xen-devel.lists.xenproject.org archive mirror
 help / color / mirror / Atom feed
From: Andrew Cooper <andrew.cooper3@citrix.com>
To: Jan Beulich <jbeulich@suse.com>
Cc: Xen-devel <xen-devel@lists.xenproject.org>,
	"Wei Liu" <wl@xen.org>, "Roger Pau Monné" <roger.pau@citrix.com>
Subject: Re: [Xen-devel] [PATCH 4/4] x86/svm: Use named (bit)fields for task switch exit info
Date: Wed, 4 Dec 2019 20:04:49 +0000	[thread overview]
Message-ID: <d4fc4cfe-b4d0-5c26-12dc-df1631f8d49b@citrix.com> (raw)
In-Reply-To: <2d347b3a-7b2e-d28a-59d0-2206bc215206@suse.com>

On 04/12/2019 10:24, Jan Beulich wrote:
> On 04.12.2019 10:43, Andrew Cooper wrote:
>> Introduce vmcb.e1.* and vmcb.e2.* to provide names to fields in exitinfo{1,2}
>> respectively.  Implement the task switch names for now, and clean up the
>> TASK_SWITCH handler.
> "e1" and "e2" look overly short - and hence possibly ambiguous -
> to me. Make them perhaps "ei1" and "ei2"?

Written on their own like that perhaps, but the ei[12] versions are
equally ambiguous.

However, they are only ever used with vmcb-> in code, so there is no
issue with ambiguity.

>  Furthermore, seeing ...
>
>> @@ -2795,19 +2792,12 @@ void svm_vmexit_handler(struct cpu_user_regs *regs)
>>          if ( insn_len < 0 && (insn_len = svm_get_task_switch_insn_len()) == 0 )
>>              goto crash_or_fault;
>>  
>> -        if ( (vmcb->exitinfo2 >> 36) & 1 )
>> -            reason = TSW_iret;
>> -        else if ( (vmcb->exitinfo2 >> 38) & 1 )
>> -            reason = TSW_jmp;
>> -        else
>> -            reason = TSW_call_or_int;
>> -        if ( (vmcb->exitinfo2 >> 44) & 1 )
>> -            errcode = (uint32_t)vmcb->exitinfo2;
>> -
>> -        hvm_task_switch(vmcb->exitinfo1, reason, errcode, insn_len,
>> -                        (vmcb->exitinfo2 & (1ul << 48)) ? X86_EFLAGS_RF : 0);
>> +        hvm_task_switch(vmcb->e1.task_switch.sel,
>> +                        vmcb->e2.task_switch.iret ? TSW_iret :
>> +                        vmcb->e2.task_switch.jmp  ? TSW_jmp : TSW_call_or_int,
>> +                        vmcb->e2.task_switch.ev ? vmcb->e2.task_switch.ec : -1,
>> +                        insn_len, vmcb->e2.task_switch.rf ? X86_EFLAGS_RF : 0);
> ... this, wouldn't it make sense to simply have "ei" covering both
> parts, no longer making it a requirement to use (and hence look up)
> the numeric suffixes at use sites?

Net delta is:

diff --git a/xen/include/asm-x86/hvm/svm/vmcb.h
b/xen/include/asm-x86/hvm/svm/vmcb.h
index 02b5e86b49..864618ccf9 100644
--- a/xen/include/asm-x86/hvm/svm/vmcb.h
+++ b/xen/include/asm-x86/hvm/svm/vmcb.h
@@ -419,17 +419,15 @@ struct vmcb_struct {
     u64 interrupt_shadow;       /* offset 0x68 */
     u64 exitcode;               /* offset 0x70 */
     union {
-        u64 exitinfo1;          /* offset 0x78 */
+        struct {
+            uint64_t exitinfo1; /* offset 0x78 */
+            uint64_t exitinfo2; /* offset 0x80 */
+        };
         union {
             struct {
                 uint16_t sel;
-            } task_switch;
-        } e1;
-    };
-    union {
-        u64 exitinfo2;          /* offset 0x80 */
-        union {
-            struct {
+                uint64_t :48;
+
                 uint32_t ec;
                 uint32_t :4;
                 bool     iret:1;
@@ -440,7 +438,7 @@ struct vmcb_struct {
                 uint32_t :3;
                 bool     rf:1;
             } task_switch;
-        } e2;
+        } ei;
     };
     intinfo_t exitintinfo;      /* offset 0x88 */
     u64 _np_enable;             /* offset 0x90 - cleanbit 4 */

LGTM.

~Andrew

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

  reply	other threads:[~2019-12-04 20:05 UTC|newest]

Thread overview: 18+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-12-04  9:43 [Xen-devel] [PATCH 0/4] x86/svm: (Post TASK_SWITCH) cleanup Andrew Cooper
2019-12-04  9:43 ` [Xen-devel] [PATCH 1/4] x86/svm: Clean up construct_vmcb() Andrew Cooper
2019-12-04 10:06   ` Jan Beulich
2019-12-04 19:21     ` Andrew Cooper
2019-12-04  9:43 ` [Xen-devel] [PATCH 2/4] x86/svm: Don't shadow variables in svm_vmexit_handler() Andrew Cooper
2019-12-04 10:10   ` Jan Beulich
2019-12-04  9:43 ` [Xen-devel] [PATCH 3/4] x86/svm: Clean up intinfo_t variables Andrew Cooper
2019-12-04 10:19   ` Jan Beulich
2019-12-04 19:22     ` Andrew Cooper
2019-12-04 19:38   ` Andrew Cooper
2019-12-05  9:11     ` Jan Beulich
2019-12-05 12:33   ` Andrew Cooper
2019-12-04  9:43 ` [Xen-devel] [PATCH 4/4] x86/svm: Use named (bit)fields for task switch exit info Andrew Cooper
2019-12-04 10:24   ` Jan Beulich
2019-12-04 20:04     ` Andrew Cooper [this message]
2019-12-05  9:05       ` Jan Beulich
2019-12-05 10:51 ` [Xen-devel] [PATCH 5/4] x86/svm: Minor cleanup to start_svm() Andrew Cooper
2019-12-05 10:53   ` Jan Beulich

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=d4fc4cfe-b4d0-5c26-12dc-df1631f8d49b@citrix.com \
    --to=andrew.cooper3@citrix.com \
    --cc=jbeulich@suse.com \
    --cc=roger.pau@citrix.com \
    --cc=wl@xen.org \
    --cc=xen-devel@lists.xenproject.org \
    /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).