Message ID | 20240101172143.14530-2-xry111@xry111.site |
---|---|
State | New |
Headers |
Return-Path: <linux-kernel+bounces-13913-ouuuleilei=gmail.com@vger.kernel.org> Delivered-To: ouuuleilei@gmail.com Received: by 2002:a05:7301:6f82:b0:100:9c79:88ff with SMTP id tb2csp4069843dyb; Mon, 1 Jan 2024 09:23:13 -0800 (PST) X-Google-Smtp-Source: AGHT+IEPFxDggkBCz9yUCoC0pnkZG6sHKxRRENkY6/mvKe5mKkW38sJCSBp5DXQvEtyx+EGNIgXH X-Received: by 2002:a17:90a:d985:b0:28b:e98d:aaeb with SMTP id d5-20020a17090ad98500b0028be98daaebmr7604438pjv.23.1704129793348; Mon, 01 Jan 2024 09:23:13 -0800 (PST) ARC-Seal: i=1; a=rsa-sha256; t=1704129793; cv=none; d=google.com; s=arc-20160816; b=AW35UQfJkYaqD8tIwAosgIyX+hzok8EkoqpPDrcdBSUWpxWX6BfG/JDeHUg8DhnZfl T63oiidsqReSGNGlbOtGDcEyZFuitbXT0j/fJUMK/ioeTam2GE4UaMamWHZlPJZ/2Syv q15AxR9P+TsOQHGkWn84ptn0xloacUpxFF8GE7ugPXQEeAVQ+HBkBHm8nl2BuHEWfAwQ cXcbr9jzrfa2hXWUY4qRE3GscCRS4IuU1d5cRjh5csVcfvotF982yUJoru2uhYtJyghp OrzXJ9c5p8PP5e6vDN6L9ObDa0NoZ31ghBG7Rf0Zq3U64mmgt0dGdYxn0IiE0OvgXh4P 25Ug== ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=arc-20160816; h=content-transfer-encoding:mime-version:list-unsubscribe :list-subscribe:list-id:precedence:message-id:date:subject:cc:to :from:dkim-signature; bh=HWkfPtmeKIwinoUYqjaH8/01M1EBAwuxmKHI9r6fzNk=; fh=bFWT0OjK1S4mdPKBqHOHbI7VZvQcJcj1GEvvcPyamM0=; b=tcDx3p3c1lVSPgD+GWqRy7g5gmR7Q3kjySt8z1cXwkBH0WueYTXjR1XdIC9VUgnc2/ kMygdNkuDV5Xhft1ddZcnPD93+xI1ccn4wtAllYDiUt1fdPBd1VXHr7tLGB4ewh+rGFK pLYECs0CmApE6P8dSbnpWqhkqx6Hvdv3ZbuujpRo5HlikKCmzSuA3FNCfvPXPtwxq2+5 uB4n6xlIF/ZshM2+9mhIodqoEY9ebdqRonj7SRMRvgl8NkYoLd2iMjLy2kejQe+cGh+r vZT+6JgpyWRItbXBbmlE7vWXNrhh+zjZkRje16NgFj+WmJ1ES0jZ8KwXPVE46niBuiLB AXhg== ARC-Authentication-Results: i=1; mx.google.com; dkim=pass header.i=@xry111.site header.s=default header.b=DyQU46Wl; spf=pass (google.com: domain of linux-kernel+bounces-13913-ouuuleilei=gmail.com@vger.kernel.org designates 2604:1380:45e3:2400::1 as permitted sender) smtp.mailfrom="linux-kernel+bounces-13913-ouuuleilei=gmail.com@vger.kernel.org"; dmarc=pass (p=REJECT sp=REJECT dis=NONE) header.from=xry111.site Received: from sv.mirrors.kernel.org (sv.mirrors.kernel.org. [2604:1380:45e3:2400::1]) by mx.google.com with ESMTPS id m8-20020a17090ab78800b0028b9b0b2bbesi21776592pjr.189.2024.01.01.09.23.13 for <ouuuleilei@gmail.com> (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Mon, 01 Jan 2024 09:23:13 -0800 (PST) Received-SPF: pass (google.com: domain of linux-kernel+bounces-13913-ouuuleilei=gmail.com@vger.kernel.org designates 2604:1380:45e3:2400::1 as permitted sender) client-ip=2604:1380:45e3:2400::1; Authentication-Results: mx.google.com; dkim=pass header.i=@xry111.site header.s=default header.b=DyQU46Wl; spf=pass (google.com: domain of linux-kernel+bounces-13913-ouuuleilei=gmail.com@vger.kernel.org designates 2604:1380:45e3:2400::1 as permitted sender) smtp.mailfrom="linux-kernel+bounces-13913-ouuuleilei=gmail.com@vger.kernel.org"; dmarc=pass (p=REJECT sp=REJECT dis=NONE) header.from=xry111.site Received: from smtp.subspace.kernel.org (wormhole.subspace.kernel.org [52.25.139.140]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by sv.mirrors.kernel.org (Postfix) with ESMTPS id 0D519281C08 for <ouuuleilei@gmail.com>; Mon, 1 Jan 2024 17:23:13 +0000 (UTC) Received: from localhost.localdomain (localhost.localdomain [127.0.0.1]) by smtp.subspace.kernel.org (Postfix) with ESMTP id 9ED2D8F61; Mon, 1 Jan 2024 17:23:00 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=xry111.site header.i=@xry111.site header.b="DyQU46Wl" X-Original-To: linux-kernel@vger.kernel.org Received: from xry111.site (xry111.site [89.208.246.23]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 7817B6AC0; Mon, 1 Jan 2024 17:22:56 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=xry111.site Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=xry111.site DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=xry111.site; s=default; t=1704129763; bh=kOVhB8eAhMECnpseuk0bXmRz3MVKmBv7ePdWReB+bcU=; h=From:To:Cc:Subject:Date:From; b=DyQU46WlQcXzDRfGJtks7zcVhy3iJMOOA7Bgez0O+zKuP3IIc2sOO2lu0ZBZb90rM CLkX1SoLUCvMlaIQ5RwaXWgvZXYCayzuCokSGZW2zXX2lk6chlYbWNNcolav/EqFWK 2F6OVaVK7+YgUrO7jOatoTY7qu1kjlsjcFAaFoKw= Received: from stargazer.. (unknown [IPv6:240e:358:11a9:2200:dc73:854d:832e:3]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature ECDSA (P-384) server-digest SHA384) (Client did not present a certificate) (Authenticated sender: xry111@xry111.site) by xry111.site (Postfix) with ESMTPSA id 270FA66C94; Mon, 1 Jan 2024 12:22:34 -0500 (EST) From: Xi Ruoyao <xry111@xry111.site> To: Huacai Chen <chenhuacai@kernel.org>, WANG Xuerui <kernel@xen0n.name> Cc: Eric Biederman <ebiederm@xmission.com>, Kees Cook <keescook@chromium.org>, Tiezhu Yang <yangtiezhu@loongson.cn>, Jinyang He <hejinyang@loongson.cn>, Jiaxun Yang <jiaxun.yang@flygoat.com>, loongarch@lists.linux.dev, linux-mm@kvack.org, linux-kernel@vger.kernel.org, Xi Ruoyao <xry111@xry111.site>, stable@vger.kernel.org Subject: [PATCH] LoongArch: Fix and simplify fcsr initialization on execve Date: Tue, 2 Jan 2024 01:21:44 +0800 Message-ID: <20240101172143.14530-2-xry111@xry111.site> X-Mailer: git-send-email 2.43.0 Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: <linux-kernel.vger.kernel.org> List-Subscribe: <mailto:linux-kernel+subscribe@vger.kernel.org> List-Unsubscribe: <mailto:linux-kernel+unsubscribe@vger.kernel.org> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-getmail-retrieved-from-mailbox: INBOX X-GMAIL-THRID: 1786909602282481319 X-GMAIL-MSGID: 1786909602282481319 |
Series |
LoongArch: Fix and simplify fcsr initialization on execve
|
|
Commit Message
Xi Ruoyao
Jan. 1, 2024, 5:21 p.m. UTC
There has been a lingering bug in LoongArch Linux systems causing some
GCC tests to intermittently fail (see Closes link). I've made a minimal
reproducer:
zsh% cat measure.s
.align 4
.globl _start
_start:
movfcsr2gr $a0, $fcsr0
bstrpick.w $a0, $a0, 16, 16
beqz $a0, .ok
break 0
.ok:
li.w $a7, 93
syscall 0
zsh% cc mesaure.s -o measure -nostdlib
zsh% echo $((1.0/3))
0.33333333333333331
zsh% while ./measure; do ; done
This while loop should not stop as POSIX is clear that execve must set
fenv to the default, where FCSR should be zero. But in fact it will
just stop after running for a while (normally less than 30 seconds).
