- Linux-stable-mirror - lists.linaro.org

request for 4.14-stable: 93b3623ada15 ("gcc-plugins: Use dynamic initializers")

by Lance Albertson

Hi Greg, I was trying to compile 4.14.67 using GCC 8.2.0 with gcc-plugins enabled and ran into this issue [1]. If I applied both of the commits mentioned, it resolved the problem so I think it should be included in 4.14. Thanks! [1] https://groups.google.com/forum/#!msg/qubes-devel/Q3cdQKQS4Tk/gKr4F52HAAAJ -- Lance Albertson Director Oregon State University | Open Source Lab --- From 93b3623ada1582379833ea442fbc7b15e6417fbc Mon Sep 17 00:00:00 2001 From: Kees Cook <keescook(a)chromium.org> Date: Mon, 5 Feb 2018 17:27:46 -0800 Subject: [PATCH 1/2] gcc-plugins: Use dynamic initializers GCC 8 changed the order of some fields and is very picky about ordering in static initializers, so instead just move to dynamic initializers, and drop the redundant already-zero field assignments. Suggested-by: Valdis Kletnieks <valdis.kletnieks(a)vt.edu> Signed-off-by: Kees Cook <keescook(a)chromium.org> --- scripts/gcc-plugins/latent_entropy_plugin.c | 17 ++---- scripts/gcc-plugins/randomize_layout_plugin.c | 75 ++++++++------------------- scripts/gcc-plugins/structleak_plugin.c | 19 +++---- 3 files changed, 33 insertions(+), 78 deletions(-) diff --git a/scripts/gcc-plugins/latent_entropy_plugin.c b/scripts/gcc-plugins/latent_entropy_plugin.c index 65264960910d..cbe1d6c4b1a5 100644 --- a/scripts/gcc-plugins/latent_entropy_plugin.c +++ b/scripts/gcc-plugins/latent_entropy_plugin.c @@ -255,21 +255,14 @@ static tree handle_latent_entropy_attribute(tree *node, tree name, return NULL_TREE; } -static struct attribute_spec latent_entropy_attr = { - .name = "latent_entropy", - .min_length = 0, - .max_length = 0, - .decl_required = true, - .type_required = false, - .function_type_required = false, - .handler = handle_latent_entropy_attribute, -#if BUILDING_GCC_VERSION >= 4007 - .affects_type_identity = false -#endif -}; +static struct attribute_spec latent_entropy_attr = { }; static void register_attributes(void *event_data __unused, void *data __unused) { + latent_entropy_attr.name = "latent_entropy"; + latent_entropy_attr.decl_required = true; + latent_entropy_attr.handler = handle_latent_entropy_attribute; + register_attribute(&latent_entropy_attr); } diff --git a/scripts/gcc-plugins/randomize_layout_plugin.c b/scripts/gcc-plugins/randomize_layout_plugin.c index 0073af326449..c4a345c3715b 100644 --- a/scripts/gcc-plugins/randomize_layout_plugin.c +++ b/scripts/gcc-plugins/randomize_layout_plugin.c @@ -580,68 +580,35 @@ static void finish_type(void *event_data, void *data) return; } -static struct attribute_spec randomize_layout_attr = { - .name = "randomize_layout", - // related to args - .min_length = 0, - .max_length = 0, - .decl_required = false, - // need type declaration - .type_required = true, - .function_type_required = false, - .handler = handle_randomize_layout_attr, -#if BUILDING_GCC_VERSION >= 4007 - .affects_type_identity = true -#endif -}; +static struct attribute_spec randomize_layout_attr = { }; +static struct attribute_spec no_randomize_layout_attr = { }; +static struct attribute_spec randomize_considered_attr = { }; +static struct attribute_spec randomize_performed_attr = { }; -static struct attribute_spec no_randomize_layout_attr = { - .name = "no_randomize_layout", - // related to args - .min_length = 0, - .max_length = 0, - .decl_required = false, - // need type declaration - .type_required = true, - .function_type_required = false, - .handler = handle_randomize_layout_attr, +static void register_attributes(void *event_data, void *data) +{ + randomize_layout_attr.name = "randomize_layout"; + randomize_layout_attr.type_required = true; + randomize_layout_attr.handler = handle_randomize_layout_attr; #if BUILDING_GCC_VERSION >= 4007 - .affects_type_identity = true + randomize_layout_attr.affects_type_identity = true; #endif -}; -static struct attribute_spec randomize_considered_attr = { - .name = "randomize_considered", - // related to args - .min_length = 0, - .max_length = 0, - .decl_required = false, - // need type declaration - .type_required = true, - .function_type_required = false, - .handler = handle_randomize_considered_attr, + no_randomize_layout_attr.name = "no_randomize_layout"; + no_randomize_layout_attr.type_required = true; + no_randomize_layout_attr.handler = handle_randomize_layout_attr; #if BUILDING_GCC_VERSION >= 4007 - .affects_type_identity = false + no_randomize_layout_attr.affects_type_identity = true; #endif -}; -static struct attribute_spec randomize_performed_attr = { - .name = "randomize_performed", - // related to args - .min_length = 0, - .max_length = 0, - .decl_required = false, - // need type declaration - .type_required = true, - .function_type_required = false, - .handler = handle_randomize_performed_attr, -#if BUILDING_GCC_VERSION >= 4007 - .affects_type_identity = false -#endif -}; + randomize_considered_attr.name = "randomize_considered"; + randomize_considered_attr.type_required = true; + randomize_considered_attr.handler = handle_randomize_considered_attr; + + randomize_performed_attr.name = "randomize_performed"; + randomize_performed_attr.type_required = true; + randomize_performed_attr.handler = handle_randomize_performed_attr; -static void register_attributes(void *event_data, void *data) -{ register_attribute(&randomize_layout_attr); register_attribute(&no_randomize_layout_attr); register_attribute(&randomize_considered_attr); diff --git a/scripts/gcc-plugins/structleak_plugin.c b/scripts/gcc-plugins/structleak_plugin.c index 3f8dd4868178..10292f791e99 100644 --- a/scripts/gcc-plugins/structleak_plugin.c +++ b/scripts/gcc-plugins/structleak_plugin.c @@ -57,21 +57,16 @@ static tree handle_user_attribute(tree *node, tree name, tree args, int flags, b return NULL_TREE; } -static struct attribute_spec user_attr = { - .name = "user", - .min_length = 0, - .max_length = 0, - .decl_required = false, - .type_required = false, - .function_type_required = false, - .handler = handle_user_attribute, -#if BUILDING_GCC_VERSION >= 4007 - .affects_type_identity = true -#endif -}; +static struct attribute_spec user_attr = { }; static void register_attributes(void *event_data, void *data) { + user_attr.name = "user"; + user_attr.handler = handle_user_attribute; +#if BUILDING_GCC_VERSION >= 4007 + user_attr.affects_type_identity = true; +#endif + register_attribute(&user_attr); } -- Lance Albertson Director Oregon State University | Open Source Lab

6 years, 11 months

2
2
0 0

[PATCH] staging: android: ion: check for kref overflow

by Daniel Rosenberg

This patch is against 4.9. It does not apply to master due to a large rework of ion in 4.12 which removed the affected functions altogther. 4c23cbff073f3b9b ("staging: android: ion: Remove import interface") Userspace can cause the kref to handles to increment arbitrarily high. Ensure it does not overflow. Signed-off-by: Daniel Rosenberg <drosen(a)google.com> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- drivers/staging/android/ion/ion.c | 17 ++++++++++++++--- 1 file changed, 14 insertions(+), 3 deletions(-) diff --git a/drivers/staging/android/ion/ion.c b/drivers/staging/android/ion/ion.c index 6f9974cb0e152..48821948fa487 100644 --- a/drivers/staging/android/ion/ion.c +++ b/drivers/staging/android/ion/ion.c @@ -15,6 +15,7 @@ * */ +#include <linux/atomic.h> #include <linux/device.h> #include <linux/err.h> #include <linux/file.h> @@ -305,6 +306,16 @@ static void ion_handle_get(struct ion_handle *handle) kref_get(&handle->ref); } +/* Must hold the client lock */ +static struct ion_handle *ion_handle_get_check_overflow( + struct ion_handle *handle) +{ + if (atomic_read(&handle->ref.refcount) + 1 == 0) + return ERR_PTR(-EOVERFLOW); + ion_handle_get(handle); + return handle; +} + int ion_handle_put_nolock(struct ion_handle *handle) { return kref_put(&handle->ref, ion_handle_destroy); @@ -347,9 +358,9 @@ struct ion_handle *ion_handle_get_by_id_nolock(struct ion_client *client, handle = idr_find(&client->idr, id); if (handle) - ion_handle_get(handle); + return ion_handle_get_check_overflow(handle); - return handle ? handle : ERR_PTR(-EINVAL); + return ERR_PTR(-EINVAL); } struct ion_handle *ion_handle_get_by_id(struct ion_client *client, @@ -1100,7 +1111,7 @@ struct ion_handle *ion_import_dma_buf(struct ion_client *client, /* if a handle exists for this buffer just take a reference to it */ handle = ion_handle_lookup(client, buffer); if (!IS_ERR(handle)) { - ion_handle_get(handle); + handle = ion_handle_get_check_overflow(handle); mutex_unlock(&client->lock); goto end; } -- 2.19.0.rc0.228.g281dcd1b4d0-goog

6 years, 11 months

2
5
0 0

Applied "regulator: Fix 'do-nothing' value for regulators without suspend state" to the regulator tree

by Mark Brown

The patch regulator: Fix 'do-nothing' value for regulators without suspend state has been applied to the regulator tree at https://git.kernel.org/pub/scm/linux/kernel/git/broonie/regulator.git All being well this means that it will be integrated into the linux-next tree (usually sometime in the next 24 hours) and sent to Linus during the next merge window (or sooner if it is a bug fix), however if problems are discovered then the patch may be dropped or reverted. You may get further e-mails resulting from automated or manual testing and review of the tree, please engage with people reporting problems and send followup patches addressing any issues that are reported if needed. If any updates are required or you are submitting further changes they should be sent as incremental updates against current git, existing patches will not be replaced. Please add any relevant lists and maintainers to the CCs when replying to this mail. Thanks, Mark >From 3edd79cf5a44b12dbb13bc320f5788aed6562b36 Mon Sep 17 00:00:00 2001 From: Marek Szyprowski <m.szyprowski(a)samsung.com> Date: Mon, 3 Sep 2018 16:49:37 +0200 Subject: [PATCH] regulator: Fix 'do-nothing' value for regulators without suspend state Some regulators don't have all states defined and in such cases regulator core should not assume anything. However in current implementation of of_get_regulation_constraints() DO_NOTHING_IN_SUSPEND enable value was set only for regulators which had suspend node defined, otherwise the default 0 value was used, what means DISABLE_IN_SUSPEND. This lead to broken system suspend/resume on boards, which had simple regulator constraints definition (without suspend state nodes). To avoid further mismatches between the default and uninitialized values of the suspend enabled/disabled states, change the values of the them, so default '0' means DO_NOTHING_IN_SUSPEND. Fixes: 72069f9957a1: regulator: leave one item to record whether regulator is enabled Signed-off-by: Marek Szyprowski <m.szyprowski(a)samsung.com> Signed-off-by: Mark Brown <broonie(a)kernel.org> Cc: stable(a)vger.kernel.org --- drivers/regulator/core.c | 2 +- drivers/regulator/of_regulator.c | 2 -- include/linux/regulator/machine.h | 6 +++--- 3 files changed, 4 insertions(+), 6 deletions(-) diff --git a/drivers/regulator/core.c b/drivers/regulator/core.c index bb1324f93143..90215f57270f 100644 --- a/drivers/regulator/core.c +++ b/drivers/regulator/core.c @@ -3161,7 +3161,7 @@ static inline int regulator_suspend_toggle(struct regulator_dev *rdev, if (!rstate->changeable) return -EPERM; - rstate->enabled = en; + rstate->enabled = (en) ? ENABLE_IN_SUSPEND : DISABLE_IN_SUSPEND; return 0; } diff --git a/drivers/regulator/of_regulator.c b/drivers/regulator/of_regulator.c index 638f17d4c848..210fc20f7de7 100644 --- a/drivers/regulator/of_regulator.c +++ b/drivers/regulator/of_regulator.c @@ -213,8 +213,6 @@ static void of_get_regulation_constraints(struct device_node *np, else if (of_property_read_bool(suspend_np, "regulator-off-in-suspend")) suspend_state->enabled = DISABLE_IN_SUSPEND; - else - suspend_state->enabled = DO_NOTHING_IN_SUSPEND; if (!of_property_read_u32(np, "regulator-suspend-min-microvolt", &pval)) diff --git a/include/linux/regulator/machine.h b/include/linux/regulator/machine.h index 3468703d663a..a459a5e973a7 100644 --- a/include/linux/regulator/machine.h +++ b/include/linux/regulator/machine.h @@ -48,9 +48,9 @@ struct regulator; * DISABLE_IN_SUSPEND - turn off regulator in suspend states * ENABLE_IN_SUSPEND - keep regulator on in suspend states */ -#define DO_NOTHING_IN_SUSPEND (-1) -#define DISABLE_IN_SUSPEND 0 -#define ENABLE_IN_SUSPEND 1 +#define DO_NOTHING_IN_SUSPEND 0 +#define DISABLE_IN_SUSPEND 1 +#define ENABLE_IN_SUSPEND 2 /* Regulator active discharge flags */ enum regulator_active_discharge { -- 2.19.0.rc1

