Re: [RFC] Add "riscv-" for embedded cross toolchains
On Mon, Jul 9, 2018 at 10:15 PM, Ben Elliston wrote: > OK--I don't terribly mind what is done as long as it is clear and > uniform. If you think riscv/riscv-* should be allowed, I am happy to > put it back. Yes, please put it back, as riscv-rtems is using it. Jim ___ config-patches mailing list config-patches@gnu.org https://lists.gnu.org/mailman/listinfo/config-patches
Re: [RFC] Add "riscv-" for embedded cross toolchains
On Mon, Jul 09, 2018 at 08:41:09AM -0700, Jim Wilson wrote: > We already merged config.sub patches into binutils and gcc, and added > riscv-* support to them. So at this point I think we just have to > allow riscv-*-* even though we don't want it. I don't expect to use > it and I don't expect to maintain it, but some stubborn people will > never give up on it, so we have to allow it. > > I don't care about Liviu, but there is a riscv-rtems port rolling out. > We had agreement to expand riscv-rtems to riscv32-rtems at config.sub > time, but then you modified config.sub to remove the riscv->riscv32 > conversion, so the riscv-rtems port went out as just riscv-rtems and > we can't drop support for that now. OK--I don't terribly mind what is done as long as it is clear and uniform. If you think riscv/riscv-* should be allowed, I am happy to put it back. Ben signature.asc Description: PGP signature ___ config-patches mailing list config-patches@gnu.org https://lists.gnu.org/mailman/listinfo/config-patches
Re: [RFC] Add "riscv-" for embedded cross toolchains
On Mon, Jul 9, 2018 at 4:07 AM, Liviu Ionescu wrote: >> On 9 Jul 2018, at 12:20, Ben Elliston wrote: >> >> On Tue, Jul 03, 2018 at 10:41:49AM -0700, Jim Wilson wrote: >> >>> Accepting riscv-linux isn't OK. That will cause complications. >>> Though I'm not sure if that can be easily avoided if we allow >>> riscv-elf. We already merged config.sub patches into binutils and gcc, and added riscv-* support to them. So at this point I think we just have to allow riscv-*-* even though we don't want it. I don't expect to use it and I don't expect to maintain it, but some stubborn people will never give up on it, so we have to allow it. I don't care about Liviu, but there is a riscv-rtems port rolling out. We had agreement to expand riscv-rtems to riscv32-rtems at config.sub time, but then you modified config.sub to remove the riscv->riscv32 conversion, so the riscv-rtems port went out as just riscv-rtems and we can't drop support for that now. Jim ___ config-patches mailing list config-patches@gnu.org https://lists.gnu.org/mailman/listinfo/config-patches
Re: [RFC] Add "riscv-" for embedded cross toolchains
> On 9 Jul 2018, at 12:20, Ben Elliston wrote: > > On Tue, Jul 03, 2018 at 10:41:49AM -0700, Jim Wilson wrote: > >> Accepting riscv-linux isn't OK. That will cause complications. >> Though I'm not sure if that can be easily avoided if we allow >> riscv-elf. > >> There are no plans for any tools to accept riscv- as a CPU, so >> adding support for this to config.sub isn't very useful. > > In light of this, I have removed riscv- from config.sub because: This is regrettable. Jim has no interest in embedded cross toolchains, but this does not mean that there are no plans to accept `riscv-`, as proven by my binary distribution. Jim anyway solved the config.gcc configuration, so the initial concerns were no longer a priority. If you decide to permanently deny `riscv-`, I'll no longer maintain the embedded RISC-V toolchain, which, in my opinion, requires it. Regards, Liviu ___ config-patches mailing list config-patches@gnu.org https://lists.gnu.org/mailman/listinfo/config-patches
Re: [RFC] Add "riscv-" for embedded cross toolchains
