get_g
Re: [Qemu-devel] [PATCH v2 6/7] hw/riscv/sifive_u: Move the uart device tree node under /soc/
On Sat, May 12, 2018 at 11:28 AM, Alistair Francis <alistair.fran...@wdc.com > wrote: > Signed-off-by: Alistair Francis <alistair.fran...@wdc.com> > Reviewed-by: Michael Clark <m...@sifive.com> > --- > hw/riscv/sifive_u.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/hw/riscv/sifive_u.c b/hw/riscv/sifive_u.c > index 50389cdc90..540d53bf2f 100644 > --- a/hw/riscv/sifive_u.c > +++ b/hw/riscv/sifive_u.c > @@ -194,7 +194,7 @@ static void create_fdt(SiFiveUState *s, const struct > MemmapEntry *memmap, > g_free(cells); > g_free(nodename); > > -nodename = g_strdup_printf("/uart@%lx", > +nodename = g_strdup_printf("/soc/uart@%lx", > (long)memmap[SIFIVE_U_UART0].base); > qemu_fdt_add_subnode(fdt, nodename); > qemu_fdt_setprop_string(fdt, nodename, "compatible", "sifive,uart0"); > -- > 2.17.0 > >
Re: [Qemu-devel] [PATCH v2 1/7] hw/riscv/sifive_u: Create a U54 SoC object
On Sat, May 12, 2018 at 11:27 AM, Alistair Francis <alistair.fran...@wdc.com > wrote: > Create a SiFive Unleashed U54 SoC and use that in the sifive_u machine. > > We leave the SoC, RAM, device tree and reset/fdt loading as part of the > machine. All the other device creation has been moved to the SoC. > > Signed-off-by: Alistair Francis <alistair.fran...@wdc.com> > This board currently supports U34 and U54 cores. The SOC names have 3 digits. e.g. FU540. We are trying to keep the SiFive U Series board generic. - Global search and replace the prefix "riscv_sifive_u54" to "riscv_sifive_u_soc" - Rename TYPE_RISCV_U54_SOC to TYPE_RISCV_U_SOC - Rename SiFiveU54State SiFiveUSOC (I don't think we need the State suffix for the SOC) Assuming we can do the renames to keep the SiFive U Series machine/SOC general: Reviewed-by: Michael Clark <m...@sifive.com> --- > hw/riscv/sifive_u.c | 90 - > include/hw/riscv/sifive_u.h | 16 ++- > 2 files changed, 82 insertions(+), 24 deletions(-) > > diff --git a/hw/riscv/sifive_u.c b/hw/riscv/sifive_u.c > index c05dcbba95..26155e932d 100644 > --- a/hw/riscv/sifive_u.c > +++ b/hw/riscv/sifive_u.c > @@ -116,10 +116,10 @@ static void create_fdt(SiFiveUState *s, const struct > MemmapEntry *memmap, > qemu_fdt_setprop_cell(fdt, "/cpus", "#size-cells", 0x0); > qemu_fdt_setprop_cell(fdt, "/cpus", "#address-cells", 0x1); > > -for (cpu = s->soc.num_harts - 1; cpu >= 0; cpu--) { > +for (cpu = s->soc.cpus.num_harts - 1; cpu >= 0; cpu--) { > nodename = g_strdup_printf("/cpus/cpu@%d", cpu); > char *intc = g_strdup_printf("/cpus/cpu@%d/interrupt-controller", > cpu); > -char *isa = riscv_isa_string(>soc.harts[cpu]); > +char *isa = riscv_isa_string(>soc.cpus.harts[cpu]); > qemu_fdt_add_subnode(fdt, nodename); > qemu_fdt_setprop_cell(fdt, nodename, "clock-frequency", >SIFIVE_U_CLOCK_FREQ); > @@ -140,8 +140,8 @@ static void create_fdt(SiFiveUState *s, const struct > MemmapEntry *memmap, > g_free(nodename); > } > > -cells = g_new0(uint32_t, s->soc.num_harts * 4); > -for (cpu = 0; cpu < s->soc.num_harts; cpu++) { > +cells = g_new0(uint32_t, s->soc.cpus.num_harts * 4); > +for (cpu = 0; cpu < s->soc.cpus.num_harts; cpu++) { > nodename = > g_strdup_printf("/cpus/cpu@%d/interrupt-controller", cpu); > uint32_t intc_phandle = qemu_fdt_get_phandle(fdt, nodename); > @@ -159,12 +159,12 @@ static void create_fdt(SiFiveUState *s, const struct > MemmapEntry *memmap, > 0x0, memmap[SIFIVE_U_CLINT].base, > 0x0, memmap[SIFIVE_U_CLINT].size); > qemu_fdt_setprop(fdt, nodename, "interrupts-extended", > -cells, s->soc.num_harts * sizeof(uint32_t) * 4); > +cells, s->soc.cpus.num_harts * sizeof(uint32_t) * 4); > g_free(cells); > g_free(nodename); > > -cells = g_new0(uint32_t, s->soc.num_harts * 4); > -for (cpu = 0; cpu < s->soc.num_harts; cpu++) { > +cells = g_new0(uint32_t, s->soc.cpus.num_harts * 4); > +for (cpu = 0; cpu < s->soc.cpus.num_harts; cpu++) { > nodename = > g_strdup_printf("/cpus/cpu@%d/interrupt-controller", cpu); > uint32_t intc_phandle = qemu_fdt_get_phandle(fdt, nodename); > @@ -181,7 +181,7 @@ static void create_fdt(SiFiveUState *s, const struct > MemmapEntry *memmap, > qemu_fdt_setprop_string(fdt, nodename, "compatible", "riscv,plic0"); > qemu_fdt_setprop(fdt, nodename, "interrupt-controller", NULL, 0); > qemu_fdt_setprop(fdt, nodename, "interrupts-extended", > -cells, s->soc.num_harts * sizeof(uint32_t) * 4); > +cells, s->soc.cpus.num_harts * sizeof(uint32_t) * 4); > qemu_fdt_setprop_cells(fdt, nodename, "reg", > 0x0, memmap[SIFIVE_U_PLIC].base, > 0x0, memmap[SIFIVE_U_PLIC].size); > @@ -217,17 +217,12 @@ static void riscv_sifive_u_init(MachineState > *machine) > SiFiveUState *s = g_new0(SiFiveUState, 1); > MemoryRegion *system_memory = get_system_memory(); > MemoryRegion *main_mem = g_new(MemoryRegion, 1); > -MemoryRegion *mask_rom = g_new(MemoryRegion, 1); > int i; > > -/* Initialize SOC */ > -object_initialize(>soc, sizeof(s->soc), TYPE_RISCV_HART_ARRAY); > +/* Initialize SoC */ > +object_initialize(>soc, sizeof(s->soc), TYPE_RISCV_U54_SOC); > object_property_add_child(OBJECT(machine), &
Re: [Qemu-devel] [PULL 0/3] RISC-V: QEMU 2.13 Minor Fixes
On Wed, 9 May 2018 at 11:14 PM, Peter Maydell <peter.mayd...@linaro.org> wrote: > On 8 May 2018 at 23:05, Michael Clark <m...@sifive.com> wrote: > > > > > > On Wed, May 9, 2018 at 8:49 AM, Peter Maydell <peter.mayd...@linaro.org> > > wrote: > >> > >> On 8 May 2018 at 21:07, Michael Clark <m...@sifive.com> wrote: > >> > The following changes since commit > >> > c8b7e627b4269a3bc3ae41d9f420547a47e6d9b9: > >> > > >> > Merge remote-tracking branch > 'remotes/ericb/tags/pull-nbd-2018-05-04' > >> > into staging (2018-05-04 14:42:46 +0100) > >> > > >> > are available in the git repository at: > >> > > >> > https://github.com/riscv/riscv-qemu.git > >> > tags/riscv-qemu-2.13-minor-fixes-2 > >> > > >> > for you to fetch changes up to > fc28db142ea9331d24d08a6418efa93e258a6876: > >> > > >> > riscv: requires libfdt (2018-05-09 07:57:57 +1200) > >> > > >> > > >> > RISC-V: QEMU 2.13 Minor Fixes > >> > > >> > * Require libfdt when configuring for 'riscv*-softmmu' > >> > * Increase HTIF priority and allow zero base address > >> > > >> > Consider backporting to stable branch for 2.12.1 > >> > >> Ah, if you want something flagged to be backported to > >> the stable branch, the way to do that is to include > >> the line "Cc: qemu-sta...@nongnu.org" in the commit > >> message. We use the same approach for flagging things > >> for stable as the Linux kernel. (Individual submaintainers > >> are generally in the best position to decide whether > >> something should go into stable, and this system leaves > >> that choice up to them. A patch submitter can also suggest > >> something for stable when you submit the initial patch for > >> review by adding the Cc: line and cc'ing qemu-stable on the > >> patchmail.) > > > > > > Okay. Given the 'Cc is missing in the commit (only on the cover) I could > > send a separate patch. > > > > Or I can respin with the 'Cc on the configure libfdt dependency fix > commit > > and remove the backport message from the tag... > > Tag messages are just informational (we use them to help > in writing up the changelog for releases), mostly. So I > would recommend respinning the pullreq with the Cc: line > in the relevant commit(s). I respun the PR with a Cc: qemu-stable in the configure libfdt dependency patch. You should have received a PR with the tag ‘riscv-qemu-2.13-minor-fixes-3’ Thanks, Michael > >
[Qemu-devel] [PULL 3/3] riscv: requires libfdt
From: KONRAD Frederic <frederic.kon...@adacore.com> When compiling on a machine without libfdt installed the configure script should try to get libfdt from the git or should die because otherwise CONFIG_LIBFDT is not set and the build process end in an error in the link phase.. eg: hw/riscv/virt.o: In function `riscv_virt_board_init': qemu/src/hw/riscv/virt.c:317: undefined reference to `qemu_fdt_setprop_cell' qemu/src/hw/riscv/virt.c:319: undefined reference to `qemu_fdt_setprop_cell' qemu/src/hw/riscv/virt.c:345: undefined reference to `qemu_fdt_dumpdtb' collect2: error: ld returned 1 exit status make[1]: *** [qemu-system-riscv64] Error 1 make: *** [subdir-riscv64-softmmu] Error 2 Cc: qemu-sta...@nongnu.org Reviewed-by: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Reviewed-by: Philippe Mathieu-Daudé <f4...@amsat.org> Reviewed-by: Michael Clark <m...@sifive.com> Signed-off-by: KONRAD Frederic <frederic.kon...@adacore.com> Signed-off-by: Michael Clark <m...@sifive.com> Message-Id: <1525360636-18229-4-git-send-email-frederic.kon...@adacore.com> --- configure | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/configure b/configure index 1443422e83f1..479e9a787e6e 100755 --- a/configure +++ b/configure @@ -3761,7 +3761,7 @@ fi fdt_required=no for target in $target_list; do case $target in - aarch64*-softmmu|arm*-softmmu|ppc*-softmmu|microblaze*-softmmu|mips64el-softmmu) + aarch64*-softmmu|arm*-softmmu|ppc*-softmmu|microblaze*-softmmu|mips64el-softmmu|riscv*-softmmu) fdt_required=yes ;; esac -- 2.7.0
[Qemu-devel] [PULL 0/3] RISC-V: QEMU 2.13 Minor Fixes
The following changes since commit c8b7e627b4269a3bc3ae41d9f420547a47e6d9b9: Merge remote-tracking branch 'remotes/ericb/tags/pull-nbd-2018-05-04' into staging (2018-05-04 14:42:46 +0100) are available in the git repository at: https://github.com/riscv/riscv-qemu.git tags/riscv-qemu-2.13-minor-fixes-3 for you to fetch changes up to a666409f0df5dce113a5bd2c4c144a0792f2a4a3: riscv: requires libfdt (2018-05-09 22:14:28 +1200) RISC-V: QEMU 2.13 Minor Fixes * Require libfdt when configuring for 'riscv*-softmmu' * Increase HTIF priority and allow zero base address KONRAD Frederic (3): riscv: spike: allow base == 0 riscv: htif: increase the priority of the htif subregion riscv: requires libfdt configure | 2 +- hw/riscv/riscv_htif.c | 12 2 files changed, 9 insertions(+), 5 deletions(-) -- 2.7.0
[Qemu-devel] [PULL 1/3] riscv: spike: allow base == 0