6 years, 11 months

1
0
0 0

[PATCH for 4.18.y] x86/dumpstack: Don't dump kernel memory based on usermode RIP

by Jann Horn

commit 342db04ae71273322f0011384a9ed414df8bdae4 upstream. show_opcodes() is used both for dumping kernel instructions and for dumping user instructions. If userspace causes #PF by jumping to a kernel address, show_opcodes() can be reached with regs->ip controlled by the user, pointing to kernel code. Make sure that userspace can't trick us into dumping kernel memory into dmesg. Manually backported: show_opcodes() has changed a bit in the meantime. I have manually tested the backport. Fixes: 7cccf0725cf7 ("x86/dumpstack: Add a show_ip() function") Cc: stable(a)vger.kernel.org Link: https://lkml.kernel.org/r/20180828154901.112726-1-jannh@google.com Signed-off-by: Jann Horn <jannh(a)google.com> --- Since I manually backported this, I have removed all other sign-off/reviewed-by lines. I hope that's correct? arch/x86/include/asm/stacktrace.h | 2 +- arch/x86/kernel/dumpstack.c | 21 +++++++++++++++------ arch/x86/mm/fault.c | 2 +- 3 files changed, 17 insertions(+), 8 deletions(-) diff --git a/arch/x86/include/asm/stacktrace.h b/arch/x86/include/asm/stacktrace.h index b6dc698f992a..f335aad404a4 100644 --- a/arch/x86/include/asm/stacktrace.h +++ b/arch/x86/include/asm/stacktrace.h @@ -111,6 +111,6 @@ static inline unsigned long caller_frame_pointer(void) return (unsigned long)frame; } -void show_opcodes(u8 *rip, const char *loglvl); +void show_opcodes(struct pt_regs *regs, const char *loglvl); void show_ip(struct pt_regs *regs, const char *loglvl); #endif /* _ASM_X86_STACKTRACE_H */ diff --git a/arch/x86/kernel/dumpstack.c b/arch/x86/kernel/dumpstack.c index 666a284116ac..1c65c648cacc 100644 --- a/arch/x86/kernel/dumpstack.c +++ b/arch/x86/kernel/dumpstack.c @@ -91,23 +91,32 @@ static void printk_stack_address(unsigned long address, int reliable, * Thus, the 2/3rds prologue and 64 byte OPCODE_BUFSIZE is just a random * guesstimate in attempt to achieve all of the above. */ -void show_opcodes(u8 *rip, const char *loglvl) +void show_opcodes(struct pt_regs *regs, const char *loglvl) { unsigned int code_prologue = OPCODE_BUFSIZE * 2 / 3; u8 opcodes[OPCODE_BUFSIZE]; - u8 *ip; + unsigned long ip; int i; + bool bad_ip; printk("%sCode: ", loglvl); - ip = (u8 *)rip - code_prologue; - if (probe_kernel_read(opcodes, ip, OPCODE_BUFSIZE)) { + ip = regs->ip - code_prologue; + + /* + * Make sure userspace isn't trying to trick us into dumping kernel + * memory by pointing the userspace instruction pointer at it. + */ + bad_ip = user_mode(regs) && + __chk_range_not_ok(ip, OPCODE_BUFSIZE, TASK_SIZE_MAX); + + if (bad_ip || probe_kernel_read(opcodes, (u8 *)ip, OPCODE_BUFSIZE)) { pr_cont("Bad RIP value.\n"); return; } for (i = 0; i < OPCODE_BUFSIZE; i++, ip++) { - if (ip == rip) + if (ip == regs->ip) pr_cont("<%02x> ", opcodes[i]); else pr_cont("%02x ", opcodes[i]); @@ -122,7 +131,7 @@ void show_ip(struct pt_regs *regs, const char *loglvl) #else printk("%sRIP: %04x:%pS\n", loglvl, (int)regs->cs, (void *)regs->ip); #endif - show_opcodes((u8 *)regs->ip, loglvl); + show_opcodes(regs, loglvl); } void show_iret_regs(struct pt_regs *regs) diff --git a/arch/x86/mm/fault.c b/arch/x86/mm/fault.c index 2aafa6ab6103..d1f1612672c7 100644 --- a/arch/x86/mm/fault.c +++ b/arch/x86/mm/fault.c @@ -838,7 +838,7 @@ show_signal_msg(struct pt_regs *regs, unsigned long error_code, printk(KERN_CONT "\n"); - show_opcodes((u8 *)regs->ip, loglvl); + show_opcodes(regs, loglvl); } static void -- 2.19.0.rc1.350.ge57e33dbd1-goog

6 years, 11 months

2
1
0 0

FAILED: patch "[PATCH] s390/lib: use expoline for all bcr instructions" failed to apply to 4.4-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 4.4-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. thanks, greg k-h ------------------ original commit in Linus's tree ------------------ >From 5eda25b10297684c1f46a14199ec00210f3c346e Mon Sep 17 00:00:00 2001 From: Martin Schwidefsky <schwidefsky(a)de.ibm.com> Date: Mon, 6 Aug 2018 13:49:47 +0200 Subject: [PATCH] s390/lib: use expoline for all bcr instructions The memove, memset, memcpy, __memset16, __memset32 and __memset64 function have an additional indirect return branch in form of a "bzr" instruction. These need to use expolines as well. Cc: <stable(a)vger.kernel.org> # v4.17+ Fixes: 97489e0663 ("s390/lib: use expoline for indirect branches") Reviewed-by: Heiko Carstens <heiko.carstens(a)de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky(a)de.ibm.com> diff --git a/arch/s390/lib/mem.S b/arch/s390/lib/mem.S index 2311f15be9cf..40c4d59c926e 100644 --- a/arch/s390/lib/mem.S +++ b/arch/s390/lib/mem.S @@ -17,7 +17,7 @@ ENTRY(memmove) ltgr %r4,%r4 lgr %r1,%r2 - bzr %r14 + jz .Lmemmove_exit aghi %r4,-1 clgr %r2,%r3 jnh .Lmemmove_forward @@ -36,6 +36,7 @@ ENTRY(memmove) .Lmemmove_forward_remainder: larl %r5,.Lmemmove_mvc ex %r4,0(%r5) +.Lmemmove_exit: BR_EX %r14 .Lmemmove_reverse: ic %r0,0(%r4,%r3) @@ -65,7 +66,7 @@ EXPORT_SYMBOL(memmove) */ ENTRY(memset) ltgr %r4,%r4 - bzr %r14 + jz .Lmemset_exit ltgr %r3,%r3 jnz .Lmemset_fill aghi %r4,-1 @@ -80,6 +81,7 @@ ENTRY(memset) .Lmemset_clear_remainder: larl %r3,.Lmemset_xc ex %r4,0(%r3) +.Lmemset_exit: BR_EX %r14 .Lmemset_fill: cghi %r4,1 @@ -115,7 +117,7 @@ EXPORT_SYMBOL(memset) */ ENTRY(memcpy) ltgr %r4,%r4 - bzr %r14 + jz .Lmemcpy_exit aghi %r4,-1 srlg %r5,%r4,8 ltgr %r5,%r5 @@ -124,6 +126,7 @@ ENTRY(memcpy) .Lmemcpy_remainder: larl %r5,.Lmemcpy_mvc ex %r4,0(%r5) +.Lmemcpy_exit: BR_EX %r14 .Lmemcpy_loop: mvc 0(256,%r1),0(%r3) @@ -145,9 +148,9 @@ EXPORT_SYMBOL(memcpy) .macro __MEMSET bits,bytes,insn ENTRY(__memset\bits) ltgr %r4,%r4 - bzr %r14 + jz .L__memset_exit\bits cghi %r4,\bytes - je .L__memset_exit\bits + je .L__memset_store\bits aghi %r4,-(\bytes+1) srlg %r5,%r4,8 ltgr %r5,%r5 @@ -163,8 +166,9 @@ ENTRY(__memset\bits) larl %r5,.L__memset_mvc\bits ex %r4,0(%r5) BR_EX %r14 -.L__memset_exit\bits: +.L__memset_store\bits: \insn %r3,0(%r2) +.L__memset_exit\bits: BR_EX %r14 .L__memset_mvc\bits: mvc \bytes(1,%r1),0(%r1)

6 years, 11 months

2
1
0 0

FAILED: patch "[PATCH] s390/lib: use expoline for all bcr instructions" failed to apply to 4.9-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 4.9-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. thanks, greg k-h ------------------ original commit in Linus's tree ------------------ >From 5eda25b10297684c1f46a14199ec00210f3c346e Mon Sep 17 00:00:00 2001 From: Martin Schwidefsky <schwidefsky(a)de.ibm.com> Date: Mon, 6 Aug 2018 13:49:47 +0200 Subject: [PATCH] s390/lib: use expoline for all bcr instructions The memove, memset, memcpy, __memset16, __memset32 and __memset64 function have an additional indirect return branch in form of a "bzr" instruction. These need to use expolines as well. Cc: <stable(a)vger.kernel.org> # v4.17+ Fixes: 97489e0663 ("s390/lib: use expoline for indirect branches") Reviewed-by: Heiko Carstens <heiko.carstens(a)de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky(a)de.ibm.com> diff --git a/arch/s390/lib/mem.S b/arch/s390/lib/mem.S index 2311f15be9cf..40c4d59c926e 100644 --- a/arch/s390/lib/mem.S +++ b/arch/s390/lib/mem.S @@ -17,7 +17,7 @@ ENTRY(memmove) ltgr %r4,%r4 lgr %r1,%r2 - bzr %r14 + jz .Lmemmove_exit aghi %r4,-1 clgr %r2,%r3 jnh .Lmemmove_forward @@ -36,6 +36,7 @@ ENTRY(memmove) .Lmemmove_forward_remainder: larl %r5,.Lmemmove_mvc ex %r4,0(%r5) +.Lmemmove_exit: BR_EX %r14 .Lmemmove_reverse: ic %r0,0(%r4,%r3) @@ -65,7 +66,7 @@ EXPORT_SYMBOL(memmove) */ ENTRY(memset) ltgr %r4,%r4 - bzr %r14 + jz .Lmemset_exit ltgr %r3,%r3 jnz .Lmemset_fill aghi %r4,-1 @@ -80,6 +81,7 @@ ENTRY(memset) .Lmemset_clear_remainder: larl %r3,.Lmemset_xc ex %r4,0(%r3) +.Lmemset_exit: BR_EX %r14 .Lmemset_fill: cghi %r4,1 @@ -115,7 +117,7 @@ EXPORT_SYMBOL(memset) */ ENTRY(memcpy) ltgr %r4,%r4 - bzr %r14 + jz .Lmemcpy_exit aghi %r4,-1 srlg %r5,%r4,8 ltgr %r5,%r5 @@ -124,6 +126,7 @@ ENTRY(memcpy) .Lmemcpy_remainder: larl %r5,.Lmemcpy_mvc ex %r4,0(%r5) +.Lmemcpy_exit: BR_EX %r14 .Lmemcpy_loop: mvc 0(256,%r1),0(%r3) @@ -145,9 +148,9 @@ EXPORT_SYMBOL(memcpy) .macro __MEMSET bits,bytes,insn ENTRY(__memset\bits) ltgr %r4,%r4 - bzr %r14 + jz .L__memset_exit\bits cghi %r4,\bytes - je .L__memset_exit\bits + je .L__memset_store\bits aghi %r4,-(\bytes+1) srlg %r5,%r4,8 ltgr %r5,%r5 @@ -163,8 +166,9 @@ ENTRY(__memset\bits) larl %r5,.L__memset_mvc\bits ex %r4,0(%r5) BR_EX %r14 -.L__memset_exit\bits: +.L__memset_store\bits: \insn %r3,0(%r2) +.L__memset_exit\bits: BR_EX %r14 .L__memset_mvc\bits: mvc \bytes(1,%r1),0(%r1)

