From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mta0.migadu.com (out-223.mta0.migadu.com [91.218.175.223]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 8B83454B1CB for ; Wed, 9 Sep 2026 13:08:23 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=91.218.175.223 ARC-Seal:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1788959306; cv=none; b=XkV7xZziLtw4MtpZ/kyoWIxXIBi0vt2lgkbofjcF/3ZQiWPZGDWSPFBtP0/OSvBi4b1/kTG3O4gPj4fjtW6Th9wCv4gb20em3hFewyZvKdUCYsFtJF5vpyH0z/c8ECQlZww0VpHKpwOioG00t5agbKWKigalcipOXk+GbY4BWks= ARC-Message-Signature:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1788959306; c=relaxed/simple; bh=VhlFjNAgQI2Jtqxdp11XUFniGZMKsYGwyqeIsjbkGdA=; h=Message-ID:Date:MIME-Version:Subject:From:To:Cc:References: In-Reply-To:Content-Type; b=GuEOZH4X/6+DFo88whzUGHy8Z5Av7utdH63gkkM7gbhklgNCl7PG8CjB8zy2aktHociy9bMym0/jvQZujRqTmsIZqx15gOlYsErlyB6MIWLHgjw4BFx1mIqptNTuenRQRMUlj3KV2jAxT/nXRCv382gJ8lR5exEJhpDoWYC4It8= ARC-Authentication-Results:i=1; smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=linux.dev; spf=pass smtp.mailfrom=linux.dev; dkim=pass (1024-bit key) header.d=linux.dev header.i=@linux.dev header.b=SqpWv98C; arc=none smtp.client-ip=91.218.175.223 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=linux.dev Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=linux.dev Authentication-Results: smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=linux.dev header.i=@linux.dev header.b="SqpWv98C" X-Envelope-To: linux-kernel@vger.kernel.org DKIM-Signature: a=rsa-sha256; bh=VhlFjNAgQI2Jtqxdp11XUFniGZMKsYGwyqeIsjbkGdA=; c=simple/simple; d=linux.dev; h=from:to:subject:date:message-id:mime-version:content-type; s=key1; t=1788959301; v=1; x=1789564101; b=SqpWv98C61R76bJ0zUrMa/dUYoDYli14Cu7Sz2NGtXopPX3r0MU39mdYsfsxVAcN0nIXkHZB iDpYgU6xxC+1nMXUk9M4EJPCHIPpCUSVYrqxdZ4RPhIqRh0+meNClDCpRd33JuUrmJn0ZAzk0IJ KsweRhAgxCuHVDoM9FQKa/ME= X-Envelope-To: linux-kernel@vger.kernel.org Received: by smtp.migadu.com with ESMTPS id 25b253702b94a6c9; Wed, 09 Sep 2026 13:08:21 +0000 X-Mizu-Trace-ID: 25b253702b94a6c9 X-Migadu-Flow: FLOW_OUT Message-ID: <7dbba23e-5530-43c4-9c44-20108cd0cef5@linux.dev> Date: Wed, 9 Sep 2026 21:08:09 +0800 Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 User-Agent: Mozilla Thunderbird Subject: Re: [PATCH v9 3/4] module: introduce SH_ENTSIZE_STANDALONE for separately allocated sections From: Hao Ge To: Petr Pavlu Cc: Luis Chamberlain , Daniel Gomez , Sami Tolvanen , Aaron Tomlin , Suren Baghdasaryan , Andrew Morton , linux-modules@vger.kernel.org, linux-kernel@vger.kernel.org, linux-mm@kvack.org, Sashiko , stable@vger.kernel.org References: <20260908092412.115953-1-hao.ge@linux.dev> <20260908092412.115953-4-hao.ge@linux.dev> <76b5edbb-8231-41d0-9e7b-f965af13c324@suse.com> Content-Language: en-US In-Reply-To: Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 8bit On 9/9/26 20:47, Hao Ge wrote: > Hi Petr > > > On 9/9/26 19:30, Petr Pavlu wrote: >> On 9/8/26 11:24 AM, Hao Ge wrote: >>> SHF_ALLOC means, per the ELF spec, that a section occupies memory >>> during process execution. Some module sections occupy memory >>> outside the regular module layout, for example the percpu section >>> with its per-CPU allocations. The loader currently excludes such >>> a section from the layout by clearing its SHF_ALLOC, which >>> overloads the flag with a loader-internal meaning. >>> apply_relocations() needs a special case for the section, and >>> find_sec(".data..percpu") returns different results before and >>> after layout_and_allocate(). >>> >>> Introduce SH_ENTSIZE_STANDALONE to mark sections with a separate >>> allocation. The percpu section is its first user. layout_sections() >>> and move_module() skip marked sections, and apply_relocations() goes >>> back to testing only SHF_ALLOC. Based on a patch by Petr Pavlu [1]. >>> >>> .data..percpu keeps SHF_ALLOC, so it would now show up under >>> /sys/module/*/sections/. The section has one instance per CPU and no >>> single address to report, and the entry never existed before, so >>> skip it in add_sect_attrs(). add_notes_attrs() indexes its attrs[] >>> array and skips it too. No functional change otherwise. >>> >>> Fixes: 4835f747d3ed ("alloc_tag: support for page allocation tag >>> compression") >>> Reported-by: Sashiko >>> Link: >>> https://lore.kernel.org/all/499bb60c-c6e3-43a3-bd92-95a0567ece5e@suse.com/ >>> [1] >>> Suggested-by: Petr Pavlu >>> Cc: stable@vger.kernel.org >>> Signed-off-by: Hao Ge >>> --- >>> [...] >>> diff --git a/kernel/module/kallsyms.c b/kernel/module/kallsyms.c >>> index 0fc11e45df9b..49190deae61e 100644 >>> --- a/kernel/module/kallsyms.c >>> +++ b/kernel/module/kallsyms.c >>> @@ -76,7 +76,7 @@ static char elf_type(const Elf_Sym *sym, const >>> struct load_info *info) >>>   } >>>     static bool is_core_symbol(const Elf_Sym *src, const Elf_Shdr >>> *sechdrs, >>> -               unsigned int shnum, unsigned int pcpundx) >>> +               unsigned int shnum) >>>   { >>>       const Elf_Shdr *sec; >>>       enum mod_mem_type type; >>> @@ -86,11 +86,6 @@ static bool is_core_symbol(const Elf_Sym *src, >>> const Elf_Shdr *sechdrs, >>>           !src->st_name) >>>           return false; >>>   -#ifdef CONFIG_KALLSYMS_ALL >>> -    if (src->st_shndx == pcpundx) >>> -        return true; >>> -#endif >>> - >>>       sec = sechdrs + src->st_shndx; >>>       type = sec->sh_entsize >> SH_ENTSIZE_TYPE_SHIFT; >>>       if (!(sec->sh_flags & SHF_ALLOC) >>> @@ -131,8 +126,7 @@ void layout_symtab(struct module *mod, struct >>> load_info *info) >>>       /* Compute total space required for the core symbols' strtab. */ >>>       for (ndst = i = 0; i < nsrc; i++) { >>>           if (i == 0 || is_livepatch_module(mod) || >>> -            is_core_symbol(src + i, info->sechdrs, info->hdr->e_shnum, >>> -                   info->index.pcpu)) { >>> +            is_core_symbol(src + i, info->sechdrs, >>> info->hdr->e_shnum)) { >>>               strtab_size += strlen(&info->strtab[src[i].st_name]) + 1; >>>               ndst++; >>>           } >>> @@ -199,8 +193,7 @@ void add_kallsyms(struct module *mod, const >>> struct load_info *info) >>>       for (ndst = i = 0; i < kallsyms->num_symtab; i++) { >>>           kallsyms->typetab[i] = elf_type(src + i, info); >>>           if (i == 0 || is_livepatch_module(mod) || >>> -            is_core_symbol(src + i, info->sechdrs, info->hdr->e_shnum, >>> -                   info->index.pcpu)) { >>> +            is_core_symbol(src + i, info->sechdrs, >>> info->hdr->e_shnum)) { >>>               ssize_t ret; >>>                 mod->core_kallsyms.typetab[ndst] = >> FTR These changes in kernel/module/kallsyms.c