From: KONRAD Frederic <frederic.kon...@adacore.com> The sanity check on base doesn't allow htif to be mapped @0. Check if the symbol exists instead so we can map it where we want. Reviewed-by: Michael Clark <m...@sifive.com> Signed-off-by: KONRAD Frederic <frederic.kon...@adacore.com> Signed-off-by: Michael Clark <m...@sifive.com> Message-Id: <1525360636-18229-2-git-send-email-frederic.kon...@adacore.com> --- hw/riscv/riscv_htif.c | 7 +-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/hw/riscv/riscv_htif.c b/hw/riscv/riscv_htif.c index 3e17f3025106..be252ec8cce9 100644 --- a/hw/riscv/riscv_htif.c +++ b/hw/riscv/riscv_htif.c @@ -41,17 +41,20 @@ } while (0) static uint64_t fromhost_addr, tohost_addr; +static int address_symbol_set; void htif_symbol_callback(const char *st_name, int st_info, uint64_t st_value, -uint64_t st_size) + uint64_t st_size) { if (strcmp("fromhost", st_name) == 0) { +address_symbol_set |= 1; fromhost_addr = st_value; if (st_size != 8) { error_report("HTIF fromhost must be 8 bytes"); exit(1); } } else if (strcmp("tohost", st_name) == 0) { +address_symbol_set |= 2; tohost_addr = st_value; if (st_size != 8) { error_report("HTIF tohost must be 8 bytes"); @@ -248,7 +251,7 @@ HTIFState *htif_mm_init(MemoryRegion *address_space, MemoryRegion *main_mem, qemu_chr_fe_init(>chr, chr, _abort); qemu_chr_fe_set_handlers(>chr, htif_can_recv, htif_recv, htif_event, htif_be_change, s, NULL, true); -if (base) { +if (address_symbol_set == 3) { memory_region_init_io(>mmio, NULL, _mm_ops, s, TYPE_HTIF_UART, size); memory_region_add_subregion(address_space, base, >mmio); -- 2.7.0
[Qemu-devel] [PULL 2/3] riscv: htif: increase the priority of the htif subregion
From: KONRAD Frederic <frederic.kon...@adacore.com> The htif device is supposed to be mapped over an other subregion. So increase its priority to one to avoid any conflict. Here is the output of info mtree: Before: (qemu) info mtree address-space: memory - (prio 0, i/o): system -000f (prio 0, i/o): riscv.htif.uart -00011fff (prio 0, ram): riscv.spike.bootrom 0200-0200 (prio 0, i/o): riscv.sifive.clint 8000-87ff (prio 0, ram): riscv.spike.ram address-space: I/O - (prio 0, i/o): io address-space: cpu-memory-0 - (prio 0, i/o): system -000f (prio 0, i/o): riscv.htif.uart -00011fff (prio 0, ram): riscv.spike.bootrom 0200-0200 (prio 0, i/o): riscv.sifive.clint 8000-87ff (prio 0, ram): riscv.spike.ram After: (qemu) info mtree address-space: memory - (prio 0, i/o): system -000f (prio 1, i/o): riscv.htif.uart -00011fff (prio 0, ram): riscv.spike.bootrom 0200-0200 (prio 0, i/o): riscv.sifive.clint 8000-87ff (prio 0, ram): riscv.spike.ram address-space: I/O - (prio 0, i/o): io address-space: cpu-memory-0 - (prio 0, i/o): system -000f (prio 1, i/o): riscv.htif.uart -00011fff (prio 0, ram): riscv.spike.bootrom 0200-0200 (prio 0, i/o): riscv.sifive.clint 8000-87ff (prio 0, ram): riscv.spike.ram Reviewed-by: Michael Clark <m...@sifive.com> Signed-off-by: KONRAD Frederic <frederic.kon...@adacore.com> Signed-off-by: Michael Clark <m...@sifive.com> Message-Id: <1525360636-18229-3-git-send-email-frederic.kon...@adacore.com> --- hw/riscv/riscv_htif.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/hw/riscv/riscv_htif.c b/hw/riscv/riscv_htif.c index be252ec8cce9..f73512941fb6 100644 --- a/hw/riscv/riscv_htif.c +++ b/hw/riscv/riscv_htif.c @@ -253,8 +253,9 @@ HTIFState *htif_mm_init(MemoryRegion *address_space, MemoryRegion *main_mem, htif_be_change, s, NULL, true); if (address_symbol_set == 3) { memory_region_init_io(>mmio, NULL, _mm_ops, s, -TYPE_HTIF_UART, size); -memory_region_add_subregion(address_space, base, >mmio); + TYPE_HTIF_UART, size); +memory_region_add_subregion_overlap(address_space, base, +>mmio, 1); } return s; -- 2.7.0
[Qemu-devel] [PATCH v1 5/6] target/riscv: convert to TranslatorOps
From: "Emilio G. Cota" <c...@braap.org> Reviewed-by: Richard Henderson <richard.hender...@linaro.org> Cc: Michael Clark <m...@sifive.com> Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Signed-off-by: Emilio G. Cota <c...@braap.org> --- target/riscv/translate.c | 158 --- 1 file changed, 80 insertions(+), 78 deletions(-) diff --git a/target/riscv/translate.c b/target/riscv/translate.c index 68979abfd7ed..1788668c6f28 100644 --- a/target/riscv/translate.c +++ b/target/riscv/translate.c @@ -1837,78 +1837,71 @@ static void decode_opc(CPURISCVState *env, DisasContext *ctx) } } -void gen_intermediate_code(CPUState *cs, TranslationBlock *tb) +static void riscv_tr_init_disas_context(DisasContextBase *dcbase, CPUState *cs) { -CPURISCVState *env = cs->env_ptr; -DisasContext ctx; -target_ulong page_start; -int num_insns; -int max_insns; - -ctx.base.pc_first = tb->pc; -ctx.base.pc_next = ctx.base.pc_first; -/* once we have GDB, the rest of the translate.c implementation should be - ready for singlestep */ -ctx.base.singlestep_enabled = cs->singlestep_enabled; -ctx.base.tb = tb; -ctx.base.is_jmp = DISAS_NEXT; - -page_start = ctx.base.pc_first & TARGET_PAGE_MASK; -ctx.pc_succ_insn = ctx.base.pc_first; -ctx.flags = tb->flags; -ctx.mem_idx = tb->flags & TB_FLAGS_MMU_MASK; -ctx.frm = -1; /* unknown rounding mode */ - -num_insns = 0; -max_insns = tb_cflags(ctx.base.tb) & CF_COUNT_MASK; -if (max_insns == 0) { -max_insns = CF_COUNT_MASK; -} -if (max_insns > TCG_MAX_INSNS) { -max_insns = TCG_MAX_INSNS; -} -gen_tb_start(tb); +DisasContext *ctx = container_of(dcbase, DisasContext, base); -while (ctx.base.is_jmp == DISAS_NEXT) { -tcg_gen_insn_start(ctx.base.pc_next); -num_insns++; +ctx->pc_succ_insn = ctx->base.pc_first; +ctx->flags = ctx->base.tb->flags; +ctx->mem_idx = ctx->base.tb->flags & TB_FLAGS_MMU_MASK; +ctx->frm = -1; /* unknown rounding mode */ +} -if (unlikely(cpu_breakpoint_test(cs, ctx.base.pc_next, BP_ANY))) { -tcg_gen_movi_tl(cpu_pc, ctx.base.pc_next); -ctx.base.is_jmp = DISAS_NORETURN; -gen_exception_debug(); -/* The address covered by the breakpoint must be included in - [tb->pc, tb->pc + tb->size) in order to for it to be - properly cleared -- thus we increment the PC here so that - the logic setting tb->size below does the right thing. */ -ctx.base.pc_next += 4; -goto done_generating; -} +static void riscv_tr_tb_start(DisasContextBase *db, CPUState *cpu) +{ +} -if (num_insns == max_insns && (tb_cflags(ctx.base.tb) & CF_LAST_IO)) { -gen_io_start(); -} +static void riscv_tr_insn_start(DisasContextBase *dcbase, CPUState *cpu) +{ +DisasContext *ctx = container_of(dcbase, DisasContext, base); + +tcg_gen_insn_start(ctx->base.pc_next); +} + +static bool riscv_tr_breakpoint_check(DisasContextBase *dcbase, CPUState *cpu, + const CPUBreakpoint *bp) +{ +DisasContext *ctx = container_of(dcbase, DisasContext, base); + +tcg_gen_movi_tl(cpu_pc, ctx->base.pc_next); +ctx->base.is_jmp = DISAS_NORETURN; +gen_exception_debug(); +/* The address covered by the breakpoint must be included in + [tb->pc, tb->pc + tb->size) in order to for it to be + properly cleared -- thus we increment the PC here so that + the logic setting tb->size below does the right thing. */ +ctx->base.pc_next += 4; +return true; +} + + +static void riscv_tr_translate_insn(DisasContextBase *dcbase, CPUState *cpu) +{ +DisasContext *ctx = container_of(dcbase, DisasContext, base); +CPURISCVState *env = cpu->env_ptr; -ctx.opcode = cpu_ldl_code(env, ctx.base.pc_next); -decode_opc(env, ); -ctx.base.pc_next = ctx.pc_succ_insn; - -if (ctx.base.is_jmp == DISAS_NEXT && -(cs->singlestep_enabled || - ctx.base.pc_next - page_start >= TARGET_PAGE_SIZE || - tcg_op_buf_full() || - num_insns >= max_insns || - singlestep)) { -ctx.base.is_jmp = DISAS_TOO_MANY; +ctx->opcode = cpu_ldl_code(env, ctx->base.pc_next); +decode_opc(env, ctx); +ctx->base.pc_next = ctx->pc_succ_insn; + +if (ctx->base.is_jmp == DISAS_NEXT) { +target_ulong page_start; + +page_start = ctx->base.pc_first & TARGET_PAGE_MASK; +if (ctx->base.pc_next - page_start >= TARGET_PAGE_SIZE) { +
[Qemu-devel] [PATCH v1 6/6] target/riscv: add misa to DisasContext
gen methods should access state from DisasContext. Add misa field to the DisasContext struct and remove CPURISCVState argument from all gen methods. [Rebased against github.com/cota/qemu/tree/trloop-conv-v3] Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Cc: Alistair Francis <alistair.fran...@wdc.com> Cc: Emilio G. Cota <c...@braap.org> Signed-off-by: Michael Clark <m...@sifive.com> --- target/riscv/translate.c | 77 ++-- 1 file changed, 42 insertions(+), 35 deletions(-) diff --git a/target/riscv/translate.c b/target/riscv/translate.c index 1788668c6f28..507d74186293 100644 --- a/target/riscv/translate.c +++ b/target/riscv/translate.c @@ -45,6 +45,7 @@ typedef struct DisasContext { target_ulong pc_succ_insn; uint32_t opcode; uint32_t flags; +uint32_t misa; uint32_t mem_idx; /* Remember the rounding mode encoded in the previous fp instruction, which we have already installed into env->fp_status. Or -1 for @@ -74,6 +75,11 @@ static const int tcg_memop_lookup[8] = { #define CASE_OP_32_64(X) case X #endif +static inline bool has_ext(DisasContext *ctx, uint32_t ext) +{ +return ctx->misa & ext; +} + static void generate_exception(DisasContext *ctx, int excp) { tcg_gen_movi_tl(cpu_pc, ctx->base.pc_next); @@ -505,14 +511,13 @@ static void gen_arith_imm(DisasContext *ctx, uint32_t opc, int rd, tcg_temp_free(source1); } -static void gen_jal(CPURISCVState *env, DisasContext *ctx, int rd, -target_ulong imm) +static void gen_jal(DisasContext *ctx, int rd, target_ulong imm) { target_ulong next_pc; /* check misaligned: */ next_pc = ctx->base.pc_next + imm; -if (!riscv_has_ext(env, RVC)) { +if (!has_ext(ctx, RVC)) { if ((next_pc & 0x3) != 0) { gen_exception_inst_addr_mis(ctx); return; @@ -526,8 +531,8 @@ static void gen_jal(CPURISCVState *env, DisasContext *ctx, int rd, ctx->base.is_jmp = DISAS_NORETURN; } -static void gen_jalr(CPURISCVState *env, DisasContext *ctx, uint32_t opc, - int rd, int rs1, target_long imm) +static void gen_jalr(DisasContext *ctx, uint32_t opc, int rd, int rs1, + target_long imm) { /* no chaining with JALR */ TCGLabel *misaligned = NULL; @@ -539,7 +544,7 @@ static void gen_jalr(CPURISCVState *env, DisasContext *ctx, uint32_t opc, tcg_gen_addi_tl(cpu_pc, cpu_pc, imm); tcg_gen_andi_tl(cpu_pc, cpu_pc, (target_ulong)-2); -if (!riscv_has_ext(env, RVC)) { +if (!has_ext(ctx, RVC)) { misaligned = gen_new_label(); tcg_gen_andi_tl(t0, cpu_pc, 0x2); tcg_gen_brcondi_tl(TCG_COND_NE, t0, 0x0, misaligned); @@ -564,8 +569,8 @@ static void gen_jalr(CPURISCVState *env, DisasContext *ctx, uint32_t opc, tcg_temp_free(t0); } -static void gen_branch(CPURISCVState *env, DisasContext *ctx, uint32_t opc, - int rs1, int rs2, target_long bimm) +static void gen_branch(DisasContext *ctx, uint32_t opc, int rs1, int rs2, + target_long bimm) { TCGLabel *l = gen_new_label(); TCGv source1, source2; @@ -602,7 +607,7 @@ static void gen_branch(CPURISCVState *env, DisasContext *ctx, uint32_t opc, gen_goto_tb(ctx, 1, ctx->pc_succ_insn); gen_set_label(l); /* branch taken */ -if (!riscv_has_ext(env, RVC) && ((ctx->base.pc_next + bimm) & 0x3)) { +if (!has_ext(ctx, RVC) && ((ctx->base.pc_next + bimm) & 0x3)) { /* misaligned */ gen_exception_inst_addr_mis(ctx); } else { @@ -1305,8 +1310,8 @@ static void gen_fp_arith(DisasContext *ctx, uint32_t opc, int rd, } } -static void gen_system(CPURISCVState *env, DisasContext *ctx, uint32_t opc, - int rd, int rs1, int csr) +static void gen_system(DisasContext *ctx, uint32_t opc, int rd, int rs1, + int csr) { TCGv source1, csr_store, dest, rs1_pass, imm_rs1; source1 = tcg_temp_new(); @@ -1348,7 +1353,7 @@ static void gen_system(CPURISCVState *env, DisasContext *ctx, uint32_t opc, gen_exception_illegal(ctx); break; case 0x102: /* SRET */ -if (riscv_has_ext(env, RVS)) { +if (has_ext(ctx, RVS)) { gen_helper_sret(cpu_pc, cpu_env, cpu_pc); tcg_gen_exit_tb(0); /* no chaining */ ctx->base.is_jmp = DISAS_NORETURN; @@ -1489,7 +1494,7 @@ static void decode_RV32_64C0(DisasContext *ctx) } } -static void decode_RV32_64C1(CPURISCVState *env, DisasContext *ctx) +static void decode_RV32_64C1(DisasContext *ctx) { uint8_t funct3 = extract32(ctx->opcode, 13, 3); uint8_t rd_rs1 = GET_C_RS1(ctx-&
[Qemu-devel] [PATCH v1 2/6] translator: merge max_insns into DisasContextBase
From: "Emilio G. Cota"While at it, use int for both num_insns and max_insns to make sure we have same-type comparisons. Reviewed-by: Richard Henderson Signed-off-by: Emilio G. Cota --- accel/tcg/translator.c | 21 ++--- include/exec/translator.h | 8 target/alpha/translate.c | 6 ++ target/arm/translate-a64.c | 8 +++- target/arm/translate.c | 11 --- target/hppa/translate.c| 7 ++- target/i386/translate.c| 5 + target/ppc/translate.c | 5 ++--- 8 files changed, 28 insertions(+), 43 deletions(-) diff --git a/accel/tcg/translator.c b/accel/tcg/translator.c index 23c6602cd921..0f9dca911399 100644 --- a/accel/tcg/translator.c +++ b/accel/tcg/translator.c @@ -34,8 +34,6 @@ void translator_loop_temp_check(DisasContextBase *db) void translator_loop(const TranslatorOps *ops, DisasContextBase *db, CPUState *cpu, TranslationBlock *tb) { -int max_insns; - /* Initialize DisasContext */ db->tb = tb; db->pc_first = tb->pc; @@ -45,18 +43,18 @@ void translator_loop(const TranslatorOps *ops, DisasContextBase *db, db->singlestep_enabled = cpu->singlestep_enabled; /* Instruction counting */ -max_insns = tb_cflags(db->tb) & CF_COUNT_MASK; -if (max_insns == 0) { -max_insns = CF_COUNT_MASK; +db->max_insns = tb_cflags(db->tb) & CF_COUNT_MASK; +if (db->max_insns == 0) { +db->max_insns = CF_COUNT_MASK; } -if (max_insns > TCG_MAX_INSNS) { -max_insns = TCG_MAX_INSNS; +if (db->max_insns > TCG_MAX_INSNS) { +db->max_insns = TCG_MAX_INSNS; } if (db->singlestep_enabled || singlestep) { -max_insns = 1; +db->max_insns = 1; } -max_insns = ops->init_disas_context(db, cpu, max_insns); +ops->init_disas_context(db, cpu); tcg_debug_assert(db->is_jmp == DISAS_NEXT); /* no early exit */ /* Reset the temp count so that we can identify leaks */ @@ -95,7 +93,8 @@ void translator_loop(const TranslatorOps *ops, DisasContextBase *db, update db->pc_next and db->is_jmp to indicate what should be done next -- either exiting this loop or locate the start of the next instruction. */ -if (db->num_insns == max_insns && (tb_cflags(db->tb) & CF_LAST_IO)) { +if (db->num_insns == db->max_insns +&& (tb_cflags(db->tb) & CF_LAST_IO)) { /* Accept I/O on the last instruction. */ gen_io_start(); ops->translate_insn(db, cpu); @@ -111,7 +110,7 @@ void translator_loop(const TranslatorOps *ops, DisasContextBase *db, /* Stop translation if the output buffer is full, or we have executed all of the allowed instructions. */ -if (tcg_op_buf_full() || db->num_insns >= max_insns) { +if (tcg_op_buf_full() || db->num_insns >= db->max_insns) { db->is_jmp = DISAS_TOO_MANY; break; } diff --git a/include/exec/translator.h b/include/exec/translator.h index e2dc2a04ae37..71e7b2c34714 100644 --- a/include/exec/translator.h +++ b/include/exec/translator.h @@ -58,6 +58,7 @@ typedef enum DisasJumpType { * disassembly). * @is_jmp: What instruction to disassemble next. * @num_insns: Number of translated instructions (including current). + * @max_insns: Maximum number of instructions to be translated in this TB. * @singlestep_enabled: "Hardware" single stepping enabled. * * Architecture-agnostic disassembly context. @@ -67,7 +68,8 @@ typedef struct DisasContextBase { target_ulong pc_first; target_ulong pc_next; DisasJumpType is_jmp; -unsigned int num_insns; +int num_insns; +int max_insns; bool singlestep_enabled; } DisasContextBase; @@ -76,7 +78,6 @@ typedef struct DisasContextBase { * @init_disas_context: * Initialize the target-specific portions of DisasContext struct. * The generic DisasContextBase has already been initialized. - * Return max_insns, modified as necessary by db->tb->flags. * * @tb_start: * Emit any code required before the start of the main loop, @@ -106,8 +107,7 @@ typedef struct DisasContextBase { * Print instruction disassembly to log. */ typedef struct TranslatorOps { -int (*init_disas_context)(DisasContextBase *db, CPUState *cpu, - int max_insns); +void (*init_disas_context)(DisasContextBase *db, CPUState *cpu); void (*tb_start)(DisasContextBase *db, CPUState *cpu); void (*insn_start)(DisasContextBase *db, CPUState *cpu); bool (*breakpoint_check)(DisasContextBase *db, CPUState *cpu, diff --git a/target/alpha/translate.c b/target/alpha/translate.c index 73a1b5e63ed1..15eca71d49f7 100644 --- a/target/alpha/translate.c +++ b/target/alpha/translate.c @@ -2919,8 +2919,7 @@ static DisasJumpType translate_one(DisasContext *ctx, uint32_t
[Qemu-devel] [PATCH v1 4/6] target/riscv: convert to DisasContextBase
From: "Emilio G. Cota" <c...@braap.org> Notes: - Did not convert {num,max}_insns, since the corresponding code will go away in the next patch. - ctx->pc becomes ctx->base.pc_next, and ctx->next_pc becomes ctx->pc_succ_insn. While at it, convert the remaining tb->cflags readers to tb_cflags(). Reviewed-by: Richard Henderson <richard.hender...@linaro.org> Cc: Michael Clark <m...@sifive.com> Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Signed-off-by: Emilio G. Cota <c...@braap.org> --- target/riscv/translate.c | 129 +++ 1 file changed, 64 insertions(+), 65 deletions(-) diff --git a/target/riscv/translate.c b/target/riscv/translate.c index 1fee5b51dc20..68979abfd7ed 100644 --- a/target/riscv/translate.c +++ b/target/riscv/translate.c @@ -40,14 +40,12 @@ static TCGv load_val; #include "exec/gen-icount.h" typedef struct DisasContext { -struct TranslationBlock *tb; -target_ulong pc; -target_ulong next_pc; +DisasContextBase base; +/* pc_succ_insn points to the instruction following base.pc_next */ +target_ulong pc_succ_insn; uint32_t opcode; uint32_t flags; uint32_t mem_idx; -int singlestep_enabled; -DisasJumpType is_jmp; /* Remember the rounding mode encoded in the previous fp instruction, which we have already installed into env->fp_status. Or -1 for no previous fp instruction. Note that we exit the TB when writing @@ -78,21 +76,21 @@ static const int tcg_memop_lookup[8] = { static void generate_exception(DisasContext *ctx, int excp) { -tcg_gen_movi_tl(cpu_pc, ctx->pc); +tcg_gen_movi_tl(cpu_pc, ctx->base.pc_next); TCGv_i32 helper_tmp = tcg_const_i32(excp); gen_helper_raise_exception(cpu_env, helper_tmp); tcg_temp_free_i32(helper_tmp); -ctx->is_jmp = DISAS_NORETURN; +ctx->base.is_jmp = DISAS_NORETURN; } static void generate_exception_mbadaddr(DisasContext *ctx, int excp) { -tcg_gen_movi_tl(cpu_pc, ctx->pc); +tcg_gen_movi_tl(cpu_pc, ctx->base.pc_next); tcg_gen_st_tl(cpu_pc, cpu_env, offsetof(CPURISCVState, badaddr)); TCGv_i32 helper_tmp = tcg_const_i32(excp); gen_helper_raise_exception(cpu_env, helper_tmp); tcg_temp_free_i32(helper_tmp); -ctx->is_jmp = DISAS_NORETURN; +ctx->base.is_jmp = DISAS_NORETURN; } static void gen_exception_debug(void) @@ -114,12 +112,12 @@ static void gen_exception_inst_addr_mis(DisasContext *ctx) static inline bool use_goto_tb(DisasContext *ctx, target_ulong dest) { -if (unlikely(ctx->singlestep_enabled)) { +if (unlikely(ctx->base.singlestep_enabled)) { return false; } #ifndef CONFIG_USER_ONLY -return (ctx->tb->pc & TARGET_PAGE_MASK) == (dest & TARGET_PAGE_MASK); +return (ctx->base.tb->pc & TARGET_PAGE_MASK) == (dest & TARGET_PAGE_MASK); #else return true; #endif @@ -131,10 +129,10 @@ static void gen_goto_tb(DisasContext *ctx, int n, target_ulong dest) /* chaining is only allowed when the jump is to the same page */ tcg_gen_goto_tb(n); tcg_gen_movi_tl(cpu_pc, dest); -tcg_gen_exit_tb((uintptr_t)ctx->tb + n); +tcg_gen_exit_tb((uintptr_t)ctx->base.tb + n); } else { tcg_gen_movi_tl(cpu_pc, dest); -if (ctx->singlestep_enabled) { +if (ctx->base.singlestep_enabled) { gen_exception_debug(); } else { tcg_gen_exit_tb(0); @@ -513,7 +511,7 @@ static void gen_jal(CPURISCVState *env, DisasContext *ctx, int rd, target_ulong next_pc; /* check misaligned: */ -next_pc = ctx->pc + imm; +next_pc = ctx->base.pc_next + imm; if (!riscv_has_ext(env, RVC)) { if ((next_pc & 0x3) != 0) { gen_exception_inst_addr_mis(ctx); @@ -521,11 +519,11 @@ static void gen_jal(CPURISCVState *env, DisasContext *ctx, int rd, } } if (rd != 0) { -tcg_gen_movi_tl(cpu_gpr[rd], ctx->next_pc); +tcg_gen_movi_tl(cpu_gpr[rd], ctx->pc_succ_insn); } -gen_goto_tb(ctx, 0, ctx->pc + imm); /* must use this for safety */ -ctx->is_jmp = DISAS_NORETURN; +gen_goto_tb(ctx, 0, ctx->base.pc_next + imm); /* must use this for safety */ +ctx->base.is_jmp = DISAS_NORETURN; } static void gen_jalr(CPURISCVState *env, DisasContext *ctx, uint32_t opc, @@ -548,7 +546,7 @@ static void gen_jalr(CPURISCVState *env, DisasContext *ctx, uint32_t opc, } if (rd != 0) { -tcg_gen_movi_tl(cpu_gpr[rd], ctx->next_pc); +tcg_gen_movi_tl(cpu_gpr[rd], ctx->pc_succ_insn); } tcg_gen_exit_tb(0); @@ -556,7 +554,7 @@ static void gen_jalr(CPURISCVState *env, DisasContext
