From: "Roger Pau Monné" <roger.pau@citrix.com>
To: Jan Beulich <jbeulich@suse.com>
Cc: "xen-devel@lists.xenproject.org" <xen-devel@lists.xenproject.org>,
Andrew Cooper <andrew.cooper3@citrix.com>, Wei Liu <wl@xen.org>
Subject: Re: [PATCH v2 1/6] x86: support cache-writeback in flush_area_local() et al
Date: Tue, 13 May 2025 15:01:54 +0200 [thread overview]
Message-ID: <aCNCwkGbhj-uy-vD@macbook.lan> (raw)
In-Reply-To: <e27ff909-93d1-b51b-ac88-20b17f5cf642@suse.com>
On Wed, May 03, 2023 at 11:44:39AM +0200, Jan Beulich wrote:
> The majority of the present callers really aren't after invalidating
> cache contents, but only after writeback. Make this available by simply
> extending the FLUSH_CACHE handling accordingly. No feature checks are
> required here: cache_writeback() falls back to cache_flush() as
> necessary, while WBNOINVD degenerates to WBINVD on older hardware.
>
> Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Roger Pau Monné <roger.pau@citrix.com>
> ---
> v2: FLUSH_WRITEBACK -> FLUSH_CACHE_WRITEBACK.
>
> --- a/xen/arch/x86/flushtlb.c
> +++ b/xen/arch/x86/flushtlb.c
> @@ -232,7 +232,7 @@ unsigned int flush_area_local(const void
> if ( flags & FLUSH_HVM_ASID_CORE )
> hvm_flush_guest_tlbs();
>
> - if ( flags & FLUSH_CACHE )
> + if ( flags & (FLUSH_CACHE | FLUSH_CACHE_WRITEBACK) )
> {
> const struct cpuinfo_x86 *c = ¤t_cpu_data;
> unsigned long sz = 0;
> @@ -245,13 +245,16 @@ unsigned int flush_area_local(const void
> c->x86_clflush_size && c->x86_cache_size && sz &&
> ((sz >> 10) < c->x86_cache_size) )
> {
> - cache_flush(va, sz);
> - flags &= ~FLUSH_CACHE;
> + if ( flags & FLUSH_CACHE )
> + cache_flush(va, sz);
> + else
> + cache_writeback(va, sz);
> + flags &= ~(FLUSH_CACHE | FLUSH_CACHE_WRITEBACK);
> }
> - else
> - {
> + else if ( flags & FLUSH_CACHE )
> wbinvd();
> - }
> + else
> + wbnoinvd();
> }
>
> if ( flags & FLUSH_ROOT_PGTBL )
> --- a/xen/arch/x86/include/asm/flushtlb.h
> +++ b/xen/arch/x86/include/asm/flushtlb.h
> @@ -135,6 +135,8 @@ void switch_cr3_cr4(unsigned long cr3, u
> #else
> # define FLUSH_NO_ASSIST 0
> #endif
> + /* Write back data cache contents */
> +#define FLUSH_CACHE_WRITEBACK 0x10000
>
> /* Flush local TLBs/caches. */
> unsigned int flush_area_local(const void *va, unsigned int flags);
> @@ -194,7 +196,11 @@ static inline int clean_and_invalidate_d
> }
> static inline int clean_dcache_va_range(const void *p, unsigned long size)
> {
> - return clean_and_invalidate_dcache_va_range(p, size);
> + unsigned int order = get_order_from_bytes(size);
> +
> + /* sub-page granularity support needs to be added if necessary */
> + flush_area_local(p, FLUSH_CACHE_WRITEBACK | FLUSH_ORDER(order));
> + return 0;
> }
I'm planning to get rid of the clean_dcache_va_range() helper on x86,
but I don't want to force you to rebase on top of that.
Thanks, Roger.
next prev parent reply other threads:[~2025-05-13 13:02 UTC|newest]
Thread overview: 29+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-05-03 9:43 [PATCH v2 0/6] x86: reduce cache flushing overhead Jan Beulich
2023-05-03 9:44 ` [PATCH v2 1/6] x86: support cache-writeback in flush_area_local() et al Jan Beulich
2025-05-13 13:01 ` Roger Pau Monné [this message]
2023-05-03 9:45 ` [PATCH v2 2/6] x86/HVM: restrict guest-induced WBINVD to cache writeback Jan Beulich
2025-05-13 13:12 ` Andrew Cooper
2025-05-13 13:24 ` Jan Beulich
2025-05-13 13:41 ` Roger Pau Monné
2025-05-13 13:54 ` Jan Beulich
2025-05-14 11:30 ` Roger Pau Monné
2025-05-14 12:46 ` Jan Beulich
2025-05-14 14:42 ` Roger Pau Monné
2025-05-14 14:53 ` Jan Beulich
2023-05-03 9:45 ` [PATCH v2 3/6] x86/PV: restrict guest-induced WBINVD (or alike) " Jan Beulich
2023-05-03 9:46 ` [PATCH v2 4/6] VT-d: restrict iommu_flush_all() " Jan Beulich
2025-05-14 11:40 ` Roger Pau Monné
2025-05-14 12:49 ` Jan Beulich
2025-05-14 14:44 ` Roger Pau Monné
2025-05-14 14:57 ` Jan Beulich
2023-05-03 9:46 ` [PATCH v2 5/6] x86: FLUSH_CACHE -> FLUSH_CACHE_EVICT Jan Beulich
2025-05-14 11:44 ` Roger Pau Monné
2025-05-14 12:52 ` Jan Beulich
2025-05-14 14:45 ` Roger Pau Monné
2025-05-15 6:46 ` Jan Beulich
2023-05-03 9:47 ` [PATCH v2 6/6] x86/HVM: limit cache writeback overhead Jan Beulich
2025-05-14 13:00 ` Roger Pau Monné
2025-05-14 13:20 ` Jan Beulich
2025-05-14 15:12 ` Roger Pau Monné
2025-05-15 6:47 ` Jan Beulich
2025-05-15 8:09 ` Roger Pau Monné
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=aCNCwkGbhj-uy-vD@macbook.lan \
--to=roger.pau@citrix.com \
--cc=andrew.cooper3@citrix.com \
--cc=jbeulich@suse.com \
--cc=wl@xen.org \
--cc=xen-devel@lists.xenproject.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.