All of lore.kernel.org
 help / color / mirror / Atom feed
From: Ingo Molnar <mingo@kernel.org>
To: linux-kernel@vger.kernel.org
Cc: Juergen Gross <jgross@suse.com>,
	"H . Peter Anvin" <hpa@zytor.com>,
	Linus Torvalds <torvalds@linux-foundation.org>,
	Peter Zijlstra <peterz@infradead.org>,
	Borislav Petkov <bp@alien8.de>,
	Thomas Gleixner <tglx@linutronix.de>
Subject: [PATCH 12/41] x86/alternatives: Remove the confusing, inaccurate & unnecessary 'temp_mm_state_t' abstraction
Date: Thu, 27 Mar 2025 21:53:25 +0100	[thread overview]
Message-ID: <20250327205355.378659-13-mingo@kernel.org> (raw)
In-Reply-To: <20250327205355.378659-1-mingo@kernel.org>

So the temp_mm_state_t abstraction used by use_temporary_mm() and
unuse_temporary_mm() is super confusing:

 - The whole machinery is about temporarily switching to the
   text_poke_mm utility MM that got allocated during bootup
   for text-patching purposes alone:

	temp_mm_state_t prev;

        /*
         * Loading the temporary mm behaves as a compiler barrier, which
         * guarantees that the PTE will be set at the time memcpy() is done.
         */
        prev = use_temporary_mm(text_poke_mm);

 - Yet the value that gets saved in the temp_mm_state_t variable
   is not the temporary MM ... but the previous MM...

 - Ie. we temporarily put the non-temporary MM into a variable
   that has the temp_mm_state_t type. This makes no sense whatsoever.

 - The confusion continues in unuse_temporary_mm():

	static inline void unuse_temporary_mm(temp_mm_state_t prev_state)

   Here we unuse an MM that is ... not the temporary MM, but the
   previous MM. :-/

Fix up all this confusion by removing the unnecessary layer of
abstraction and using a bog-standard 'struct mm_struct *prev_mm'
variable to save the MM to.

Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/kernel/alternative.c | 24 ++++++++++--------------
 1 file changed, 10 insertions(+), 14 deletions(-)

diff --git a/arch/x86/kernel/alternative.c b/arch/x86/kernel/alternative.c
index 7d14c8abd3aa..557ee2546177 100644
--- a/arch/x86/kernel/alternative.c
+++ b/arch/x86/kernel/alternative.c
@@ -2139,10 +2139,6 @@ void __init_or_module text_poke_early(void *addr, const void *opcode,
 	}
 }
 
-typedef struct {
-	struct mm_struct *mm;
-} temp_mm_state_t;
-
 /*
  * Using a temporary mm allows to set temporary mappings that are not accessible
  * by other CPUs. Such mappings are needed to perform sensitive memory writes
@@ -2156,9 +2152,9 @@ typedef struct {
  *          loaded, thereby preventing interrupt handler bugs from overriding
  *          the kernel memory protection.
  */
-static inline temp_mm_state_t use_temporary_mm(struct mm_struct *mm)
+static inline struct mm_struct *use_temporary_mm(struct mm_struct *temp_mm)
 {
-	temp_mm_state_t temp_state;
+	struct mm_struct *prev_mm;
 
 	lockdep_assert_irqs_disabled();
 
@@ -2170,8 +2166,8 @@ static inline temp_mm_state_t use_temporary_mm(struct mm_struct *mm)
 	if (this_cpu_read(cpu_tlbstate_shared.is_lazy))
 		leave_mm();
 
-	temp_state.mm = this_cpu_read(cpu_tlbstate.loaded_mm);
-	switch_mm_irqs_off(NULL, mm, current);
+	prev_mm = this_cpu_read(cpu_tlbstate.loaded_mm);
+	switch_mm_irqs_off(NULL, temp_mm, current);
 
 	/*
 	 * If breakpoints are enabled, disable them while the temporary mm is
@@ -2187,17 +2183,17 @@ static inline temp_mm_state_t use_temporary_mm(struct mm_struct *mm)
 	if (hw_breakpoint_active())
 		hw_breakpoint_disable();
 
-	return temp_state;
+	return prev_mm;
 }
 
 __ro_after_init struct mm_struct *text_poke_mm;
 __ro_after_init unsigned long text_poke_addr;
 
-static inline void unuse_temporary_mm(temp_mm_state_t prev_state)
+static inline void unuse_temporary_mm(struct mm_struct *prev_mm)
 {
 	lockdep_assert_irqs_disabled();
 
-	switch_mm_irqs_off(NULL, prev_state.mm, current);
+	switch_mm_irqs_off(NULL, prev_mm, current);
 