[Qemu-devel] [PATCH v1 3/6] target/riscv: convert to DisasJumpType
From: "Emilio G. Cota" <c...@braap.org> Reviewed-by: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Reviewed-by: Richard Henderson <richard.hender...@linaro.org> Cc: Michael Clark <m...@sifive.com> Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Signed-off-by: Emilio G. Cota <c...@braap.org> --- target/riscv/translate.c | 72 +++- 1 file changed, 28 insertions(+), 44 deletions(-) diff --git a/target/riscv/translate.c b/target/riscv/translate.c index a98033ca77ca..1fee5b51dc20 100644 --- a/target/riscv/translate.c +++ b/target/riscv/translate.c @@ -26,6 +26,7 @@ #include "exec/helper-proto.h" #include "exec/helper-gen.h" +#include "exec/translator.h" #include "exec/log.h" #include "instmap.h" @@ -46,7 +47,7 @@ typedef struct DisasContext { uint32_t flags; uint32_t mem_idx; int singlestep_enabled; -int bstate; +DisasJumpType is_jmp; /* Remember the rounding mode encoded in the previous fp instruction, which we have already installed into env->fp_status. Or -1 for no previous fp instruction. Note that we exit the TB when writing @@ -55,13 +56,6 @@ typedef struct DisasContext { int frm; } DisasContext; -enum { -BS_NONE = 0, /* When seen outside of translation while loop, indicates - need to exit tb due to end of page. */ -BS_STOP = 1, /* Need to exit tb for syscall, sret, etc. */ -BS_BRANCH = 2, /* Need to exit tb for branch, jal, etc. */ -}; - /* convert riscv funct3 to qemu memop for load/store */ static const int tcg_memop_lookup[8] = { [0 ... 7] = -1, @@ -88,7 +82,7 @@ static void generate_exception(DisasContext *ctx, int excp) TCGv_i32 helper_tmp = tcg_const_i32(excp); gen_helper_raise_exception(cpu_env, helper_tmp); tcg_temp_free_i32(helper_tmp); -ctx->bstate = BS_BRANCH; +ctx->is_jmp = DISAS_NORETURN; } static void generate_exception_mbadaddr(DisasContext *ctx, int excp) @@ -98,7 +92,7 @@ static void generate_exception_mbadaddr(DisasContext *ctx, int excp) TCGv_i32 helper_tmp = tcg_const_i32(excp); gen_helper_raise_exception(cpu_env, helper_tmp); tcg_temp_free_i32(helper_tmp); -ctx->bstate = BS_BRANCH; +ctx->is_jmp = DISAS_NORETURN; } static void gen_exception_debug(void) @@ -531,7 +525,7 @@ static void gen_jal(CPURISCVState *env, DisasContext *ctx, int rd, } gen_goto_tb(ctx, 0, ctx->pc + imm); /* must use this for safety */ -ctx->bstate = BS_BRANCH; +ctx->is_jmp = DISAS_NORETURN; } static void gen_jalr(CPURISCVState *env, DisasContext *ctx, uint32_t opc, @@ -562,7 +556,7 @@ static void gen_jalr(CPURISCVState *env, DisasContext *ctx, uint32_t opc, gen_set_label(misaligned); gen_exception_inst_addr_mis(ctx); } -ctx->bstate = BS_BRANCH; +ctx->is_jmp = DISAS_NORETURN; break; default: @@ -616,7 +610,7 @@ static void gen_branch(CPURISCVState *env, DisasContext *ctx, uint32_t opc, } else { gen_goto_tb(ctx, 0, ctx->pc + bimm); } -ctx->bstate = BS_BRANCH; +ctx->is_jmp = DISAS_NORETURN; } static void gen_load(DisasContext *ctx, uint32_t opc, int rd, int rs1, @@ -1344,12 +1338,12 @@ static void gen_system(CPURISCVState *env, DisasContext *ctx, uint32_t opc, /* always generates U-level ECALL, fixed in do_interrupt handler */ generate_exception(ctx, RISCV_EXCP_U_ECALL); tcg_gen_exit_tb(0); /* no chaining */ -ctx->bstate = BS_BRANCH; +ctx->is_jmp = DISAS_NORETURN; break; case 0x1: /* EBREAK */ generate_exception(ctx, RISCV_EXCP_BREAKPOINT); tcg_gen_exit_tb(0); /* no chaining */ -ctx->bstate = BS_BRANCH; +ctx->is_jmp = DISAS_NORETURN; break; #ifndef CONFIG_USER_ONLY case 0x002: /* URET */ @@ -1359,7 +1353,7 @@ static void gen_system(CPURISCVState *env, DisasContext *ctx, uint32_t opc, if (riscv_has_ext(env, RVS)) { gen_helper_sret(cpu_pc, cpu_env, cpu_pc); tcg_gen_exit_tb(0); /* no chaining */ -ctx->bstate = BS_BRANCH; +ctx->is_jmp = DISAS_NORETURN; } else { gen_exception_illegal(ctx); } @@ -1370,7 +1364,7 @@ static void gen_system(CPURISCVState *env, DisasContext *ctx, uint32_t opc, case 0x302: /* MRET */ gen_helper_mret(cpu_pc, cpu_env, cpu_pc); tcg_gen_exit_tb(0); /* no chaining */ -ctx->bstate = BS_BRANCH; +ctx->is_jmp = DISAS_NORETURN; break; case 0x7b2
[Qemu-devel] [PATCH v1 1/6] target/riscv: avoid integer overflow in next_page PC check
From: "Emilio G. Cota" <c...@braap.org> If the PC is in the last page of the address space, next_page_start overflows to 0. Fix it. Reported-by: Richard Henderson <richard.hender...@linaro.org> Suggested-by: Richard Henderson <richard.hender...@linaro.org> Reviewed-by: Richard Henderson <richard.hender...@linaro.org> Reviewed-by: Michael Clark <m...@sifive.com> Acked-by: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Cc: Michael Clark <m...@sifive.com> Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Signed-off-by: Emilio G. Cota <c...@braap.org> --- target/riscv/translate.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/target/riscv/translate.c b/target/riscv/translate.c index c0e6a044d383..a98033ca77ca 100644 --- a/target/riscv/translate.c +++ b/target/riscv/translate.c @@ -1850,11 +1850,11 @@ void gen_intermediate_code(CPUState *cs, TranslationBlock *tb) CPURISCVState *env = cs->env_ptr; DisasContext ctx; target_ulong pc_start; -target_ulong next_page_start; +target_ulong page_start; int num_insns; int max_insns; pc_start = tb->pc; -next_page_start = (pc_start & TARGET_PAGE_MASK) + TARGET_PAGE_SIZE; +page_start = pc_start & TARGET_PAGE_MASK; ctx.pc = pc_start; /* once we have GDB, the rest of the translate.c implementation should be @@ -1904,7 +1904,7 @@ void gen_intermediate_code(CPUState *cs, TranslationBlock *tb) if (cs->singlestep_enabled) { break; } -if (ctx.pc >= next_page_start) { +if (ctx.pc - page_start >= TARGET_PAGE_SIZE) { break; } if (tcg_op_buf_full()) { -- 2.7.0
[Qemu-devel] [PATCH v1 0/6] Translation loop conversion for riscv
This is the riscv subset of Emilio G. Cota's translation loop conversion patches forward ported to master, with a patch from https://github.com/riscv/riscv-qemu/tree/qemu-2.13-for-upstream The last patch in the series adds misa to DisasContext and removes CPURISCVState from gen methods. The intention is to add additional patches to this series to check for the misa M, A, F and D extensions in multiply/divide, atomics, single and double precision floating point methods to throw illegal instruction exceptions for extensions that are not enabled. The first step is to move misa into DisasContext. The misa change has been rebased against Emilio G. Cota's translation loop series so that target/riscv/translate.c merges can be done more easily and so that patches can be moved between trees. Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Cc: Alistair Francis <alistair.fran...@wdc.com> Cc: Emilio G. Cota <c...@braap.org> Signed-off-by: Michael Clark <m...@sifive.com> Emilio G. Cota (5): target/riscv: avoid integer overflow in next_page PC check translator: merge max_insns into DisasContextBase target/riscv: convert to DisasJumpType target/riscv: convert to DisasContextBase target/riscv: convert to TranslatorOps Michael Clark (1): target/riscv: add misa to DisasContext accel/tcg/translator.c | 21 ++- include/exec/translator.h | 8 +- target/alpha/translate.c | 6 +- target/arm/translate-a64.c | 8 +- target/arm/translate.c | 11 +- target/hppa/translate.c| 7 +- target/i386/translate.c| 5 +- target/ppc/translate.c | 5 +- target/riscv/translate.c | 326 ++--- 9 files changed, 187 insertions(+), 210 deletions(-) -- 2.7.0
Re: [Qemu-devel] [PATCH 18/18] target/riscv: convert to TranslatorOps
On Sat, Apr 21, 2018 at 6:55 AM, Emilio G. Cota <c...@braap.org> wrote: > Reviewed-by: Richard Henderson <richard.hender...@linaro.org> > Cc: Michael Clark <m...@sifive.com> > Cc: Palmer Dabbelt <pal...@sifive.com> > Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> > Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> > Signed-off-by: Emilio G. Cota <c...@braap.org> > Reviewed-by: Michael Clark <m...