Note that "$((1.0/3))" is needed to reproduce the issue because it
raises FE_INVALID and makes fcsr0 non-zero.
The problem is we are relying on SET_PERSONALITY2 to reset
current->thread.fpu.fcsr. But SET_PERSONALITY2 is executed before
start_thread which calls lose_fpu(0). We can see if kernel preempt is
enabled, we may switch to another thread after SET_PERSONALITY2 but
before lose_fpu(0). Then bad thing happens: during the thread switch
the value of the fcsr0 register is stored into current->thread.fpu.fcsr,
making it dirty again.
The issue can be fixed by setting current->thread.fpu.fcsr after
lose_fpu(0) because lose_fpu clears TIF_USEDFPU, then the thread
switch won't touch current->thread.fpu.fcsr.
The only other architecture setting FCSR in SET_PERSONALITY2 is MIPS.
They do this for supporting different FP flavors (NaN encodings etc).
which do not exist on LoongArch. I'm not sure how MIPS evades the issue
(or maybe it's just buggy too) as I don't have a running MIPS hardware
now.
So for LoongArch, just remove the current->thread.fpu.fcsr setting from
SET_PERSONALITY2 and do it in start_thread, after lose_fpu(0). And we
just set it to 0, instead of boot_cpu_data.fpu_csr0 (because we should
provide the userspace a consistent configuration, no matter how hardware
and firmware behave).
The while loop failing with the mainline kernel has survived one hour
after this change.
Closes: https://github.com/loongson-community/discussions/issues/7
Fixes: 803b0fc5c3f2 ("LoongArch: Add process management")
Cc: stable@vger.kernel.org
Signed-off-by: Xi Ruoyao <xry111@xry111.site>
---
arch/loongarch/include/asm/elf.h | 5 -----
arch/loongarch/kernel/elf.c | 5 -----
arch/loongarch/kernel/process.c | 1 +
3 files changed, 1 insertion(+), 10 deletions(-)
Comments
Hi, Ruoyao, On Tue, Jan 2, 2024 at 1:23 AM Xi Ruoyao <xry111@xry111.site> wrote: > > There has been a lingering bug in LoongArch Linux systems causing some > GCC tests to intermittently fail (see Closes link). I've made a minimal > reproducer: > > zsh% cat measure.s > .align 4 > .globl _start > _start: > movfcsr2gr $a0, $fcsr0 > bstrpick.w $a0, $a0, 16, 16 > beqz $a0, .ok > break 0 > .ok: > li.w $a7, 93 > syscall 0 > zsh% cc mesaure.s -o measure -nostdlib > zsh% echo $((1.0/3)) > 0.33333333333333331 > zsh% while ./measure; do ; done > > This while loop should not stop as POSIX is clear that execve must set > fenv to the default, where FCSR should be zero. But in fact it will > just stop after running for a while (normally less than 30 seconds). > Note that "$((1.0/3))" is needed to reproduce the issue because it > raises FE_INVALID and makes fcsr0 non-zero. > > The problem is we are relying on SET_PERSONALITY2 to reset > current->thread.fpu.fcsr. But SET_PERSONALITY2 is executed before > start_thread which calls lose_fpu(0). We can see if kernel preempt is > enabled, we may switch to another thread after SET_PERSONALITY2 but > before lose_fpu(0). Then bad thing happens: during the thread switch > the value of the fcsr0 register is stored into current->thread.fpu.fcsr, > making it dirty again. > > The issue can be fixed by setting current->thread.fpu.fcsr