have a conflict with the >> series "Ignore local labels and mapping symbols during module load" [1], >> which is currently queued on modules-next, but it should be >> straightforward to resolve. >> >>> diff --git a/kernel/module/sysfs.c b/kernel/module/sysfs.c >>> index 01c65d608873..f64170344e69 100644 >>> --- a/kernel/module/sysfs.c >>> +++ b/kernel/module/sysfs.c >>> @@ -62,6 +62,15 @@ static void free_sect_attrs(struct >>> module_sect_attrs *sect_attrs) >>>       kfree(sect_attrs); >>>   } >>>   +/* >>> + * .data..percpu has a separate allocation per CPU and no single >>> + * address to report. >>> + */ >>> +static bool sect_visible(const struct load_info *info, unsigned int i) >>> +{ >>> +    return !sect_empty(&info->sechdrs[i]) && i != info->index.pcpu; >>> +} >>> + >>>   static int add_sect_attrs(struct module *mod, const struct >>> load_info *info) >>>   { >>>       struct module_sect_attrs *sect_attrs; >>> @@ -72,7 +81,7 @@ static int add_sect_attrs(struct module *mod, >>> const struct load_info *info) >>>         /* Count loaded sections and allocate structures */ >>>       for (i = 0; i < info->hdr->e_shnum; i++) >>> -        if (!sect_empty(&info->sechdrs[i])) >>> +        if (sect_visible(info, i)) >>>               nloaded++; >>>       sect_attrs = kzalloc_flex(*sect_attrs, attrs, nloaded); >>>       if (!sect_attrs) >>> @@ -92,7 +101,7 @@ static int add_sect_attrs(struct module *mod, >>> const struct load_info *info) >>>       for (i = 0; i < info->hdr->e_shnum; i++) { >>>           Elf_Shdr *sec = &info->sechdrs[i]; >>>   -        if (sect_empty(sec)) >>> +        if (!sect_visible(info, i)) >>>               continue; >>>           sysfs_bin_attr_init(sattr); >>>           sattr->attr.name = >>> @@ -181,7 +190,7 @@ static int add_notes_attrs(struct module *mod, >>> const struct load_info *info) >>>         nattr = ¬es_attrs->attrs[0]; >>>       for (loaded = i = 0; i < info->hdr->e_shnum; ++i) { >>> -        if (sect_empty(&info->sechdrs[i])) >>> +        if (!sect_visible(info, i)) >>>               continue; >>>           if (info->sechdrs[i].sh_type == SHT_NOTE) { >>>               sysfs_bin_attr_init(nattr); >> add_notes_attrs() has two sect_empty() calls. Both should be changed to >> sect_visible(). > > > I kept this part unmodified to preserve the loop's original intent. > > This loop counts SHT_NOTE sections. > > SHT_NOTE refers to ELF note sections, which hold non-executable > > metadata such as build ID and ABI info. > > I wonder if we could keep the current implementation. > > As noted in the comment above, the top part counts SHT_NOTE sections > > and allocates structures, while the lower logic handles control of > node attributes. > > > WDYT? > > Sorry, I've reconsidered this. I think changing it to sect_visible would be better. sect_visible stands for the count of externally visible note attributes, so the code above and below can align with each other. Sorry for the noise. I'll rebase onto modules-next, apply this change and push a new revision. I'll also add your Reviewed-by tag. Thanks Best Regards Hao > Thanks > > Best Regards > > Hao > >> >> With this fixed, the patch looks ok to me. Feel free to add: >> >> Reviewed-by: Petr Pavlu >> >> [1] >> https://lore.kernel.org/linux-modules/20260820125007.22943-1-yangtiezhu@loongson.cn/ >>