Message ID | 20240110040203.1920924-1-hchauhan@ventanamicro.com |
---|---|
Headers | show |
Series | Export debug triggers as an extension | expand |
Himanshu, We spoke offline but let's make everyone aware: - 'sdtrig' should be marked with 'x-' and be an experimental extension since the spec isn't yet frozen; - Alvin sent a patch to the ML adding the 'mcontext' CSR for 'sdtrig' some time ago: "[PATCH v2] target/riscv: Implement optional CSR mcontext of debug Sdtrig extension" It would be good to put his patch on top of this series to ease the review for everyone. The changes done in patch 2 would also be applicable to the mcontext CSR; - last but probably the most important: the existing 'debug' flag seems to be acting as the actual 'sdtrig' extension due to how the flag is gating trigger code, e.g.: if (cpu->cfg.debug) { riscv_trigger_realize(&cpu->env); } and if (cpu->cfg.debug) { riscv_trigger_reset_hold(env); } If that's really the case, all the checks with cpu->cfg.debug will need to also include cpu->cfg.ext_sdtrig (one or the other). And now we'll have to make an option: do we leave the debug triggers (i.e. the 'debug' flag) as always enabled? If it's up to me I would make 'debug' as default 'false' and deprecate it. Users will need to enable the debug triggers via x-sdtrig=true from now on. This will break existing behavior, but the way it is now we're always enabling an extension (via the debug flag) that isn't even frozen, so we're already in the wrong. Alistair, any thoughts? Thanks, Daniel On 1/10/24 01:02, Himanshu Chauhan wrote: > All the CPUs may or may not implement the debug trigger (sdtrig) > extension. The presence of it should be dynamically detectable. > This patch exports the debug triggers as an extension which > can be turned on or off by sdtrig=<true/false> option. It is > turned on by default. > > "sdtrig" is concatenated to ISA string when it is enabled. > Like so: > rv64imafdch_zicbom_*_sdtrig_*_sstc_svadu > > > Himanshu Chauhan (2): > target/riscv: Export sdtrig as an extension and ISA string > target/riscv: Raise an exception when sdtrig is turned off > > target/riscv/cpu.c | 2 ++ > target/riscv/cpu_cfg.h | 1 + > target/riscv/csr.c | 20 ++++++++++++++++++++ > 3 files changed, 23 insertions(+) >
On Thu, Jan 11, 2024 at 5:20 AM Daniel Henrique Barboza <dbarboza@ventanamicro.com> wrote: > > Himanshu, > > We spoke offline but let's make everyone aware: > > - 'sdtrig' should be marked with 'x-' and be an experimental extension since > the spec isn't yet frozen; > > - Alvin sent a patch to the ML adding the 'mcontext' CSR for 'sdtrig' some time > ago: > > "[PATCH v2] target/riscv: Implement optional CSR mcontext of debug Sdtrig extension" > > It would be good to put his patch on top of this series to ease the review for everyone. > The changes done in patch 2 would also be applicable to the mcontext CSR; > > > - last but probably the most important: the existing 'debug' flag seems to be acting as > the actual 'sdtrig' extension due to how the flag is gating trigger code, e.g.: > > if (cpu->cfg.debug) { > riscv_trigger_realize(&cpu->env); > } > > and > > if (cpu->cfg.debug) { > riscv_trigger_reset_hold(env); > } > > > If that's really the case, all the checks with cpu->cfg.debug will need to also include > cpu->cfg.ext_sdtrig (one or the other). And now we'll have to make an option: do we leave > the debug triggers (i.e. the 'debug' flag) as always enabled? From memory the "debug" property is for the original debug spec: https://github.com/riscv/riscv-debug-spec/releases/tag/task_group_vote That was ratified and is an official extension. AFAIK this is what is in physical hardware as well. The actual PDF says draft though, I'm not sure what's going on there. The debug spec doesn't have a Z* name, so it's just "debug", at least AFAIK. "sdtrig" seems to be a new backwards-incompatible extension doing basically the same thing. What a mess > > If it's up to me I would make 'debug' as default 'false' and deprecate it. Users will need I don't think that's the right approach. It's a ratified extension that we are supporting and is in hardware. I think we are stuck supporting it > to enable the debug triggers via x-sdtrig=true from now on. This will break existing behavior, > but the way it is now we're always enabling an extension (via the debug flag) that isn't even > frozen, so we're already in the wrong. Maybe the debug group can chime in and say what they expect users to do? It seems like there are conflicting specs Alistair