 	/* Clear the cpumask, to indicate no TLB flushing is needed anywhere */
 	cpumask_clear_cpu(raw_smp_processor_id(), mm_cpumask(text_poke_mm));
@@ -2228,7 +2224,7 @@ static void *__text_poke(text_poke_f func, void *addr, const void *src, size_t l
 {
 	bool cross_page_boundary = offset_in_page(addr) + len > PAGE_SIZE;
 	struct page *pages[2] = {NULL};
-	temp_mm_state_t prev;
+	struct mm_struct *prev_mm;
 	unsigned long flags;
 	pte_t pte, *ptep;
 	spinlock_t *ptl;
@@ -2286,7 +2282,7 @@ static void *__text_poke(text_poke_f func, void *addr, const void *src, size_t l
 	 * Loading the temporary mm behaves as a compiler barrier, which
 	 * guarantees that the PTE will be set at the time memcpy() is done.
 	 */
-	prev = use_temporary_mm(text_poke_mm);
+	prev_mm = use_temporary_mm(text_poke_mm);
 
 	kasan_disable_current();
 	func((u8 *)text_poke_addr + offset_in_page(addr), src, len);
@@ -2307,7 +2303,7 @@ static void *__text_poke(text_poke_f func, void *addr, const void *src, size_t l
 	 * instruction that already allows the core to see the updated version.
 	 * Xen-PV is assumed to serialize execution in a similar manner.
 	 */
-	unuse_temporary_mm(prev);
+	unuse_temporary_mm(prev_mm);
 
 	/*
 	 * Flushing the TLB might involve IPIs, which would require enabled
-- 
2.45.2


  parent reply	other threads:[~2025-03-27 20:54 UTC|newest]