6 years, 11 months

2
1
0 0

Applied "spi: tegra20-slink: explicitly enable/disable clock" to the spi tree

by Mark Brown

The patch spi: tegra20-slink: explicitly enable/disable clock has been applied to the spi tree at https://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi.git All being well this means that it will be integrated into the linux-next tree (usually sometime in the next 24 hours) and sent to Linus during the next merge window (or sooner if it is a bug fix), however if problems are discovered then the patch may be dropped or reverted. You may get further e-mails resulting from automated or manual testing and review of the tree, please engage with people reporting problems and send followup patches addressing any issues that are reported if needed. If any updates are required or you are submitting further changes they should be sent as incremental updates against current git, existing patches will not be replaced. Please add any relevant lists and maintainers to the CCs when replying to this mail. Thanks, Mark >From 7001cab1dabc0b72b2b672ef58a90ab64f5e2343 Mon Sep 17 00:00:00 2001 From: Marcel Ziswiler <marcel.ziswiler(a)toradex.com> Date: Wed, 29 Aug 2018 08:47:57 +0200 Subject: [PATCH] spi: tegra20-slink: explicitly enable/disable clock Depending on the SPI instance one may get an interrupt storm upon requesting resp. interrupt unless the clock is explicitly enabled beforehand. This has been observed trying to bring up instance 4 on T20. Signed-off-by: Marcel Ziswiler <marcel.ziswiler(a)toradex.com> Signed-off-by: Mark Brown <broonie(a)kernel.org> Cc: stable(a)vger.kernel.org --- drivers/spi/spi-tegra20-slink.c | 31 +++++++++++++++++++++++-------- 1 file changed, 23 insertions(+), 8 deletions(-) diff --git a/drivers/spi/spi-tegra20-slink.c b/drivers/spi/spi-tegra20-slink.c index 6f7b946b5ced..1427f343b39a 100644 --- a/drivers/spi/spi-tegra20-slink.c +++ b/drivers/spi/spi-tegra20-slink.c @@ -1063,6 +1063,24 @@ static int tegra_slink_probe(struct platform_device *pdev) goto exit_free_master; } + /* disabled clock may cause interrupt storm upon request */ + tspi->clk = devm_clk_get(&pdev->dev, NULL); + if (IS_ERR(tspi->clk)) { + ret = PTR_ERR(tspi->clk); + dev_err(&pdev->dev, "Can not get clock %d\n", ret); + goto exit_free_master; + } + ret = clk_prepare(tspi->clk); + if (ret < 0) { + dev_err(&pdev->dev, "Clock prepare failed %d\n", ret); + goto exit_free_master; + } + ret = clk_enable(tspi->clk); + if (ret < 0) { + dev_err(&pdev->dev, "Clock enable failed %d\n", ret); + goto exit_free_master; + } + spi_irq = platform_get_irq(pdev, 0); tspi->irq = spi_irq; ret = request_threaded_irq(tspi->irq, tegra_slink_isr, @@ -1071,14 +1089,7 @@ static int tegra_slink_probe(struct platform_device *pdev) if (ret < 0) { dev_err(&pdev->dev, "Failed to register ISR for IRQ %d\n", tspi->irq); - goto exit_free_master; - } - - tspi->clk = devm_clk_get(&pdev->dev, NULL); - if (IS_ERR(tspi->clk)) { - dev_err(&pdev->dev, "can not get clock\n"); - ret = PTR_ERR(tspi->clk); - goto exit_free_irq; + goto exit_clk_disable; } tspi->rst = devm_reset_control_get_exclusive(&pdev->dev, "spi"); @@ -1138,6 +1149,8 @@ static int tegra_slink_probe(struct platform_device *pdev) tegra_slink_deinit_dma_param(tspi, true); exit_free_irq: free_irq(spi_irq, tspi); +exit_clk_disable: + clk_disable(tspi->clk); exit_free_master: spi_master_put(master); return ret; @@ -1150,6 +1163,8 @@ static int tegra_slink_remove(struct platform_device *pdev) free_irq(tspi->irq, tspi); + clk_disable(tspi->clk); + if (tspi->tx_dma_chan) tegra_slink_deinit_dma_param(tspi, false); -- 2.19.0.rc1

6 years, 11 months

1
0
0 0

Applied "ASoC: rsnd: fixup not to call clk_get/set under non-atomic" to the asoc tree

by Mark Brown

The patch ASoC: rsnd: fixup not to call clk_get/set under non-atomic has been applied to the asoc tree at https://git.kernel.org/pub/scm/linux/kernel/git/broonie/sound.git All being well this means that it will be integrated into the linux-next tree (usually sometime in the next 24 hours) and sent to Linus during the next merge window (or sooner if it is a bug fix), however if problems are discovered then the patch may be dropped or reverted. You may get further e-mails resulting from automated or manual testing and review of the tree, please engage with people reporting problems and send followup patches addressing any issues that are reported if needed. If any updates are required or you are submitting further changes they should be sent as incremental updates against current git, existing patches will not be replaced. Please add any relevant lists and maintainers to the CCs when replying to this mail. Thanks, Mark >From 4d230d12710646788af581ba0155d83ab48b955c Mon Sep 17 00:00:00 2001 From: Jiada Wang <jiada_wang(a)mentor.com> Date: Mon, 3 Sep 2018 07:08:58 +0000 Subject: [PATCH] ASoC: rsnd: fixup not to call clk_get/set under non-atomic Clocking operations clk_get/set_rate, are non-atomic, they shouldn't be called in soc_pcm_trigger() which is atomic. Following issue was found due to execution of clk_get_rate() causes sleep in soc_pcm_trigger(), which shouldn't be blocked. We can reproduce this issue by following > enable CONFIG_DEBUG_ATOMIC_SLEEP=y > compile, and boot > mount -t debugfs none /sys/kernel/debug > while true; do cat /sys/kernel/debug/clk/clk_summary > /dev/null; done & > while true; do aplay xxx; done This patch adds support to .prepare callback, and moves non-atomic clocking operations to it. As .prepare is non-atomic, it is always called before trigger_start/trigger_stop. BUG: sleeping function called from invalid context at kernel/locking/mutex.c:620 in_atomic(): 1, irqs_disabled(): 128, pid: 2242, name: aplay INFO: lockdep is turned off. irq event stamp: 5964 hardirqs last enabled at (5963): [<ffff200008e59e40>] mutex_lock_nested+0x6e8/0x6f0 hardirqs last disabled at (5964): [<ffff200008e623f0>] _raw_spin_lock_irqsave+0x24/0x68 softirqs last enabled at (5502): [<ffff200008081838>] __do_softirq+0x560/0x10c0 softirqs last disabled at (5495): [<ffff2000080c2e78>] irq_exit+0x160/0x25c Preemption disabled at:[ 62.904063] [<ffff200008be4d48>] snd_pcm_stream_lock+0xb4/0xc0 CPU: 2 PID: 2242 Comm: aplay Tainted: G B C 4.9.54+ #186 Hardware name: Renesas Salvator-X board based on r8a7795 (DT) Call trace: [<ffff20000808fe48>] dump_backtrace+0x0/0x37c [<ffff2000080901d8>] show_stack+0x14/0x1c [<ffff2000086f4458>] dump_stack+0xfc/0x154 [<ffff2000081134a0>] ___might_sleep+0x57c/0x58c [<ffff2000081136b8>] __might_sleep+0x208/0x21c [<ffff200008e5980c>] mutex_lock_nested+0xb4/0x6f0 [<ffff2000087cac74>] clk_prepare_lock+0xb0/0x184 [<ffff2000087cb094>] clk_core_get_rate+0x14/0x54 [<ffff2000087cb0f4>] clk_get_rate+0x20/0x34 [<ffff20000113aa00>] rsnd_adg_ssi_clk_try_start+0x158/0x4f8 [snd_soc_rcar] [<ffff20000113da00>] rsnd_ssi_init+0x668/0x7a0 [snd_soc_rcar] [<ffff200001133ff4>] rsnd_soc_dai_trigger+0x4bc/0xcf8 [snd_soc_rcar] [<ffff200008c1af24>] soc_pcm_trigger+0x2a4/0x2d4 Fixes: e7d850dd10f4 ("ASoC: rsnd: use mod base common method on SSI-parent") Signed-off-by: Jiada Wang <jiada_wang(a)mentor.com> Signed-off-by: Timo Wischer <twischer(a)de.adit-jv.com> [Kuninori: tidyup for upstream] Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx(a)renesas.com> Tested-by: Hiroyuki Yokoyama <hiroyuki.yokoyama.vx(a)renesas.com> Signed-off-by: Mark Brown <broonie(a)kernel.org> Cc: stable(a)vger.kernel.org --- sound/soc/sh/rcar/core.c | 11 +++++++++++ sound/soc/sh/rcar/rsnd.h | 7 +++++++ sound/soc/sh/rcar/ssi.c | 16 ++++++++++------ 3 files changed, 28 insertions(+), 6 deletions(-) diff --git a/sound/soc/sh/rcar/core.c b/sound/soc/sh/rcar/core.c index f8425d8b44d2..b35f5509cfe2 100644 --- a/sound/soc/sh/rcar/core.c +++ b/sound/soc/sh/rcar/core.c @@ -958,12 +958,23 @@ static void rsnd_soc_dai_shutdown(struct snd_pcm_substream *substream, rsnd_dai_stream_quit(io); } +static int rsnd_soc_dai_prepare(struct snd_pcm_substream *substream, + struct snd_soc_dai *dai) +{ + struct rsnd_priv *priv = rsnd_dai_to_priv(dai); + struct rsnd_dai *rdai = rsnd_dai_to_rdai(dai); + struct rsnd_dai_stream *io = rsnd_rdai_to_io(rdai, substream); + + return rsnd_dai_call(prepare, io, priv); +} + static const struct snd_soc_dai_ops rsnd_soc_dai_ops = { .startup = rsnd_soc_dai_startup, .shutdown = rsnd_soc_dai_shutdown, .trigger = rsnd_soc_dai_trigger, .set_fmt = rsnd_soc_dai_set_fmt, .set_tdm_slot = rsnd_soc_set_dai_tdm_slot, + .prepare = rsnd_soc_dai_prepare, }; void rsnd_parse_connect_common(struct rsnd_dai *rdai, diff --git a/sound/soc/sh/rcar/rsnd.h b/sound/soc/sh/rcar/rsnd.h index 96d93330b1e1..8f7a0abfa751 100644 --- a/sound/soc/sh/rcar/rsnd.h +++ b/sound/soc/sh/rcar/rsnd.h @@ -280,6 +280,9 @@ struct rsnd_mod_ops { int (*nolock_stop)(struct rsnd_mod *mod, struct rsnd_dai_stream *io, struct rsnd_priv *priv); + int (*prepare)(struct rsnd_mod *mod, + struct rsnd_dai_stream *io, + struct rsnd_priv *priv); }; struct rsnd_dai_stream; @@ -309,6 +312,7 @@ struct rsnd_mod { * H 0: fallback * H 0: hw_params * H 0: pointer + * H 0: prepare */ #define __rsnd_mod_shift_nolock_start 0 #define __rsnd_mod_shift_nolock_stop 0 @@ -323,6 +327,7 @@ struct rsnd_mod { #define __rsnd_mod_shift_fallback 28 /* always called */ #define __rsnd_mod_shift_hw_params 28 /* always called */ #define __rsnd_mod_shift_pointer 28 /* always called */ +#define __rsnd_mod_shift_prepare 28 /* always called */ #define __rsnd_mod_add_probe 0 #define __rsnd_mod_add_remove 0 @@ -337,6 +342,7 @@ struct rsnd_mod { #define __rsnd_mod_add_fallback 0 #define __rsnd_mod_add_hw_params 0 #define __rsnd_mod_add_pointer 0 +#define __rsnd_mod_add_prepare 0 #define __rsnd_mod_call_probe 0 #define __rsnd_mod_call_remove 0 @@ -351,6 +357,7 @@ struct rsnd_mod { #define __rsnd_mod_call_pointer 0 #define __rsnd_mod_call_nolock_start 0 #define __rsnd_mod_call_nolock_stop 1 +#define __rsnd_mod_call_prepare 0 #define rsnd_mod_to_priv(mod) ((mod)->priv) #define rsnd_mod_name(mod) ((mod)->ops->name) diff --git a/sound/soc/sh/rcar/ssi.c b/sound/soc/sh/rcar/ssi.c index 8304e4ec9242..3f880ec66459 100644 --- a/sound/soc/sh/rcar/ssi.c +++ b/sound/soc/sh/rcar/ssi.c @@ -283,7 +283,7 @@ static int rsnd_ssi_master_clk_start(struct rsnd_mod *mod, if (rsnd_ssi_is_multi_slave(mod, io)) return 0; - if (ssi->usrcnt > 1) { + if (ssi->rate) { if (ssi->rate != rate) { dev_err(dev, "SSI parent/child should use same rate\n"); return -EINVAL; @@ -434,7 +434,6 @@ static int rsnd_ssi_init(struct rsnd_mod *mod, struct rsnd_priv *priv) { struct rsnd_ssi *ssi = rsnd_mod_to_ssi(mod); - int ret; if (!rsnd_ssi_is_run_mods(mod, io)) return 0; @@ -443,10 +442,6 @@ static int rsnd_ssi_init(struct rsnd_mod *mod, rsnd_mod_power_on(mod); - ret = rsnd_ssi_master_clk_start(mod, io); - if (ret < 0) - return ret; - rsnd_ssi_config_init(mod, io); rsnd_ssi_register_setup(mod); @@ -852,6 +847,13 @@ static int rsnd_ssi_pio_pointer(struct rsnd_mod *mod, return 0; } +static int rsnd_ssi_prepare(struct rsnd_mod *mod, + struct rsnd_dai_stream *io, + struct rsnd_priv *priv) +{ + return rsnd_ssi_master_clk_start(mod, io); +} + static struct rsnd_mod_ops rsnd_ssi_pio_ops = { .name = SSI_NAME, .probe = rsnd_ssi_common_probe, @@ -864,6 +866,7 @@ static struct rsnd_mod_ops rsnd_ssi_pio_ops = { .pointer = rsnd_ssi_pio_pointer, .pcm_new = rsnd_ssi_pcm_new, .hw_params = rsnd_ssi_hw_params, + .prepare = rsnd_ssi_prepare, }; static int rsnd_ssi_dma_probe(struct rsnd_mod *mod, @@ -940,6 +943,7 @@ static struct rsnd_mod_ops rsnd_ssi_dma_ops = { .pcm_new = rsnd_ssi_pcm_new, .fallback = rsnd_ssi_fallback, .hw_params = rsnd_ssi_hw_params, + .prepare = rsnd_ssi_prepare, }; int rsnd_ssi_is_dma_mode(struct rsnd_mod *mod) -- 2.19.0.rc1