On Tue, Jul 03, 2018 at 10:41:49AM -0700, Jim Wilson wrote: > Accepting riscv-linux isn't OK. That will cause complications. > Though I'm not sure if that can be easily avoided if we allow > riscv-elf. > There are no plans for any tools to accept riscv- as a CPU, so > adding support for this to config.sub isn't very useful. In light of this, I have removed riscv- from config.sub because: > Most GNU tools accept --program-transform-name as a configure > option, so it is easy to strip off the 32 or 64 in the program name > at install time if you wish. That should solve Liviu's problem. .. I had forgotten about --program-transform-name. That will indeed solve the problem _if_ you want your toolchain binaries prefixed with riscv instead of riscv32/riscv64. Cheers, Ben PS: nice to hear from you, Jim. signature.asc Description: PGP signature ___ config-patches mailing list config-patches@gnu.org https://lists.gnu.org/mailman/listinfo/config-patches
Re: [RFC] Add "riscv-" for embedded cross toolchains
On Tue, Jul 3, 2018 at 10:54 AM, Liviu Ionescu wrote: >> I think Liviu's argument has a flaw. While it is true that both riscv32 and >> risvc64 can emit both 32-bit and 64-bit code, they have different defaults. >> riscv32 emits 32-bit by default, and riscv64 emits 64-bit by default. That >> is a property we lose if we allow riscv, because it won't be obvious if >> 32-bit or 64-bit code is the default. > > The defaults you mention apply to the Linux compilers, which are not the > topic of this thread. Ben's latest patch suggestion adds riscv-linux support, which makes linux concerns relevant to this thread. > To minimise uncertainty, the embedded cross compilers should not provide such > defaults, the `-march` option should be mandatory, preferably the `-mabi` > option too. This would break the embedded cross compilers for most people. I don't think it would be wise to do this. I'm not aware of any other GNU toolchain that works this way. There is always a default architecture and ABI. Jim ___ config-patches mailing list config-patches@gnu.org https://lists.gnu.org/mailman/listinfo/config-patches
Re: [RFC] Add "riscv-" for embedded cross toolchains
> On 3 Jul 2018, at 20:41, Jim Wilson wrote: > > I think Liviu's argument has a flaw. While it is true that both riscv32 and > risvc64 can emit both 32-bit and 64-bit code, they have different defaults. > riscv32 emits 32-bit by default, and riscv64 emits 64-bit by default. That > is a property we lose if we allow riscv, because it won't be obvious if > 32-bit or 64-bit code is the default. The defaults you mention apply to the Linux compilers, which are not the topic of this thread. To minimise uncertainty, the embedded cross compilers should not provide such defaults, the `-march` option should be mandatory, preferably the `-mabi` option too. It would be good if the RISC-V compiler would implement this. Regards, Liviu ___ config-patches mailing list config-patches@gnu.org https://lists.gnu.org/mailman/listinfo/config-patches
Re: [RFC] Add "riscv-" for embedded cross toolchains
On 07/03/2018 03:29 AM, Ben Elliston wrote: +riscv-linuxriscv-unknown-linux-gnu Accepting riscv-linux isn't OK. That will cause complications. Though I'm not sure if that can be easily avoided if we allow riscv-elf. There are no plans for any tools to accept riscv- as a CPU, so adding support for this to config.sub isn't very useful. Personally, I think Palmer's suggestion of mapping riscv to riscv32 is the best compromise. I realize that Liviu isn't happy about that, but we can't do much about that. Most GNU tools accept --program-transform-name as a configure option, so it is easy to strip off the 32 or 64 in the program name at install time if you wish. That should solve Liviu's problem. I think Liviu's argument has a flaw. While it is true that both riscv32 and risvc64 can emit both 32-bit and 64-bit code, they have different defaults. riscv32 emits 32-bit by default, and riscv64 emits 64-bit by default. That is a property we lose if we allow riscv, because it won't be obvious if 32-bit or 64-bit code is the default. Jim ___ config-patches mailing list config-patches@gnu.org https://lists.gnu.org/mailman/listinfo/config-patches
Re: [RFC] Add "riscv-" for embedded cross toolchains
> On 3 Jul 2018, at 13:29, Ben Elliston wrote: > > OK, how about this? This patch makes everything uniform for my embedded toolchain it seems ok. however please note that I did not try to update my `riscv-none-embed-` build scripts to use this new structure; actually my scripts are still based on the 7.x version of the RISC-V repositories, so they might need some adjustments. but I guess I can manage this. regards, Liviu ___ config-patches mailing list config-patches@gnu.org https://lists.gnu.org/mailman/listinfo/config-patches
