Received: by 2002:a05:6a10:206:0:0:0:0 with SMTP id 6csp995685pxj; Wed, 2 Jun 2021 17:35:31 -0700 (PDT) X-Google-Smtp-Source: ABdhPJxCKMkDv5QXEtTWaovwXw383r+p5brq37mliJWknFC2LFrdOpcaQL0VRX4XIqire5xfczNk X-Received: by 2002:a05:6402:da:: with SMTP id i26mr42136297edu.379.1622680531325; Wed, 02 Jun 2021 17:35:31 -0700 (PDT) ARC-Seal: i=1; a=rsa-sha256; t=1622680531; cv=none; d=google.com; s=arc-20160816; b=1GqvexGoT9UiGGdiC2AIOr9KS9h4fDecIj530+p+xth+DgpSEri4CDdY0IAscSZXuI 1FC+jtU2prvZbAvd0ktLI5cUk3n9CJ2apxIXW7KUABtx1fpzm33A4OGMVZ46SvrFaYKE 1dabLND1RuxEccKr4IUzphIbeDlV3Gk81KX2NKJnl8SdJsfk75os5jpzDz7y6jBMlTDI SSo8zYbhWN15egO2kEyzVwnHeG9ltP3RgTqotk/4SR39cgilC9iHIuUKmbpRlKEcHYTM C8Qj+Wbx5/NqlmZ6tf31AewXVNAp4RUQfxu+KfjZG0AYNS1PMrgapkWhRhVSUekjy7rh zsyA== ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=arc-20160816; h=list-id:precedence:content-transfer-encoding:mime-version :references:in-reply-to:message-id:subject:cc:to:from:date :dkim-signature; bh=jVxNTE81NIObaNox3cqDt2/qjUY4VPS8v0d/5qhjP6A=; b=egW/CpnCmxqViZUM/8szcZWOUrIPt3BR8b81bZgH8cl0AHqtueBzqwgc+/NUTq+bCh L91FkWiuP8VWxBVOQ4/g1xJjJno2FFSVrFufsBEghAiJyckdncIYh/TtJ3yVfQ5X9XJh /CQ9vBzBsqkLUB/7b2gtnYN9koH24Q/PHScjmBkYIq2ypt2ewsRga+hjej9RZg0+G19f qNGNRCGTM7KhaRfHRw8L6yeYL7gqibngTarvGJl/GDwpwyIeQnYHs1mQqublEyjjmkiw pisKTz847m8b26tNzbNwhEeYs6oQSnNzHzq5WrGl5sQaieHu/v+KGTQt/XgT3a6UciPW G/+Q== ARC-Authentication-Results: i=1; mx.google.com; dkim=pass header.i=@mail.ustc.edu.cn header.s=dkim header.b="AF4gk9/b"; spf=pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) smtp.mailfrom=linux-kernel-owner@vger.kernel.org; dmarc=pass (p=QUARANTINE sp=QUARANTINE dis=NONE) header.from=mail.ustc.edu.cn Return-Path: Received: from vger.kernel.org (vger.kernel.org. [23.128.96.18]) by mx.google.com with ESMTP id c7si1048030ejb.100.2021.06.02.17.35.08; Wed, 02 Jun 2021 17:35:31 -0700 (PDT) Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18; Authentication-Results: mx.google.com; dkim=pass header.i=@mail.ustc.edu.cn header.s=dkim header.b="AF4gk9/b"; spf=pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) smtp.mailfrom=linux-kernel-owner@vger.kernel.org; dmarc=pass (p=QUARANTINE sp=QUARANTINE dis=NONE) header.from=mail.ustc.edu.cn Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S229567AbhFCAfJ (ORCPT + 99 others); Wed, 2 Jun 2021 20:35:09 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:59520 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S229541AbhFCAfI (ORCPT ); Wed, 2 Jun 2021 20:35:08 -0400 Received: from ustc.edu.cn (email6.ustc.edu.cn [IPv6:2001:da8:d800::8]) by lindbergh.monkeyblade.net (Postfix) with ESMTP id 38741C06174A for ; Wed, 2 Jun 2021 17:33:22 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=mail.ustc.edu.cn; s=dkim; h=Received:Date:From:To:Cc:Subject: Message-ID:In-Reply-To:References:MIME-Version:Content-Type: Content-Transfer-Encoding; bh=jVxNTE81NIObaNox3cqDt2/qjUY4VPS8v0 d/5qhjP6A=; b=AF4gk9/bZe79S98kgtIWqrp+7jHR3xBJvt3mjdM0uO32YFXn38 