@sifive.com> > --- > target/riscv/translate.c | 158 -- > - > 1 file changed, 80 insertions(+), 78 deletions(-) > > diff --git a/target/riscv/translate.c b/target/riscv/translate.c > index 18ec1a7..fc9c659 100644 > --- a/target/riscv/translate.c > +++ b/target/riscv/translate.c > @@ -1836,78 +1836,71 @@ static void decode_opc(CPURISCVState *env, > DisasContext *ctx) > } > } > > -void gen_intermediate_code(CPUState *cs, TranslationBlock *tb) > +static void riscv_tr_init_disas_context(DisasContextBase *dcbase, > CPUState *cs) > { > -CPURISCVState *env = cs->env_ptr; > -DisasContext ctx; > -target_ulong page_start; > -int num_insns; > -int max_insns; > - > -ctx.base.pc_first = tb->pc; > -ctx.base.pc_next = ctx.base.pc_first; > -/* once we have GDB, the rest of the translate.c implementation > should be > - ready for singlestep */ > -ctx.base.singlestep_enabled = cs->singlestep_enabled; > -ctx.base.tb = tb; > -ctx.base.is_jmp = DISAS_NEXT; > - > -page_start = ctx.base.pc_first & TARGET_PAGE_MASK; > -ctx.pc_succ_insn = ctx.base.pc_first; > -ctx.flags = tb->flags; > -ctx.mem_idx = tb->flags & TB_FLAGS_MMU_MASK; > -ctx.frm = -1; /* unknown rounding mode */ > - > -num_insns = 0; > -max_insns = tb_cflags(ctx.base.tb) & CF_COUNT_MASK; > -if (max_insns == 0) { > -max_insns = CF_COUNT_MASK; > -} > -if (max_insns > TCG_MAX_INSNS) { > -max_insns = TCG_MAX_INSNS; > -} > -gen_tb_start(tb); > +DisasContext *ctx = container_of(dcbase, DisasContext, base); > > -while (ctx.base.is_jmp == DISAS_NEXT) { > -tcg_gen_insn_start(ctx.base.pc_next); > -num_insns++; > +ctx->pc_succ_insn = ctx->base.pc_first; > +ctx->flags = ctx->base.tb->flags; > +ctx->mem_idx = ctx->base.tb->flags & TB_FLAGS_MMU_MASK; > +ctx->frm = -1; /* unknown rounding mode */ > +} > > -if (unlikely(cpu_breakpoint_test(cs, ctx.base.pc_next, BP_ANY))) > { > -tcg_gen_movi_tl(cpu_pc, ctx.base.pc_next); > -ctx.base.is_jmp = DISAS_NORETURN; > -gen_exception_debug(); > -/* The address covered by the breakpoint must be included in > - [tb->pc, tb->pc + tb->size) in order to for it to be > - properly cleared -- thus we increment the PC here so that > - the logic setting tb->size below does the right thing. */ > -ctx.base.pc_next += 4; > -goto done_generating; > -} > +static void riscv_tr_tb_start(DisasContextBase *db, CPUState *cpu) > +{ > +} > > -if (num_insns == max_insns && (tb_cflags(ctx.base.tb) & > CF_LAST_IO)) { > -gen_io_start(); > -} > +static void riscv_tr_insn_start(DisasContextBase *dcbase, CPUState *cpu) > +{ > +DisasContext *ctx = container_of(dcbase, DisasContext, base); > + > +tcg_gen_insn_start(ctx->base.pc_next); > +} > + > +static bool riscv_tr_breakpoint_check(DisasContextBase *dcbase, CPUState > *cpu, > + const CPUBreakpoint *bp) > +{ > +DisasContext *ctx = container_of(dcbase, DisasContext, base); > + > +tcg_gen_movi_tl(cpu_pc, ctx->base.pc_next); > +ctx->base.is_jmp = DISAS_NORETURN; > +gen_exception_debug(); > +/* The address covered by the breakpoint must be included in > + [tb->pc, tb->pc + tb->size) in order to for it to be > + properly cleared -- thus we increment the PC here so that > + the logic setting tb->size below does the right thing. */ > +ctx->base.pc_next += 4; > +return true; > +} > + > + > +static void riscv_tr_translate_insn(DisasContextBase *dcbase, CPUState > *cpu) > +{ > +DisasContext *ctx = container_of(dcbase, DisasContext, base); > +CPURISCVState *env = cpu->env_ptr; > + > +ctx->opcode = cpu_ldl_code(env, ctx->base.pc_next); > +decode_opc(env, ctx); > +ctx->base.pc_next = ctx->pc_succ_insn; > + > +if (ctx->base.is_
Re: [Qemu-devel] [PATCH 01/18] translator: merge max_insns into DisasContextBase
On Sat, Apr 21, 2018 at 6:55 AM, Emilio G. Cota <c...@braap.org> wrote: > While at it, use int for both num_insns and max_insns to make > sure we have same-type comparisons. > > Reviewed-by: Richard Henderson <richard.hender...@linaro.org> > Signed-off-by: Emilio G. Cota <c...@braap.org> > Reviewed-by: Michael Clark <m...@sifive.com> > --- > accel/tcg/translator.c | 21 ++--- > include/exec/translator.h | 8 > target/alpha/translate.c | 6 ++ > target/arm/translate-a64.c | 8 +++- > target/arm/translate.c | 9 +++-- > target/hppa/translate.c| 7 ++- > target/i386/translate.c| 5 + > target/ppc/translate.c | 5 ++--- > 8 files changed, 27 insertions(+), 42 deletions(-) > > diff --git a/accel/tcg/translator.c b/accel/tcg/translator.c > index 23c6602..0f9dca9 100644 > --- a/accel/tcg/translator.c > +++ b/accel/tcg/translator.c > @@ -34,8 +34,6 @@ void translator_loop_temp_check(DisasContextBase *db) > void translator_loop(const TranslatorOps *ops, DisasContextBase *db, > CPUState *cpu, TranslationBlock *tb) > { > -int max_insns; > - > /* Initialize DisasContext */ > db->tb = tb; > db->pc_first = tb->pc; > @@ -45,18 +43,18 @@ void translator_loop(const TranslatorOps *ops, > DisasContextBase *db, > db->singlestep_enabled = cpu->singlestep_enabled; > > /* Instruction counting */ > -max_insns = tb_cflags(db->tb) & CF_COUNT_MASK; > -if (max_insns == 0) { > -max_insns = CF_COUNT_MASK; > +db->max_insns = tb_cflags(db->tb) & CF_COUNT_MASK; > +if (db->max_insns == 0) { > +db->max_insns = CF_COUNT_MASK; > } > -if (max_insns > TCG_MAX_INSNS) { > -max_insns = TCG_MAX_INSNS; > +if (db->max_insns > TCG_MAX_INSNS) { > +db->max_insns = TCG_MAX_INSNS; > } > if (db->singlestep_enabled || singlestep) { > -max_insns = 1; > +db->max_insns = 1; > } > > -max_insns = ops->init_disas_context(db, cpu, max_insns); > +ops->init_disas_context(db, cpu); > tcg_debug_assert(db->is_jmp == DISAS_NEXT); /* no early exit */ > > /* Reset the temp count so that we can identify leaks */ > @@ -95,7 +93,8 @@ void translator_loop(const TranslatorOps *ops, > DisasContextBase *db, > update db->pc_next and db->is_jmp to indicate what should be > done next -- either exiting this loop or locate the start of > the next instruction. */ > -if (db->num_insns == max_insns && (tb_cflags(db->tb) & > CF_LAST_IO)) { > +if (db->num_insns == db->max_insns > +&& (tb_cflags(db->tb) & CF_LAST_IO)) { > /* Accept I/O on the last instruction. */ > gen_io_start(); > ops->translate_insn(db, cpu); > @@ -111,7 +110,7 @@ void translator_loop(const TranslatorOps *ops, > DisasContextBase *db, > > /* Stop translation if the output buffer is full, > or we have executed all of the allowed instructions. */ > -if (tcg_op_buf_full() || db->num_insns >= max_insns) { > +if (tcg_op_buf_full() || db->num_insns >= db->max_insns) { > db->is_jmp = DISAS_TOO_MANY; > break; > } > diff --git a/include/exec/translator.h b/include/exec/translator.h > index e2dc2a0..71e7b2c 100644 > --- a/include/exec/translator.h > +++ b/include/exec/translator.h > @@ -58,6 +58,7 @@ typedef enum DisasJumpType { > * disassembly). > * @is_jmp: What instruction to disassemble next. > * @num_insns: Number of translated instructions (including current). > + * @max_insns: Maximum number of instructions to be translated in this TB. > * @singlestep_enabled: "Hardware" single stepping enabled. > * > * Architecture-agnostic disassembly context. > @@ -67,7 +68,8 @@ typedef struct DisasContextBase { > target_ulong pc_first; > target_ulong pc_next; > DisasJumpType is_jmp; > -unsigned int num_insns; > +int num_insns; > +int max_insns; > bool singlestep_enabled; > } DisasContextBase; > > @@ -76,7 +78,6 @@ typedef struct DisasContextBase { > * @init_disas_context: > * Initialize the target-specific portions of DisasContext struct. > * The generic DisasContextBase has already been initialized. > - * Return max_insns, modified as necessary by db->tb->flags. > * > * @tb_start: > * Emit any code required before the start of the main l
[Qemu-devel] [PATCH v1] RISC-V: Add misa to DisasContext
gen methods should access state from DisasContext. Add misa field to the DisasContext struct and remove CPURISCVState argument from all gen methods. Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Cc: Alistair Francis <alistair.fran...@wdc.com> Cc: Emilio G. Cota <c...@braap.org> Signed-off-by: Michael Clark <m...@sifive.com> --- target/riscv/translate.c | 74 ++-- 1 file changed, 40 insertions(+), 34 deletions(-) diff --git a/target/riscv/translate.c b/target/riscv/translate.c index c0e6a044d383..5b9e5fe343d2 100644 --- a/target/riscv/translate.c +++ b/target/riscv/translate.c @@ -44,6 +44,7 @@ typedef struct DisasContext { target_ulong next_pc; uint32_t opcode; uint32_t flags; +uint32_t misa; uint32_t mem_idx; int singlestep_enabled; int bstate; @@ -82,6 +83,11 @@ static const int tcg_memop_lookup[8] = { #define CASE_OP_32_64(X) case X #endif +static inline bool has_ext(DisasContext *ctx, uint32_t ext) +{ +return ctx->misa & ext; +} + static void generate_exception(DisasContext *ctx, int excp) { tcg_gen_movi_tl(cpu_pc, ctx->pc); @@ -513,14 +519,13 @@ static void gen_arith_imm(DisasContext *ctx, uint32_t opc, int rd, tcg_temp_free(source1); } -static void gen_jal(CPURISCVState *env, DisasContext *ctx, int rd, -target_ulong imm) +static void gen_jal(DisasContext *ctx, int rd, target_ulong imm) { target_ulong next_pc; /* check misaligned: */ next_pc = ctx->pc + imm; -if (!riscv_has_ext(env, RVC)) { +if (!has_ext(ctx, RVC)) { if ((next_pc & 0x3) != 0) { gen_exception_inst_addr_mis(ctx); return; @@ -534,8 +539,8 @@ static void gen_jal(CPURISCVState *env, DisasContext *ctx, int rd, ctx->bstate = BS_BRANCH; } -static void gen_jalr(CPURISCVState *env, DisasContext *ctx, uint32_t opc, - int rd, int rs1, target_long imm) +static void gen_jalr(DisasContext *ctx, uint32_t opc, int rd, int rs1, + target_long imm) { /* no chaining with JALR */ TCGLabel *misaligned = NULL; @@ -547,7 +552,7 @@ static void gen_jalr(CPURISCVState *env, DisasContext *ctx, uint32_t opc, tcg_gen_addi_tl(cpu_pc, cpu_pc, imm); tcg_gen_andi_tl(cpu_pc, cpu_pc, (target_ulong)-2); -if (!riscv_has_ext(env, RVC)) { +if (!has_ext(ctx, RVC)) { misaligned = gen_new_label(); tcg_gen_andi_tl(t0, cpu_pc, 0x2); tcg_gen_brcondi_tl(TCG_COND_NE, t0, 0x0, misaligned); @@ -572,8 +577,8 @@ static void gen_jalr(CPURISCVState *env, DisasContext *ctx, uint32_t opc, tcg_temp_free(t0); } -static void gen_branch(CPURISCVState *env, DisasContext *ctx, uint32_t opc, - int rs1, int rs2, target_long bimm) +static void gen_branch(DisasContext *ctx, uint32_t opc, int rs1, int rs2, + target_long bimm) { TCGLabel *l = gen_new_label(); TCGv source1, source2; @@ -610,7 +615,7 @@ static void gen_branch(CPURISCVState *env, DisasContext *ctx, uint32_t opc, gen_goto_tb(ctx, 1, ctx->next_pc); gen_set_label(l); /* branch taken */ -if (!riscv_has_ext(env, RVC) && ((ctx->pc + bimm) & 0x3)) { +if (!has_ext(ctx, RVC) && ((ctx->pc + bimm) & 0x3)) { /* misaligned */ gen_exception_inst_addr_mis(ctx); } else { @@ -1313,8 +1318,8 @@ static void gen_fp_arith(DisasContext *ctx, uint32_t opc, int rd, } } -static void gen_system(CPURISCVState *env, DisasContext *ctx, uint32_t opc, - int rd, int rs1, int csr) +static void gen_system(DisasContext *ctx, uint32_t opc, int rd, int rs1, + int csr) { TCGv source1, csr_store, dest, rs1_pass, imm_rs1; source1 = tcg_temp_new(); @@ -1356,7 +1361,7 @@ static void gen_system(CPURISCVState *env, DisasContext *ctx, uint32_t opc, gen_exception_illegal(ctx); break; case 0x102: /* SRET */ -if (riscv_has_ext(env, RVS)) { +if (has_ext(ctx, RVS)) { gen_helper_sret(cpu_pc, cpu_env, cpu_pc); tcg_gen_exit_tb(0); /* no chaining */ ctx->bstate = BS_BRANCH; @@ -1497,7 +1502,7 @@ static void decode_RV32_64C0(DisasContext *ctx) } } -static void decode_RV32_64C1(CPURISCVState *env, DisasContext *ctx) +static void decode_RV32_64C1(DisasContext *ctx) { uint8_t funct3 = extract32(ctx->opcode, 13, 3); uint8_t rd_rs1 = GET_C_RS1(ctx->opcode); @@ -1517,7 +1522,7 @@ static void decode_RV32_64C1(CPURISCVState *env, DisasContext *ctx) GET_C_IMM(ctx->opcode)); #else /* C.JAL(RV32) -> jal x1, offset[11:1] */ -gen_jal(env, ctx, 1, GET_C_J_