after > lose_fpu(0) because lose_fpu clears TIF_USEDFPU, then the thread > switch won't touch current->thread.fpu.fcsr. > > The only other architecture setting FCSR in SET_PERSONALITY2 is MIPS. > They do this for supporting different FP flavors (NaN encodings etc). > which do not exist on LoongArch. I'm not sure how MIPS evades the issue > (or maybe it's just buggy too) as I don't have a running MIPS hardware > now. I think you can use QEMU. :) > > So for LoongArch, just remove the current->thread.fpu.fcsr setting from > SET_PERSONALITY2 and do it in start_thread, after lose_fpu(0). And we > just set it to 0, instead of boot_cpu_data.fpu_csr0 (because we should > provide the userspace a consistent configuration, no matter how hardware > and firmware behave). I still prefer to set fcsr to boot_cpu_data.fpu_csr0, because we will add LoongArch32 later, not sure whether something will change. Huacai > > The while loop failing with the mainline kernel has survived one hour > after this change. > > Closes: https://github.com/loongson-community/discussions/issues/7 > Fixes: 803b0fc5c3f2 ("LoongArch: Add process management") > Cc: stable@vger.kernel.org > Signed-off-by: Xi Ruoyao <xry111@xry111.site> > --- > arch/loongarch/include/asm/elf.h | 5 ----- > arch/loongarch/kernel/elf.c | 5 ----- > arch/loongarch/kernel/process.c | 1 + > 3 files changed, 1 insertion(+), 10 deletions(-) > > diff --git a/arch/loongarch/include/asm/elf.h b/arch/loongarch/include/asm/elf.h > index 9b16a3b8e706..f16bd42456e4 100644 > --- a/arch/loongarch/include/asm/elf.h > +++ b/arch/loongarch/include/asm/elf.h > @@ -241,8 +241,6 @@ void loongarch_dump_regs64(u64 *uregs, const struct pt_regs *regs); > do { \ > current->thread.vdso = &vdso_info; \ > \ > - loongarch_set_personality_fcsr(state); \ > - \ > if (personality(current->personality) != PER_LINUX) \ > set_personality(PER_LINUX); \ > } while (0) > @@ -259,7 +257,6 @@ do { \ > clear_thread_flag(TIF_32BIT_ADDR); \ > \ > current->thread.vdso = &vdso_info; \ > - loongarch_set_personality_fcsr(state); \ > \ > p = personality(current->personality); \ > if (p != PER_LINUX32 && p != PER_LINUX) \ > @@ -340,6 +337,4 @@ extern int arch_elf_pt_proc(void *ehdr, void *phdr, struct file *elf, > extern int arch_check_elf(void *ehdr, bool has_interpreter, void *interp_ehdr, > struct arch_elf_state *state); > > -extern void loongarch_set_personality_fcsr(struct arch_elf_state *state); > - > #endif /* _ASM_ELF_H */ > diff --git a/arch/loongarch/kernel/elf.c b/arch/loongarch/kernel/elf.c > index 183e94fc9c69..0fa81ced28dc 100644 > --- a/arch/loongarch/kernel/elf.c > +++ b/arch/loongarch/kernel/elf.c > @@ -23,8 +23,3 @@ int arch_check_elf(void *_ehdr, bool has_interpreter, void *_interp_ehdr, > { > return 0; > } > - > -void loongarch_set_personality_fcsr(struct arch_elf_state *state) > -{ > - current->thread.fpu.fcsr = boot_cpu_data.fpu_csr0; > -} > diff --git a/arch/loongarch/kernel/process.c b/arch/loongarch/kernel/process.c > index 767d94cce0de..caed58770650 100644 > --- a/arch/loongarch/kernel/process.c > +++ b/arch/loongarch/kernel/process.c > @@ -92,6 +92,7 @@ void start_thread(struct pt_regs *regs, unsigned long pc, unsigned long sp) > clear_used_math(); > regs->csr_era = pc; > regs->regs[3] = sp; > + current->thread.fpu.fcsr = 0; > } > > void flush_thread(void) > -- > 2.43.0 > >