On Fri, 2024-01-12 at 13:52 +1000, Alistair Francis wrote: > On Thu, Jan 11, 2024 at 5:20 AM Daniel Henrique Barboza > <dbarboza@ventanamicro.com> wrote: > > > > Himanshu, > > > > We spoke offline but let's make everyone aware: > > > > - 'sdtrig' should be marked with 'x-' and be an experimental > > extension since > > the spec isn't yet frozen; > > > > - Alvin sent a patch to the ML adding the 'mcontext' CSR for > > 'sdtrig' some time > > ago: > > > > "[PATCH v2] target/riscv: Implement optional CSR mcontext of debug > > Sdtrig extension" > > > > It would be good to put his patch on top of this series to ease the > > review for everyone. > > The changes done in patch 2 would also be applicable to the > > mcontext CSR; > > > > > > - last but probably the most important: the existing 'debug' flag > > seems to be acting as > > the actual 'sdtrig' extension due to how the flag is gating trigger > > code, e.g.: > > > > if (cpu->cfg.debug) { > > riscv_trigger_realize(&cpu->env); > > } > > > > and > > > > if (cpu->cfg.debug) { > > riscv_trigger_reset_hold(env); > > } > > > > > > If that's really the case, all the checks with cpu->cfg.debug will > > need to also include > > cpu->cfg.ext_sdtrig (one or the other). And now we'll have to make > > an option: do we leave > > the debug triggers (i.e. the 'debug' flag) as always enabled? > > From memory the "debug" property is for the original debug spec: > https://github.com/riscv/riscv-debug-spec/releases/tag/task_group_vote > > That was ratified and is an official extension. AFAIK this is what is > in physical hardware as well. > > The actual PDF says draft though, I'm not sure what's going on there. > > The debug spec doesn't have a Z* name, so it's just "debug", at least > AFAIK. > > "sdtrig" seems to be a new backwards-incompatible extension doing > basically the same thing. What a mess > > > > > If it's up to me I would make 'debug' as default 'false' and > > deprecate it. Users will need > > I don't think that's the right approach. It's a ratified extension > that we are supporting and is in hardware. I think we are stuck > supporting it > > I've done a bit of digging and I agree things are quite messy. Here are my discoveries: The debug option and the code for triggers was added in these commits: c9711bd778 target/riscv: cpu: Enable native debug feature 38b4e781a4 target/riscv: machine: Add debug state description b6092544fc target/riscv: csr: Hook debug CSR read/write 1acdb3b013 target/riscv: cpu: Add a config option for native debug 95799e36c1 target/riscv: Add initial support for the Sdtrig extension In March 2022 - since the commit refers to the Sdtrig extension name and from the date this was an implementation not of the ratified 0.13 debug spec (which did not have Sdtrig as a separate extension) but rather a version of the in development 1.0 debug spec. It's not trivial to tell if it's closer to the ratified 0.13 version or the (hopefully soon to be frozen) 1.0 version. As the only part of the debug specification to be implemented is the triggers then effectively the debug option is x-sdtrig. I don't think there is any way for code running on the machine to identify what version of the debug is implemented - the appropriate register is only available for external debug. Once 1.0 is frozen then the presence of Sdtrig isa string would indicate 1.0 trigger support is available. According to JIRA - https://jira.riscv.org/browse/RVS-981 the debug specification should freeze this month. How about considering this as a solution: - Add a new x-sdtrig option that defaults to false - Deprecate debug option - but retain it with default on - Add warning if triggers are used and x-sdtrig is not enabled - Update the trigger implementation to match frozen spec There is potentially a chance that some use cases will be broken but I don't think triggers are being widely use - the SBI support only just got merged: https://github.com/riscv-software-src/opensbi/commit/97f234f15c9657c6ec69fa6ed745be8107bf6ae2 Hope this is helpful, Rob
On 1/12/24 10:34, Rob Bradford wrote: > On Fri, 2024-01-12 at 13:52 +1000, Alistair Francis wrote: >> On Thu, Jan 11, 2024 at 5:20 AM Daniel Henrique Barboza >> <dbarboza@ventanamicro.com> wrote: >>> >>> Himanshu, >>> >>> We spoke offline but let's make everyone aware: >>> >>> - 'sdtrig' should be marked with 'x-' and be an experimental >>> extension since >>> the spec isn't yet frozen; >>> >>> - Alvin sent a patch to the ML adding the 'mcontext' CSR for >>> 'sdtrig' some time >>> ago: >>> >>> "[PATCH v2] target/riscv: Implement optional CSR mcontext of debug >>> Sdtrig extension" >>> >>> It would be good to put his patch on top of this series to ease the >>> review for everyone. >>> The changes done in patch 2 would also be applicable to the >>> mcontext CSR; >>> >>> >>> - last but probably the most important: the existing 'debug' flag >>> seems to be acting as >>> the actual 'sdtrig' extension due to how the flag is gating trigger >>> code, e.g.: >>> >>> if (cpu->cfg.debug) { >>> riscv_trigger_realize(&cpu->env); >>> } >>> >>> and >>> >>> if (cpu->cfg.debug) { >>> riscv_trigger_reset_hold(env); >>> } >>> >>> >>> If that's really the case, all the checks with cpu->cfg.debug will >>> need to also include >>> cpu->cfg.ext_sdtrig (one or the other). And now we'll have to make >>> an option: do we leave >>> the debug triggers (i.e. the 'debug' flag) as always enabled? >> >> From memory the "debug" property is for the original debug spec: >> https://github.com/riscv/riscv-debug-spec/releases/tag/task_group_vote >> >> That was ratified and is an official extension. AFAIK this is what is >> in physical hardware as well. >> >> The actual PDF says draft though, I'm not sure what's going on there. >> >> The debug spec doesn't have a Z* name, so it's just "debug", at least >> AFAIK. >> >> "sdtrig" seems to be a new backwards-incompatible extension doing >> basically the same thing. What a mess >> >>> >>> If it's up to me I would make 'debug' as default 'false' and >>> deprecate it. Users will need >> >> I don't think that's the right approach. It's a ratified extension >> that we are supporting and is in hardware. I think we are stuck >> supporting it >> >> > > I've done a bit of digging and I agree things are quite messy. Here are > my discoveries: > > The debug option and the code for triggers was added in these commits: > > c9711bd778 target/riscv: cpu: Enable native debug feature > 38b4e781a4 target/riscv: machine: Add debug state description > b6092544fc target/riscv: csr: Hook debug CSR read/write > 1acdb3b013 target/riscv: cpu: Add a config option for native debug > 95799e36c1 target/riscv: Add initial support for the Sdtrig extension > > In March 2022 - since the commit refers to the Sdtrig extension name > and from the date this was an implementation not of the ratified 0.13 > debug spec (which did not have Sdtrig as a separate extension) but > rather a version of the in development 1.0 debug spec. > > It's not trivial to tell if it's closer to the ratified 0.13 version or > the (hopefully soon to be frozen) 1.0 version. > > As the only part of the debug specification to be implemented is the > triggers then effectively the debug option is x-sdtrig. > > I don't think there is any way for code running on the machine to > identify what version of the debug is implemented - the appropriate > register is only available for external debug. Once 1.0 is frozen then > the presence of Sdtrig isa string would indicate 1.0 trigger support is > available. > > According to JIRA - https://jira.riscv.org/browse/RVS-981 the debug > specification should freeze this month. > > How about considering this as a solution: > > - Add a new x-sdtrig option that defaults to false > - Deprecate debug option - but retain it with default on > - Add warning if triggers are used and x-sdtrig is not enabled > - Update the trigger implementation to match frozen spec If x-sdtrig is 'false' and 'debug' is on, and then we warn if debug=true and x-sdtrig is false, we'll warn every time using the defaults. Given what you said here: > > There is potentially a chance that some use cases will be broken but I > don't think triggers are being widely use - the SBI support only just > got merged: > https://github.com/riscv-software-src/opensbi/commit/97f234f15c9657c6ec69fa6ed745be8107bf6a> I believe we can deprecate 'debug' and simply ignore its existence. Do everything else with x-sdtrig. So if an user sets any 'debug' value we'll: - warn that the flag is deprecated - set x-sdtrig to whatever value the user set to 'debug' 'debug' will become just an alternate way to set x-sdtrig. The logic should just check for x-sdtrig. Thanks, Daniel > Hope this is helpful, > > Rob >
> On 17-Jan-2024, at 10:29 PM, Daniel Henrique Barboza <dbarboza@ventanamicro.com> wrote: > > > > On 1/12/24 10:34, Rob Bradford wrote: >> On Fri, 2024-01-12 at 13:52 +1000, Alistair Francis wrote: >>> On Thu, Jan 11, 2024 at 5:20 AM Daniel Henrique Barboza >>> <dbarboza@ventanamicro.com> wrote: >>>> >>>> Himanshu, >>>> >>>> We spoke offline but let's make everyone aware: >>>> >>>> - 'sdtrig' should be marked with 'x-' and be an experimental >>>> extension since >>>> the spec isn't yet frozen; >>>> >>>> - Alvin sent a patch to the ML adding the 'mcontext' CSR for >>>> 'sdtrig' some time >>>> ago: >>>> >>>> "[PATCH v2] target/riscv: Implement optional CSR mcontext of debug >>>> Sdtrig extension" >>>> >>>> It would be good to put his patch on top of this series to ease the >>>> review for everyone. >>>> The changes done in patch 2 would also be applicable to the >>>> mcontext CSR; >>>> >>>> >>>> - last but probably the most important: the existing 'debug' flag >>>> seems to be acting as >>>> the actual 'sdtrig' extension due to how the flag is gating trigger >>>> code, e.g.: >>>> >>>> if (cpu->cfg.debug) { >>>> riscv_trigger_realize(&cpu->env); >>>> } >>>> >>>> and >>>> >>>> if (cpu->cfg.debug) { >>>> riscv_trigger_reset_hold(env); >>>> } >>>> >>>> >>>> If that's really the case, all the checks with cpu->cfg.debug will >>>> need to also include >>>> cpu->cfg.ext_sdtrig (one or the other). And now we'll have to make >>>> an option: do we leave >>>> the debug triggers (i.e. the 'debug' flag) as always enabled? >>> >>> From memory the "debug" property is for the original debug spec: >>> https://github.com/riscv/riscv-debug-spec/releases/tag/task_group_vote >>> >>> That was ratified and is an official extension. AFAIK this is what is >>> in physical hardware as well. >>> >>> The actual PDF says draft though, I'm not sure what's going on there. >>> >>> The debug spec doesn't have a Z* name, so it's just "debug", at least >>> AFAIK. >>> >>> "sdtrig" seems to be a new backwards-incompatible extension doing >>> basically the same thing. What a mess >>> >>>> >>>> If it's up to me I would make 'debug' as default 'false' and >>>> deprecate it. Users will need >>> >>> I don't think that's the right approach. It's a ratified extension >>> that