Re: [Qemu-devel] [PATCH 00/10] Avoid integer overflow in next_page_start
On Thu, Apr 12, 2018 at 11:56 AM, Richard Henderson < richard.hender...@linaro.org> wrote: > On 04/12/2018 01:29 AM, Emilio G. Cota wrote: > > To ease an eventual merge I'll be updating the patches' R-b tags as > > they come in this branch: > > https://github.com/cota/qemu/tree/next_page_overflow-r-b > > > > BTW to avoid conflicts we should merge this before the translator loop > > conversion series; I'll make that clear when I send a new version > > of that patch set. > > Right-o. Thanks. We'll get these in right away once development starts > again > so that you don't have to carry it long. Emilio, Is this your latest branch for these changes? - https://github.com/cota/qemu/tree/trloop-conv-v3 I see that this branch has this change "target/riscv: avoid integer overflow in next_page PC check" I have some changes for target/riscv/translate.c and was thinking of basing them on your patches so we don't have conflicts. Thanks, Michael
Re: [Qemu-devel] [PULL 00/20] RISC-V: QEMU 2.13 Privileged ISA emulation updates
On Wed, May 9, 2018 at 1:22 AM, Peter Maydell <peter.mayd...@linaro.org> wrote: > On 6 May 2018 at 00:35, Michael Clark <m...@sifive.com> wrote: > > The following changes since commit c8b7e627b4269a3bc3ae41d9f42054 > 7a47e6d9b9: > > > > Merge remote-tracking branch 'remotes/ericb/tags/pull-nbd-2018-05-04' > into staging (2018-05-04 14:42:46 +0100) > > > > are available in the git repository at: > > > > https://github.com/riscv/riscv-qemu.git tags/riscv-qemu-2.13-pull- > 20180506 > > > > for you to fetch changes up to 5aec3247c190f10654250203a1742490ae7343a2: > > > > RISC-V: Mark ROM read-only after copying in code (2018-05-06 10:54:21 > +1200) > > > > > > RISC-V: QEMU 2.13 Privileged ISA emulation updates > > > > Several code cleanups, minor specification conformance changes, > > fixes to make ROM read-only and add device-tree size checks. > > > > * Honour privileged ISA v1.10 counter enable CSRs. > > * Implements WARL behavior for CSRs that don't support writes > > * Past behavior of raising traps was non-conformant > > with the RISC-V Privileged ISA Specification v1.10. > > * Allow S-mode access to sstatus.MXR when priv ISA >= v1.10 > > * Sets mtval/stval to zero on exceptions without addresses > > * Past behavior of leaving the last value was non-conformant > > with the RISC-V Privileged ISA Specition v1.10. mtval/stval > > must be set on all exceptions; to zero if not supported. > > * Make ROMs read-only and implement device-tree size checks > > * Uses memory_region_init_rom and rom_add_blob_fixed_as > > * Adds hexidecimal instruction bytes to disassembly output. > > * Fixes missing break statement for rv128 disassembly. > > * Several code cleanups > > * Replacing hard-coded constants with enums > > * Dead-code elimination > > > > This is an incremental pull that contains 20 reviewed changes out > > of 38 changes currently queued in the qemu-2.13-for-upstream branch. > > > > Applied, thanks. > Thanks! I didn't include qemu-stable on this series as I simply hadn't spent the time to locally test the issue with 'sifive_u' that Alastair mentioned. I've now locally reproduced Alastair's results confirming that the last commit in this series, namely "RISC-V: Mark ROM read-only after copying in code", fixes the 'sifive_u' board. - https://git.qemu.org/?p=qemu.git;a=commit;h=5aec3247c190f10654250203a1742490ae7343a2 There are however several other usability issues and bugs fixess in this series that could be considered for 'qemu-stable'. - "RISC-V: Mark ROM read-only after copying in code" fixes the 'sifive_u' board issue (bug fix) - "RISC-V: No traps on writes to misa,minstret,mcycle" fixes a case where an illegal instruction should not be generated (bug fix - misa write bug was hit by Andes Technology) - "RISC-V: Make mtvec/stvec ignore vectored traps" fixes a case where an illegal instruction should not be generated (bug fix - not hit yet, but nevertheless is a fix) - "RISC-V: Use [ms]counteren CSRs when priv ISA >= v1.10" fixes the counter enable CSRs, so they actually work (bug fix - hit by someone working on 'perf' for riscv-linux) - "RISC-V: Include instruction hex in disassembly" adds instruction bytes to disassembly to improbe usability, and was on the list before soft-freeze (small usability issue) - "RISC-V: Make virt board description match spike" makes the -help listing use consistent naming for privileged ISA version (small usability issue) The rest of the series are the innocuous fixes and code cleanups (replacing constants with enums, etc) that had been on the list mid-March, with the exception of "RISC-V: Fix missing break statement in disassembler" which fixes the rv128 disassembly issue picked up recently by coverity and the "RISC-V: Add mcycle/minstret support for -icount auto" is a small new feature. It would be safe to import the tag as the cleanup commits or I can selectively backport changes. The issue is the fixes depending on context of the cleanup commits early in the series. The sifive_u fix: "RISC-V: Mark ROM read-only after copying in code" depends on diff context from "RISC-V: Remove identity_translate from load_elf" and "RISC-V: Use ROM base address and size from memmap" which were both code cleanups. The question is whether code cleanups can come into qemu-stable (i.e. merge this tag to stable) or whether we should cherry-pick and solve merge conflicts to get the minimal set of commits with what we classify as important fixes or usability issue into qemu-stable. The most essential would be t
Re: [Qemu-devel] [PATCH] RISC-V: Remove unnecessary header include
On Wed, May 9, 2018 at 2:42 AM, Philippe Mathieu-Daudé <f4...@amsat.org> wrote: > Signed-off-by: Philippe Mathieu-Daudé <f4...@amsat.org> > Reviewed-by: Michael Clark <m...@sifive.com> --- > hw/riscv/riscv_htif.c | 1 - > hw/riscv/sifive_e.c | 1 - > hw/riscv/sifive_u.c | 1 - > hw/riscv/spike.c | 1 - > hw/riscv/virt.c | 1 - > 5 files changed, 5 deletions(-) > > diff --git a/hw/riscv/riscv_htif.c b/hw/riscv/riscv_htif.c > index 3e17f30251..43533f9dc2 100644 > --- a/hw/riscv/riscv_htif.c > +++ b/hw/riscv/riscv_htif.c > @@ -29,7 +29,6 @@ > #include "chardev/char-fe.h" > #include "hw/riscv/riscv_htif.h" > #include "qemu/timer.h" > -#include "exec/address-spaces.h" > #include "qemu/error-report.h" > > #define RISCV_DEBUG_HTIF 0 > diff --git a/hw/riscv/sifive_e.c b/hw/riscv/sifive_e.c > index e4ecb7aa4b..d3d11e849b 100644 > --- a/hw/riscv/sifive_e.c > +++ b/hw/riscv/sifive_e.c > @@ -46,7 +46,6 @@ > #include "hw/riscv/sifive_e.h" > #include "chardev/char.h" > #include "sysemu/arch_init.h" > -#include "exec/address-spaces.h" > #include "elf.h" > > static const struct MemmapEntry { > diff --git a/hw/riscv/sifive_u.c b/hw/riscv/sifive_u.c > index c05dcbba95..07f8644695 100644 > --- a/hw/riscv/sifive_u.c > +++ b/hw/riscv/sifive_u.c > @@ -44,7 +44,6 @@ > #include "chardev/char.h" > #include "sysemu/arch_init.h" > #include "sysemu/device_tree.h" > -#include "exec/address-spaces.h" > #include "elf.h" > > #include > diff --git a/hw/riscv/spike.c b/hw/riscv/spike.c > index f94e2b6707..aecdf296cc 100644 > --- a/hw/riscv/spike.c > +++ b/hw/riscv/spike.c > @@ -39,7 +39,6 @@ > #include "chardev/char.h" > #include "sysemu/arch_init.h" > #include "sysemu/device_tree.h" > -#include "exec/address-spaces.h" > #include "elf.h" > > #include > diff --git a/hw/riscv/virt.c b/hw/riscv/virt.c > index ad03113e0f..d967624bcc 100644 > --- a/hw/riscv/virt.c > +++ b/hw/riscv/virt.c > @@ -37,7 +37,6 @@ > #include "chardev/char.h" > #include "sysemu/arch_init.h" > #include "sysemu/device_tree.h" > -#include "exec/address-spaces.h" > #include "elf.h" > > #include > -- > 2.17.0 > >
Re: [Qemu-devel] [PULL 0/3] RISC-V: QEMU 2.13 Minor Fixes
On Wed, May 9, 2018 at 8:49 AM, Peter Maydell <peter.mayd...@linaro.org> wrote: > On 8 May 2018 at 21:07, Michael Clark <m...@sifive.com> wrote: > > The following changes since commit c8b7e627b4269a3bc3ae41d9f42054 > 7a47e6d9b9: > > > > Merge remote-tracking branch 'remotes/ericb/tags/pull-nbd-2018-05-04' > into staging (2018-05-04 14:42:46 +0100) > > > > are available in the git repository at: > > > > https://github.com/riscv/riscv-qemu.git tags/riscv-qemu-2.13-minor- > fixes-2 > > > > for you to fetch changes up to fc28db142ea9331d24d08a6418efa93e258a6876: > > > > riscv: requires libfdt (2018-05-09 07:57:57 +1200) > > > > > > RISC-V: QEMU 2.13 Minor Fixes > > > > * Require libfdt when configuring for 'riscv*-softmmu' > > * Increase HTIF priority and allow zero base address > > > > Consider backporting to stable branch for 2.12.1 > > Ah, if you want something flagged to be backported to > the stable branch, the way to do that is to include > the line "Cc: qemu-sta...@nongnu.org" in the commit > message. We use the same approach for flagging things > for stable as the Linux kernel. (Individual submaintainers > are generally in the best position to decide whether > something should go into stable, and this system leaves > that choice up to them. A patch submitter can also suggest > something for stable when you submit the initial patch for > review by adding the Cc: line and cc'ing qemu-stable on the > patchmail.) > Okay. Given the 'Cc is missing in the commit (only on the cover) I could send a separate patch. Or I can respin with the 'Cc on the configure libfdt dependency fix commit and remove the backport message from the tag...