6 years, 11 months

1
0
0 0

[PATCH AUTOSEL 4.4 01/47] misc: mic: SCIF Fix scif_get_new_port() error handling

by Sasha Levin

From: Dan Carpenter <dan.carpenter(a)oracle.com> [ Upstream commit a39284ae9d2ad09975c8ae33f1bd0f05fbfbf6ee ] There are only 2 callers of scif_get_new_port() and both appear to get the error handling wrong. Both treat zero returns as error, but it actually returns negative error codes and >= 0 on success. Fixes: e9089f43c9a7 ("misc: mic: SCIF open close bind and listen APIs") Signed-off-by: Dan Carpenter <dan.carpenter(a)oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> Signed-off-by: Sasha Levin <alexander.levin(a)microsoft.com> --- drivers/misc/mic/scif/scif_api.c | 20 +++++++++----------- 1 file changed, 9 insertions(+), 11 deletions(-) diff --git a/drivers/misc/mic/scif/scif_api.c b/drivers/misc/mic/scif/scif_api.c index ddc9e4b08b5c..56efa9d18a9a 100644 --- a/drivers/misc/mic/scif/scif_api.c +++ b/drivers/misc/mic/scif/scif_api.c @@ -370,11 +370,10 @@ int scif_bind(scif_epd_t epd, u16 pn) goto scif_bind_exit; } } else { - pn = scif_get_new_port(); - if (!pn) { - ret = -ENOSPC; + ret = scif_get_new_port(); + if (ret < 0) goto scif_bind_exit; - } + pn = ret; } ep->state = SCIFEP_BOUND; @@ -648,13 +647,12 @@ int __scif_connect(scif_epd_t epd, struct scif_port_id *dst, bool non_block) err = -EISCONN; break; case SCIFEP_UNBOUND: - ep->port.port = scif_get_new_port(); - if (!ep->port.port) { - err = -ENOSPC; - } else { - ep->port.node = scif_info.nodeid; - ep->conn_async_state = ASYNC_CONN_IDLE; - } + err = scif_get_new_port(); + if (err < 0) + break; + ep->port.port = err; + ep->port.node = scif_info.nodeid; + ep->conn_async_state = ASYNC_CONN_IDLE; /* Fall through */ case SCIFEP_BOUND: /* -- 2.17.1

6 years, 11 months

2
47
0 0

[PATCH] Revert "ARM: imx_v6_v7_defconfig: Select ULPI support"

by Fabio Estevam

From: Fabio Estevam <fabio.estevam(a)nxp.com> This reverts commit 0d0af17ae83d6feb29d676c72423461419df5110. This commit causes reboot to fail on imx6 wandboard, so let's revert it. Cc: <stable(a)vger.kernel.org> #4.4 Reported-by: Rasmus Villemoes <rasmus.villemoes(a)prevas.dk> Signed-off-by: Fabio Estevam <fabio.estevam(a)nxp.com> --- arch/arm/configs/imx_v6_v7_defconfig | 2 -- 1 file changed, 2 deletions(-) diff --git a/arch/arm/configs/imx_v6_v7_defconfig b/arch/arm/configs/imx_v6_v7_defconfig index b3490c1..4187f69 100644 --- a/arch/arm/configs/imx_v6_v7_defconfig +++ b/arch/arm/configs/imx_v6_v7_defconfig @@ -261,7 +261,6 @@ CONFIG_USB_STORAGE=y CONFIG_USB_CHIPIDEA=y CONFIG_USB_CHIPIDEA_UDC=y CONFIG_USB_CHIPIDEA_HOST=y -CONFIG_USB_CHIPIDEA_ULPI=y CONFIG_USB_SERIAL=m CONFIG_USB_SERIAL_GENERIC=y CONFIG_USB_SERIAL_FTDI_SIO=m @@ -288,7 +287,6 @@ CONFIG_USB_G_NCM=m CONFIG_USB_GADGETFS=m CONFIG_USB_MASS_STORAGE=m CONFIG_USB_G_SERIAL=m -CONFIG_USB_ULPI_BUS=y CONFIG_MMC=y CONFIG_MMC_SDHCI=y CONFIG_MMC_SDHCI_PLTFM=y -- 2.7.4

6 years, 11 months

1
0
0 0

[PATCH] Revert "ARM: imx_v6_v7_defconfig: Select ULPI support"

by Fabio Estevam

From: Fabio Estevam <fabio.estevam(a)nxp.com> This reverts commit 721476147fd2571309b6aa6daa695b39170602ef. This commit causes reboot to fail on imx6 wandboard, so let's revert it. Cc: <stable(a)vger.kernel.org> #4.9 Reported-by: Rasmus Villemoes <rasmus.villemoes(a)prevas.dk> Signed-off-by: Fabio Estevam <fabio.estevam(a)nxp.com> --- arch/arm/configs/imx_v6_v7_defconfig | 2 -- 1 file changed, 2 deletions(-) diff --git a/arch/arm/configs/imx_v6_v7_defconfig b/arch/arm/configs/imx_v6_v7_defconfig index 6b7d4f5..8ec4dbb 100644 --- a/arch/arm/configs/imx_v6_v7_defconfig +++ b/arch/arm/configs/imx_v6_v7_defconfig @@ -271,7 +271,6 @@ CONFIG_USB_STORAGE=y CONFIG_USB_CHIPIDEA=y CONFIG_USB_CHIPIDEA_UDC=y CONFIG_USB_CHIPIDEA_HOST=y -CONFIG_USB_CHIPIDEA_ULPI=y CONFIG_USB_SERIAL=m CONFIG_USB_SERIAL_GENERIC=y CONFIG_USB_SERIAL_FTDI_SIO=m @@ -308,7 +307,6 @@ CONFIG_USB_GADGETFS=m CONFIG_USB_FUNCTIONFS=m CONFIG_USB_MASS_STORAGE=m CONFIG_USB_G_SERIAL=m -CONFIG_USB_ULPI_BUS=y CONFIG_MMC=y CONFIG_MMC_SDHCI=y CONFIG_MMC_SDHCI_PLTFM=y -- 2.7.4

6 years, 11 months

1
0
0 0

[PATCH] Revert "ARM: imx_v6_v7_defconfig: Select ULPI support"

by Fabio Estevam

From: Fabio Estevam <fabio.estevam(a)nxp.com> This reverts commit 2059e527a659cf16d6bb709f1c8509f7a7623fc4. This commit causes reboot to fail on imx6 wandboard, so let's revert it. Cc: <stable(a)vger.kernel.org> #4.14 Reported-by: Rasmus Villemoes <rasmus.villemoes(a)prevas.dk> Signed-off-by: Fabio Estevam <fabio.estevam(a)nxp.com> --- arch/arm/configs/imx_v6_v7_defconfig | 2 -- 1 file changed, 2 deletions(-) diff --git a/arch/arm/configs/imx_v6_v7_defconfig b/arch/arm/configs/imx_v6_v7_defconfig index 21ac9f0..32acac9 100644 --- a/arch/arm/configs/imx_v6_v7_defconfig +++ b/arch/arm/configs/imx_v6_v7_defconfig @@ -289,7 +289,6 @@ CONFIG_USB_STORAGE=y CONFIG_USB_CHIPIDEA=y CONFIG_USB_CHIPIDEA_UDC=y CONFIG_USB_CHIPIDEA_HOST=y -CONFIG_USB_CHIPIDEA_ULPI=y CONFIG_USB_SERIAL=m CONFIG_USB_SERIAL_GENERIC=y CONFIG_USB_SERIAL_FTDI_SIO=m @@ -326,7 +325,6 @@ CONFIG_USB_GADGETFS=m CONFIG_USB_FUNCTIONFS=m CONFIG_USB_MASS_STORAGE=m CONFIG_USB_G_SERIAL=m -CONFIG_USB_ULPI_BUS=y CONFIG_MMC=y CONFIG_MMC_SDHCI=y CONFIG_MMC_SDHCI_PLTFM=y -- 2.7.4

6 years, 11 months

1
0
0 0

[PATCH v2 02/10] usb: roles: Handle driver reference counting

by Heikki Krogerus

This fixes potential "BUG: unable to handle kernel paging request at ..." from happening. Fixes: fde0aa6c175a ("usb: common: Small class for USB role switches") Cc: <stable(a)vger.kernel.org> Signed-off-by: Heikki Krogerus <heikki.krogerus(a)linux.intel.com> --- drivers/usb/common/roles.c | 15 ++++++++++++--- 1 file changed, 12 insertions(+), 3 deletions(-) diff --git a/drivers/usb/common/roles.c b/drivers/usb/common/roles.c index 15cc76e22123..3d8a776e55ee 100644 --- a/drivers/usb/common/roles.c +++ b/drivers/usb/common/roles.c @@ -109,8 +109,15 @@ static void *usb_role_switch_match(struct device_connection *con, int ep, */ struct usb_role_switch *usb_role_switch_get(struct device *dev) { - return device_connection_find_match(dev, "usb-role-switch", NULL, - usb_role_switch_match); + struct usb_role_switch *sw; + + sw = device_connection_find_match(dev, "usb-role-switch", NULL, + usb_role_switch_match); + + if (!IS_ERR(sw)) + WARN_ON(!try_module_get(sw->dev.parent->driver->owner)); + + return sw; } EXPORT_SYMBOL_GPL(usb_role_switch_get); @@ -122,8 +129,10 @@ EXPORT_SYMBOL_GPL(usb_role_switch_get); */ void usb_role_switch_put(struct usb_role_switch *sw) { - if (!IS_ERR_OR_NULL(sw)) + if (!IS_ERR_OR_NULL(sw)) { put_device(&sw->dev); + module_put(sw->dev.parent->driver->owner); + } } EXPORT_SYMBOL_GPL(usb_role_switch_put); -- 2.18.0

6 years, 11 months

1
0
0 0

[PATCH v2 01/10] usb: typec: Take care of driver module reference counting

by Heikki Krogerus

Functions typec_mux_get() and typec_switch_get() already make sure that the mux device reference count is incremented, but the same must be done to the driver module as well to prevent the drivers from being unloaded in the middle of operation. This fixes a potential "BUG: unable to handle kernel paging request at ..." from happening. Fixes: 93dd2112c7b2 ("usb: typec: mux: Get the mux identifier from function parameter") Cc: <stable(a)vger.kernel.org> Signed-off-by: Heikki Krogerus <heikki.krogerus(a)linux.intel.com> --- drivers/usb/typec/mux.c | 17 +++++++++++++---- 1 file changed, 13 insertions(+), 4 deletions(-) diff --git a/drivers/usb/typec/mux.c b/drivers/usb/typec/mux.c index ddaac63ecf12..d990aa510fab 100644 --- a/drivers/usb/typec/mux.c +++ b/drivers/usb/typec/mux.c @@ -9,6 +9,7 @@ #include <linux/device.h> #include <linux/list.h> +#include <linux/module.h> #include <linux/mutex.h> #include <linux/usb/typec_mux.h> @@ -49,8 +50,10 @@ struct typec_switch *typec_switch_get(struct device *dev) mutex_lock(&switch_lock); sw = device_connection_find_match(dev, "typec-switch", NULL, typec_switch_match); - if (!IS_ERR_OR_NULL(sw)) + if (!IS_ERR_OR_NULL(sw)) { + WARN_ON(!try_module_get(sw->dev->driver->owner)); get_device(sw->dev); + } mutex_unlock(&switch_lock); return sw; @@ -65,8 +68,10 @@ EXPORT_SYMBOL_GPL(typec_switch_get); */ void typec_switch_put(struct typec_switch *sw) { - if (!IS_ERR_OR_NULL(sw)) + if (!IS_ERR_OR_NULL(sw)) { + module_put(sw->dev->driver->owner); put_device(sw->dev); + } } EXPORT_SYMBOL_GPL(typec_switch_put); @@ -136,8 +141,10 @@ struct typec_mux *typec_mux_get(struct device *dev, const char *name) mutex_lock(&mux_lock); mux = device_connection_find_match(dev, name, NULL, typec_mux_match); - if (!IS_ERR_OR_NULL(mux)) + if (!IS_ERR_OR_NULL(mux)) { + WARN_ON(!try_module_get(mux->dev->driver->owner)); get_device(mux->dev); + } mutex_unlock(&mux_lock); return mux; @@ -152,8 +159,10 @@ EXPORT_SYMBOL_GPL(typec_mux_get); */ void typec_mux_put(struct typec_mux *mux) { - if (!IS_ERR_OR_NULL(mux)) + if (!IS_ERR_OR_NULL(mux)) { + module_put(mux->dev->driver->owner); put_device(mux->dev); + } } EXPORT_SYMBOL_GPL(typec_mux_put); -- 2.18.0