we are supporting and is in hardware. I think we are stuck >>> supporting it >>> >>> >> I've done a bit of digging and I agree things are quite messy. Here are >> my discoveries: >> The debug option and the code for triggers was added in these commits: >> c9711bd778 target/riscv: cpu: Enable native debug feature >> 38b4e781a4 target/riscv: machine: Add debug state description >> b6092544fc target/riscv: csr: Hook debug CSR read/write >> 1acdb3b013 target/riscv: cpu: Add a config option for native debug >> 95799e36c1 target/riscv: Add initial support for the Sdtrig extension >> In March 2022 - since the commit refers to the Sdtrig extension name >> and from the date this was an implementation not of the ratified 0.13 >> debug spec (which did not have Sdtrig as a separate extension) but >> rather a version of the in development 1.0 debug spec. >> It's not trivial to tell if it's closer to the ratified 0.13 version or >> the (hopefully soon to be frozen) 1.0 version. >> As the only part of the debug specification to be implemented is the >> triggers then effectively the debug option is x-sdtrig. >> I don't think there is any way for code running on the machine to >> identify what version of the debug is implemented - the appropriate >> register is only available for external debug. Once 1.0 is frozen then >> the presence of Sdtrig isa string would indicate 1.0 trigger support is >> available. >> According to JIRA - https://jira.riscv.org/browse/RVS-981 the debug >> specification should freeze this month. >> How about considering this as a solution: >> - Add a new x-sdtrig option that defaults to false >> - Deprecate debug option - but retain it with default on >> - Add warning if triggers are used and x-sdtrig is not enabled >> - Update the trigger implementation to match frozen spec > > If x-sdtrig is 'false' and 'debug' is on, and then we warn if debug=true and x-sdtrig > is false, we'll warn every time using the defaults. > > Given what you said here: > >> There is potentially a chance that some use cases will be broken but I >> don't think triggers are being widely use - the SBI support only just >> got merged: >> https://github.com/riscv-software-src/opensbi/commit/97f234f15c9657c6ec69fa6ed745be8107bf6a> > > I believe we can deprecate 'debug' and simply ignore its existence. Do everything else with > x-sdtrig. So if an user sets any 'debug' value we'll: > > - warn that the flag is deprecated > - set x-sdtrig to whatever value the user set to 'debug' > > 'debug' will become just an alternate way to set x-sdtrig. The logic should just check for > x-sdtrig. My v2 patchset which I posted a few ours back, gets rid of ‘debug’. It is replaced with x-sdtrig. Keeping them both doesn’t make sense. Here is v2: https://mail.gnu.org/archive/html/qemu-riscv/2024-01/msg00570.html Regards Himanshu > > > Thanks, > > > Daniel > >> Hope this is helpful, >> Rob
On Fri, Jan 12, 2024 at 11:34 PM Rob Bradford <rbradford@rivosinc.com> wrote: > > On Fri, 2024-01-12 at 13:52 +1000, Alistair Francis wrote: > > On Thu, Jan 11, 2024 at 5:20 AM Daniel Henrique Barboza > > <dbarboza@ventanamicro.com> wrote: > > > > > > Himanshu, > > > > > > We spoke offline but let's make everyone aware: > > > > > > - 'sdtrig' should be marked with 'x-' and be an experimental > > > extension since > > > the spec isn't yet frozen; > > > > > > - Alvin sent a patch to the ML adding the 'mcontext' CSR for > > > 'sdtrig' some time > > > ago: > > > > > > "[PATCH v2] target/riscv: Implement optional CSR mcontext of debug > > > Sdtrig extension" > > > > > > It would be good to put his patch on top of this series to ease the > > > review for everyone. > > > The changes done in patch 2 would also be applicable to the > > > mcontext CSR; > > > > > > > > > - last but probably the most important: the existing 'debug' flag > > > seems to be acting as > > > the actual 'sdtrig' extension due to how the flag is gating trigger > > > code, e.g.: > > > > > > if (cpu->cfg.debug) { > > > riscv_trigger_realize(&cpu->env); > > > } > > > > > > and > > > > > > if (cpu->cfg.debug) { > > > riscv_trigger_reset_hold(env); > > > } > > > > > > > > > If that's really the case, all the checks with cpu->cfg.debug will > > > need to also include > > > cpu->cfg.ext_sdtrig (one or the other). And now we'll have to make > > > an option: do we leave > > > the debug triggers (i.e. the 'debug' flag) as always enabled? > > > > From memory the "debug" property is for the original debug spec: > > https://github.com/riscv/riscv-debug-spec/releases/tag/task_group_vote > > > > That was ratified and is an official extension. AFAIK this is what is > > in physical hardware as well. > > > > The actual PDF says draft though, I'm not sure what's going on there. > > > > The debug spec doesn't have a Z* name, so it's just "debug", at least > > AFAIK. > > > > "sdtrig" seems to be a new backwards-incompatible extension doing > > basically the same thing. What a mess > > > > > > > > If it's up to me I would make 'debug' as default 'false' and > > > deprecate it. Users will need > > > > I don't think that's the right approach. It's a ratified extension > > that we are supporting and