All of lore.kernel.org
 help / color / mirror / Atom feed
From: Sean Christopherson <seanjc@google.com>
To: Varad Gautam <varad.gautam@suse.com>
Cc: kvm@vger.kernel.org, pbonzini@redhat.com, drjones@redhat.com,
	marcorr@google.com, zxwang42@gmail.com, erdemaktas@google.com,
	rientjes@google.com, brijesh.singh@amd.com,
	Thomas.Lendacky@amd.com, jroedel@suse.de, bp@suse.de
Subject: Re: [kvm-unit-tests PATCH v2 06/10] x86: efi: Stop using UEFI-provided %gs for percpu storage
Date: Wed, 13 Apr 2022 16:23:42 +0000	[thread overview]
Message-ID: <Ylb5DmIjc8NPjzQ9@google.com> (raw)
In-Reply-To: <20220412173407.13637-7-varad.gautam@suse.com>

On Tue, Apr 12, 2022, Varad Gautam wrote:
> UEFI tests do not update MSR_GS_BASE during bringup, and continue
> using the GS_BASE set up by the UEFI implementation for percpu
> storage.
> 
> Update this MSR during setup_segments64() to allow storing percpu
> data at a sane location reserved by the testcase, and ensure that
> this happens before any operation that ends up storing to the percpu
> space

It's worth noting in the changelog that reset_apic() needs to be moved below
setup_gdt_tss() as it depends on per-cpu setup.  That definitely won't be obvious
to most people.

> Signed-off-by: Varad Gautam <varad.gautam@suse.com>
> ---
>  lib/x86/setup.c      | 9 ++++++---
>  x86/efi/efistart64.S | 7 +++++++
>  2 files changed, 13 insertions(+), 3 deletions(-)
> 
> diff --git a/lib/x86/setup.c b/lib/x86/setup.c
> index 7dd6677..5d32d3f 100644
> --- a/lib/x86/setup.c
> +++ b/lib/x86/setup.c
> @@ -170,7 +170,8 @@ void setup_multiboot(struct mbi_bootinfo *bi)
>  #ifdef CONFIG_EFI
>  
>  /* From x86/efi/efistart64.S */
> -extern void setup_segments64(void);
> +extern void setup_segments64(u64 gs_base);
> +extern u8 stacktop;
>  
>  static efi_status_t setup_memory_allocator(efi_bootinfo_t *efi_bootinfo)
>  {
> @@ -271,12 +272,14 @@ static void setup_page_table(void)
>  static void setup_gdt_tss(void)
>  {
>  	size_t tss_offset;
> +	u64 gs_base;
>  
>  	/* 64-bit setup_tss does not use the stacktop argument.  */
>  	tss_offset = setup_tss(NULL);
>  	load_gdt_tss(tss_offset);
>  
> -	setup_segments64();
> +	gs_base = (u64)(&stacktop) - (PAGE_SIZE * (pre_boot_apic_id() + 1));

Rather than follow the (IMO awful) non-EFI behavior of hijacking a chunk of the
stack, which is a symptom of doing everything in asm, since this is now C we
can declare a proper percpu array and index that.  Disclaimer, this has only been
tested with smp=1 at this point, haven't reached the end of the series :-)

diff --git a/lib/x86/setup.c b/lib/x86/setup.c
index 6131ea2..91a06f7 100644
--- a/lib/x86/setup.c
+++ b/lib/x86/setup.c
@@ -169,6 +169,8 @@ void setup_multiboot(struct mbi_bootinfo *bi)

 #ifdef CONFIG_EFI

+static struct percpu_data __percpu_data[MAX_TEST_CPUS];
+
 static efi_status_t setup_memory_allocator(efi_bootinfo_t *efi_bootinfo)
 {
        int i;
@@ -285,6 +287,8 @@ static void setup_gdt_tss(void)
                     "1:"
                     :: "r" ((u64)KERNEL_DS), "i" (KERNEL_CS)
        );
+
+       wrmsr(MSR_GS_BASE, (u64)&__percpu_data[pre_boot_apic_id()]);
 }

 efi_status_t setup_efi(efi_bootinfo_t *efi_bootinfo)
@@ -326,8 +330,8 @@ efi_status_t setup_efi(efi_bootinfo_t *efi_bootinfo)
                return status;
        }

