From: Dave Hansen <dave.hansen@intel.com>
To: Kai Huang <kai.huang@intel.com>,
linux-kernel@vger.kernel.org, kvm@vger.kernel.org
Cc: seanjc@google.com, pbonzini@redhat.com, len.brown@intel.com,
tony.luck@intel.com, rafael.j.wysocki@intel.com,
reinette.chatre@intel.com, dan.j.williams@intel.com,
peterz@infradead.org, ak@linux.intel.com,
kirill.shutemov@linux.intel.com,
sathyanarayanan.kuppuswamy@linux.intel.com,
isaku.yamahata@intel.com
Subject: Re: [PATCH v5 08/22] x86/virt/tdx: Shut down TDX module in case of error
Date: Mon, 27 Jun 2022 13:46:48 -0700 [thread overview]
Message-ID: <77c90075-79d4-7cc7-f266-1b67e586513b@intel.com> (raw)
In-Reply-To: <cc90e5f8be0c6f48a144240d4569b15bd4b75dd8.camel@intel.com>
On 6/26/22 22:26, Kai Huang wrote:
> On Fri, 2022-06-24 at 11:50 -0700, Dave Hansen wrote:
>> So, the last patch was called:
>>
>> Implement SEAMCALL function
>>
>> and yet, in this patch, we have a "seamcall()" function. That's a bit
>> confusing and not covered at *all* in this subject.
>>
>> Further, seamcall() is the *ONLY* caller of __seamcall() that I see in
>> this series. That makes its presence here even more odd.
>>
>> The seamcall() bits should either be in their own patch, or mashed in
>> with __seamcall().
>
> Right. The reason I didn't put the seamcall() into previous patch was it is
> only used in this tdx.c, so it should be static. But adding a static function
> w/o using it in previous patch will trigger a compile warning. So I introduced
> here where it is first used.
>
> One option is I can introduce seamcall() as a static inline function in tdx.h in
> previous patch so there won't be a warning. I'll change to use this way.
> Please let me know if you have any comments.
Does a temporary __unused get rid of the warning?
>>> /*
>>> * Detect and initialize the TDX module.
>>> *
>>> @@ -138,7 +195,10 @@ static int init_tdx_module(void)
>>>
>>> static void shutdown_tdx_module(void)
>>> {
>>> - /* TODO: Shut down the TDX module */
>>> + struct seamcall_ctx sc = { .fn = TDH_SYS_LP_SHUTDOWN };
>>> +
>>> + seamcall_on_each_cpu(&sc);
>>> +
>>> tdx_module_status = TDX_MODULE_SHUTDOWN;
>>> }
>>>
>>> @@ -221,6 +281,9 @@ bool platform_tdx_enabled(void)
>>> * CPU hotplug is temporarily disabled internally to prevent any cpu
>>> * from going offline.
>>> *
>>> + * Caller also needs to guarantee all CPUs are in VMX operation during
>>> + * this function, otherwise Oops may be triggered.
>>
>> I would *MUCH* rather have this be a:
>>
>> if (!cpu_feature_enabled(X86_FEATURE_VMX))
>> WARN_ONCE("VMX should be on blah blah\n");
>>
>> than just plain oops. Even a pr_err() that preceded the oops would be
>> nicer than an oops that someone has to go decode and then grumble when
>> their binutils is too old that it can't disassemble the TDCALL.
>
> I can add this to seamcall():
>
> /*
> * SEAMCALL requires CPU being in VMX operation otherwise it causes
> #UD.
> * Sanity check and return early to avoid Oops. Note cpu_vmx_enabled()
> * actually only checks whether VMX is enabled but doesn't check
> whether
> * CPU is in VMX operation (VMXON is done). There's no way to check
> * whether VMXON has been done, but currently enabling VMX and doing
> * VMXON are always done together.
> */
> if (!cpu_vmx_enabled()) {
> WARN_ONCE("CPU is not in VMX operation before making
> SEAMCALL");
> return -EINVAL;
> }
>
> The reason I didn't do is I'd like to make seamcall() simple, that it only
> returns TDX_SEAMCALL_VMFAILINVALID or the actual SEAMCALL leaf error. With
> above, this function also returns kernel error code, which isn't good.
I think you're missing the point. You wasted two lines of code on a
*COMMENT* that doesn't actually help anyone decode an oops. You could
have, instead, spent two lines on actual code that would have been just
as good or better than a comment *AND* help folks looking at an oops.
It's almost always better to do something actionable in code than to
comment it, unless it's in some crazy fast path.
> Alternatively, we can always add EXTABLE to TDX_MODULE_CALL macro to handle #UD
> and #GP by returning dedicated error codes (please also see my reply to previous
> patch for the code needed to handle), in which case we don't need such check
> here.