is in hardware. I think we are stuck > > supporting it > > > > > > I've done a bit of digging and I agree things are quite messy. Here are > my discoveries: > > The debug option and the code for triggers was added in these commits: > > c9711bd778 target/riscv: cpu: Enable native debug feature > 38b4e781a4 target/riscv: machine: Add debug state description > b6092544fc target/riscv: csr: Hook debug CSR read/write > 1acdb3b013 target/riscv: cpu: Add a config option for native debug > 95799e36c1 target/riscv: Add initial support for the Sdtrig extension > > In March 2022 - since the commit refers to the Sdtrig extension name > and from the date this was an implementation not of the ratified 0.13 > debug spec (which did not have Sdtrig as a separate extension) but > rather a version of the in development 1.0 debug spec. Yeah... We used the "stable" from master. That is our mistake there. I'm pretty sure we targeted the 0.13. The "Sdtrig" was only added in the v4 as the changelog says: "mention Sdtrig extension in the commit" > > It's not trivial to tell if it's closer to the ratified 0.13 version or > the (hopefully soon to be frozen) 1.0 version. > > As the only part of the debug specification to be implemented is the > triggers then effectively the debug option is x-sdtrig. > > I don't think there is any way for code running on the machine to > identify what version of the debug is implemented - the appropriate > register is only available for external debug. Once 1.0 is frozen then > the presence of Sdtrig isa string would indicate 1.0 trigger support is > available. > > According to JIRA - https://jira.riscv.org/browse/RVS-981 the debug > specification should freeze this month. > > How about considering this as a solution: > > - Add a new x-sdtrig option that defaults to false > - Deprecate debug option - but retain it with default on We can't deprecate a ratified spec. The 0.13 just seems to call it "debug" so that's what we are stuck with > - Add warning if triggers are used and x-sdtrig is not enabled > - Update the trigger implementation to match frozen spec We will need to support two versions, as there are two ratified specs. Alistair > > There is potentially a chance that some use cases will be broken but I > don't think triggers are being widely use - the SBI support only just > got merged: > https://github.com/riscv-software-src/opensbi/commit/97f234f15c9657c6ec69fa6ed745be8107bf6ae2 > > Hope this is helpful, > > Rob >
On Mon, Jan 22, 2024 at 03:42:10PM +1000, Alistair Francis wrote: > > > From memory the "debug" property is for the original debug spec: > > > https://github.com/riscv/riscv-debug-spec/releases/tag/task_group_vote > > > > > > That was ratified and is an official extension. AFAIK this is what is > > > in physical hardware as well. > > > > > > The actual PDF says draft though, I'm not sure what's going on there. > > > > > > The debug spec doesn't have a Z* name, so it's just "debug", at least > > > AFAIK. > > > > > > "sdtrig" seems to be a new backwards-incompatible extension doing > > > basically the same thing. What a mess ... > > > > I've done a bit of digging and I agree things are quite messy. Here are > > my discoveries: > > > > The debug option and the code for triggers was added in these commits: > > > > c9711bd778 target/riscv: cpu: Enable native debug feature > > 38b4e781a4 target/riscv: machine: Add debug state description > > b6092544fc target/riscv: csr: Hook debug CSR read/write > > 1acdb3b013 target/riscv: cpu: Add a config option for native debug > > 95799e36c1 target/riscv: Add initial support for the Sdtrig extension > > > > In March 2022 - since the commit refers to the Sdtrig extension name > > and from the date this was an implementation not of the ratified 0.13 > > debug spec (which did not have Sdtrig as a separate extension) but > > rather a version of the in development 1.0 debug spec. > > Yeah... We used the "stable" from master. That is our mistake there. > > I'm pretty sure we targeted the 0.13. The "Sdtrig" was only added in > the v4 as the changelog says: "mention Sdtrig extension in the commit" > > > > > It's not trivial to tell if it's closer to the ratified 0.13 version or > > the (hopefully soon to be frozen) 1.0 version. > > > > As the only part of the debug specification to be implemented is the > > triggers then effectively the debug option is x-sdtrig. > > > > I don't think there is any way for code running on the machine to > > identify what version of the debug is implemented - the appropriate > > register is only available for external debug. Once 1.0 is frozen then > > the presence of Sdtrig isa string would indicate 1.0 trigger support is > > available. > > > > According to JIRA - https://jira.riscv.org/browse/RVS-981 the debug > > specification should freeze this month. > > > > How about considering this as a solution: > > > > - Add a new x-sdtrig option that defaults to false > > - Deprecate debug option - but retain it with default on > > We can't deprecate a ratified spec. The 0.13 just seems to call it > "debug" so that's what we are stuck with > > > - Add warning if triggers are used and x-sdtrig is not enabled > > - Update the trigger implementation to match frozen spec > > We will need to support two versions, as there are two ratified specs. > We'll likely want to be allowed to deprecate ratified extensions as riscv evolves. Despite best intentions, extensions may be designed and ratified which ultimately fail to be of much utility, and new extensions will supersede old extensions. If QEMU keeps every extension it adds, then we'll slow progress on new extensions by maintaining old extension code. The old extensions will also bitrot or waste CI resources getting tested for no reason. I don't know the history of 'debug' and 'sdtrig', other than what I've read above, but, to me, it looks like 'debug' might be one of the first extensions which should be deprecated. Assuming we have a long enough deprecation period, then I think it's always safe to attempt a deprecation. If somebody shouts, then it can always be taken back off the chopping block. Thanks, drew
On Mon, Jan 22, 2024 at 7:16 PM Andrew Jones <ajones@ventanamicro.com> wrote: > > On Mon, Jan 22, 2024 at 03:42:10PM +1000, Alistair Francis wrote: > > > > From memory the "debug" property is for the original debug spec: > > > > https://github.com/riscv/riscv-debug-spec/releases/tag/task_group_vote > > > > > > > > That was ratified and is an official extension. AFAIK this is what is > > > > in physical hardware as well. > > > > > > > > The actual PDF says draft though, I'm not sure what's going on there. > > > > > > > > The debug spec doesn't have a Z* name, so it's just "debug", at least > > > > AFAIK. > > > > > > > > "sdtrig" seems to be a new backwards-incompatible extension doing > > > > basically the same thing. What a mess > ... > > > > > > I've done a bit of digging and I agree things are quite messy. Here are > > > my discoveries: > > > > > > The debug option and the code for triggers was added in these commits: > > > > > > c9711bd778 target/riscv: cpu: Enable native debug feature > > > 38b4e781a4 target/riscv: machine: Add debug state description > > > b6092544fc target/riscv: csr: Hook debug CSR read/write > > > 1acdb3b013 target/riscv: cpu: Add a config option for native debug > > > 95799e36c1 target/riscv: Add initial support for the Sdtrig extension > > > > > > In March 2022 - since the commit refers to the Sdtrig extension name > > > and from the date this was an implementation not of the ratified 0.13 > > > debug spec (which did not have Sdtrig as a separate extension) but > > > rather a version of the in development 1.0 debug spec. > > > > Yeah... We used the "stable" from master. That is our mistake there. > > > > I'm pretty sure we targeted the 0.13. The "Sdtrig" was only added in > > the v4 as the changelog says: "mention Sdtrig extension in the commit" > > > > > > > > It's not trivial to tell if it's closer to the ratified 0.13 version or > > > the (hopefully soon to be frozen) 1.0 version. > > > > > > As the only part of the debug specification to be implemented is the > > > triggers then effectively the debug option is x-sdtrig. > > > > > > I don't think there is any way for code running on the machine to > > > identify what version of the debug is implemented - the appropriate > > > register is only available for external debug. Once 1.0 is frozen then > > > the presence of Sdtrig isa string would indicate 1.0 trigger support is > > > available. > > > > > > According to JIRA - https://jira.riscv.org/browse/RVS-981 the debug > > > specification should freeze this month. > > > > > > How about considering this as a solution: > > > > > > - Add a new x-sdtrig option that defaults to false > > > - Deprecate debug option - but retain it with default on > > > > We can't deprecate a ratified spec. The 0.13 just seems to call it > > "debug" so that's what we are stuck with > > > > > - Add warning if triggers are used and x-sdtrig is not enabled > > > - Update the trigger implementation to match frozen spec > > > > We will need to support two versions, as there are two ratified specs. > > > > We'll likely want to be allowed to deprecate ratified extensions as riscv > evolves. Despite best intentions, extensions may be designed and ratified > which ultimately fail to be of much utility, and new extensions will > supersede old extensions. If QEMU keeps every extension it adds, then > we'll slow progress on new extensions by maintaining old extension code. > The old extensions will also bitrot or waste CI resources getting tested > for no reason. I agree that we might need to deprecate extensions. I'm not sure the debug extension is there though. The debug extension is used in current shipping hardware and has been ratified. The Sdtrig isn't even ratified yet (https://lists.riscv.org/g/tech-announce/message/320) Right now I feel that we should at least wait for hardware that supports Sdtrig to start to come out. Then we can look at deprecating debug. Deprecating it now seems a bit premature. Alistair > > I don't know the history of 'debug' and 'sdtrig', other than what I've > read above, but, to me, it looks like 'debug' might be one of the first > extensions which should be deprecated. Assuming we have a long enough > deprecation period, then I think it's always safe to attempt a > deprecation. If somebody shouts, then it can always be taken back off the > chopping block. > > Thanks, > drew >