nOk4DqtpUk0B1CMQeTkdObg9+qDrIzkTEAeaCa6s2nM3SbyTAHKg2wNjUVZEhoh/ Pz4aUGAdBkJsb5+v09yYB3ySpMmsgWpG71i9aH7Dp4HThsAZvCw1ZkwjU= Received: from xhacker (unknown [101.86.20.15]) by newmailweb.ustc.edu.cn (Coremail) with SMTP id LkAmygBXyqIrI7hg9GOEAA--.43861S2; Thu, 03 Jun 2021 08:32:43 +0800 (CST) Date: Thu, 3 Jun 2021 08:27:14 +0800 From: Jisheng Zhang To: Nathan Chancellor Cc: Paul Walmsley , Palmer Dabbelt , Albert Ou , Alexandre Ghiti , linux-riscv@lists.infradead.org, linux-kernel@vger.kernel.org, clang-built-linux@googlegroups.com Subject: Re: [PATCH] riscv: mm: init: Consolidate vars, functions Message-ID: <20210603082714.6e524f4c@xhacker> In-Reply-To: References: <20210516211556.43c00055@xhacker> <20210602231226.0e45524b@xhacker> <20210602235851.001a0d41@xhacker> MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit X-CM-TRANSID: LkAmygBXyqIrI7hg9GOEAA--.43861S2 X-Coremail-Antispam: 1UD129KBjvAXoWfJryDKF15GF18AF13tr1fJFb_yoW8Jw18Ko WrJr17Jw4xGryUJrn3A345XFWUGw1qyFsxuryUAr4DWrnrta1jyFWUGr4rZa98tw1rGFZr Cr1Utwn7Aa1kJrn7n29KB7ZKAUJUUUUU529EdanIXcx71UUUUU7v73VFW2AGmfu7bjvjm3 AaLaJ3UjIYCTnIWjp_UUU5A7k0a2IF6w4kM7kC6x804xWl14x267AKxVW8JVW5JwAFc2x0 x2IEx4CE42xK8VAvwI8IcIk0rVWrJVCq3wAFIxvE14AKwVWUJVWUGwA2ocxC64kIII0Yj4 1l84x0c7CEw4AK67xGY2AK021l84ACjcxK6xIIjxv20xvE14v26ryj6F1UM28EF7xvwVC0 I7IYx2IY6xkF7I0E14v26r4j6F4UM28EF7xvwVC2z280aVAFwI0_Cr1j6rxdM28EF7xvwV C2z280aVCY1x0267AKxVW0oVCq3wAS0I0E0xvYzxvE52x082IY62kv0487Mc02F40EFcxC 0VAKzVAqx4xG6I80ewAv7VC0I7IYx2IY67AKxVWUJVWUGwAv7VC2z280aVAFwI0_Jr0_Gr 1lOx8S6xCaFVCjc4AY6r1j6r4UM4x0Y48IcVAKI48JMxAIw28IcxkI7VAKI48JMxC20s02 6xCaFVCjc4AY6r1j6r4UMI8I3I0E5I8CrVAFwI0_Jr0_Jr4lx2IqxVCjr7xvwVAFwI0_Jr I_JrWlx4CE17CEb7AF67AKxVWUtVW8ZwCIc40Y0x0EwIxGrwCI42IY6xIIjxv20xvE14v2 6r1j6r1xMIIF0xvE2Ix0cI8IcVCY1x0267AKxVWUJVW8JwCI42IY6xAIw20EY4v20xvaj4 0_WFyUJVCq3wCI42IY6I8E87Iv67AKxVWUJVW8JwCI42IY6I8E87Iv6xkF7I0E14v26r4j 6r4UJbIYCTnIWIevJa73UjIFyTuYvjxUqEoXUUUUU X-CM-SenderInfo: xmv2xttqjtqzxdloh3xvwfhvlgxou0/ Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Wed, 2 Jun 2021 12:54:40 -0700 Nathan Chancellor wrote: > On Wed, Jun 02, 2021 at 11:58:51PM +0800, Jisheng Zhang wrote: > > On Wed, 2 Jun 2021 23:12:26 +0800 > > Jisheng Zhang wrote: > > > > > On Tue, 1 Jun 2021 13:21:05 -0700 > > > Nathan Chancellor wrote: > > > > > > > Hi Jisheng, > > > > > > Hi Nathan, > > > > > > > > > > > On Sun, May 16, 2021 at 09:15:56PM +0800, Jisheng Zhang wrote: > > > > > From: Jisheng Zhang > > > > > > > > > > Consolidate the following items in init.c > > > > > > > > > > Staticize global vars as much as possible; > > > > > Add __initdata mark if the global var isn't needed after init > > > > > Add __init mark if the func isn't needed after init > > > > > Add __ro_after_init if the global var is read only after init > > > > > > > > > > Signed-off-by: Jisheng Zhang > > > > > --- > > > > > arch/riscv/include/asm/set_memory.h | 2 +- > > > > > arch/riscv/mm/init.c | 36 +++++++++++++++-------------- > > > > > 2 files changed, 20 insertions(+), 18 deletions(-) > > > > > > > > > > diff --git a/arch/riscv/include/asm/set_memory.h b/arch/riscv/include/asm/set_memory.h > > > > > index 086f757e8ba3..9d4d455726d4 100644 > > > > > --- a/arch/riscv/include/asm/set_memory.h > > > > > +++ b/arch/riscv/include/asm/set_memory.h > > > > > @@ -27,7 +27,7 @@ static inline int set_memory_rw_nx(unsigned long addr, int numpages) { return 0; > > > > > #endif > > > > > > > > > > #if defined(CONFIG_64BIT) && defined(CONFIG_STRICT_KERNEL_RWX) > > > > > -void protect_kernel_linear_mapping_text_rodata(void); > > > > > +void __init protect_kernel_linear_mapping_text_rodata(void); > > > > > #else > > > > > static inline void protect_kernel_linear_mapping_text_rodata(void) {} > > > > > #endif > > > > > diff --git a/arch/riscv/mm/init.c b/arch/riscv/mm/init.c > > > > > index 4c4c92ce0bb8..eac2d5c27b3e 100644 > > > > > --- a/arch/riscv/mm/init.c > > > > > +++ b/arch/riscv/mm/init.c > > > > > @@ -53,7 +53,7 @@ struct pt_alloc_ops { > > > > > #endif > > > > > }; > > > > > > > > > > -static phys_addr_t dma32_phys_limit __ro_after_init; > > > > > +static phys_addr_t dma32_phys_limit __initdata; > > > > > > > > > > static void __init zone_sizes_init(void) > > > > > { > > > > > @@ -184,7 +184,7 @@ extern char _sdata[], _edata[]; > > > > > #endif /* CONFIG_XIP_KERNEL */ > > > > > > > > > > #ifdef CONFIG_MMU > > > > > -static struct pt_alloc_ops _pt_ops __ro_after_init; > > > > > +static struct pt_alloc_ops _pt_ops __initdata; > > > > > > > > > > #ifdef CONFIG_XIP_KERNEL > > > > > #define pt_ops (*(struct pt_alloc_ops *)XIP_FIXUP(&_pt_ops)) > > > > > @@ -200,13 +200,13 @@ EXPORT_SYMBOL(va_pa_offset); > > > > > #endif > > > > > /* Offset between kernel mapping virtual address and kernel load address */ > > > > > #ifdef CONFIG_64BIT > > > > > -unsigned long va_kernel_pa_offset; > > > > > +unsigned long va_kernel_pa_offset __ro_after_init; > > > > > EXPORT_SYMBOL(va_kernel_pa_offset); > > > > > #endif > > > > > #ifdef CONFIG_XIP_KERNEL > > > > > #define va_kernel_pa_offset (*((unsigned long *)XIP_FIXUP(&va_kernel_pa_offset))) > > > > > #endif > > > > > -unsigned long va_kernel_xip_pa_offset; > > > > > +unsigned long va_kernel_xip_pa_offset __ro_after_init; > > > > > EXPORT_SYMBOL(va_kernel_xip_pa_offset); > > > > > #ifdef CONFIG_XIP_KERNEL > > > > > #define va_kernel_xip_pa_offset (*((unsigned long *)XIP_FIXUP(&va_kernel_xip_pa_offset))) > > > > > @@ -216,7 +216,7 @@ EXPORT_SYMBOL(pfn_base); > > > > > > > > > > pgd_t swapper_pg_dir[PTRS_PER_PGD] __page_aligned_bss; > > > > > pgd_t trampoline_pg_dir[PTRS_PER_PGD] __page_aligned_bss; > > > > > -pte_t fixmap_pte[PTRS_PER_PTE] __page_aligned_bss; > > > > > +static pte_t fixmap_pte[PTRS_PER_PTE] __page_aligned_bss; > > > > > > > > > > pgd_t early_pg_dir[PTRS_PER_PGD] __initdata __aligned(PAGE_SIZE); > > > > > > > > > > @@ -253,7 +253,7 @@ static inline pte_t *__init get_pte_virt_fixmap(phys_addr_t pa) > > > > > return (pte_t *)set_fixmap_offset(FIX_PTE, pa); > > > > > } > > > > > > > > > > -static inline