6 years, 11 months

1
0
0 0

FAILED: patch "[PATCH] watchdog: Mark watchdog touch functions as notrace" failed to apply to 4.9-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 4.9-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. thanks, greg k-h ------------------ original commit in Linus's tree ------------------ >From cb9d7fd51d9fbb329d182423bd7b92d0f8cb0e01 Mon Sep 17 00:00:00 2001 From: Vincent Whitchurch <vincent.whitchurch(a)axis.com> Date: Tue, 21 Aug 2018 17:25:07 +0200 Subject: [PATCH] watchdog: Mark watchdog touch functions as notrace Some architectures need to use stop_machine() to patch functions for ftrace, and the assumption is that the stopped CPUs do not make function calls to traceable functions when they are in the stopped state. Commit ce4f06dcbb5d ("stop_machine: Touch_nmi_watchdog() after MULTI_STOP_PREPARE") added calls to the watchdog touch functions from the stopped CPUs and those functions lack notrace annotations. This leads to crashes when enabling/disabling ftrace on ARM kernels built with the Thumb-2 instruction set. Fix it by adding the necessary notrace annotations. Fixes: ce4f06dcbb5d ("stop_machine: Touch_nmi_watchdog() after MULTI_STOP_PREPARE") Signed-off-by: Vincent Whitchurch <vincent.whitchurch(a)axis.com> Signed-off-by: Thomas Gleixner <tglx(a)linutronix.de> Cc: Peter Zijlstra <peterz(a)infradead.org> Cc: oleg(a)redhat.com Cc: tj(a)kernel.org Cc: stable(a)vger.kernel.org Link: https://lkml.kernel.org/r/20180821152507.18313-1-vincent.whitchurch@axis.com diff --git a/kernel/watchdog.c b/kernel/watchdog.c index 5470dce212c0..977918d5d350 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -261,7 +261,7 @@ static void __touch_watchdog(void) * entering idle state. This should only be used for scheduler events. * Use touch_softlockup_watchdog() for everything else. */ -void touch_softlockup_watchdog_sched(void) +notrace void touch_softlockup_watchdog_sched(void) { /* * Preemption can be enabled. It doesn't matter which CPU's timestamp @@ -270,7 +270,7 @@ void touch_softlockup_watchdog_sched(void) raw_cpu_write(watchdog_touch_ts, 0); } -void touch_softlockup_watchdog(void) +notrace void touch_softlockup_watchdog(void) { touch_softlockup_watchdog_sched(); wq_watchdog_touch(raw_smp_processor_id()); diff --git a/kernel/watchdog_hld.c b/kernel/watchdog_hld.c index 1f7020d65d0a..71381168dede 100644 --- a/kernel/watchdog_hld.c +++ b/kernel/watchdog_hld.c @@ -29,7 +29,7 @@ static struct cpumask dead_events_mask; static unsigned long hardlockup_allcpu_dumped; static atomic_t watchdog_cpus = ATOMIC_INIT(0); -void arch_touch_nmi_watchdog(void) +notrace void arch_touch_nmi_watchdog(void) { /* * Using __raw here because some code paths have diff --git a/kernel/workqueue.c b/kernel/workqueue.c index 60e80198c3df..0280deac392e 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -5574,7 +5574,7 @@ static void wq_watchdog_timer_fn(struct timer_list *unused) mod_timer(&wq_watchdog_timer, jiffies + thresh); } -void wq_watchdog_touch(int cpu) +notrace void wq_watchdog_touch(int cpu) { if (cpu >= 0) per_cpu(wq_watchdog_touched_cpu, cpu) = jiffies;

6 years, 11 months

1
0
0 0

FAILED: patch "[PATCH] power: generic-adc-battery: fix out-of-bounds write when" failed to apply to 4.4-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 4.4-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. thanks, greg k-h ------------------ original commit in Linus's tree ------------------ >From 932d47448c3caa0fa99e84d7f5bc302aa286efd8 Mon Sep 17 00:00:00 2001 From: "H. Nikolaus Schaller" <hns(a)goldelico.com> Date: Tue, 26 Jun 2018 15:28:29 +0200 Subject: [PATCH] power: generic-adc-battery: fix out-of-bounds write when copying channel properties We did have sporadic problems in the pinctrl framework during boot where a pin group name unexpectedly became NULL leading to a NULL dereference in strcmp. Detailled analysis of the failing cases did reveal that there were two devm allocated objects close to each other. The second one was the affected group_desc in pinmux and the first one was the psy_desc->properties buffer of the gab driver. Review of the gab code showed that the address calculation for one memcpy() is wrong. It does properties + sizeof(type) * index but C is defined to do the index multiplication already for pointer + integer additions. Hence the factor was applied twice and the memcpy() does write outside of the properties buffer. Sometimes it happened to be the pinctrl and triggered the strcmp(NULL). Anyways, it is overkill to use a memcpy() here instead of a simple assignment, which is easier to read and has less risk for wrong address calculations. So we change code to a simple assignment. If we initialize the index to the first free location, we can even remove the local variable 'properties'. This bug seems to exist right from the beginning in 3.7-rc1 in commit e60fea794e6e ("power: battery: Generic battery driver using IIO") Signed-off-by: H. Nikolaus Schaller <hns(a)goldelico.com> Cc: stable(a)vger.kernel.org Fixes: e60fea794e6e ("power: battery: Generic battery driver using IIO") Signed-off-by: Sebastian Reichel <sebastian.reichel(a)collabora.co.uk> diff --git a/drivers/power/supply/generic-adc-battery.c b/drivers/power/supply/generic-adc-battery.c index 28dc056eaafa..11f59275bff4 100644 --- a/drivers/power/supply/generic-adc-battery.c +++ b/drivers/power/supply/generic-adc-battery.c @@ -241,10 +241,9 @@ static int gab_probe(struct platform_device *pdev) struct power_supply_desc *psy_desc; struct power_supply_config psy_cfg = {}; struct gab_platform_data *pdata = pdev->dev.platform_data; - enum power_supply_property *properties; int ret = 0; int chan; - int index = 0; + int index = ARRAY_SIZE(gab_props); adc_bat = devm_kzalloc(&pdev->dev, sizeof(*adc_bat), GFP_KERNEL); if (!adc_bat) { @@ -278,8 +277,6 @@ static int gab_probe(struct platform_device *pdev) } memcpy(psy_desc->properties, gab_props, sizeof(gab_props)); - properties = (enum power_supply_property *) - ((char *)psy_desc->properties + sizeof(gab_props)); /* * getting channel from iio and copying the battery properties @@ -293,15 +290,12 @@ static int gab_probe(struct platform_device *pdev) adc_bat->channel[chan] = NULL; } else { /* copying properties for supported channels only */ - memcpy(properties + sizeof(*(psy_desc->properties)) * index, - &gab_dyn_props[chan], - sizeof(gab_dyn_props[chan])); - index++; + psy_desc->properties[index++] = gab_dyn_props[chan]; } } /* none of the channels are supported so let's bail out */ - if (index == 0) { + if (index == ARRAY_SIZE(gab_props)) { ret = -ENODEV; goto second_mem_fail; } @@ -312,7 +306,7 @@ static int gab_probe(struct platform_device *pdev) * as come channels may be not be supported by the device.So * we need to take care of that. */ - psy_desc->num_properties = ARRAY_SIZE(gab_props) + index; + psy_desc->num_properties = index; adc_bat->psy = power_supply_register(&pdev->dev, psy_desc, &psy_cfg); if (IS_ERR(adc_bat->psy)) {

6 years, 11 months

1
0
0 0

FAILED: patch "[PATCH] power: generic-adc-battery: check for duplicate properties" failed to apply to 4.4-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 4.4-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. thanks, greg k-h ------------------ original commit in Linus's tree ------------------ >From a427503edaaed9b75ed9746a654cece7e93e60a8 Mon Sep 17 00:00:00 2001 From: "H. Nikolaus Schaller" <hns(a)goldelico.com> Date: Tue, 26 Jun 2018 15:28:30 +0200 Subject: [PATCH] power: generic-adc-battery: check for duplicate properties copied from iio channels If an iio channel defines a basic property, there are duplicate entries in /sys/class/power/*/uevent. So add a check to avoid duplicates. Since all channels may be duplicates, we have to modify the related error check. Signed-off-by: H. Nikolaus Schaller <hns(a)goldelico.com> Cc: stable(a)vger.kernel.org Fixes: e60fea794e6e ("power: battery: Generic battery driver using IIO") Signed-off-by: Sebastian Reichel <sebastian.reichel(a)collabora.co.uk> diff --git a/drivers/power/supply/generic-adc-battery.c b/drivers/power/supply/generic-adc-battery.c index 11f59275bff4..bc462d1ec963 100644 --- a/drivers/power/supply/generic-adc-battery.c +++ b/drivers/power/supply/generic-adc-battery.c @@ -244,6 +244,7 @@ static int gab_probe(struct platform_device *pdev) int ret = 0; int chan; int index = ARRAY_SIZE(gab_props); + bool any = false; adc_bat = devm_kzalloc(&pdev->dev, sizeof(*adc_bat), GFP_KERNEL); if (!adc_bat) { @@ -290,12 +291,22 @@ static int gab_probe(struct platform_device *pdev) adc_bat->channel[chan] = NULL; } else { /* copying properties for supported channels only */ - psy_desc->properties[index++] = gab_dyn_props[chan]; + int index2; + + for (index2 = 0; index2 < index; index2++) { + if (psy_desc->properties[index2] == + gab_dyn_props[chan]) + break; /* already known */ + } + if (index2 == index) /* really new */ + psy_desc->properties[index++] = + gab_dyn_props[chan]; + any = true; } } /* none of the channels are supported so let's bail out */ - if (index == ARRAY_SIZE(gab_props)) { + if (!any) { ret = -ENODEV; goto second_mem_fail; }

6 years, 11 months

1
0
0 0

FAILED: patch "[PATCH] mei: bus: type promotion bug in mei_nfc_if_version()" failed to apply to 4.4-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 4.4-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. thanks, greg k-h ------------------ original commit in Linus's tree ------------------ >From b40b3e9358fbafff6a4ba0f4b9658f6617146f9c Mon Sep 17 00:00:00 2001 From: Dan Carpenter <dan.carpenter(a)oracle.com> Date: Wed, 11 Jul 2018 15:29:31 +0300 Subject: [PATCH] mei: bus: type promotion bug in mei_nfc_if_version() We accidentally removed the check for negative returns without considering the issue of type promotion. The "if_version_length" variable is type size_t so if __mei_cl_recv() returns a negative then "bytes_recv" is type promoted to a high positive value and treated as success. Cc: <stable(a)vger.kernel.org> Fixes: 582ab27a063a ("mei: bus: fix received data size check in NFC fixup") Signed-off-by: Dan Carpenter <dan.carpenter(a)oracle.com> Signed-off-by: Tomas Winkler <tomas.winkler(a)intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> diff --git a/drivers/misc/mei/bus-fixup.c b/drivers/misc/mei/bus-fixup.c index e45fe826d87d..65e28be3c8cc 100644 --- a/drivers/misc/mei/bus-fixup.c +++ b/drivers/misc/mei/bus-fixup.c @@ -341,7 +341,7 @@ static int mei_nfc_if_version(struct mei_cl *cl, ret = 0; bytes_recv = __mei_cl_recv(cl, (u8 *)reply, if_version_length, 0, 0); - if (bytes_recv < if_version_length) { + if (bytes_recv < 0 || bytes_recv < if_version_length) { dev_err(bus->dev, "Could not read IF version\n"); ret = -EIO; goto err;