On Mon, Feb 5, 2024 at 9:36 AM Alistair Francis <alistair23@gmail.com> wrote: > > On Mon, Jan 22, 2024 at 7:16 PM Andrew Jones <ajones@ventanamicro.com> wrote: > > > > On Mon, Jan 22, 2024 at 03:42:10PM +1000, Alistair Francis wrote: > > > > > From memory the "debug" property is for the original debug spec: > > > > > https://github.com/riscv/riscv-debug-spec/releases/tag/task_group_vote > > > > > > > > > > That was ratified and is an official extension. AFAIK this is what is > > > > > in physical hardware as well. > > > > > > > > > > The actual PDF says draft though, I'm not sure what's going on there. > > > > > > > > > > The debug spec doesn't have a Z* name, so it's just "debug", at least > > > > > AFAIK. > > > > > > > > > > "sdtrig" seems to be a new backwards-incompatible extension doing > > > > > basically the same thing. What a mess > > ... > > > > > > > > I've done a bit of digging and I agree things are quite messy. Here are > > > > my discoveries: > > > > > > > > The debug option and the code for triggers was added in these commits: > > > > > > > > c9711bd778 target/riscv: cpu: Enable native debug feature > > > > 38b4e781a4 target/riscv: machine: Add debug state description > > > > b6092544fc target/riscv: csr: Hook debug CSR read/write > > > > 1acdb3b013 target/riscv: cpu: Add a config option for native debug > > > > 95799e36c1 target/riscv: Add initial support for the Sdtrig extension > > > > > > > > In March 2022 - since the commit refers to the Sdtrig extension name > > > > and from the date this was an implementation not of the ratified 0.13 > > > > debug spec (which did not have Sdtrig as a separate extension) but > > > > rather a version of the in development 1.0 debug spec. > > > > > > Yeah... We used the "stable" from master. That is our mistake there. > > > > > > I'm pretty sure we targeted the 0.13. The "Sdtrig" was only added in > > > the v4 as the changelog says: "mention Sdtrig extension in the commit" > > > > > > > > > > > It's not trivial to tell if it's closer to the ratified 0.13 version or > > > > the (hopefully soon to be frozen) 1.0 version. > > > > > > > > As the only part of the debug specification to be implemented is the > > > > triggers then effectively the debug option is x-sdtrig. > > > > > > > > I don't think there is any way for code running on the machine to > > > > identify what version of the debug is implemented - the appropriate > > > > register is only available for external debug. Once 1.0 is frozen then > > > > the presence of Sdtrig isa string would indicate 1.0 trigger support is > > > > available. > > > > > > > > According to JIRA - https://jira.riscv.org/browse/RVS-981 the debug > > > > specification should freeze this month. > > > > > > > > How about considering this as a solution: > > > > > > > > - Add a new x-sdtrig option that defaults to false > > > > - Deprecate debug option - but retain it with default on > > > > > > We can't deprecate a ratified spec. The 0.13 just seems to call it > > > "debug" so that's what we are stuck with > > > > > > > - Add warning if triggers are used and x-sdtrig is not enabled > > > > - Update the trigger implementation to match frozen spec > > > > > > We will need to support two versions, as there are two ratified specs. > > > > > > > We'll likely want to be allowed to deprecate ratified extensions as riscv > > evolves. Despite best intentions, extensions may be designed and ratified > > which ultimately fail to be of much utility, and new extensions will > > supersede old extensions. If QEMU keeps every extension it adds, then > > we'll slow progress on new extensions by maintaining old extension code. > > The old extensions will also bitrot or waste CI resources getting tested > > for no reason. > > I agree that we might need to deprecate extensions. > > I'm not sure the debug extension is there though. The debug extension > is used in current shipping hardware and has been ratified. The Sdtrig > isn't even ratified yet > (https://lists.riscv.org/g/tech-announce/message/320) Is shipping real hardware OR ratification a requirement of QEMU patch acceptance ? Regards, Anup > > Right now I feel that we should at least wait for hardware that > supports Sdtrig to start to come out. Then we can look at deprecating > debug. Deprecating it now seems a bit premature. > > Alistair > > > > > I don't know the history of 'debug' and 'sdtrig', other than what I've > > read above, but, to me, it looks like 'debug' might be one of the first > > extensions which should be deprecated. Assuming we have a long enough > > deprecation period, then I think it's always safe to attempt a > > deprecation. If somebody shouts, then it can always be taken back off the > > chopping block. > > > > Thanks, > > drew > > >
