Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.33 as permitted sender) client-ip=23.128.96.33;
Message-ID: <eae782dc-17da-4d2b-9840-f2b027d5b192@intel.com>
Date:   Thu, 2 Nov 2023 09:02:44 -0700
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
Subject: Re: [PATCH] x86/mm/ident_map: Use gbpages only where full GB page
 should be mapped.
Content-Language: en-US
To:     Steve Wahl <steve.wahl@hpe.com>, rja_direct@groups.int.hpe.com,
        Dave Hansen <dave.hansen@linux.intel.com>,
        Andy Lutomirski <luto@kernel.org>,
        Peter Zijlstra <peterz@infradead.org>,
        Thomas Gleixner <tglx@linutronix.de>,
        Ingo Molnar <mingo@redhat.com>, Borislav Petkov <bp@alien8.de>,
        x86@kernel.org, "H. Peter Anvin" <hpa@zytor.com>,
        linux-kernel@vger.kernel.org
References: <20231031195049.2075561-1-steve.wahl@hpe.com>
From:   Dave Hansen <dave.hansen@intel.com>
Autocrypt: addr=dave.hansen@intel.com; keydata=
 xsFNBE6HMP0BEADIMA3XYkQfF3dwHlj58Yjsc4E5y5G67cfbt8dvaUq2fx1lR0K9h1bOI6fC
 oAiUXvGAOxPDsB/P6UEOISPpLl5IuYsSwAeZGkdQ5g6m1xq7AlDJQZddhr/1DC/nMVa/2BoY
 2UnKuZuSBu7lgOE193+7Uks3416N2hTkyKUSNkduyoZ9F5twiBhxPJwPtn/wnch6n5RsoXsb
 ygOEDxLEsSk/7eyFycjE+btUtAWZtx+HseyaGfqkZK0Z9bT1lsaHecmB203xShwCPT49Blxz
 VOab8668QpaEOdLGhtvrVYVK7x4skyT3nGWcgDCl5/Vp3TWA4K+IofwvXzX2ON/Mj7aQwf5W
 iC+3nWC7q0uxKwwsddJ0Nu+dpA/UORQWa1NiAftEoSpk5+nUUi0WE+5DRm0H+TXKBWMGNCFn
 c6+EKg5zQaa8KqymHcOrSXNPmzJuXvDQ8uj2J8XuzCZfK4uy1+YdIr0yyEMI7mdh4KX50LO1
 pmowEqDh7dLShTOif/7UtQYrzYq9cPnjU2ZW4qd5Qz2joSGTG9eCXLz5PRe5SqHxv6ljk8mb
 ApNuY7bOXO/A7T2j5RwXIlcmssqIjBcxsRRoIbpCwWWGjkYjzYCjgsNFL6rt4OL11OUF37wL
 QcTl7fbCGv53KfKPdYD5hcbguLKi/aCccJK18ZwNjFhqr4MliQARAQABzUVEYXZpZCBDaHJp
 c3RvcGhlciBIYW5zZW4gKEludGVsIFdvcmsgQWRkcmVzcykgPGRhdmUuaGFuc2VuQGludGVs
 LmNvbT7CwXgEEwECACIFAlQ+9J0CGwMGCwkIBwMCBhUIAgkKCwQWAgMBAh4BAheAAAoJEGg1
 lTBwyZKwLZUP/0dnbhDc229u2u6WtK1s1cSd9WsflGXGagkR6liJ4um3XCfYWDHvIdkHYC1t
 MNcVHFBwmQkawxsYvgO8kXT3SaFZe4ISfB4K4CL2qp4JO+nJdlFUbZI7cz/Td9z8nHjMcWYF
 IQuTsWOLs/LBMTs+ANumibtw6UkiGVD3dfHJAOPNApjVr+M0P/lVmTeP8w0uVcd2syiaU5jB
 aht9CYATn+ytFGWZnBEEQFnqcibIaOrmoBLu2b3fKJEd8Jp7NHDSIdrvrMjYynmc6sZKUqH2
 I1qOevaa8jUg7wlLJAWGfIqnu85kkqrVOkbNbk4TPub7VOqA6qG5GCNEIv6ZY7HLYd/vAkVY