On Tue, 2024-01-02 at 10:35 +0800, Huacai Chen wrote: /* snip */ > > The only other architecture setting FCSR in SET_PERSONALITY2 is MIPS. > > They do this for supporting different FP flavors (NaN encodings etc). > > which do not exist on LoongArch. I'm not sure how MIPS evades the issue > > (or maybe it's just buggy too) as I don't have a running MIPS hardware > > now. > I think you can use QEMU. :) I'll investigate it later. > > So for LoongArch, just remove the current->thread.fpu.fcsr setting from > > SET_PERSONALITY2 and do it in start_thread, after lose_fpu(0). And we > > just set it to 0, instead of boot_cpu_data.fpu_csr0 (because we should > > provide the userspace a consistent configuration, no matter how hardware > > and firmware behave). > I still prefer to set fcsr to boot_cpu_data.fpu_csr0, because we will > add LoongArch32 later, not sure whether something will change. I just seen fpu_csr0 is initialized to FPU_CSR_RN which is just 0 for LA64, so my concern about firmware & hardware leaving non-zero FCSR is not valid. I'll send v2 to keep using boot_cpu_data.fpu_csr0 then. > >
diff --git a/arch/loongarch/include/asm/elf.h b/arch/loongarch/include/asm/elf.h index 9b16a3b8e706..f16bd42456e4 100644 --- a/arch/loongarch/include/asm/elf.h +++ b/arch/loongarch/include/asm/elf.h @@ -241,8 +241,6 @@ void loongarch_dump_regs64(u64 *uregs, const struct pt_regs *regs); do { \ current->thread.vdso = &vdso_info; \ \ - loongarch_set_personality_fcsr(state); \ - \ if (personality(current->personality) != PER_LINUX) \ set_personality(PER_LINUX); \ } while (0) @@ -259,7 +257,6 @@ do { \ clear_thread_flag(TIF_32BIT_ADDR); \ \ current->thread.vdso = &vdso_info; \ - loongarch_set_personality_fcsr(state); \ \ p = personality(current->personality); \ if (p != PER_LINUX32 && p != PER_LINUX) \ @@ -340,6 +337,4 @@ extern int arch_elf_pt_proc(void *ehdr, void *phdr, struct file *elf, extern int arch_check_elf(void *ehdr, bool has_interpreter, void *interp_ehdr, struct arch_elf_state *state); -extern void loongarch_set_personality_fcsr(struct arch_elf_state *state); - #endif /* _ASM_ELF_H */ diff --git a/arch/loongarch/kernel/elf.c b/arch/loongarch/kernel/elf.c index 183e94fc9c69..0fa81ced28dc 100644 --- a/arch/loongarch/kernel/elf.c +++ b/arch/loongarch/kernel/elf.c @@ -23,8 +23,3 @@ int arch_check_elf(void *_ehdr, bool has_interpreter, void *_interp_ehdr, { return 0; } - -void loongarch_set_personality_fcsr(struct arch_elf_state *state) -{ - current->thread.fpu.fcsr = boot_cpu_data.fpu_csr0; -} diff --git a/arch/loongarch/kernel/process.c b/arch/loongarch/kernel/process.c index 767d94cce0de..caed58770650 100644 --- a/arch/loongarch/kernel/process.c +++ b/arch/loongarch/kernel/process.c @@ -92,6 +92,7 @@ void start_thread(struct pt_regs *regs, unsigned long pc, unsigned long sp) clear_used_math(); regs->csr_era = pc; regs->regs[3] = sp; + current->thread.fpu.fcsr = 0; } void flush_thread(void)