Re: [PATCH v7 02/19] xen/riscv: disable unnecessary configs

[Jan Beulich <jbeulich@xxxxxxxx>]

On 03.04.2024 12:54, Oleksii wrote:
> On Wed, 2024-04-03 at 12:28 +0200, Jan Beulich wrote:
>> On 03.04.2024 12:19, Oleksii Kurochko wrote:
>>> This patch disables unnecessary configs for two cases:
>>> 1. By utilizing EXTRA_FIXED_RANDCONFIG for randconfig builds
>>> (GitLab CI jobs).
>>> 2. By using tiny64_defconfig for non-randconfig builds.
>>>
>>> Only configs which lead to compilation issues were disabled.
>>>
>>> Signed-off-by: Oleksii Kurochko <oleksii.kurochko@xxxxxxxxx>
>>> ---
>>> Changes in V7:
>>>  - Disable only configs which cause compilation issues.
>>
>> Since the description doesn't go into details: While I can see that
>> PERF_COUNTERS and LIVEPATCH may require (a little / some more) extra
>> work, are HYPFS, ARGO, and XSM really causing issues?
> For Argo, I recieved the following compilation errors:
>    common/argo.c:1416:5: error: unknown type name 'p2m_type_t'; did you
>    mean 'hvmmem_type_t'?
>     1416 |     p2m_type_t p2mt;
>          |     ^~~~~~~~~~
>          |     hvmmem_type_t
>    common/argo.c:1419:11: error: implicit declaration of function
>    'check_get_page_from_gfn' [-Werror=implicit-function-declaration]
>     1419 |     ret = check_get_page_from_gfn(d, gfn, false, &p2mt, &page);
>          |           ^~~~~~~~~~~~~~~~~~~~~~~
>    common/argo.c:1427:10: error: 'p2m_ram_rw' undeclared (first use in
>    this function)
>     1427 |     case p2m_ram_rw:
>    
> It seems it should be included xen/p2m-common.h and asm/p2m.h in
> common/argo.c.
> 
> For CONFIG_HYPFS_CONFIG ( there is no issue with CONFIG_HYPFS,
> overlooked that ):
>    common/config_data.S:1:10: fatal error: asm/asm_defns.h: No such file
>    or directory
>        1 | #include <asm/asm_defns.h>
>    
> 
> For XSM, I recieved the following error:
> 
>    xsm/xsm_core.c:79:19: error: 'xsm_core_init' defined but not used [-
>    Werror=unused-function]
>       79 | static int __init xsm_core_init(const void *policy_buffer,
>    size_t policy_size)
> 
> I'll add an information with compilation errors to the commit message.

No need to quote full compiler diagnostics, but a hint at the problems
at least. That said, perhaps we want to rather sort the issues than
disable building stuff that sooner or later you will want to build
anyway. For hypfs we look to have an approach already. For Argo what
you suggest makes sense to me; it might be nice to understand where
the P2M headers needed are coming from on x86 and Arm. Ideally common
code .c files wouldn't include asm/*.h.

For XSM I'm a little puzzled: Shouldn't RISC-V have HAS_DEVICE_TREE=y?
Then xsm_core_init() would have a caller.

Jan