[Qemu-devel] [PULL 2/3] riscv: htif: increase the priority of the htif subregion
From: KONRAD Frederic <frederic.kon...@adacore.com> The htif device is supposed to be mapped over an other subregion. So increase its priority to one to avoid any conflict. Here is the output of info mtree: Before: (qemu) info mtree address-space: memory - (prio 0, i/o): system -000f (prio 0, i/o): riscv.htif.uart -00011fff (prio 0, ram): riscv.spike.bootrom 0200-0200 (prio 0, i/o): riscv.sifive.clint 8000-87ff (prio 0, ram): riscv.spike.ram address-space: I/O - (prio 0, i/o): io address-space: cpu-memory-0 - (prio 0, i/o): system -000f (prio 0, i/o): riscv.htif.uart -00011fff (prio 0, ram): riscv.spike.bootrom 0200-0200 (prio 0, i/o): riscv.sifive.clint 8000-87ff (prio 0, ram): riscv.spike.ram After: (qemu) info mtree address-space: memory - (prio 0, i/o): system -000f (prio 1, i/o): riscv.htif.uart -00011fff (prio 0, ram): riscv.spike.bootrom 0200-0200 (prio 0, i/o): riscv.sifive.clint 8000-87ff (prio 0, ram): riscv.spike.ram address-space: I/O - (prio 0, i/o): io address-space: cpu-memory-0 - (prio 0, i/o): system -000f (prio 1, i/o): riscv.htif.uart -00011fff (prio 0, ram): riscv.spike.bootrom 0200-0200 (prio 0, i/o): riscv.sifive.clint 8000-87ff (prio 0, ram): riscv.spike.ram Reviewed-by: Michael Clark <m...@sifive.com> Signed-off-by: KONRAD Frederic <frederic.kon...@adacore.com> Signed-off-by: Michael Clark <m...@sifive.com> Message-Id: <1525360636-18229-3-git-send-email-frederic.kon...@adacore.com> --- hw/riscv/riscv_htif.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/hw/riscv/riscv_htif.c b/hw/riscv/riscv_htif.c index be252ec8cce9..f73512941fb6 100644 --- a/hw/riscv/riscv_htif.c +++ b/hw/riscv/riscv_htif.c @@ -253,8 +253,9 @@ HTIFState *htif_mm_init(MemoryRegion *address_space, MemoryRegion *main_mem, htif_be_change, s, NULL, true); if (address_symbol_set == 3) { memory_region_init_io(>mmio, NULL, _mm_ops, s, -TYPE_HTIF_UART, size); -memory_region_add_subregion(address_space, base, >mmio); + TYPE_HTIF_UART, size); +memory_region_add_subregion_overlap(address_space, base, +>mmio, 1); } return s; -- 2.7.0
[Qemu-devel] [PULL 1/3] riscv: spike: allow base == 0
From: KONRAD Frederic <frederic.kon...@adacore.com> The sanity check on base doesn't allow htif to be mapped @0. Check if the symbol exists instead so we can map it where we want. Reviewed-by: Michael Clark <m...@sifive.com> Signed-off-by: KONRAD Frederic <frederic.kon...@adacore.com> Signed-off-by: Michael Clark <m...@sifive.com> Message-Id: <1525360636-18229-2-git-send-email-frederic.kon...@adacore.com> --- hw/riscv/riscv_htif.c | 7 +-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/hw/riscv/riscv_htif.c b/hw/riscv/riscv_htif.c index 3e17f3025106..be252ec8cce9 100644 --- a/hw/riscv/riscv_htif.c +++ b/hw/riscv/riscv_htif.c @@ -41,17 +41,20 @@ } while (0) static uint64_t fromhost_addr, tohost_addr; +static int address_symbol_set; void htif_symbol_callback(const char *st_name, int st_info, uint64_t st_value, -uint64_t st_size) + uint64_t st_size) { if (strcmp("fromhost", st_name) == 0) { +address_symbol_set |= 1; fromhost_addr = st_value; if (st_size != 8) { error_report("HTIF fromhost must be 8 bytes"); exit(1); } } else if (strcmp("tohost", st_name) == 0) { +address_symbol_set |= 2; tohost_addr = st_value; if (st_size != 8) { error_report("HTIF tohost must be 8 bytes"); @@ -248,7 +251,7 @@ HTIFState *htif_mm_init(MemoryRegion *address_space, MemoryRegion *main_mem, qemu_chr_fe_init(>chr, chr, _abort); qemu_chr_fe_set_handlers(>chr, htif_can_recv, htif_recv, htif_event, htif_be_change, s, NULL, true); -if (base) { +if (address_symbol_set == 3) { memory_region_init_io(>mmio, NULL, _mm_ops, s, TYPE_HTIF_UART, size); memory_region_add_subregion(address_space, base, >mmio); -- 2.7.0
[Qemu-devel] [PULL 3/3] riscv: requires libfdt
From: KONRAD Frederic <frederic.kon...@adacore.com> When compiling on a machine without libfdt installed the configure script should try to get libfdt from the git or should die because otherwise CONFIG_LIBFDT is not set and the build process end in an error in the link phase.. eg: hw/riscv/virt.o: In function `riscv_virt_board_init': qemu/src/hw/riscv/virt.c:317: undefined reference to `qemu_fdt_setprop_cell' qemu/src/hw/riscv/virt.c:319: undefined reference to `qemu_fdt_setprop_cell' qemu/src/hw/riscv/virt.c:345: undefined reference to `qemu_fdt_dumpdtb' collect2: error: ld returned 1 exit status make[1]: *** [qemu-system-riscv64] Error 1 make: *** [subdir-riscv64-softmmu] Error 2 Reviewed-by: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Reviewed-by: Philippe Mathieu-Daudé <f4...@amsat.org> Reviewed-by: Michael Clark <m...@sifive.com> Signed-off-by: KONRAD Frederic <frederic.kon...@adacore.com> Signed-off-by: Michael Clark <m...@sifive.com> Message-Id: <1525360636-18229-4-git-send-email-frederic.kon...@adacore.com> --- configure | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/configure b/configure index 1443422e83f1..479e9a787e6e 100755 --- a/configure +++ b/configure @@ -3761,7 +3761,7 @@ fi fdt_required=no for target in $target_list; do case $target in - aarch64*-softmmu|arm*-softmmu|ppc*-softmmu|microblaze*-softmmu|mips64el-softmmu) + aarch64*-softmmu|arm*-softmmu|ppc*-softmmu|microblaze*-softmmu|mips64el-softmmu|riscv*-softmmu) fdt_required=yes ;; esac -- 2.7.0
[Qemu-devel] [PULL 0/3] RISC-V: QEMU 2.13 Minor Fixes
The following changes since commit c8b7e627b4269a3bc3ae41d9f420547a47e6d9b9: Merge remote-tracking branch 'remotes/ericb/tags/pull-nbd-2018-05-04' into staging (2018-05-04 14:42:46 +0100) are available in the git repository at: https://github.com/riscv/riscv-qemu.git tags/riscv-qemu-2.13-minor-fixes-2 for you to fetch changes up to fc28db142ea9331d24d08a6418efa93e258a6876: riscv: requires libfdt (2018-05-09 07:57:57 +1200) RISC-V: QEMU 2.13 Minor Fixes * Require libfdt when configuring for 'riscv*-softmmu' * Increase HTIF priority and allow zero base address Consider backporting to stable branch for 2.12.1 KONRAD Frederic (3): riscv: spike: allow base == 0 riscv: htif: increase the priority of the htif subregion riscv: requires libfdt configure | 2 +- hw/riscv/riscv_htif.c | 12 2 files changed, 9 insertions(+), 5 deletions(-) -- 2.7.0
Re: [Qemu-devel] [PULL 0/3] RISC-V: QEMU 2.13 Minor Fixes
On Wed, May 9, 2018 at 4:05 AM, Peter Maydell <peter.mayd...@linaro.org> wrote: > On 8 May 2018 at 00:14, Michael Clark <m...@sifive.com> wrote: > > The following changes since commit c8b7e627b4269a3bc3ae41d9f42054 > 7a47e6d9b9: > > > > Merge remote-tracking branch 'remotes/ericb/tags/pull-nbd-2018-05-04' > into staging (2018-05-04 14:42:46 +0100) > > > > are available in the git repository at: > > > > https://github.com/riscv/riscv-qemu.git tags/riscv-qemu-2.13-minor- > fixes-1 > > > > for you to fetch changes up to d4a63373c0e6349c6f36f56a6290ad639bfa20cb: > > > > riscv: requires libfdt (2018-05-08 10:57:21 +1200) > > > > > > RISC-V: QEMU 2.13 Minor Fixes > > > > * Require libfdt when configuring for 'riscv*-softmmu' > > * Increase HTIF priority and allow zero address > > > > > > KONRAD Frederic (3): > > riscv: spike: allow base == 0 > > riscv: htif: increase the priority of the htif subregion > > riscv: requires libfdt > > Hi Michael -- these patches from Fred need your Signed-off-by: > line in the commit messages as well as his. (This indicates that > they've passed through your hands to get to master; see item 3 > in https://wiki.qemu.org/Contribute/SubmitAPullRequest ) > I was thinking about that but was not sure. This is my first pull with changes that originated outside the riscv tree. > I did a build test and they passed that fine, so if you can just > respin this pullreq with the signoff lines added to each commit > message I should be able to apply it. > Okay, thanks. I've boot tested Linux kernel in the 'spike' machines. These changes are very low risk. I'll respin with my "Signed-off-by" added. The configure libfdt patch is the main one to cherry-pick for 2.12.1, however, the other changes are innocuous enough if you want to merge the whole PR to stable, as there are 3 commits. BTW I still need to test 'sifive_u' in stable 2.12.0 to figure out the changes that need to be cherry-picked or I can backport changes if necessary.
[Qemu-devel] [PULL 3/3] riscv: requires libfdt
From: KONRAD Frederic <frederic.kon...@adacore.com> When compiling on a machine without libfdt installed the configure script should try to get libfdt from the git or should die because otherwise CONFIG_LIBFDT is not set and the build process end in an error in the link phase.. eg: hw/riscv/virt.o: In function `riscv_virt_board_init': qemu/src/hw/riscv/virt.c:317: undefined reference to `qemu_fdt_setprop_cell' qemu/src/hw/riscv/virt.c:319: undefined reference to `qemu_fdt_setprop_cell' qemu/src/hw/riscv/virt.c:345: undefined reference to `qemu_fdt_dumpdtb' collect2: error: ld returned 1 exit status make[1]: *** [qemu-system-riscv64] Error 1 make: *** [subdir-riscv64-softmmu] Error 2 Reviewed-by: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Reviewed-by: Philippe Mathieu-Daudé <f4...@amsat.org> Reviewed-by: Michael Clark <m...@sifive.com> Signed-off-by: KONRAD Frederic <frederic.kon...@adacore.com> Message-Id: <1525360636-18229-4-git-send-email-frederic.kon...@adacore.com> --- configure | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/configure b/configure index 1443422e83f1..479e9a787e6e 100755 --- a/configure +++ b/configure @@ -3761,7 +3761,7 @@ fi fdt_required=no for target in $target_list; do case $target in - aarch64*-softmmu|arm*-softmmu|ppc*-softmmu|microblaze*-softmmu|mips64el-softmmu) + aarch64*-softmmu|arm*-softmmu|ppc*-softmmu|microblaze*-softmmu|mips64el-softmmu|riscv*-softmmu) fdt_required=yes ;; esac -- 2.7.0
[Qemu-devel] [PULL 2/3] riscv: htif: increase the priority of the htif subregion
From: KONRAD Frederic <frederic.kon...@adacore.com> The htif device is supposed to be mapped over an other subregion. So increase its priority to one to avoid any conflict. Here is the output of info mtree: Before: (qemu) info mtree address-space: memory - (prio 0, i/o): system -000f (prio 0, i/o): riscv.htif.uart -00011fff (prio 0, ram): riscv.spike.bootrom 0200-0200 (prio 0, i/o): riscv.sifive.clint 8000-87ff (prio 0, ram): riscv.spike.ram address-space: I/O - (prio 0, i/o): io address-space: cpu-memory-0 - (prio 0, i/o): system -000f (prio 0, i/o): riscv.htif.uart -00011fff (prio 0, ram): riscv.spike.bootrom 0200-0200 (prio 0, i/o): riscv.sifive.clint 8000-87ff (prio 0, ram): riscv.spike.ram After: (qemu) info mtree address-space: memory - (prio 0, i/o): system -000f (prio 1, i/o): riscv.htif.uart -00011fff (prio 0, ram): riscv.spike.bootrom 0200-0200 (prio 0, i/o): riscv.sifive.clint 8000-87ff (prio 0, ram): riscv.spike.ram address-space: I/O - (prio 0, i/o): io address-space: cpu-memory-0 - (prio 0, i/o): system -000f (prio 1, i/o): riscv.htif.uart -00011fff (prio 0, ram): riscv.spike.bootrom 0200-0200 (prio 0, i/o): riscv.sifive.clint 8000-87ff (prio 0, ram): riscv.spike.ram Reviewed-by: Michael Clark <m...@sifive.com> Signed-off-by: KONRAD Frederic <frederic.kon...@adacore.com> Message-Id: <1525360636-18229-3-git-send-email-frederic.kon...@adacore.com> --- hw/riscv/riscv_htif.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/hw/riscv/riscv_htif.c b/hw/riscv/riscv_htif.c index be252ec8cce9..f73512941fb6 100644 --- a/hw/riscv/riscv_htif.c +++ b/hw/riscv/riscv_htif.c @@ -253,8 +253,9 @@ HTIFState *htif_mm_init(MemoryRegion *address_space, MemoryRegion *main_mem, htif_be_change, s, NULL, true); if (address_symbol_set == 3) { memory_region_init_io(>mmio, NULL, _mm_ops, s, -TYPE_HTIF_UART, size); -memory_region_add_subregion(address_space, base, >mmio); + TYPE_HTIF_UART, size); +memory_region_add_subregion_overlap(address_space, base, +>mmio, 1); } return s; -- 2.7.0