pte_t *get_pte_virt_late(phys_addr_t pa) > > > > > +static inline pte_t *__init get_pte_virt_late(phys_addr_t pa) > > > > > { > > > > > return (pte_t *) __va(pa); > > > > > } > > > > > @@ -272,7 +272,7 @@ static inline phys_addr_t __init alloc_pte_fixmap(uintptr_t va) > > > > > return memblock_phys_alloc(PAGE_SIZE, PAGE_SIZE); > > > > > } > > > > > > > > > > -static phys_addr_t alloc_pte_late(uintptr_t va) > > > > > +static phys_addr_t __init alloc_pte_late(uintptr_t va) > > > > > { > > > > > unsigned long vaddr; > > > > > > > > > > @@ -296,10 +296,10 @@ static void __init create_pte_mapping(pte_t *ptep, > > > > > > > > > > #ifndef __PAGETABLE_PMD_FOLDED > > > > > > > > > > -pmd_t trampoline_pmd[PTRS_PER_PMD] __page_aligned_bss; > > > > > -pmd_t fixmap_pmd[PTRS_PER_PMD] __page_aligned_bss; > > > > > -pmd_t early_pmd[PTRS_PER_PMD] __initdata __aligned(PAGE_SIZE); > > > > > -pmd_t early_dtb_pmd[PTRS_PER_PMD] __initdata __aligned(PAGE_SIZE); > > > > > +static pmd_t trampoline_pmd[PTRS_PER_PMD] __page_aligned_bss; > > > > > +static pmd_t fixmap_pmd[PTRS_PER_PMD] __page_aligned_bss; > > > > > +static pmd_t early_pmd[PTRS_PER_PMD] __initdata __aligned(PAGE_SIZE); > > > > > +static pmd_t early_dtb_pmd[PTRS_PER_PMD] __initdata __aligned(PAGE_SIZE); > > > > > > > > > > #ifdef CONFIG_XIP_KERNEL > > > > > #define trampoline_pmd ((pmd_t *)XIP_FIXUP(trampoline_pmd)) > > > > > @@ -319,7 +319,7 @@ static pmd_t *__init get_pmd_virt_fixmap(phys_addr_t pa) > > > > > return (pmd_t *)set_fixmap_offset(FIX_PMD, pa); > > > > > } > > > > > > > > > > -static pmd_t *get_pmd_virt_late(phys_addr_t pa) > > > > > +static pmd_t *__init get_pmd_virt_late(phys_addr_t pa) > > > > > { > > > > > return (pmd_t *) __va(pa); > > > > > } > > > > > @@ -336,7 +336,7 @@ static phys_addr_t __init alloc_pmd_fixmap(uintptr_t va) > > > > > return memblock_phys_alloc(PAGE_SIZE, PAGE_SIZE); > > > > > } > > > > > > > > > > -static phys_addr_t alloc_pmd_late(uintptr_t va) > > > > > +static phys_addr_t __init alloc_pmd_late(uintptr_t va) > > > > > { > > > > > unsigned long vaddr; > > > > > > > > > > @@ -454,14 +454,16 @@ asmlinkage void __init __copy_data(void) > > > > > #error "setup_vm() is called from head.S before relocate so it should not use absolute addressing." > > > > > #endif > > > > > > > > > > -uintptr_t load_pa, load_sz; > > > > > +static uintptr_t load_pa __initdata; > > > > > > > > Making load_pa static causing clang built kernels to no longer boot, > > > > hanging after just a few lines of output in the console: > > > > > > > > https://github.com/ClangBuiltLinux/continuous-integration2/runs/2717606254?check_suite_focus=true > > > > > > > > I am not sure why that would make a difference nor why GCC is okay with > > > > it. If it is a clang bug, it appears to be there for a while, given that > > > > it reproduces back to clang-11. > > > > > > I can reproduce the issue. Here are my findindings: > > > > > > * gcc + binutils can't reproduce it > > > * clang + llvm-utils + ias can reproduce