 E8Plzq/NwLAuOWxvGrOl7OPuwVeR4hBDfcrNb990MFPpjGgACzAZyjdmYoMu8j3/MAEW4P0z
 F5+EYJAOZ+z212y1pchNNauehORXgjrNKsZwxwKpPY9qb84E3O9KYpwfATsqOoQ6tTgr+1BR
 CCwP712H+E9U5HJ0iibN/CDZFVPL1bRerHziuwuQuvE0qWg0+0SChFe9oq0KAwEkVs6ZDMB2
 P16MieEEQ6StQRlvy2YBv80L1TMl3T90Bo1UUn6ARXEpcbFE0/aORH/jEXcRteb+vuik5UGY
 5TsyLYdPur3TXm7XDBdmmyQVJjnJKYK9AQxj95KlXLVO38lczsFNBFRjzmoBEACyAxbvUEhd
 GDGNg0JhDdezyTdN8C9BFsdxyTLnSH31NRiyp1QtuxvcqGZjb2trDVuCbIzRrgMZLVgo3upr
 MIOx1CXEgmn23Zhh0EpdVHM8IKx9Z7V0r+rrpRWFE8/wQZngKYVi49PGoZj50ZEifEJ5qn/H
 Nsp2+Y+bTUjDdgWMATg9DiFMyv8fvoqgNsNyrrZTnSgoLzdxr89FGHZCoSoAK8gfgFHuO54B
 lI8QOfPDG9WDPJ66HCodjTlBEr/Cwq6GruxS5i2Y33YVqxvFvDa1tUtl+iJ2SWKS9kCai2DR
 3BwVONJEYSDQaven/EHMlY1q8Vln3lGPsS11vSUK3QcNJjmrgYxH5KsVsf6PNRj9mp8Z1kIG
 qjRx08+nnyStWC0gZH6NrYyS9rpqH3j+hA2WcI7De51L4Rv9pFwzp161mvtc6eC/GxaiUGuH
 BNAVP0PY0fqvIC68p3rLIAW3f97uv4ce2RSQ7LbsPsimOeCo/5vgS6YQsj83E+AipPr09Caj
 0hloj+hFoqiticNpmsxdWKoOsV0PftcQvBCCYuhKbZV9s5hjt9qn8CE86A5g5KqDf83Fxqm/
 vXKgHNFHE5zgXGZnrmaf6resQzbvJHO0Fb0CcIohzrpPaL3YepcLDoCCgElGMGQjdCcSQ+Ci
 FCRl0Bvyj1YZUql+ZkptgGjikQARAQABwsFfBBgBAgAJBQJUY85qAhsMAAoJEGg1lTBwyZKw
 l4IQAIKHs/9po4spZDFyfDjunimEhVHqlUt7ggR1Hsl/tkvTSze8pI1P6dGp2XW6AnH1iayn
 yRcoyT0ZJ+Zmm4xAH1zqKjWplzqdb/dO28qk0bPso8+1oPO8oDhLm1+tY+cOvufXkBTm+whm
 +AyNTjaCRt6aSMnA/QHVGSJ8grrTJCoACVNhnXg/R0g90g8iV8Q+IBZyDkG0tBThaDdw1B2l
 asInUTeb9EiVfL/Zjdg5VWiF9LL7iS+9hTeVdR09vThQ/DhVbCNxVk+DtyBHsjOKifrVsYep
 WpRGBIAu3bK8eXtyvrw1igWTNs2wazJ71+0z2jMzbclKAyRHKU9JdN6Hkkgr2nPb561yjcB8
 sIq1pFXKyO+nKy6SZYxOvHxCcjk2fkw6UmPU6/j/nQlj2lfOAgNVKuDLothIxzi8pndB8Jju
 KktE5HJqUUMXePkAYIxEQ0mMc8Po7tuXdejgPMwgP7x65xtfEqI0RuzbUioFltsp1jUaRwQZ
 MTsCeQDdjpgHsj+P2ZDeEKCbma4m6Ez/YWs4+zDm1X8uZDkZcfQlD9NldbKDJEXLIjYWo1PH
 hYepSffIWPyvBMBTW2W5FRjJ4vLRrJSUoEfJuPQ3vW9Y73foyo/qFoURHO48AinGPZ7PC7TF
 vUaNOTjKedrqHkaOcqB185ahG2had0xnFsDPlx5y
In-Reply-To: <20231031195049.2075561-1-steve.wahl@hpe.com>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 10/31/23 12:50, Steve Wahl wrote:
> Instead of using gbpages for all memory regions, use them only when
> map creation requests include the full GB page of space; descend to
> using smaller 2M pages when only portions of a GB page are requested.
...