>
> Always handling #UD in TDX_MODULE_CALL macro also has another advantage: there
> will be no Oops for #UD regardless the issue that "there's no way to check
> whether VMXON has been done" in the above comment.
>
> What's your opinion?
I think you should explore using the EXTABLE. Let's see how it looks.
next prev parent reply other threads:[~2022-06-27 20:47 UTC|newest]
Thread overview: 114+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-06-22 11:15 [PATCH v5 00/22] TDX host kernel support Kai Huang
2022-06-22 11:15 ` [PATCH v5 01/22] x86/virt/tdx: Detect TDX during kernel boot Kai Huang
2022-06-23 5:57 ` Chao Gao
2022-06-23 9:23 ` Kai Huang
2022-08-02 2:01 ` [PATCH v5 1/22] " Wu, Binbin
2022-08-03 9:25 ` Kai Huang
2022-06-22 11:15 ` [PATCH v5 02/22] cc_platform: Add new attribute to prevent ACPI CPU hotplug Kai Huang
2022-06-22 11:42 ` Rafael J. Wysocki
2022-06-23 0:01 ` Kai Huang
2022-06-27 8:01 ` Igor Mammedov
2022-06-28 10:04 ` Kai Huang
2022-06-28 11:52 ` Igor Mammedov
2022-06-28 17:33 ` Rafael J. Wysocki
2022-06-28 23:41 ` Kai Huang
2022-06-24 18:57 ` Dave Hansen
2022-06-27 5:05 ` Kai Huang
2022-07-13 11:09 ` Kai Huang
2022-07-19 17:46 ` Dave Hansen
2022-07-19 23:54 ` Kai Huang
2022-08-03 3:40 ` Binbin Wu
2022-08-03 9:20 ` Kai Huang
2022-06-29 5:33 ` Christoph Hellwig
2022-06-29 9:09 ` Kai Huang
2022-08-03 3:55 ` Binbin Wu
2022-08-03 9:21 ` Kai Huang
2022-06-22 11:15 ` [PATCH v5 03/22] cc_platform: Add new attribute to prevent ACPI memory hotplug Kai Huang
2022-06-22 11:45 ` Rafael J. Wysocki
2022-06-23 0:08 ` Kai Huang
2022-06-28 17:55 ` Rafael J. Wysocki
2022-06-28 12:01 ` Igor Mammedov
2022-06-28 23:49 ` Kai Huang
2022-06-29 8:48 ` Igor Mammedov
2022-06-29 9:13 ` Kai Huang
2022-06-22 11:16 ` [PATCH v5 04/22] x86/virt/tdx: Prevent ACPI CPU hotplug and " Kai Huang
2022-06-24 1:41 ` Chao Gao
2022-06-24 11:21 ` Kai Huang
2022-06-29 8:35 ` Yuan Yao
2022-06-29 9:17 ` Kai Huang
2022-06-29 14:22 ` Dave Hansen
2022-06-29 23:02 ` Kai Huang
2022-06-30 15:44 ` Dave Hansen
2022-06-30 22:45 ` Kai Huang
2022-06-22 11:16 ` [PATCH v5 05/22] x86/virt/tdx: Prevent hot-add driver managed memory Kai Huang
2022-06-24 2:12 ` Chao Gao
2022-06-24 11:23 ` Kai Huang
2022-06-24 19:01 ` Dave Hansen
2022-06-27 5:27 ` Kai Huang
2022-06-22 11:16 ` [PATCH v5 06/22] x86/virt/tdx: Add skeleton to initialize TDX on demand Kai Huang
2022-06-24 2:39 ` Chao Gao
2022-06-24 11:27 ` Kai Huang
2022-06-22 11:16 ` [PATCH v5 07/22] x86/virt/tdx: Implement SEAMCALL function Kai Huang
2022-06-24 18:38 ` Dave Hansen
2022-06-27 5:23 ` Kai Huang
2022-06-27 20:58 ` Dave Hansen
2022-06-27 22:10 ` Kai Huang
2022-07-19 19:39 ` Dan Williams
2022-07-19 23:28 ` Kai Huang
2022-07-20 10:18 ` Kai Huang
2022-07-20 16:48 ` Dave Hansen
2022-07-21 1:52 ` Kai Huang
2022-07-27 0:34 ` Kai Huang
2022-07-27 0:50 ` Dave Hansen
2022-07-27 12:46 ` Kai Huang
2022-08-03 2:37 ` Kai Huang
2022-08-03 14:20 ` Dave Hansen
2022-08-03 22:35 ` Kai Huang
2022-08-04 10:06 ` Kai Huang
2022-06-22 11:16 ` [PATCH v5 08/22] x86/virt/tdx: Shut down TDX module in case of error Kai Huang