6 years, 11 months

1
0
0 0

FAILED: patch "[PATCH] mei: bus: type promotion bug in mei_nfc_if_version()" failed to apply to 4.9-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 4.9-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. thanks, greg k-h ------------------ original commit in Linus's tree ------------------ >From b40b3e9358fbafff6a4ba0f4b9658f6617146f9c Mon Sep 17 00:00:00 2001 From: Dan Carpenter <dan.carpenter(a)oracle.com> Date: Wed, 11 Jul 2018 15:29:31 +0300 Subject: [PATCH] mei: bus: type promotion bug in mei_nfc_if_version() We accidentally removed the check for negative returns without considering the issue of type promotion. The "if_version_length" variable is type size_t so if __mei_cl_recv() returns a negative then "bytes_recv" is type promoted to a high positive value and treated as success. Cc: <stable(a)vger.kernel.org> Fixes: 582ab27a063a ("mei: bus: fix received data size check in NFC fixup") Signed-off-by: Dan Carpenter <dan.carpenter(a)oracle.com> Signed-off-by: Tomas Winkler <tomas.winkler(a)intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> diff --git a/drivers/misc/mei/bus-fixup.c b/drivers/misc/mei/bus-fixup.c index e45fe826d87d..65e28be3c8cc 100644 --- a/drivers/misc/mei/bus-fixup.c +++ b/drivers/misc/mei/bus-fixup.c @@ -341,7 +341,7 @@ static int mei_nfc_if_version(struct mei_cl *cl, ret = 0; bytes_recv = __mei_cl_recv(cl, (u8 *)reply, if_version_length, 0, 0); - if (bytes_recv < if_version_length) { + if (bytes_recv < 0 || bytes_recv < if_version_length) { dev_err(bus->dev, "Could not read IF version\n"); ret = -EIO; goto err;

6 years, 11 months

1
0
0 0

FAILED: patch "[PATCH] mei: bus: type promotion bug in mei_nfc_if_version()" failed to apply to 4.18-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 4.18-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. thanks, greg k-h ------------------ original commit in Linus's tree ------------------ >From b40b3e9358fbafff6a4ba0f4b9658f6617146f9c Mon Sep 17 00:00:00 2001 From: Dan Carpenter <dan.carpenter(a)oracle.com> Date: Wed, 11 Jul 2018 15:29:31 +0300 Subject: [PATCH] mei: bus: type promotion bug in mei_nfc_if_version() We accidentally removed the check for negative returns without considering the issue of type promotion. The "if_version_length" variable is type size_t so if __mei_cl_recv() returns a negative then "bytes_recv" is type promoted to a high positive value and treated as success. Cc: <stable(a)vger.kernel.org> Fixes: 582ab27a063a ("mei: bus: fix received data size check in NFC fixup") Signed-off-by: Dan Carpenter <dan.carpenter(a)oracle.com> Signed-off-by: Tomas Winkler <tomas.winkler(a)intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> diff --git a/drivers/misc/mei/bus-fixup.c b/drivers/misc/mei/bus-fixup.c index e45fe826d87d..65e28be3c8cc 100644 --- a/drivers/misc/mei/bus-fixup.c +++ b/drivers/misc/mei/bus-fixup.c @@ -341,7 +341,7 @@ static int mei_nfc_if_version(struct mei_cl *cl, ret = 0; bytes_recv = __mei_cl_recv(cl, (u8 *)reply, if_version_length, 0, 0); - if (bytes_recv < if_version_length) { + if (bytes_recv < 0 || bytes_recv < if_version_length) { dev_err(bus->dev, "Could not read IF version\n"); ret = -EIO; goto err;

6 years, 11 months

1
0
0 0

FAILED: patch "[PATCH] scsi: mpt3sas: Fix calltrace observed while running IO &" failed to apply to 4.14-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 4.14-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. thanks, greg k-h ------------------ original commit in Linus's tree ------------------ >From e70183143cc472960bc60dfee1b7bbe1949feffb Mon Sep 17 00:00:00 2001 From: Sreekanth Reddy <sreekanth.reddy(a)broadcom.com> Date: Thu, 12 Jul 2018 12:53:29 -0400 Subject: [PATCH] scsi: mpt3sas: Fix calltrace observed while running IO & reset Below kernel BUG was observed while running IOs with host reset (issued from application), mpt3sas_cm0: diag reset: SUCCESS ------------[ cut here ]------------ WARNING: CPU: 12 PID: 4336 at drivers/scsi/mpt3sas/mpt3sas_base.c:3282 mpt3sas_base_clear_st+0x3d/0x40 [mpt3sas] Modules linked in: macsec tcp_diag udp_diag inet_diag unix_diag af_packet_diag netlink_diag binfmt_misc fuse xt_CHECKSUM ipt_MASQUERADE nf_nat_masquerade_ipv4 tun devlink ip6t_rpfilter ipt_REJECT nf_reject_ipv4 ip6t_REJECT nf_reject_ipv6 xt_conntrack ip_set nfnetlink ebtable_nat ebtable_broute bridge stp llc ip6table_nat nf_conntrack_ipv6 nf_defrag_ipv6 nf_nat_ipv6 ip6table_mangle ip6table_security ip6table_raw iptable_nat nf_conntrack_ipv4 nf_defrag_ipv4 nf_nat_ipv4 nf_nat nf_conntrack iptable_mangle iptable_security iptable_raw ebtable_filter ebtables ip6table_filter ip6_tables iptable_filter sunrpc vfat fat sb_edac intel_powerclamp coretemp intel_rapl iosf_mbi kvm_intel kvm irqbypass crc32_pclmul ghash_clmulni_intel aesni_intel lrw gf128mul glue_helper ablk_helper cryptd iTCO_wdt iTCO_vendor_support dcdbas pcspkr joydev ipmi_ssif ses enclosure sg ipmi_devintf acpi_pad ipmi_msghandler acpi_power_meter mei_me lpc_ich wmi mei shpchp ip_tables xfs libcrc32c sd_mod crc_t10dif crct10dif_generic ata_generic pata_acpi uas usb_storage mgag200 i2c_algo_bit drm_kms_helper syscopyarea sysfillrect sysimgblt fb_sys_fops ttm drm ata_piix mpt3sas libata crct10dif_pclmul crct10dif_common tg3 crc32c_intel i2c_core raid_class ptp scsi_transport_sas pps_core dm_mirror dm_region_hash dm_log dm_mod CPU: 12 PID: 4336 Comm: python Kdump: loaded Tainted: G W ------------ 3.10.0-875.el7.brdc.x86_64 #1 Hardware name: Dell Inc. PowerEdge R820/0YWR73, BIOS 1.5.0 03/08/2013 Call Trace: [<ffffffff9cf16583>] dump_stack+0x19/0x1b [<ffffffff9c891698>] __warn+0xd8/0x100 [<ffffffff9c8917dd>] warn_slowpath_null+0x1d/0x20 [<ffffffffc04f3f4d>] mpt3sas_base_clear_st+0x3d/0x40 [mpt3sas] [<ffffffffc05047d2>] _scsih_flush_running_cmds+0x92/0xe0 [mpt3sas] [<ffffffffc05095db>] mpt3sas_scsih_reset_handler+0x43b/0xaf0 [mpt3sas] [<ffffffff9c894829>] ? vprintk_default+0x29/0x40 [<ffffffff9cf10531>] ? printk+0x60/0x77 [<ffffffffc04f06c8>] ? _base_diag_reset+0x238/0x340 [mpt3sas] [<ffffffffc04f794d>] mpt3sas_base_hard_reset_handler+0x1ad/0x420 [mpt3sas] [<ffffffffc05132b9>] _ctl_ioctl_main.isra.12+0x11b9/0x1200 [mpt3sas] [<ffffffffc068d585>] ? xfs_file_aio_write+0x155/0x1b0 [xfs] [<ffffffff9ca1a4e3>] ? do_sync_write+0x93/0xe0 [<ffffffffc051337a>] _ctl_ioctl+0x1a/0x20 [mpt3sas] [<ffffffff9ca2fe90>] do_vfs_ioctl+0x350/0x560 [<ffffffff9ca1dec1>] ? __sb_end_write+0x31/0x60 [<ffffffff9ca30141>] SyS_ioctl+0xa1/0xc0 [<ffffffff9cf28715>] ? system_call_after_swapgs+0xa2/0x146 [<ffffffff9cf287d5>] system_call_fastpath+0x1c/0x21 [<ffffffff9cf28721>] ? system_call_after_swapgs+0xae/0x146 ---[ end trace 5dac5b98d89aaa3c ]--- ------------[ cut here ]------------ kernel BUG at block/blk-core.c:1476! invalid opcode: 0000 [#1] SMP Modules linked in: macsec tcp_diag udp_diag inet_diag unix_diag af_packet_diag netlink_diag binfmt_misc fuse xt_CHECKSUM ipt_MASQUERADE nf_nat_masquerade_ipv4 tun devlink ip6t_rpfilter ipt_REJECT nf_reject_ipv4 ip6t_REJECT nf_reject_ipv6 xt_conntrack ip_set nfnetlink ebtable_nat ebtable_broute bridge stp llc ip6table_nat nf_conntrack_ipv6 nf_defrag_ipv6 nf_nat_ipv6 ip6table_mangle ip6table_security ip6table_raw iptable_nat nf_conntrack_ipv4 nf_defrag_ipv4 nf_nat_ipv4 nf_nat nf_conntrack iptable_mangle iptable_security iptable_raw ebtable_filter ebtables ip6table_filter ip6_tables iptable_filter sunrpc vfat fat sb_edac intel_powerclamp coretemp intel_rapl iosf_mbi kvm_intel kvm irqbypass crc32_pclmul ghash_clmulni_intel aesni_intel lrw gf128mul glue_helper ablk_helper cryptd iTCO_wdt iTCO_vendor_support dcdbas pcspkr joydev ipmi_ssif ses enclosure sg ipmi_devintf acpi_pad ipmi_msghandler acpi_power_meter mei_me lpc_ich wmi mei shpchp ip_tables xfs libcrc32c sd_mod crc_t10dif crct10dif_generic ata_generic pata_acpi uas usb_storage mgag200 i2c_algo_bit drm_kms_helper syscopyarea sysfillrect sysimgblt fb_sys_fops ttm drm ata_piix mpt3sas libata crct10dif_pclmul crct10dif_common tg3 crc32c_intel i2c_core raid_class ptp scsi_transport_sas pps_core dm_mirror dm_region_hash dm_log dm_mod CPU: 12 PID: 4336 Comm: python Kdump: loaded Tainted: G W ------------ 3.10.0-875.el7.brdc.x86_64 #1 Hardware name: Dell Inc. PowerEdge R820/0YWR73, BIOS 1.5.0 03/08/2013 task: ffff903fc96e0fd0 ti: ffff903fb1eec000 task.ti: ffff903fb1eec000 RIP: 0010:[<ffffffff9cb19ec0>] [<ffffffff9cb19ec0>] blk_requeue_request+0x90/0xa0 RSP: 0018:ffff903c6b783dc0 EFLAGS: 00010087 RAX: ffff903bb67026d0 RBX: ffff903b7d6a6140 RCX: dead000000000200 RDX: ffff903bb67026d0 RSI: ffff903bb6702580 RDI: ffff903bb67026d0 RBP: ffff903c6b783dd8 R08: ffff903bb67026d0 R09: ffffd97e80000000 R10: ffff903c658bac00 R11: 0000000000000000 R12: ffff903bb6702580 R13: ffff903fa9a292f0 R14: 0000000000000246 R15: 0000000000001057 FS: 00007f7026f5b740(0000) GS:ffff903c6b780000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 00007f298877c004 CR3: 00000000caf36000 CR4: 00000000000607e0 Call Trace: <IRQ> [<ffffffff9cca68ff>] __scsi_queue_insert+0xbf/0x110 [<ffffffff9cca79ca>] scsi_io_completion+0x5da/0x6a0 [<ffffffff9cc9ca3c>] scsi_finish_command+0xdc/0x140 [<ffffffff9cca6aa2>] scsi_softirq_done+0x132/0x160 [<ffffffff9cb240c6>] blk_done_softirq+0x96/0xc0 [<ffffffff9c89a905>] __do_softirq+0xf5/0x280 [<ffffffff9cf2bd2c>] call_softirq+0x1c/0x30 [<ffffffff9c82d625>] do_softirq+0x65/0xa0 [<ffffffff9c89ac85>] irq_exit+0x105/0x110 [<ffffffff9cf2d0a8>] smp_apic_timer_interrupt+0x48/0x60 [<ffffffff9cf297f2>] apic_timer_interrupt+0x162/0x170 <EOI> [<ffffffff9cca5f41>] ? scsi_done+0x21/0x60 [<ffffffff9cb5ac18>] ? delay_tsc+0x38/0x60 [<ffffffff9cb5ab5d>] __const_udelay+0x2d/0x30 [<ffffffffc04effde>] _base_handshake_req_reply_wait+0x8e/0x4a0 [mpt3sas] [<ffffffffc04f0b13>] _base_get_ioc_facts+0x123/0x590 [mpt3sas] [<ffffffffc04f06c8>] ? _base_diag_reset+0x238/0x340 [mpt3sas] [<ffffffffc04f7993>] mpt3sas_base_hard_reset_handler+0x1f3/0x420 [mpt3sas] [<ffffffffc05132b9>] _ctl_ioctl_main.isra.12+0x11b9/0x1200 [mpt3sas] [<ffffffffc068d585>] ? xfs_file_aio_write+0x155/0x1b0 [xfs] [<ffffffff9ca1a4e3>] ? do_sync_write+0x93/0xe0 [<ffffffffc051337a>] _ctl_ioctl+0x1a/0x20 [mpt3sas] [<ffffffff9ca2fe90>] do_vfs_ioctl+0x350/0x560 [<ffffffff9ca1dec1>] ? __sb_end_write+0x31/0x60 [<ffffffff9ca30141>] SyS_ioctl+0xa1/0xc0 [<ffffffff9cf28715>] ? system_call_after_swapgs+0xa2/0x146 [<ffffffff9cf287d5>] system_call_fastpath+0x1c/0x21 [<ffffffff9cf28721>] ? system_call_after_swapgs+0xae/0x146 Code: 83 c3 10 4c 89 e2 4c 89 ee e8 8d 21 04 00 48 8b 03 48 85 c0 75 e5 41 f6 44 24 4a 10 74 ad 4c 89 e6 4c 89 ef e8 b2 42 00 00 eb a0 <0f> 0b 0f 1f 40 00 66 2e 0f 1f 84 00 00 00 00 00 66 66 66 66 90 RIP [<ffffffff9cb19ec0>] blk_requeue_request+0x90/0xa0 RSP <ffff903c6b783dc0> As a part of host reset operation, driver will flushout all IOs outstanding at driver level with "DID_RESET" result. To find which are all commands outstanding at the driver level, driver loops with smid starting from one to HBA queue depth and calls mpt3sas_scsih_scsi_lookup_get() to get scmd as shown below for (smid = 1; smid <= ioc->scsiio_depth; smid++) { scmd = mpt3sas_scsih_scsi_lookup_get(ioc, smid); if (!scmd) continue; But in mpt3sas_scsih_scsi_lookup_get() function, driver returns some scsi cmnds which are not outstanding at the driver level (possibly request is constructed at block layer since QUEUE_FLAG_QUIESCED is not set. Even if driver uses scsi_block_requests and scsi_unblock_requests, issue still persists as they will be just blocking further IO from scsi layer and not from block layer) and these commands are flushed with DID_RESET host bytes thus resulting into above kernel BUG. This issue got introduced by commit dbec4c9040ed ("scsi: mpt3sas: lockless command submission"). To fix this issue, we have modified the mpt3sas_scsih_scsi_lookup_get() to check for smid equals to zero (note: whenever any scsi cmnd is processing at the driver level then smid for that scsi cmnd will be non-zero, always it starts from one) before it returns the scmd pointer to the caller. If smid is zero then this function returns scmd pointer as NULL and driver won't flushout those scsi cmnds at driver level with DID_RESET host byte thus this issue will not be observed. [mkp: amended with updated fix from Sreekanth] Signed-off-by: Sreekanth Reddy <sreekanth.reddy(a)broadcom.com> Fixes: dbec4c9040ed ("scsi: mpt3sas: lockless command submission") Cc: stable(a)vger.kernel.org # v4.16+ Reviewed-by: Tomas Henzl <thenzl(a)redhat.com> Reviewed-by: Bart Van Assche <bart.vanassche(a)wdc.com> Signed-off-by: Martin K. Petersen <martin.petersen(a)oracle.com> diff --git a/drivers/scsi/mpt3sas/mpt3sas_base.c b/drivers/scsi/mpt3sas/mpt3sas_base.c index 17b6b0a3455f..836b2e14e3b7 100644 --- a/drivers/scsi/mpt3sas/mpt3sas_base.c +++ b/drivers/scsi/mpt3sas/mpt3sas_base.c @@ -3264,6 +3264,7 @@ void mpt3sas_base_clear_st(struct MPT3SAS_ADAPTER *ioc, st->cb_idx = 0xFF; st->direct_io = 0; atomic_set(&ioc->chain_lookup[st->smid - 1].chain_offset, 0); + st->smid = 0; } /** diff --git a/drivers/scsi/mpt3sas/mpt3sas_scsih.c b/drivers/scsi/mpt3sas/mpt3sas_scsih.c index 061736f8025b..8dd3d6720217 100644 --- a/drivers/scsi/mpt3sas/mpt3sas_scsih.c +++ b/drivers/scsi/mpt3sas/mpt3sas_scsih.c @@ -1482,7 +1482,7 @@ mpt3sas_scsih_scsi_lookup_get(struct MPT3SAS_ADAPTER *ioc, u16 smid) scmd = scsi_host_find_tag(ioc->shost, unique_tag); if (scmd) { st = scsi_cmd_priv(scmd); - if (st->cb_idx == 0xFF) + if (st->cb_idx == 0xFF || st->smid == 0) scmd = NULL; } }

