All of lore.kernel.org
 help / color / mirror / Atom feed
From: Ard Biesheuvel <ard.biesheuvel@linaro.org>
To: linux-kernel@vger.kernel.org,
	linux-arm-kernel@lists.infradead.org, mpe@ellerman.id.au,
	jeyu@redhat.com
Cc: will.deacon@arm.com, rusty@rustcorp.com.au,
	akpm@linux-foundation.org, benh@kernel.crashing.org,
	paulus@samba.org, Ard Biesheuvel <ard.biesheuvel@linaro.org>
Subject: [PATCH v3 1/3] powerpc/reloc32: fix corrupted modversion CRCs
Date: Thu, 27 Oct 2016 17:27:09 +0100	[thread overview]
Message-ID: <1477585631-18574-2-git-send-email-ard.biesheuvel@linaro.org> (raw)
In-Reply-To: <1477585631-18574-1-git-send-email-ard.biesheuvel@linaro.org>

Commit 0e0ed6406e61 ("powerpc/modules: Module CRC relocation fix causes
perf issues") fixed an issue with relocatable PIE kernels in a way that
essentially reintroduced the issue again for 32-bit builds.

Since the chosen approach does is not applicable to 32-bit, fix the
issue by updating the runtime relocation routine to ignore the load
offset for the interval [__start___kcrctab, __stop___kcrctab_gpl_future),
which is where the CRCs reside. This ensures that the values of the CRC
pseudo-symbols are no longer made dependent on the runtime load offset.

Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
---
 arch/powerpc/kernel/reloc_32.S | 36 +++++++++++++++++---
 1 file changed, 32 insertions(+), 4 deletions(-)

diff --git a/arch/powerpc/kernel/reloc_32.S b/arch/powerpc/kernel/reloc_32.S
index f366fedb0872..150686b9febb 100644
--- a/arch/powerpc/kernel/reloc_32.S
+++ b/arch/powerpc/kernel/reloc_32.S
@@ -87,12 +87,12 @@ eodyn:				/* End of Dyn Table scan */
 	 * Work out the current offset from the link time address of .rela
 	 * section.
 	 *  cur_offset[r7] = rela.run[r9] - rela.link [r7]
-	 *  _stext.link[r12] = _stext.run[r10] - cur_offset[r7]
-	 *  final_offset[r3] = _stext.final[r3] - _stext.link[r12]
+	 *  _stext.link[r11] = _stext.run[r10] - cur_offset[r7]
+	 *  final_offset[r3] = _stext.final[r3] - _stext.link[r11]
 	 */
 	subf	r7, r7, r9	/* cur_offset */
-	subf	r12, r7, r10
-	subf	r3, r12, r3	/* final_offset */
+	subf	r11, r7, r10
+	subf	r3, r11, r3	/* final_offset */
 
 	subf	r8, r6, r8	/* relaz -= relaent */
 	/*
@@ -101,6 +101,21 @@ eodyn:				/* End of Dyn Table scan */
 	 * r13	- points to the symbol table
 	 */
 
+#ifdef CONFIG_MODVERSIONS
+	/*
+	 * Treat R_PPC_RELATIVE relocations differently when they target the
+	 * interval [__start___kcrctab, __stop___kcrctab_gpl_future): in this
+	 * case, the relocated quantities are CRC pseudo-symbols, which should
+	 * be preserved as-is, rather than be modified to take the runtime
+	 * offset into account.
+	 */
+	lwz	r10, (p_kcrc_start - 0b)(r12)
+	lwz	r11, (p_kcrc_stop - 0b)(r12)
+	subf	r12, r7, r12			/* link time addr of 0b */
+	add	r10, r10, r12
+	add	r11, r11, r12
+#endif
+
 	/*
 	 * Check if we have a relocation based on symbol
 	 * r5 will hold the value of the symbol.
@@ -135,7 +150,15 @@ get_type:
 	bne	hi16
 	lwz	r4, 0(r9)	/* r_offset */
 	lwz	r0, 8(r9)	/* r_addend */
+#ifdef CONFIG_MODVERSIONS
+	cmplw	r4, r10
+	blt	do_add
+	cmplw	r4, r11
+	blt	skip_add
+do_add:
+#endif
 	add	r0, r0, r3	/* final addend */
+skip_add:
 	stwx	r0, r4, r7	/* memory[r4+r7]) = (u32)r0 */
 	b	nxtrela		/* continue */
 