2022-06-24 18:50 ` Dave Hansen
2022-06-27 5:26 ` Kai Huang
2022-06-27 20:46 ` Dave Hansen [this message]
2022-06-27 22:34 ` Kai Huang
2022-06-27 22:56 ` Dave Hansen
2022-06-27 23:59 ` Kai Huang
2022-06-28 0:03 ` Dave Hansen
2022-06-28 0:11 ` Kai Huang
2022-06-22 11:16 ` [PATCH v5 09/22] x86/virt/tdx: Detect TDX module by doing module global initialization Kai Huang
2022-06-22 11:16 ` [PATCH v5 10/22] x86/virt/tdx: Do logical-cpu scope TDX module initialization Kai Huang
2022-06-22 11:17 ` [PATCH v5 11/22] x86/virt/tdx: Get information about TDX module and TDX-capable memory Kai Huang
2022-06-22 11:17 ` [PATCH v5 12/22] x86/virt/tdx: Convert all memory regions in memblock to TDX memory Kai Huang
2022-06-24 19:40 ` Dave Hansen
2022-06-27 6:16 ` Kai Huang
2022-07-07 2:37 ` Kai Huang
2022-07-07 14:26 ` Dave Hansen
2022-07-07 14:36 ` Juergen Gross
2022-07-07 23:42 ` Kai Huang
2022-07-07 23:34 ` Kai Huang
2022-08-03 1:30 ` Kai Huang
2022-08-03 14:22 ` Dave Hansen
2022-08-03 22:14 ` Kai Huang
2022-06-22 11:17 ` [PATCH v5 13/22] x86/virt/tdx: Add placeholder to construct TDMRs based on memblock Kai Huang
2022-06-22 11:17 ` [PATCH v5 14/22] x86/virt/tdx: Create TDMRs to cover all memblock memory regions Kai Huang
2022-06-22 11:17 ` [PATCH v5 15/22] x86/virt/tdx: Allocate and set up PAMTs for TDMRs Kai Huang
2022-06-24 20:13 ` Dave Hansen
2022-06-27 10:31 ` Kai Huang
2022-06-27 20:41 ` Dave Hansen
2022-06-27 22:50 ` Kai Huang
2022-06-27 22:57 ` Dave Hansen
2022-06-27 23:05 ` Kai Huang
2022-06-28 0:48 ` Xiaoyao Li
2022-06-28 17:03 ` Dave Hansen
2022-08-17 22:46 ` Sagi Shahar
2022-08-17 23:43 ` Huang, Kai
2022-06-22 11:17 ` [PATCH v5 16/22] x86/virt/tdx: Set up reserved areas for all TDMRs Kai Huang
2022-06-22 11:17 ` [PATCH v5 17/22] x86/virt/tdx: Reserve TDX module global KeyID Kai Huang
2022-06-22 11:17 ` [PATCH v5 18/22] x86/virt/tdx: Configure TDX module with TDMRs and " Kai Huang
2022-06-22 11:17 ` [PATCH v5 19/22] x86/virt/tdx: Configure global KeyID on all packages Kai Huang
2022-06-22 11:17 ` [PATCH v5 20/22] x86/virt/tdx: Initialize all TDMRs Kai Huang
2022-06-22 11:17 ` [PATCH v5 21/22] x86/virt/tdx: Support kexec() Kai Huang
2022-06-22 11:17 ` [PATCH v5 22/22] Documentation/x86: Add documentation for TDX host support Kai Huang
2022-08-18 4:07 ` Bagas Sanjaya
2022-08-18 9:33 ` Huang, Kai
2022-06-24 19:47 ` [PATCH v5 00/22] TDX host kernel support Dave Hansen
2022-06-27 4:09 ` Kai Huang
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=77c90075-79d4-7cc7-f266-1b67e586513b@intel.com \
--to=dave.hansen@intel.com \
--cc=ak@linux.intel.com \
--cc=dan.j.williams@intel.com \
--cc=isaku.yamahata@intel.com \
--cc=kai.huang@intel.com \
--cc=kirill.shutemov@linux.intel.com \
--cc=kvm@vger.kernel.org \
--cc=len.brown@intel.com \
--cc=linux-kernel@vger.kernel.org \
--cc=pbonzini@redhat.com \
--cc=peterz@infradead.org \
--cc=rafael.j.wysocki@intel.com \
--cc=reinette.chatre@intel.com \
--cc=sathyanarayanan.kuppuswamy@linux.intel.com \
--cc=seanjc@google.com \
--cc=tony.luck@intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).