Thread overview: 47+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-03-27 20:53 [PATCH 00/41] Simplify, reorganize and clean up the x86 INT3 based batch-patching code (alternative.c) Ingo Molnar
2025-03-27 20:53 ` [PATCH 01/41] x86/alternatives: Rename 'struct bp_patching_desc' to 'struct int3_patching_desc' Ingo Molnar
2025-03-27 20:53 ` [PATCH 02/41] x86/alternatives: Rename 'bp_refs' to 'int3_refs' Ingo Molnar
2025-03-27 20:53 ` [PATCH 03/41] x86/alternatives: Rename 'text_poke_bp_batch()' to 'text_poke_int3_batch()' Ingo Molnar
2025-03-27 20:53 ` [PATCH 04/41] x86/alternatives: Rename 'text_poke_bp()' to 'text_poke_int3()' Ingo Molnar
2025-03-27 20:53 ` [PATCH 05/41] x86/alternatives: Rename 'poke_int3_handler()' to 'text_poke_int3_handler()' Ingo Molnar
2025-03-27 20:53 ` [PATCH 06/41] x86/alternatives: Rename 'poking_mm' to 'text_poke_mm' Ingo Molnar
2025-03-27 20:53 ` [PATCH 07/41] x86/alternatives: Rename 'text_poke_addr' to 'text_poke_int3_addr' Ingo Molnar
2025-03-27 20:53 ` [PATCH 08/41] x86/alternatives: Rename 'poking_addr' to 'text_poke_addr' Ingo Molnar
2025-03-27 20:53 ` [PATCH 09/41] x86/alternatives: Rename 'bp_desc' to 'int3_desc' Ingo Molnar
2025-03-27 20:53 ` [PATCH 10/41] x86/alternatives: Remove duplicate 'text_poke_early()' prototype Ingo Molnar
2025-03-27 20:53 ` [PATCH 11/41] x86/alternatives: Update comments in int3_emulate_push() Ingo Molnar
2025-03-27 20:53 ` Ingo Molnar [this message]
2025-03-27 20:53 ` [PATCH 13/41] x86/alternatives: Rename 'text_poke_flush()' to 'text_poke_int3_flush()' Ingo Molnar
2025-03-27 20:53 ` [PATCH 14/41] x86/alternatives: Rename 'text_poke_finish()' to 'text_poke_int3_finish()' Ingo Molnar
2025-03-27 20:53 ` [PATCH 15/41] x86/alternatives: Rename 'text_poke_queue()' to 'text_poke_int3_queue()' Ingo Molnar
2025-03-27 20:53 ` [PATCH 16/41] x86/alternatives: Rename 'text_poke_loc_init()' to 'text_poke_int3_loc_init()' Ingo Molnar
2025-03-27 20:53 ` [PATCH 17/41] x86/alternatives: Rename 'struct text_poke_loc' to 'struct text_poke_int3_loc' Ingo Molnar
2025-03-27 20:53 ` [PATCH 18/41] x86/alternatives: Rename 'struct int3_patching_desc' to 'struct text_poke_int3_vec' Ingo Molnar
2025-03-27 20:53 ` [PATCH 19/41] x86/alternatives: Rename 'int3_desc' to 'int3_vec' Ingo Molnar
2025-03-27 20:53 ` [PATCH 20/41] x86/alternatives: Add text_mutex) assert to text_poke_int3_flush() Ingo Molnar
2025-03-27 20:53 ` [PATCH 21/41] x86/alternatives: Assert that text_poke_int3_handler() can only ever handle 'tp_vec[]' based requests Ingo Molnar
2025-03-27 20:53 ` [PATCH 22/41] x86/alternatives: Use non-inverted logic instead of 'tp_order_fail()' Ingo Molnar
2025-03-27 20:53 ` [PATCH 23/41] x86/alternatives: Remove the 'addr == NULL means forced-flush' hack from text_poke_int3_finish()/text_poke_int3_flush()/tp_addr_ordered() Ingo Molnar
2025-03-27 20:53 ` [PATCH 24/41] x86/alternatives: Simplify text_poke_int3() by using tp_vec and existing APIs Ingo Molnar
2025-03-27 20:53 ` [PATCH 25/41] x86/alternatives: Assert input parameters in text_poke_int3_batch() Ingo Molnar
2025-03-27 20:53 ` [PATCH 26/41] x86/alternatives: Introduce 'struct text_poke_int3_array' and move tp_vec and tp_vec_nr to it Ingo Molnar
2025-03-27 20:53 ` [PATCH 27/41] x86/alternatives: Remove the tp_vec indirection Ingo Molnar
2025-03-27 20:53 ` [PATCH 28/41] x86/alternatives: Rename 'try_get_desc()' to 'try_get_tp_array()' Ingo Molnar
2025-03-27 20:53 ` [PATCH 29/41] x86/alternatives: Rename 'put_desc()' to 'put_tp_array()' Ingo Molnar
2025-03-27 20:53 ` [PATCH 30/41] x86/alternatives: Simplify try_get_tp_array() Ingo Molnar
2025-03-27 20:53 ` [PATCH 31/41] x86/alternatives: Simplify text_poke_int3_handler() Ingo Molnar
2025-03-27 20:53 ` [PATCH 32/41] x86/alternatives: Simplify text_poke_int3_batch() Ingo Molnar
2025-03-27 20:53 ` [PATCH 33/41] x86/alternatives: Rename 'text_poke_int3_batch()' to 'text_poke_int3_batch_process()' Ingo Molnar
2025-03-27 20:53 ` [PATCH 34/41] x86/alternatives: Rename 'int3_refs' to 'tp_array_refs' Ingo Molnar
2025-03-27 20:53 ` [PATCH 35/41] x86/alternatives: Move the tp_array manipulation into text_poke_int3_loc_init() and rename it to text_poke_int3_loc_add() Ingo Molnar
2025-03-27 20:53 ` [PATCH 36/41] x86/alternatives: Remove the mixed-patching restriction on text_poke_int3() Ingo Molnar
2025-03-27 20:53 ` [PATCH 37/41] x86/alternatives: Rename 'text_poke_int3()' to 'text_poke_int3_now()' Ingo Molnar
2025-03-27 20:53 ` [PATCH 38/41] x86/alternatives: Add documentation for text_poke_int3_queue() Ingo Molnar
2025-03-27 20:53 ` [PATCH 39/41] x86/alternatives: Move tp_array completion from text_poke_int3_finish() and text_poke_int3_flush() to text_poke_int3_batch_process() Ingo Molnar
2025-03-27 20:53 ` [PATCH 40/41] x86/alternatives: Rename 'text_poke_sync()' to 'text_poke_sync_each_cpu()' Ingo Molnar
2025-04-02  4:10   ` H. Peter Anvin
2025-04-03 15:05     ` Ingo Molnar
2025-03-27 20:53 ` [PATCH 41/41] x86/alternatives: Simplify tp_addr_ordered() Ingo Molnar
2025-03-27 22:19 ` [PATCH 00/41] Simplify, reorganize and clean up the x86 INT3 based batch-patching code (alternative.c) Linus Torvalds
2025-03-28 10:10   ` Ingo Molnar
2025-04-01 14:55   ` Peter Zijlstra

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=20250327205355.378659-13-mingo@kernel.org \
    --to=mingo@kernel.org \
    --cc=bp@alien8.de \
    --cc=hpa@zytor.com \
    --cc=jgross@suse.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=peterz@infradead.org \
    --cc=tglx@linutronix.de \
    --cc=torvalds@linux-foundation.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.