The logic here is sound: we obviously don't want systems rebooting
because speculation went wonky.

> diff --git a/arch/x86/mm/ident_map.c b/arch/x86/mm/ident_map.c
> index 968d7005f4a7..b63a1ffcfe9f 100644
> --- a/arch/x86/mm/ident_map.c
> +++ b/arch/x86/mm/ident_map.c
> @@ -31,18 +31,26 @@ static int ident_pud_init(struct x86_mapping_info *info, pud_t *pud_page,
>  		if (next > end)
>  			next = end;
>  
> -		if (info->direct_gbpages) {
> +		/*
> +		 * if gbpages allowed, this entry not yet present, and
> +		 * the full gbpage range is requested (both ends are
> +		 * correctly aligned), create a gbpage.
> +		 */
> +		if (info->direct_gbpages
> +		    && !pud_present(*pud)
> +		    && !(addr & ~PUD_MASK)
> +		    && !(next & ~PUD_MASK)) {

This is a _bit_ too subtle for my taste.

Let's say someone asks for mapping of 2MB@5G, then later asks for 1G@5G.
 The first call in here will do the 2MB mapping with small (pud)
entries.  The second call will see the new pud_present() check and
*also* skip large pud creation.

That's a regression.  It might not matter _much_, but it's a place where
the old code creates large PUDs and the new code creates small ones.
It's the kind of behavior change that at least needs to be noted in the
changelog.

Off the top of my head, I can't think of why we'd get overlapping
requests in here, though.  Did you think through that case?  Is it common?

>  			pud_t pudval;
>  
> -			if (pud_present(*pud))
> -				continue;
> -
> -			addr &= PUD_MASK;
>  			pudval = __pud((addr - info->offset) | info->page_flag);
>  			set_pud(pud, pudval);
>  			continue;
>  		}
>  
> +		/* if this is already a gbpage, this portion is already mapped */
> +		if (pud_large(*pud))
> +			continue;

I'd probably move this check up to above the large PUD creation code.
It would make it more obvious that any PUD that's encountered down below
is a small PUD.

>  		if (pud_present(*pud)) {
>  			pmd = pmd_offset(pud, 0);
>  			ident_pmd_init(info, pmd, addr, next);

That would end up looking something like this:

	bool do_gbpages = true;
	...

	// Is the whole range already mapped?
	if (pud_large(*pud))
		continue;

	/* PUD is either empty or small */

	// GB pages allowed?
	do_gbpages &= info->direct_gbpages;
	// Addresses aligned for GB pages?
	do_gbpages &= ~(addr & ~PUD_MASK);
	do_gbpages &= ~(next & ~PUD_MASK);
	// Check for existing mapping using a small PUD
	do_gbpages &= !pud_present(*pud);

	if (do_gbpages) {
		set_pud(pud, __pud((addr - info->offset) |
					info->page_flag));
		continue
	}