On Mon, Feb 5, 2024 at 2:50 PM Anup Patel <anup@brainfault.org> wrote: > > On Mon, Feb 5, 2024 at 9:36 AM Alistair Francis <alistair23@gmail.com> wrote: > > > > On Mon, Jan 22, 2024 at 7:16 PM Andrew Jones <ajones@ventanamicro.com> wrote: > > > > > > On Mon, Jan 22, 2024 at 03:42:10PM +1000, Alistair Francis wrote: > > > > > > From memory the "debug" property is for the original debug spec: > > > > > > https://github.com/riscv/riscv-debug-spec/releases/tag/task_group_vote > > > > > > > > > > > > That was ratified and is an official extension. AFAIK this is what is > > > > > > in physical hardware as well. > > > > > > > > > > > > The actual PDF says draft though, I'm not sure what's going on there. > > > > > > > > > > > > The debug spec doesn't have a Z* name, so it's just "debug", at least > > > > > > AFAIK. > > > > > > > > > > > > "sdtrig" seems to be a new backwards-incompatible extension doing > > > > > > basically the same thing. What a mess > > > ... > > > > > > > > > > I've done a bit of digging and I agree things are quite messy. Here are > > > > > my discoveries: > > > > > > > > > > The debug option and the code for triggers was added in these commits: > > > > > > > > > > c9711bd778 target/riscv: cpu: Enable native debug feature > > > > > 38b4e781a4 target/riscv: machine: Add debug state description > > > > > b6092544fc target/riscv: csr: Hook debug CSR read/write > > > > > 1acdb3b013 target/riscv: cpu: Add a config option for native debug > > > > > 95799e36c1 target/riscv: Add initial support for the Sdtrig extension > > > > > > > > > > In March 2022 - since the commit refers to the Sdtrig extension name > > > > > and from the date this was an implementation not of the ratified 0.13 > > > > > debug spec (which did not have Sdtrig as a separate extension) but > > > > > rather a version of the in development 1.0 debug spec. > > > > > > > > Yeah... We used the "stable" from master. That is our mistake there. > > > > > > > > I'm pretty sure we targeted the 0.13. The "Sdtrig" was only added in > > > > the v4 as the changelog says: "mention Sdtrig extension in the commit" > > > > > > > > > > > > > > It's not trivial to tell if it's closer to the ratified 0.13 version or > > > > > the (hopefully soon to be frozen) 1.0 version. > > > > > > > > > > As the only part of the debug specification to be implemented is the > > > > > triggers then effectively the debug option is x-sdtrig. > > > > > > > > > > I don't think there is any way for code running on the machine to > > > > > identify what version of the debug is implemented - the appropriate > > > > > register is only available for external debug. Once 1.0 is frozen then > > > > > the presence of Sdtrig isa string would indicate 1.0 trigger support is > > > > > available. > > > > > > > > > > According to JIRA - https://jira.riscv.org/browse/RVS-981 the debug > > > > > specification should freeze this month. > > > > > > > > > > How about considering this as a solution: > > > > > > > > > > - Add a new x-sdtrig option that defaults to false > > > > > - Deprecate debug option - but retain it with default on > > > > > > > > We can't deprecate a ratified spec. The 0.13 just seems to call it > > > > "debug" so that's what we are stuck with > > > > > > > > > - Add warning if triggers are used and x-sdtrig is not enabled > > > > > - Update the trigger implementation to match frozen spec > > > > > > > > We will need to support two versions, as there are two ratified specs. > > > > > > > > > > We'll likely want to be allowed to deprecate ratified extensions as riscv > > > evolves. Despite best intentions, extensions may be designed and ratified > > > which ultimately fail to be of much utility, and new extensions will > > > supersede old extensions. If QEMU keeps every extension it adds, then > > > we'll slow progress on new extensions by maintaining old extension code. > > > The old extensions will also bitrot or waste CI resources getting tested > > > for no reason. > > > > I agree that we might need to deprecate extensions. > > > > I'm not sure the debug extension is there though. The debug extension > > is used in current shipping hardware and has been ratified. The Sdtrig > > isn't even ratified yet > > (https://lists.riscv.org/g/tech-announce/message/320) > > Is shipping real hardware OR ratification a requirement of > QEMU patch acceptance ? We will accept an extension when it is ratified. The question here is, what if two ratified extensions conflict? The answer to me seems that we need to support both them. *Maybe* at some point in the future we can then drop the debug extension. That would require the Sdtrig extension to be widely used and debug not used (which is why I brought up shipping hardware). Alistair > > Regards, > Anup > > > > > Right now I feel that we should at least wait for hardware that > > supports Sdtrig to start to come out. Then we can look at deprecating > > debug. Deprecating it now seems a bit premature. > > > > Alistair > > > > > > > > I don't know the history of 'debug' and 'sdtrig', other than what I've > > > read above, but, to me, it looks like 'debug' might be one of the first > > > extensions which should be deprecated. Assuming we have a long enough > > > deprecation period, then I think it's always safe to attempt a > > > deprecation. If somebody shouts, then it can always be taken back off the > > > chopping block. > > > > > > Thanks, > > > drew > > > > >