Re: [RFC] Add "riscv-" for embedded cross toolchains
OK, how about this? This patch makes everything uniform (see test cases below for an illustration of behaviour): diff --git a/config.sub b/config.sub index 072700f..c95acc6 100755 --- a/config.sub +++ b/config.sub @@ -625,7 +625,7 @@ case $basic_machine in | powerpc | powerpc64 | powerpc64le | powerpcle \ | pru \ | pyramid \ - | riscv32 | riscv64 \ + | riscv | riscv32 | riscv64 \ | rl78 | rx \ | score \ | sh | sh[1234] | sh[24]a | sh[24]aeb | sh[23]e | sh[234]eb | sheb | shbe | shle | sh[1234]le | sh3ele \ @@ -752,7 +752,7 @@ case $basic_machine in | powerpc-* | powerpc64-* | powerpc64le-* | powerpcle-* \ | pru-* \ | pyramid-* \ - | riscv32-* | riscv64-* \ + | riscv-* | riscv32-* | riscv64-* \ | rl78-* | romp-* | rs6000-* | rx-* \ | sh-* | sh[1234]-* | sh[24]a-* | sh[24]aeb-* | sh[23]e-* | sh[34]eb-* | sheb-* | shbe-* \ | shle-* | sh[1234]le-* | sh3ele-* | sh64-* | sh64le-* \ @@ -1125,12 +1125,6 @@ case $basic_machine in ps2) basic_machine=i386-ibm ;; - riscv) - basic_machine=riscv32-unknown - ;; - riscv-*) - basic_machine=`echo "$basic_machine" | sed 's/^riscv/riscv32/'` - ;; rm[46]00) basic_machine=mips-siemens ;; diff --git a/testsuite/config-sub.data b/testsuite/config-sub.data index 7ad1f76..5ab1807 100644 --- a/testsuite/config-sub.data +++ b/testsuite/config-sub.data @@ -422,11 +422,11 @@ riscv64-elf riscv64-unknown-elf riscv64-hcos riscv64-unknown-hcos riscv64-linux riscv64-unknown-linux-gnu riscv64riscv64-unknown-none -riscv-company-elf riscv32-company-elf -riscv-company-hcos riscv32-company-hcos -riscv-elf riscv32-unknown-elf -riscv-linuxriscv32-unknown-linux-gnu -riscv riscv32-unknown-none +riscv-company-elf riscv-company-elf +riscv-company-hcos riscv-company-hcos +riscv-elf riscv-unknown-elf +riscv-linuxriscv-unknown-linux-gnu +riscv riscv-unknown-none rl78-linux rl78-unknown-linux-gnu rl78 rl78-unknown-none rx-linux rx-unknown-linux-gnu signature.asc Description: PGP signature ___ config-patches mailing list config-patches@gnu.org https://lists.gnu.org/mailman/listinfo/config-patches
Re: [RFC] Add "riscv-" for embedded cross toolchains
> On 3 Jul 2018, at 07:45, Ben Elliston wrote: > > How is this? > > - | riscv32-* | riscv64-* \ > + | riscv-* | riscv32-* | riscv64-* \ I think this is ok > riscv-*) > basic_machine=`echo "$basic_machine" | sed 's/^riscv/riscv32/'` I'm not sure I understand this, but automatically replacing riscv- by riscv32- is not ok. Regards, Liviu ___ config-patches mailing list config-patches@gnu.org https://lists.gnu.org/mailman/listinfo/config-patches
Re: [RFC] Add "riscv-" for embedded cross toolchains
How is this? 2018-07-03 Liviu Ionescu Ben Elliston * config.sub: Recognise riscv-*. * testsuite/config-sub.data: Adjust tests. diff --git a/config.sub b/config.sub index 072700f..76c4640 100755 --- a/config.sub +++ b/config.sub @@ -752,7 +752,7 @@ case $basic_machine in | powerpc-* | powerpc64-* | powerpc64le-* | powerpcle-* \ | pru-* \ | pyramid-* \ - | riscv32-* | riscv64-* \ + | riscv-* | riscv32-* | riscv64-* \ | rl78-* | romp-* | rs6000-* | rx-* \ | sh-* | sh[1234]-* | sh[24]a-* | sh[24]aeb-* | sh[23]e-* | sh[34]eb-* | sheb-* | shbe-* \ | shle-* | sh[1234]le-* | sh3ele-* | sh64-* | sh64le-* \ @@ -1126,7 +1126,7 @@ case $basic_machine in basic_machine=i386-ibm ;; riscv) - basic_machine=riscv32-unknown + basic_machine=riscv-unknown ;; riscv-*) basic_machine=`echo "$basic_machine" | sed 's/^riscv/riscv32/'` diff --git a/testsuite/config-sub.data b/testsuite/config-sub.data index 7ad1f76..5ab1807 100644 --- a/testsuite/config-sub.data +++ b/testsuite/config-sub.data @@ -422,11 +422,11 @@ riscv64-elf riscv64-unknown-elf riscv64-hcos riscv64-unknown-hcos riscv64-linux riscv64-unknown-linux-gnu riscv64riscv64-unknown-none -riscv-company-elf riscv32-company-elf -riscv-company-hcos riscv32-company-hcos -riscv-elf riscv32-unknown-elf -riscv-linuxriscv32-unknown-linux-gnu -riscv riscv32-unknown-none +riscv-company-elf riscv-company-elf +riscv-company-hcos riscv-company-hcos +riscv-elf riscv-unknown-elf +riscv-linuxriscv-unknown-linux-gnu +riscv riscv-unknown-none rl78-linux rl78-unknown-linux-gnu rl78 rl78-unknown-none rx-linux rx-unknown-linux-gnu signature.asc Description: PGP signature ___ config-patches mailing list config-patches@gnu.org https://lists.gnu.org/mailman/listinfo/config-patches