@@ -207,3 +230,8 @@ p_dyn:		.long	__dynamic_start - 0b
 p_rela:		.long	__rela_dyn_start - 0b
 p_sym:		.long	__dynamic_symtab - 0b
 p_st:		.long	_stext - 0b
+
+#ifdef CONFIG_MODVERSIONS
+p_kcrc_start:	.long	__start___kcrctab - 0b
+p_kcrc_stop:	.long	__stop___kcrctab_gpl_future - 0b
+#endif
-- 
2.7.4

WARNING: multiple messages have this Message-ID (diff)
From: ard.biesheuvel@linaro.org (Ard Biesheuvel)
To: linux-arm-kernel@lists.infradead.org
Subject: [PATCH v3 1/3] powerpc/reloc32: fix corrupted modversion CRCs
Date: Thu, 27 Oct 2016 17:27:09 +0100	[thread overview]
Message-ID: <1477585631-18574-2-git-send-email-ard.biesheuvel@linaro.org> (raw)
In-Reply-To: <1477585631-18574-1-git-send-email-ard.biesheuvel@linaro.org>

Commit 0e0ed6406e61 ("powerpc/modules: Module CRC relocation fix causes
perf issues") fixed an issue with relocatable PIE kernels in a way that
essentially reintroduced the issue again for 32-bit builds.

Since the chosen approach does is not applicable to 32-bit, fix the
issue by updating the runtime relocation routine to ignore the load
offset for the interval [__start___kcrctab, __stop___kcrctab_gpl_future),
which is where the CRCs reside. This ensures that the values of the CRC
pseudo-symbols are no longer made dependent on the runtime load offset.

Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
---
 arch/powerpc/kernel/reloc_32.S | 36 +++++++++++++++++---
 1 file changed, 32 insertions(+), 4 deletions(-)

diff --git a/arch/powerpc/kernel/reloc_32.S b/arch/powerpc/kernel/reloc_32.S
index f366fedb0872..150686b9febb 100644
--- a/arch/powerpc/kernel/reloc_32.S
+++ b/arch/powerpc/kernel/reloc_32.S
@@ -87,12 +87,12 @@ eodyn:				/* End of Dyn Table scan */
 	 * Work out the current offset from the link time address of .rela
 	 * section.
 	 *  cur_offset[r7] = rela.run[r9] - rela.link [r7]
-	 *  _stext.link[r12] = _stext.run[r10] - cur_offset[r7]
-	 *  final_offset[r3] = _stext.final[r3] - _stext.link[r12]
+	 *  _stext.link[r11] = _stext.run[r10] - cur_offset[r7]
+	 *  final_offset[r3] = _stext.final[r3] - _stext.link[r11]
 	 */
 	subf	r7, r7, r9	/* cur_offset */
-	subf	r12, r7, r10
-	subf	r3, r12, r3	/* final_offset */
+	subf	r11, r7, r10
+	subf	r3, r11, r3	/* final_offset */
 
 	subf	r8, r6, r8	/* relaz -= relaent */
 	/*
@@ -101,6 +101,21 @@ eodyn:				/* End of Dyn Table scan */
 	 * r13	- points to the symbol table
 	 */
 
+#ifdef CONFIG_MODVERSIONS
+	/*
+	 * Treat R_PPC_RELATIVE relocations differently when they target the
+	 * interval [__start___kcrctab, __stop___kcrctab_gpl_future): in this
+	 * case, the relocated quantities are CRC pseudo-symbols, which should
+	 * be preserved as-is, rather than be modified to take the runtime
+	 * offset into account.
+	 */
+	lwz	r10, (p_kcrc_start - 0b)(r12)
+	lwz	r11, (p_kcrc_stop - 0b)(r12)
+	subf	r12, r7, r12			/* link time addr of 0b */
+	add	r10, r10, r12
+	add	r11, r11, r12
+#endif
+
 	/*
 	 * Check if we have a relocation based on symbol
 	 * r5 will hold the value of the symbol.
@@ -135,7 +150,15 @@ get_type:
 	bne	hi16
 	lwz	r4, 0(r9)	/* r_offset */
 	lwz	r0, 8(r9)	/* r_addend */
+#ifdef CONFIG_MODVERSIONS
+	cmplw	r4, r10
+	blt	do_add
+	cmplw	r4, r11
+	blt	skip_add
+do_add:
+#endif
 	add	r0, r0, r3	/* final addend */
+skip_add:
 	stwx	r0, r4, r7	/* memory[r4+r7]) = (u32)r0 */
 	b	nxtrela		/* continue */
 