-       reset_apic();
        setup_gdt_tss();
+       reset_apic();
        setup_idt();
        load_idt();
        mask_pic_interrupts();

  reply	other threads:[~2022-04-13 16:23 UTC|newest]

Thread overview: 27+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-04-12 17:33 [kvm-unit-tests PATCH v2 00/10] SMP Support for x86 UEFI Tests Varad Gautam
2022-04-12 17:33 ` [kvm-unit-tests PATCH v2 01/10] x86: Move ap_init() to smp.c Varad Gautam
2022-04-13 15:16   ` Sean Christopherson
2022-04-13 18:44   ` Sean Christopherson
2022-04-13 18:55     ` Sean Christopherson
2022-04-13 18:59       ` Sean Christopherson
2022-04-13 19:01         ` Sean Christopherson
2022-04-13 19:04           ` Sean Christopherson
2022-04-12 17:33 ` [kvm-unit-tests PATCH v2 02/10] x86: Move load_idt() to desc.c Varad Gautam
2022-04-13 15:19   ` Sean Christopherson
2022-04-12 17:34 ` [kvm-unit-tests PATCH v2 03/10] x86: desc: Split IDT entry setup into a generic helper Varad Gautam
2022-04-13 16:35   ` Sean Christopherson
2022-04-12 17:34 ` [kvm-unit-tests PATCH v2 04/10] x86: Move load_gdt_tss() to desc.c Varad Gautam
2022-04-13 15:58   ` Sean Christopherson
2022-04-12 17:34 ` [kvm-unit-tests PATCH v2 05/10] x86: efi: Stop using UEFI-provided stack Varad Gautam
2022-04-13 16:20   ` Sean Christopherson
2022-04-12 17:34 ` [kvm-unit-tests PATCH v2 06/10] x86: efi: Stop using UEFI-provided %gs for percpu storage Varad Gautam
2022-04-13 16:23   ` Sean Christopherson [this message]
2022-04-12 17:34 ` [kvm-unit-tests PATCH v2 07/10] x86: efi, smp: Transition APs from 16-bit to 32-bit mode Varad Gautam
2022-04-13 19:17   ` Sean Christopherson
2022-04-12 17:34 ` [kvm-unit-tests PATCH v2 08/10] x86: Move 32-bit bringup routines to start32.S Varad Gautam
2022-04-13 19:33   ` Sean Christopherson
2022-04-12 17:34 ` [kvm-unit-tests PATCH v2 09/10] x86: efi, smp: Transition APs from 32-bit to 64-bit mode Varad Gautam
2022-04-12 17:34 ` [kvm-unit-tests PATCH v2 10/10] x86: Provide a common 64-bit AP entrypoint for EFI and non-EFI Varad Gautam
2022-04-13 19:55   ` Sean Christopherson
2022-04-13 19:57 ` [kvm-unit-tests PATCH v2 00/10] SMP Support for x86 UEFI Tests Sean Christopherson
2022-04-26 11:51   ` Varad Gautam

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=Ylb5DmIjc8NPjzQ9@google.com \
    --to=seanjc@google.com \
    --cc=Thomas.Lendacky@amd.com \
    --cc=bp@suse.de \
    --cc=brijesh.singh@amd.com \
    --cc=drjones@redhat.com \
    --cc=erdemaktas@google.com \
    --cc=jroedel@suse.de \
    --cc=kvm@vger.kernel.org \
    --cc=marcorr@google.com \
    --cc=pbonzini@redhat.com \
    --cc=rientjes@google.com \
    --cc=varad.gautam@suse.com \
    --cc=zxwang42@gmail.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.