it > > > * clang + binutils can reproduce it > > Yes, this seems like something strictly related to clang so that is what > I have been testing with. > > > > All below tests are done with clang + binutils. > > > > > > Then I applied below modification: > > > > > > -static uintptr_t load_pa __initdata; > > > +uintptr_t load_pa __initdata; > > > > > > I got below panic log: > > > [ 0.015418] Unable to handle kernel paging request at virtual address fffffffffffffff9 > > > [ 0.016432] Oops [#1] > > > [ 0.016679] Modules linked in: > > > [ 0.017103] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 5.13.0-rc1+ #5 > > > [ 0.017711] Hardware name: riscv-virtio,qemu (DT) > > > [ 0.018201] epc : trace_hardirqs_on+0x60/0xb2 > > > [ 0.018582] ra : restore_all+0xe/0x66 > > > [ 0.018879] epc : ffffffff800cb09a ra : ffffffff800027b8 sp : ffffffff80c03dd0 > > > [ 0.019376] gp : ffffffff80d001c8 tp : ffffffff80c0c180 t0 : 0000000000000020 > > > [ 0.019870] t1 : ffffffff80006e40 t2 : ffffffff800d2e0a s0 : ffffffff80c03e00 > > > [ 0.020346] s1 : 0000000000000001 a0 : 0000000000000001 a1 : 0000000000000001 > > > [ 0.020800] a2 : 0000000000000001 a3 : 0000000000000000 a4 : 0000000000000000 > > > [ 0.021243] a5 : 0000000000000000 a6 : 0000000000000000 a7 : 0000000054494d45 > > > [ 0.021717] s2 : ffffffff800027b8 s3 : ffffffff80d35968 s4 : ffffffff8061e1d8 > > > [ 0.022179] s5 : ffffffff80c0c180 s6 : ffffffe000e34b50 s7 : 00000000800130f0 > > > [ 0.022674] s8 : 000000000000007f s9 : 0000000080012010 s10: 0000000000000000 > > > [ 0.023176] s11: 0000000000000000 t3 : ffffffff80d00108 t4 : ffffffff80006e40 > > > [ 0.023693] t5 : ffffffff80006e40 t6 : ffffffff800d2e0a > > > [ 0.024153] status: 0000000000000100 badaddr: fffffffffffffff9 cause: 000000000000000d > > > [ 0.025367] Call Trace: > > > [ 0.025749] [] trace_hardirqs_on+0x60/0xb2 > > > [ 0.026402] [] restore_all+0xe/0x66 > > > [ 0.027261] Unable to handle kernel paging request at virtual address fffffffffffffffa > > > [ 0.027827] Oops [#2] > > > [ 0.028013] Modules linked in: > > > [ 0.028321] CPU: 0 PID: 0 Comm: swapper/0 Tainted: G D 5.13.0-rc1+ #5 > > > [ 0.028839] Hardware name: riscv-virtio,qemu (DT) > > > [ 0.029166] epc : trace_hardirqs_on+0x60/0xb2 > > > [ 0.029505] ra : restore_all+0xe/0x66 > > > [ 0.029785] epc : ffffffff800cb09a ra : ffffffff800027b8 sp : ffffffff80c03a80 > > > [ 0.030266] gp : ffffffff80d001c8 tp : ffffffff80c0c180 t0 : 0000000000000020 > > > [ 0.030748] t1 : ffffffff80006e40 t2 : ffffffff800d2e0a s0 : ffffffff80c03ab0 > > > [ 0.031227] s1 : 0000000000000001 a0 : 0000000000000002 a1 : 0000000000000002 > > > [ 0.031717] a2 : 0000000000000001 a3 : 0000000000000000 a4 : 0000000000000000 > > > [ 0.032199] a5 : 0000000000000000 a6 : 0000000000000000 a7 : 0000000054494d45 > > > [ 0.032680] s2 : ffffffff800027b8 s3 : ffffffff80d35968 s4 : ffffffff8061e1d8 > > > [ 0.033160] s5 : ffffffff80c0c180 s6 : ffffffe000e34b50 s7 : 00000000800130f0 > > > [ 0.033642] s8 : 000000000000007f s9 : 