[RFC] Add "riscv-" for embedded cross toolchains
Hi, Thank you all for the valuable feedback on the previous related thread (http://lists.gnu.org/archive/html/config-patches/2018-06/threads.html). However I think the topic needs a reset. The initial assumption for the patch was: > Most RISC-V toolchains can generate code for all RISC-V targets, but are > currently named "riscv32-*" and "riscv64-*". Users have suggested that > these names are unintuitive and would like tuples to be "riscv-*". Well, I don't know other users who suggested this, so I'll speak only for myself, based on the experience with the GNU MCU Eclipse RISC-V Embedded GCC distribution. First, I'll state what this proposal **is not**: - it **is not** about renaming the Linux RISC-V compiler, in any Linux distribution - it **is not** about automatically renaming `riscv-` to anything else - generally it **is not** about disturbing any native RISC-V compiler, in any Linux distribution, in any way; from my point of view the naming scheme for the native compilers is ok, and I see no reason to change it. However, the use case for native compilers is different from the use case for embedded cross toolchains, since embedded cross toolchains: - do not need to be 'guessed' for capabilities in configure scripts, - are not related, in terms of 32/64-bit capabilities, to the host platform where they run, - generally are distributed as a single executable that is capable of generating binaries for multiple architectures/ABIs, with multilibs. The RISC-V non-Linux toolchain is no exception, a single executable can generate binaries for many architectures/ABIs based on separate `-march` & `-mabi` options (in their RISC-V GCC distribution, Microsemi identified 39(!) different multilib combinations). In this context, it is more natural to invoke the compile using something like: riscv-xxx-gcc -march=rv32i -mabi=ilp32 <- generate a 32-bit binary riscv-xxx-gcc -march=rv64i -mabi=lp64 <- generate a 64-bit binary Since the current config.sub file does not accept `riscv-`, and only explicit `riscv32-` and `riscv64-` are allowed, this leads to confusing invocations, like: riscv32-unknown-elf-gcc -march=rv64i -mabi=lp64 <- generate a 64-bit binary The current proposal adds exactly this, the `riscv-*` tuple prefix alongside the existing `riscv32-*` and `riscv64-*`. I think that this can be an incremental step towards a more elaborate mechanism that John defined as "config-for-target vs config-for-toolchain". I agree that "exact-target" and "toolchain" use-cases are fundamentally different, and one specifics should not be imposed on the other. I definitely welcome John volunteering to implement such a mechanism. Regards, Liviu p.s. For completeness, the proposed patch, as already used in GNU MCU Eclipse RISC-V Embedded GCC is: ``` diff --git a/config.sub b/config.sub index d1f5b54..cc4d9ed 100755 --- a/config.sub +++ b/config.sub @@ -752,7 +752,7 @@ case $basic_machine in | powerpc-* | powerpc64-* | powerpc64le-* | powerpcle-* \ | pru-* \ | pyramid-* \ - | riscv32-* | riscv64-* \ + | riscv32-* | riscv64-* | riscv-* \ | rl78-* | romp-* | rs6000-* | rx-* \ | sh-* | sh[1234]-* | sh[24]a-* | sh[24]aeb-* | sh[23]e-* | sh[34]eb-* | sheb-* | shbe-* \ | shle-* | sh[1234]le-* | sh3ele-* | sh64-* | sh64le-* \ @@ -1522,6 +1522,8 @@ case $os in ;; none) ;; + embed) + ;; *-eabi) ;; *) ``` In addition, this patch adds `embed` as an acceptable suffix for the tuple, for explicit use in embedded toolchains that do not have a separate embedded ABI (RISC-V case for now). ___ config-patches mailing list config-patches@gnu.org https://lists.gnu.org/mailman/listinfo/config-patches