@@ -207,3 +230,8 @@ p_dyn:		.long	__dynamic_start - 0b
 p_rela:		.long	__rela_dyn_start - 0b
 p_sym:		.long	__dynamic_symtab - 0b
 p_st:		.long	_stext - 0b
+
+#ifdef CONFIG_MODVERSIONS
+p_kcrc_start:	.long	__start___kcrctab - 0b
+p_kcrc_stop:	.long	__stop___kcrctab_gpl_future - 0b
+#endif
-- 
2.7.4

  reply	other threads:[~2016-10-27 16:27 UTC|newest]

Thread overview: 34+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-10-27 16:27 [PATCH v3 0/3] modversions: Fix CRC mangling under CONFIG_RELOCATABLE=y Ard Biesheuvel
2016-10-27 16:27 ` Ard Biesheuvel
2016-10-27 16:27 ` Ard Biesheuvel [this message]
2016-10-27 16:27   ` [PATCH v3 1/3] powerpc/reloc32: fix corrupted modversion CRCs Ard Biesheuvel
2016-10-28 10:27   ` Suzuki K Poulose
2016-10-28 10:27     ` Suzuki K Poulose
2016-10-27 16:27 ` [PATCH v3 2/3] powerpc/reloc64: add support for 32-bit CRC pseudo-symbols Ard Biesheuvel
2016-10-27 16:27   ` Ard Biesheuvel
2016-11-25 11:29   ` Michael Ellerman
2016-11-25 11:29     ` Michael Ellerman
2016-11-25 12:48     ` Ard Biesheuvel
2016-11-25 12:48       ` Ard Biesheuvel
2016-12-01  9:39       ` Michael Ellerman
2016-12-01  9:39         ` Michael Ellerman
2016-12-01  9:45         ` Ard Biesheuvel
2016-12-01  9:45           ` Ard Biesheuvel
2016-12-01 16:28           ` Ard Biesheuvel
2016-12-01 16:28             ` Ard Biesheuvel
2016-10-27 16:27 ` [PATCH v3 3/3] modversions: treat symbol CRCs as 32 bit quantities on 64 bit archs Ard Biesheuvel
2016-10-27 16:27   ` Ard Biesheuvel
2016-11-04  9:55 ` [PATCH v3 0/3] modversions: Fix CRC mangling under CONFIG_RELOCATABLE=y Ard Biesheuvel
2016-11-04  9:55   ` Ard Biesheuvel
2016-11-10  4:22   ` Michael Ellerman
2016-11-10  4:22     ` Michael Ellerman
2016-11-15  9:13     ` Ard Biesheuvel
2016-11-15  9:13       ` Ard Biesheuvel
2016-11-25  8:44       ` Ard Biesheuvel
2016-11-25  8:44         ` Ard Biesheuvel
2016-11-25 11:12         ` Michael Ellerman
2016-11-25 11:12           ` Michael Ellerman
2016-11-16 19:23 ` Uwe Kleine-König
2016-11-16 19:23   ` Uwe Kleine-König
2016-11-16 20:29   ` Ard Biesheuvel
2016-11-16 20:29     ` Ard Biesheuvel

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=1477585631-18574-2-git-send-email-ard.biesheuvel@linaro.org \
    --to=ard.biesheuvel@linaro.org \
    --cc=akpm@linux-foundation.org \
    --cc=benh@kernel.crashing.org \
    --cc=jeyu@redhat.com \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mpe@ellerman.id.au \
    --cc=paulus@samba.org \
    --cc=rusty@rustcorp.com.au \
    --cc=will.deacon@arm.com \
    /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.