0000000080012010 s10: 0000000000000000 > > > [ 0.034123] s11: 0000000000000000 t3 : ffffffff80d00108 t4 : ffffffff80006e40 > > > [ 0.034601] t5 : ffffffff80006e40 t6 : ffffffff800d2e0a > > > [ 0.034965] status: 0000000000000100 badaddr: fffffffffffffffa cause: 000000000000000d > > > [ 0.035492] Call Trace: > > > [ 0.035682] [] trace_hardirqs_on+0x60/0xb2 > > > [ 0.036077] [] restore_all+0xe/0x66 > > > [ 0.036545] ---[ end trace 7f4fbff09d927668 ]--- > > > [ 0.037188] Kernel panic - not syncing: Attempted to kill the idle task! > > > [ 0.038107] ---[ end Kernel panic - not syncing: Attempted to kill the idle task! ]--- > > > > > > Then I checked 5.13-rc1, above panic log can be reproduced too. So the issue > > > should exist there for a while. I never tried clang with riscv, did you remember > > > which last commit or version clang works, I may try to bisect. > V> > > > > > More findings: > > > > *The above panic issue can also be seen from 5.12-rc2. If disable FTRACE, then > > the panic disappears, kernel can boot > > > > *so I retested riscv next tree w/ FTRACE disabled, kernel can boot w/ below > > modification: > > Yeah, I do not enable CONFIG_FTRACE because it is not enabled in > ARCH=riscv defconfig by default. > > > -static uintptr_t load_pa __initdata; > > +uintptr_t load_pa __initdata; > > > > This is a weird issue. Any clue is appreciated. > > Unfortunately, this is outside of my realm of expertise, as I am > unfamiliar with RISC-V at this level. Maybe Palmer has some ideas. I > would think that changing this variable to static would be fine given > that the symbol is only used in this translation unit but clearly not. I > have attempted to debug this in gdb but that does not really get me > anywhere: I cannot break on start_kernel() for whatever reason and the > kernel never gets to my breakpoint in setup_vm(). > > I did decide to through a BUG() around arch/riscv/mm/init.c to see > exactly which statement causes everything to hang. I landed on: > > csr_write(CSR_SATP, PFN_DOWN(__pa_symbol(swapper_pg_dir)) | SATP_MODE); > > in setup_vm_final(). No idea how that is relevant to this. > > Some people in the ClangBuiltLinux bi-weekly meeting today pointed out > that since load_pa is assigned to the linker defined symbol _start, > there could be some optimization that goes awry here. I have keyed that Thanks for the inspiration. Below patch fixes the hang issue, but I didn't go through all necessary WRITE_ONCE convertions. diff --git a/arch/riscv/mm/init.c b/arch/riscv/mm/init.c index ae32f78207f0..fa9336a2583f 100644 --- a/arch/riscv/mm/init.c +++ b/arch/riscv/mm/init.c @@ -504,7 +504,7 @@ asmlinkage void __init setup_vm(uintptr_t dtb_pa) va_kernel_xip_pa_offset = kernel_virt_addr - xiprom; #else - load_pa = (uintptr_t)(&_start); + WRITE_ONCE(load_pa, (uintptr_t)(&_start)); load_sz = (uintptr_t)(&_end) - load_pa; #endif > list in now. For full context: > > https://lore.kernel.org/r/YLaWseLdg5JYElVx@Ryzen-9-3900X.localdomain/ > > Cheers, > Nathan