6 years, 11 months

1
0
0 0

FAILED: patch "[PATCH] tpm: separate cmd_ready/go_idle from runtime_pm" failed to apply to 4.14-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 4.14-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. thanks, greg k-h ------------------ original commit in Linus's tree ------------------ >From 627448e85c766587f6fdde1ea3886d6615081c77 Mon Sep 17 00:00:00 2001 From: Tomas Winkler <tomas.winkler(a)intel.com> Date: Thu, 28 Jun 2018 18:13:33 +0300 Subject: [PATCH] tpm: separate cmd_ready/go_idle from runtime_pm Fix tpm ptt initialization error: tpm tpm0: A TPM error (378) occurred get tpm pcr allocation. We cannot use go_idle cmd_ready commands via runtime_pm handles as with the introduction of localities this is no longer an optional feature, while runtime pm can be not enabled. Though cmd_ready/go_idle provides a power saving, it's also a part of TPM2 protocol and should be called explicitly. This patch exposes cmd_read/go_idle via tpm class ops and removes runtime pm support as it is not used by any driver. When calling from nested context always use both flags: TPM_TRANSMIT_UNLOCKED and TPM_TRANSMIT_RAW. Both are needed to resolve tpm spaces and locality request recursive calls to tpm_transmit(). TPM_TRANSMIT_RAW should never be used standalone as it will fail on double locking. While TPM_TRANSMIT_UNLOCKED standalone should be called from non-recursive locked contexts. New wrappers are added tpm_cmd_ready() and tpm_go_idle() to streamline tpm_try_transmit code. tpm_crb no longer needs own power saving functions and can drop using tpm_pm_suspend/resume. This patch cannot be really separated from the locality fix. Fixes: 888d867df441 (tpm: cmd_ready command can be issued only after granting locality) Cc: stable(a)vger.kernel.org Fixes: 888d867df441 (tpm: cmd_ready command can be issued only after granting locality) Signed-off-by: Tomas Winkler <tomas.winkler(a)intel.com> Tested-by: Jarkko Sakkinen <jarkko.sakkinen(a)linux.intel.com> Reviewed-by: Jarkko Sakkinen <jarkko.sakkinen(a)linux.intel.com> Signed-off-by: Jarkko Sakkinen <jarkko.sakkinen(a)linux.intel.com> diff --git a/drivers/char/tpm/tpm-interface.c b/drivers/char/tpm/tpm-interface.c index e32f6e85dc6d..31b86e027f9d 100644 --- a/drivers/char/tpm/tpm-interface.c +++ b/drivers/char/tpm/tpm-interface.c @@ -29,7 +29,6 @@ #include <linux/mutex.h> #include <linux/spinlock.h> #include <linux/freezer.h> -#include <linux/pm_runtime.h> #include <linux/tpm_eventlog.h> #include "tpm.h" @@ -369,10 +368,13 @@ err_len: return -EINVAL; } -static int tpm_request_locality(struct tpm_chip *chip) +static int tpm_request_locality(struct tpm_chip *chip, unsigned int flags) { int rc; + if (flags & TPM_TRANSMIT_RAW) + return 0; + if (!chip->ops->request_locality) return 0; @@ -385,10 +387,13 @@ static int tpm_request_locality(struct tpm_chip *chip) return 0; } -static void tpm_relinquish_locality(struct tpm_chip *chip) +static void tpm_relinquish_locality(struct tpm_chip *chip, unsigned int flags) { int rc; + if (flags & TPM_TRANSMIT_RAW) + return; + if (!chip->ops->relinquish_locality) return; @@ -399,6 +404,28 @@ static void tpm_relinquish_locality(struct tpm_chip *chip) chip->locality = -1; } +static int tpm_cmd_ready(struct tpm_chip *chip, unsigned int flags) +{ + if (flags & TPM_TRANSMIT_RAW) + return 0; + + if (!chip->ops->cmd_ready) + return 0; + + return chip->ops->cmd_ready(chip); +} + +static int tpm_go_idle(struct tpm_chip *chip, unsigned int flags) +{ + if (flags & TPM_TRANSMIT_RAW) + return 0; + + if (!chip->ops->go_idle) + return 0; + + return chip->ops->go_idle(chip); +} + static ssize_t tpm_try_transmit(struct tpm_chip *chip, struct tpm_space *space, u8 *buf, size_t bufsiz, @@ -449,14 +476,15 @@ static ssize_t tpm_try_transmit(struct tpm_chip *chip, /* Store the decision as chip->locality will be changed. */ need_locality = chip->locality == -1; - if (!(flags & TPM_TRANSMIT_RAW) && need_locality) { - rc = tpm_request_locality(chip); + if (need_locality) { + rc = tpm_request_locality(chip, flags); if (rc < 0) goto out_no_locality; } - if (chip->dev.parent) - pm_runtime_get_sync(chip->dev.parent); + rc = tpm_cmd_ready(chip, flags); + if (rc) + goto out; rc = tpm2_prepare_space(chip, space, ordinal, buf); if (rc) @@ -516,13 +544,16 @@ out_recv: } rc = tpm2_commit_space(chip, space, ordinal, buf, &len); + if (rc) + dev_err(&chip->dev, "tpm2_commit_space: error %d\n", rc); out: - if (chip->dev.parent) - pm_runtime_put_sync(chip->dev.parent); + rc = tpm_go_idle(chip, flags); + if (rc) + goto out; if (need_locality) - tpm_relinquish_locality(chip); + tpm_relinquish_locality(chip, flags); out_no_locality: if (chip->ops->clk_enable != NULL) diff --git a/drivers/char/tpm/tpm.h b/drivers/char/tpm/tpm.h index 9824cccb2c76..17833ef63f6c 100644 --- a/drivers/char/tpm/tpm.h +++ b/drivers/char/tpm/tpm.h @@ -512,9 +512,17 @@ extern const struct file_operations tpm_fops; extern const struct file_operations tpmrm_fops; extern struct idr dev_nums_idr; +/** + * enum tpm_transmit_flags + * + * @TPM_TRANSMIT_UNLOCKED: used to lock sequence of tpm_transmit calls. + * @TPM_TRANSMIT_RAW: prevent recursive calls into setup steps + * (go idle, locality,..). Always use with UNLOCKED + * as it will fail on double locking. + */ enum tpm_transmit_flags { - TPM_TRANSMIT_UNLOCKED = BIT(0), - TPM_TRANSMIT_RAW = BIT(1), + TPM_TRANSMIT_UNLOCKED = BIT(0), + TPM_TRANSMIT_RAW = BIT(1), }; ssize_t tpm_transmit(struct tpm_chip *chip, struct tpm_space *space, diff --git a/drivers/char/tpm/tpm2-space.c b/drivers/char/tpm/tpm2-space.c index 6122d3276f72..11c85ed8c113 100644 --- a/drivers/char/tpm/tpm2-space.c +++ b/drivers/char/tpm/tpm2-space.c @@ -39,7 +39,8 @@ static void tpm2_flush_sessions(struct tpm_chip *chip, struct tpm_space *space) for (i = 0; i < ARRAY_SIZE(space->session_tbl); i++) { if (space->session_tbl[i]) tpm2_flush_context_cmd(chip, space->session_tbl[i], - TPM_TRANSMIT_UNLOCKED); + TPM_TRANSMIT_UNLOCKED | + TPM_TRANSMIT_RAW); } } @@ -84,7 +85,7 @@ static int tpm2_load_context(struct tpm_chip *chip, u8 *buf, tpm_buf_append(&tbuf, &buf[*offset], body_size); rc = tpm_transmit_cmd(chip, NULL, tbuf.data, PAGE_SIZE, 4, - TPM_TRANSMIT_UNLOCKED, NULL); + TPM_TRANSMIT_UNLOCKED | TPM_TRANSMIT_RAW, NULL); if (rc < 0) { dev_warn(&chip->dev, "%s: failed with a system error %d\n", __func__, rc); @@ -133,7 +134,7 @@ static int tpm2_save_context(struct tpm_chip *chip, u32 handle, u8 *buf, tpm_buf_append_u32(&tbuf, handle); rc = tpm_transmit_cmd(chip, NULL, tbuf.data, PAGE_SIZE, 0, - TPM_TRANSMIT_UNLOCKED, NULL); + TPM_TRANSMIT_UNLOCKED | TPM_TRANSMIT_RAW, NULL); if (rc < 0) { dev_warn(&chip->dev, "%s: failed with a system error %d\n", __func__, rc); @@ -170,7 +171,8 @@ static void tpm2_flush_space(struct tpm_chip *chip) for (i = 0; i < ARRAY_SIZE(space->context_tbl); i++) if (space->context_tbl[i] && ~space->context_tbl[i]) tpm2_flush_context_cmd(chip, space->context_tbl[i], - TPM_TRANSMIT_UNLOCKED); + TPM_TRANSMIT_UNLOCKED | + TPM_TRANSMIT_RAW); tpm2_flush_sessions(chip, space); } @@ -377,7 +379,8 @@ static int tpm2_map_response_header(struct tpm_chip *chip, u32 cc, u8 *rsp, return 0; out_no_slots: - tpm2_flush_context_cmd(chip, phandle, TPM_TRANSMIT_UNLOCKED); + tpm2_flush_context_cmd(chip, phandle, + TPM_TRANSMIT_UNLOCKED | TPM_TRANSMIT_RAW); dev_warn(&chip->dev, "%s: out of slots for 0x%08X\n", __func__, phandle); return -ENOMEM; @@ -465,7 +468,8 @@ static int tpm2_save_space(struct tpm_chip *chip) return rc; tpm2_flush_context_cmd(chip, space->context_tbl[i], - TPM_TRANSMIT_UNLOCKED); + TPM_TRANSMIT_UNLOCKED | + TPM_TRANSMIT_RAW); space->context_tbl[i] = ~0; } diff --git a/drivers/char/tpm/tpm_crb.c b/drivers/char/tpm/tpm_crb.c index 34fbc6cb097b..36952ef98f90 100644 --- a/drivers/char/tpm/tpm_crb.c +++ b/drivers/char/tpm/tpm_crb.c @@ -132,7 +132,7 @@ static bool crb_wait_for_reg_32(u32 __iomem *reg, u32 mask, u32 value, } /** - * crb_go_idle - request tpm crb device to go the idle state + * __crb_go_idle - request tpm crb device to go the idle state * * @dev: crb device * @priv: crb private data @@ -147,7 +147,7 @@ static bool crb_wait_for_reg_32(u32 __iomem *reg, u32 mask, u32 