From: Nicholas Piggin <npiggin@gmail.com>
To: qemu-ppc@nongnu.org
Cc: "Cédric Le Goater" <clg@kaod.org>,
qemu-devel@nongnu.org, "Nicholas Piggin" <npiggin@gmail.com>,
"David Gibson" <david@gibson.dropbear.id.au>
Subject: [PATCH v1 2/3] spapr: Set LPCR to current AIL mode when starting a new CPU
Date: Wed, 26 May 2021 19:16:25 +1000 [thread overview]
Message-ID: <20210526091626.3388262-3-npiggin@gmail.com> (raw)
In-Reply-To: <20210526091626.3388262-1-npiggin@gmail.com>
TCG does not keep track of AIL mode in a central place, it's based on
the current LPCR[AIL] bits. Synchronize the new CPU's LPCR to the
current LPCR in rtas_start_cpu(), similarly to the way the ILE bit is
synchronized.
Open-code the ILE setting as well now that the caller's LPCR is
available directly, there is no need for the indirection.
Without this, under both TCG and KVM, adding a POWER8/9/10 class CPU
with a new core ID after a modern Linux has booted results in the new
CPU's LPCR missing the LPCR[AIL]=0b11 setting that the other CPUs have.
This can cause crashes and unexpected behaviour.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
hw/ppc/spapr_rtas.c | 14 +++++++++-----
1 file changed, 9 insertions(+), 5 deletions(-)
diff --git a/hw/ppc/spapr_rtas.c b/hw/ppc/spapr_rtas.c
index 63d96955c0..b476382ae6 100644
--- a/hw/ppc/spapr_rtas.c
+++ b/hw/ppc/spapr_rtas.c
@@ -132,8 +132,8 @@ static void rtas_start_cpu(PowerPCCPU *callcpu, SpaprMachineState *spapr,
target_ulong id, start, r3;
PowerPCCPU *newcpu;
CPUPPCState *env;
- PowerPCCPUClass *pcc;
target_ulong lpcr;
+ target_ulong caller_lpcr;
if (nargs != 3 || nret != 1) {
rtas_st(rets, 0, RTAS_OUT_PARAM_ERROR);
@@ -152,7 +152,6 @@ static void rtas_start_cpu(PowerPCCPU *callcpu, SpaprMachineState *spapr,
}
env = &newcpu->env;
- pcc = POWERPC_CPU_GET_CLASS(newcpu);
if (!CPU(newcpu)->halted) {
rtas_st(rets, 0, RTAS_OUT_HW_ERROR);
@@ -164,10 +163,15 @@ static void rtas_start_cpu(PowerPCCPU *callcpu, SpaprMachineState *spapr,
env->msr = (1ULL << MSR_SF) | (1ULL << MSR_ME);
hreg_compute_hflags(env);
+ caller_lpcr = callcpu->env.spr[SPR_LPCR];
lpcr = env->spr[SPR_LPCR];
- if (!pcc->interrupts_big_endian(callcpu)) {
- lpcr |= LPCR_ILE;
- }
+
+ /* Set ILE the same way */
+ lpcr = (lpcr & ~LPCR_ILE) | (caller_lpcr & LPCR_ILE);
+
+ /* Set AIL the same way */
+ lpcr = (lpcr & ~LPCR_AIL) | (caller_lpcr & LPCR_AIL);
+
if (env->mmu_model == POWERPC_MMU_3_00) {
/*
* New cpus are expected to start in the same radix/hash mode
--
2.23.0
next prev parent reply other threads:[~2021-05-26 9:22 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-05-26 9:16 [PATCH v1 0/3] ppc: LPCR synchronisation fixes Nicholas Piggin
2021-05-26 9:16 ` [PATCH v1 1/3] spapr: Remove stale comment about power-saving LPCR bits Nicholas Piggin
2021-05-26 15:18 ` Cédric Le Goater
2021-05-26 15:42 ` Greg Kurz
2021-05-27 1:28 ` David Gibson
2021-05-26 9:16 ` Nicholas Piggin [this message]
2021-05-26 15:24 ` [PATCH v1 2/3] spapr: Set LPCR to current AIL mode when starting a new CPU Cédric Le Goater
2021-05-26 16:03 ` Greg Kurz
2021-05-27 1:29 ` David Gibson
2021-05-26 9:16 ` [PATCH v1 3/3] target/ppc: Synchronize with KVM's LPCR value when creating a vCPU Nicholas Piggin
2021-05-27 1:48 ` David Gibson
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=20210526091626.3388262-3-npiggin@gmail.com \
--to=npiggin@gmail.com \
--cc=clg@kaod.org \
--cc=david@gibson.dropbear.id.au \
--cc=qemu-devel@nongnu.org \
--cc=qemu-ppc@nongnu.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).