[Qemu-devel] [PULL 1/3] riscv: spike: allow base == 0
From: KONRAD Frederic <frederic.kon...@adacore.com> The sanity check on base doesn't allow htif to be mapped @0. Check if the symbol exists instead so we can map it where we want. Reviewed-by: Michael Clark <m...@sifive.com> Signed-off-by: KONRAD Frederic <frederic.kon...@adacore.com> Message-Id: <1525360636-18229-2-git-send-email-frederic.kon...@adacore.com> --- hw/riscv/riscv_htif.c | 7 +-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/hw/riscv/riscv_htif.c b/hw/riscv/riscv_htif.c index 3e17f3025106..be252ec8cce9 100644 --- a/hw/riscv/riscv_htif.c +++ b/hw/riscv/riscv_htif.c @@ -41,17 +41,20 @@ } while (0) static uint64_t fromhost_addr, tohost_addr; +static int address_symbol_set; void htif_symbol_callback(const char *st_name, int st_info, uint64_t st_value, -uint64_t st_size) + uint64_t st_size) { if (strcmp("fromhost", st_name) == 0) { +address_symbol_set |= 1; fromhost_addr = st_value; if (st_size != 8) { error_report("HTIF fromhost must be 8 bytes"); exit(1); } } else if (strcmp("tohost", st_name) == 0) { +address_symbol_set |= 2; tohost_addr = st_value; if (st_size != 8) { error_report("HTIF tohost must be 8 bytes"); @@ -248,7 +251,7 @@ HTIFState *htif_mm_init(MemoryRegion *address_space, MemoryRegion *main_mem, qemu_chr_fe_init(>chr, chr, _abort); qemu_chr_fe_set_handlers(>chr, htif_can_recv, htif_recv, htif_event, htif_be_change, s, NULL, true); -if (base) { +if (address_symbol_set == 3) { memory_region_init_io(>mmio, NULL, _mm_ops, s, TYPE_HTIF_UART, size); memory_region_add_subregion(address_space, base, >mmio); -- 2.7.0
[Qemu-devel] [PULL 0/3] RISC-V: QEMU 2.13 Minor Fixes
The following changes since commit c8b7e627b4269a3bc3ae41d9f420547a47e6d9b9: Merge remote-tracking branch 'remotes/ericb/tags/pull-nbd-2018-05-04' into staging (2018-05-04 14:42:46 +0100) are available in the git repository at: https://github.com/riscv/riscv-qemu.git tags/riscv-qemu-2.13-minor-fixes-1 for you to fetch changes up to d4a63373c0e6349c6f36f56a6290ad639bfa20cb: riscv: requires libfdt (2018-05-08 10:57:21 +1200) RISC-V: QEMU 2.13 Minor Fixes * Require libfdt when configuring for 'riscv*-softmmu' * Increase HTIF priority and allow zero address KONRAD Frederic (3): riscv: spike: allow base == 0 riscv: htif: increase the priority of the htif subregion riscv: requires libfdt configure | 2 +- hw/riscv/riscv_htif.c | 12 2 files changed, 9 insertions(+), 5 deletions(-) -- 2.7.0
[Qemu-devel] [PULL 19/20] RISC-V: No traps on writes to misa, minstret, mcycle
These fields are marked WARL (Write Any Values, Reads Legal Values) in the RISC-V Privileged Architecture Specification so instead of raising exceptions, illegal writes are silently dropped. Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Alistair Francis <alistair.fran...@wdc.com> Signed-off-by: Michael Clark <m...@sifive.com> --- target/riscv/op_helper.c | 25 + 1 file changed, 13 insertions(+), 12 deletions(-) diff --git a/target/riscv/op_helper.c b/target/riscv/op_helper.c index af0c52a48418..3abf52453cfc 100644 --- a/target/riscv/op_helper.c +++ b/target/riscv/op_helper.c @@ -213,17 +213,19 @@ void csr_write_helper(CPURISCVState *env, target_ulong val_to_write, break; } case CSR_MINSTRET: -qemu_log_mask(LOG_UNIMP, "CSR_MINSTRET: write not implemented"); -goto do_illegal; +/* minstret is WARL so unsupported writes are ignored */ +break; case CSR_MCYCLE: -qemu_log_mask(LOG_UNIMP, "CSR_MCYCLE: write not implemented"); -goto do_illegal; +/* mcycle is WARL so unsupported writes are ignored */ +break; +#if defined(TARGET_RISCV32) case CSR_MINSTRETH: -qemu_log_mask(LOG_UNIMP, "CSR_MINSTRETH: write not implemented"); -goto do_illegal; +/* minstreth is WARL so unsupported writes are ignored */ +break; case CSR_MCYCLEH: -qemu_log_mask(LOG_UNIMP, "CSR_MCYCLEH: write not implemented"); -goto do_illegal; +/* mcycleh is WARL so unsupported writes are ignored */ +break; +#endif case CSR_MUCOUNTEREN: if (env->priv_ver <= PRIV_VERSION_1_09_1) { env->scounteren = val_to_write; @@ -337,10 +339,9 @@ void csr_write_helper(CPURISCVState *env, target_ulong val_to_write, case CSR_MBADADDR: env->mbadaddr = val_to_write; break; -case CSR_MISA: { -qemu_log_mask(LOG_UNIMP, "CSR_MISA: misa writes not supported"); -goto do_illegal; -} +case CSR_MISA: +/* misa is WARL so unsupported writes are ignored */ +break; case CSR_PMPCFG0: case CSR_PMPCFG1: case CSR_PMPCFG2: -- 2.7.0
[Qemu-devel] [PULL 17/20] RISC-V: Add mcycle/minstret support for -icount auto
Previously the mycycle/minstret CSRs and rdcycle/rdinstret psuedo instructions would return the time as a proxy for an increasing instruction counter in the absence of having a precise instruction count. If QEMU is invoked with -icount, the mcycle/minstret CSRs and rdcycle/rdinstret psuedo instructions will return the instruction count. Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Alistair Francis <alistair.fran...@wdc.com> Signed-off-by: Michael Clark <m...@sifive.com> Reviewed-by: Alistair Francis <alistair.fran...@wdc.com> --- target/riscv/op_helper.c | 28 ++-- target/riscv/translate.c | 2 ++ 2 files changed, 28 insertions(+), 2 deletions(-) diff --git a/target/riscv/op_helper.c b/target/riscv/op_helper.c index 7416412b187c..3512462f4fd8 100644 --- a/target/riscv/op_helper.c +++ b/target/riscv/op_helper.c @@ -434,25 +434,49 @@ target_ulong csr_read_helper(CPURISCVState *env, target_ulong csrno) case CSR_INSTRET: case CSR_CYCLE: if (ctr_ok) { +#if !defined(CONFIG_USER_ONLY) +if (use_icount) { +return cpu_get_icount(); +} else { +return cpu_get_host_ticks(); +} +#else return cpu_get_host_ticks(); +#endif } break; #if defined(TARGET_RISCV32) case CSR_INSTRETH: case CSR_CYCLEH: if (ctr_ok) { +#if !defined(CONFIG_USER_ONLY) +if (use_icount) { +return cpu_get_icount() >> 32; +} else { +return cpu_get_host_ticks() >> 32; +} +#else return cpu_get_host_ticks() >> 32; +#endif } break; #endif #ifndef CONFIG_USER_ONLY case CSR_MINSTRET: case CSR_MCYCLE: -return cpu_get_host_ticks(); +if (use_icount) { +return cpu_get_icount(); +} else { +return cpu_get_host_ticks(); +} case CSR_MINSTRETH: case CSR_MCYCLEH: #if defined(TARGET_RISCV32) -return cpu_get_host_ticks() >> 32; +if (use_icount) { +return cpu_get_icount() >> 32; +} else { +return cpu_get_host_ticks() >> 32; +} #endif break; case CSR_MUCOUNTEREN: diff --git a/target/riscv/translate.c b/target/riscv/translate.c index c3a029afefd9..c0e6a044d383 100644 --- a/target/riscv/translate.c +++ b/target/riscv/translate.c @@ -1390,6 +1390,7 @@ static void gen_system(CPURISCVState *env, DisasContext *ctx, uint32_t opc, break; default: tcg_gen_movi_tl(imm_rs1, rs1); +gen_io_start(); switch (opc) { case OPC_RISC_CSRRW: gen_helper_csrrw(dest, cpu_env, source1, csr_store); @@ -1413,6 +1414,7 @@ static void gen_system(CPURISCVState *env, DisasContext *ctx, uint32_t opc, gen_exception_illegal(ctx); return; } +gen_io_end(); gen_set_gpr(rd, dest); /* end tb since we may be changing priv modes, to get mmu_index right */ tcg_gen_movi_tl(cpu_pc, ctx->next_pc); -- 2.7.0
[Qemu-devel] [PULL 10/20] RISC-V: Remove EM_RISCV ELF_MACHINE indirection
Pointless indirection. Other ports use EM_ constants directly. Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Signed-off-by: Michael Clark <m...@sifive.com> Signed-off-by: Palmer Dabbelt <pal...@sifive.com> Reviewed-by: Philippe Mathieu-Daudé <f4...@amsat.org> Reviewed-by: Alistair Francis <alistair.fran...@wdc.com> --- hw/riscv/sifive_e.c | 2 +- hw/riscv/sifive_u.c | 2 +- hw/riscv/spike.c| 2 +- hw/riscv/virt.c | 2 +- target/riscv/cpu.h | 1 - 5 files changed, 4 insertions(+), 5 deletions(-) diff --git a/hw/riscv/sifive_e.c b/hw/riscv/sifive_e.c index 22dc526713c2..6fa223818502 100644 --- a/hw/riscv/sifive_e.c +++ b/hw/riscv/sifive_e.c @@ -88,7 +88,7 @@ static uint64_t load_kernel(const char *kernel_filename) if (load_elf(kernel_filename, NULL, NULL, _entry, NULL, _high, - 0, ELF_MACHINE, 1, 0) < 0) { + 0, EM_RISCV, 1, 0) < 0) { error_report("qemu: could not load kernel '%s'", kernel_filename); exit(1); } diff --git a/hw/riscv/sifive_u.c b/hw/riscv/sifive_u.c index 5bb495ab9a6c..84afed4c3b0e 100644 --- a/hw/riscv/sifive_u.c +++ b/hw/riscv/sifive_u.c @@ -74,7 +74,7 @@ static uint64_t load_kernel(const char *kernel_filename) if (load_elf(kernel_filename, NULL, NULL, _entry, NULL, _high, - 0, ELF_MACHINE, 1, 0) < 0) { + 0, EM_RISCV, 1, 0) < 0) { error_report("qemu: could not load kernel '%s'", kernel_filename); exit(1); } diff --git a/hw/riscv/spike.c b/hw/riscv/spike.c index 44eab94e17ef..9e18c618bfbb 100644 --- a/hw/riscv/spike.c +++ b/hw/riscv/spike.c @@ -64,7 +64,7 @@ static uint64_t load_kernel(const char *kernel_filename) uint64_t kernel_entry, kernel_high; if (load_elf_ram_sym(kernel_filename, NULL, NULL, -_entry, NULL, _high, 0, ELF_MACHINE, 1, 0, +_entry, NULL, _high, 0, EM_RISCV, 1, 0, NULL, true, htif_symbol_callback) < 0) { error_report("qemu: could not load kernel '%s'", kernel_filename); exit(1); diff --git a/hw/riscv/virt.c b/hw/riscv/virt.c index 88b9ad509315..7ef9ba26debc 100644 --- a/hw/riscv/virt.c +++ b/hw/riscv/virt.c @@ -68,7 +68,7 @@ static uint64_t load_kernel(const char *kernel_filename) if (load_elf(kernel_filename, NULL, NULL, _entry, NULL, _high, - 0, ELF_MACHINE, 1, 0) < 0) { + 0, EM_RISCV, 1, 0) < 0) { error_report("qemu: could not load kernel '%s'", kernel_filename); exit(1); } diff --git a/target/riscv/cpu.h b/target/riscv/cpu.h index 41e06ac0f91d..9871e6feb1de 100644 --- a/target/riscv/cpu.h +++ b/target/riscv/cpu.h @@ -34,7 +34,6 @@ #define TCG_GUEST_DEFAULT_MO 0 -#define ELF_MACHINE EM_RISCV #define CPUArchState struct CPURISCVState #include "qemu-common.h" -- 2.7.0
[Qemu-devel] [PULL 13/20] RISC-V: Hardwire satp to 0 for no-mmu case
satp is WARL so it should not trap on illegal writes, rather it can be hardwired to zero and silently ignore illegal writes. It seems the RISC-V WARL behaviour is preferred to having to trap overhead versus simply reading back the value and checking if the write took (saves hundreds of cycles and more complex trap handling code). Cc: Sagar Karandikar <sag...@eecs.berkeley.edu> Cc: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> Cc: Palmer Dabbelt <pal...@sifive.com> Cc: Alistair Francis <alistair.fran...@wdc.com> Signed-off-by: Michael Clark <m...@sifive.com> Reviewed-by: Alistair Francis <alistair.fran...@wdc.com> --- target/riscv/op_helper.c | 7 +-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/target/riscv/op_helper.c b/target/riscv/op_helper.c index 7c6068bac958..101dac1ee8dc 100644 --- a/target/riscv/op_helper.c +++ b/target/riscv/op_helper.c @@ -255,7 +255,7 @@ void csr_write_helper(CPURISCVState *env, target_ulong val_to_write, } case CSR_SATP: /* CSR_SPTBR */ { if (!riscv_feature(env, RISCV_FEATURE_MMU)) { -goto do_illegal; +break; } if (env->priv_ver <= PRIV_VERSION_1_09_1 && (val_to_write ^ env->sptbr)) { @@ -465,7 +465,10 @@ target_ulong csr_read_helper(CPURISCVState *env, target_ulong csrno) return env->scounteren; case CSR_SCAUSE: return env->scause; -case CSR_SPTBR: +case CSR_SATP: /* CSR_SPTBR */ +if (!riscv_feature(env, RISCV_FEATURE_MMU)) { +return 0; +} if (env->priv_ver >= PRIV_VERSION_1_10_0) { return env->satp; } else { -- 2.7.0