value, * * Return: 0 always */ -static int crb_go_idle(struct device *dev, struct crb_priv *priv) +static int __crb_go_idle(struct device *dev, struct crb_priv *priv) { if ((priv->sm == ACPI_TPM2_START_METHOD) || (priv->sm == ACPI_TPM2_COMMAND_BUFFER_WITH_START_METHOD) || @@ -163,11 +163,20 @@ static int crb_go_idle(struct device *dev, struct crb_priv *priv) dev_warn(dev, "goIdle timed out\n"); return -ETIME; } + return 0; } +static int crb_go_idle(struct tpm_chip *chip) +{ + struct device *dev = &chip->dev; + struct crb_priv *priv = dev_get_drvdata(dev); + + return __crb_go_idle(dev, priv); +} + /** - * crb_cmd_ready - request tpm crb device to enter ready state + * __crb_cmd_ready - request tpm crb device to enter ready state * * @dev: crb device * @priv: crb private data @@ -181,7 +190,7 @@ static int crb_go_idle(struct device *dev, struct crb_priv *priv) * * Return: 0 on success -ETIME on timeout; */ -static int crb_cmd_ready(struct device *dev, struct crb_priv *priv) +static int __crb_cmd_ready(struct device *dev, struct crb_priv *priv) { if ((priv->sm == ACPI_TPM2_START_METHOD) || (priv->sm == ACPI_TPM2_COMMAND_BUFFER_WITH_START_METHOD) || @@ -200,6 +209,14 @@ static int crb_cmd_ready(struct device *dev, struct crb_priv *priv) return 0; } +static int crb_cmd_ready(struct tpm_chip *chip) +{ + struct device *dev = &chip->dev; + struct crb_priv *priv = dev_get_drvdata(dev); + + return __crb_cmd_ready(dev, priv); +} + static int __crb_request_locality(struct device *dev, struct crb_priv *priv, int loc) { @@ -401,6 +418,8 @@ static const struct tpm_class_ops tpm_crb = { .send = crb_send, .cancel = crb_cancel, .req_canceled = crb_req_canceled, + .go_idle = crb_go_idle, + .cmd_ready = crb_cmd_ready, .request_locality = crb_request_locality, .relinquish_locality = crb_relinquish_locality, .req_complete_mask = CRB_DRV_STS_COMPLETE, @@ -520,7 +539,7 @@ static int crb_map_io(struct acpi_device *device, struct crb_priv *priv, * PTT HW bug w/a: wake up the device to access * possibly not retained registers. */ - ret = crb_cmd_ready(dev, priv); + ret = __crb_cmd_ready(dev, priv); if (ret) goto out_relinquish_locality; @@ -565,7 +584,7 @@ out: if (!ret) priv->cmd_size = cmd_size; - crb_go_idle(dev, priv); + __crb_go_idle(dev, priv); out_relinquish_locality: @@ -628,32 +647,7 @@ static int crb_acpi_add(struct acpi_device *device) chip->acpi_dev_handle = device->handle; chip->flags = TPM_CHIP_FLAG_TPM2; - rc = __crb_request_locality(dev, priv, 0); - if (rc) - return rc; - - rc = crb_cmd_ready(dev, priv); - if (rc) - goto out; - - pm_runtime_get_noresume(dev); - pm_runtime_set_active(dev); - pm_runtime_enable(dev); - - rc = tpm_chip_register(chip); - if (rc) { - crb_go_idle(dev, priv); - pm_runtime_put_noidle(dev); - pm_runtime_disable(dev); - goto out; - } - - pm_runtime_put_sync(dev); - -out: - __crb_relinquish_locality(dev, priv, 0); - - return rc; + return tpm_chip_register(chip); } static int crb_acpi_remove(struct acpi_device *device) @@ -663,52 +657,11 @@ static int crb_acpi_remove(struct acpi_device *device) tpm_chip_unregister(chip); - pm_runtime_disable(dev); - return 0; } -static int __maybe_unused crb_pm_runtime_suspend(struct device *dev) -{ - struct tpm_chip *chip = dev_get_drvdata(dev); - struct crb_priv *priv = dev_get_drvdata(&chip->dev); - - return crb_go_idle(dev, priv); -} - -static int __maybe_unused crb_pm_runtime_resume(struct device *dev) -{ - struct tpm_chip *chip = dev_get_drvdata(dev); - struct crb_priv *priv = dev_get_drvdata(&chip->dev); - - return crb_cmd_ready(dev, priv); -} - -static int __maybe_unused crb_pm_suspend(struct device *dev) -{ - int ret; - - ret = tpm_pm_suspend(dev); - if (ret) - return ret; - - return crb_pm_runtime_suspend(dev); -} - -static int __maybe_unused crb_pm_resume(struct device *dev) -{ - int ret; - - ret = crb_pm_runtime_resume(dev); - if (ret) - return ret; - - return tpm_pm_resume(dev); -} - static const struct dev_pm_ops crb_pm = { - SET_SYSTEM_SLEEP_PM_OPS(crb_pm_suspend, crb_pm_resume) - SET_RUNTIME_PM_OPS(crb_pm_runtime_suspend, crb_pm_runtime_resume, NULL) + SET_SYSTEM_SLEEP_PM_OPS(tpm_pm_suspend, tpm_pm_resume) }; static const struct acpi_device_id crb_device_ids[] = { diff --git a/include/linux/tpm.h b/include/linux/tpm.h index 06639fb6ab85..8eb5e5ebe136 100644 --- a/include/linux/tpm.h +++ b/include/linux/tpm.h @@ -43,6 +43,8 @@ struct tpm_class_ops { u8 (*status) (struct tpm_chip *chip); bool (*update_timeouts)(struct tpm_chip *chip, unsigned long *timeout_cap); + int (*go_idle)(struct tpm_chip *chip); + int (*cmd_ready)(struct tpm_chip *chip); int (*request_locality)(struct tpm_chip *chip, int loc); int (*relinquish_locality)(struct tpm_chip *chip, int loc); void (*clk_enable)(struct tpm_chip *chip, bool value);

6 years, 11 months

1
0
0 0

FAILED: patch "[PATCH] MIPS: Schedule on CPUs we need to lose FPU for a mode switch" failed to apply to 4.4-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 4.4-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. thanks, greg k-h ------------------ original commit in Linus's tree ------------------ >From 8c8d953c28000045e5e823f3398319f04d49a7f1 Mon Sep 17 00:00:00 2001 From: Paul Burton <paul.burton(a)mips.com> Date: Tue, 19 Dec 2017 15:11:08 -0800 Subject: [PATCH] MIPS: Schedule on CPUs we need to lose FPU for a mode switch Commit 6b8322576e9d ("MIPS: Force CPUs to lose FP context during mode switches") ensures that we react to PR_SET_FP_MODE prctl syscalls quickly by broadcasting an IPI in order to cause CPUs to lose FPU access when necessary. Whilst it achieves that, unfortunately it causes all sorts of strange race conditions because: 1) The IPI may arrive at a point where the FPU is in the process of being enabled, but that process is not yet complete leading to a state we aren't prepared to handle. For example: [ 370.215903] do_cpu invoked from kernel context![#1]: [ 370.221064] CPU: 0 PID: 963 Comm: fp-prctl Not tainted 4.9.0-rc5-00323-g210db32-dirty #226 [ 370.229420] task: a8000000fd672e00 task.stack: a8000000fd630000 [ 370.235399] $ 0 : 0000000000000000 0000000000000001 0000000000000001 a8000000fd630000 [ 370.243882] $ 4 : a8000000fd672e00 0000000000000000 0000000000000453 0000000000000000 [ 370.252317] $ 8 : 0000000000000000 a8000000fd637c28 1000000000000000 0000000000000010 [ 370.260753] $12 : 00000000140084e0 ffffffff80109c00 0000000000000000 0000000000000002 [ 370.269179] $16 : ffffffff8092f080 a8000000fd672e00 ffffffff80107fe8 a8000000fd485000 [ 370.277612] $20 : ffffffff8084d328 ffffffff80940000 0000000000000009 ffffffff80930000 [ 370.286038] $24 : 0000000000000000 900000001612048c [ 370.294476] $28 : a8000000fd630000 a8000000fd637ac0 ffffffff80937300 ffffffff8010807c [ 370.302909] Hi : 0000000000000000 [ 370.306595] Lo : 0000000000000200 [ 370.310376] epc : ffffffff80115d38 _save_fp+0x10/0xa0 [ 370.315784] ra : ffffffff8010807c prepare_for_fp_mode_switch+0x94/0x1b0 [ 370.322707] Status: 140084e2 KX SX UX KERNEL EXL [ 370.327980] Cause : 1080002c (ExcCode 0b) [ 370.332091] PrId : 0001a428 (MIPS P6600) [ 370.336179] Modules linked in: [ 370.339486] Process fp-prctl (pid: 963, threadinfo=a8000000fd630000, task=a8000000fd672e00, tls=00000000756e67d0) [ 370.349724] Stack : 0000000000000000 a8000000fd557dc0 0000000000000000 ffffffff801ca8e0 [ 370.358161] 0000000000000000 a8000000fd637b9c 0000000000000009 ffffffff80923780 [ 370.366575] ffffffff80850000 ffffffff8011610c 00000000000000b8 ffffffff801a5084 [ 370.374989] ffffffff8084a370 ffffffff8084a388 ffffffff80923780 ffffffff80923828 [ 370.383395] 0000000000010000 ffffffff809237a8 0000000000020000 ffffffff80a40000 [ 370.391817] 000000000000007c 00000000004a0000 00000000756dedd0 ffffffff801a5188 [ 370.400230] a800000002014900 0000000000000001 ffffffff80923780 0000000080923828 [ 370.408644] ffffffff80923780 ffffffff80923780 ffffffff80923828 ffffffff801a521c [ 370.417066] ffffffff80923780 ffffffff80923828 0000000000010000 ffffffff801a8f84 [ 370.425472] ffffffff80a40000 a8000000fd637c20 ffffffff80a39240 0000000000000001 [ 370.433885] ... [ 370.436562] Call Trace: [ 370.439222] [<ffffffff80115d38>] _save_fp+0x10/0xa0 [ 370.444305] [<ffffffff8010807c>] prepare_for_fp_mode_switch+0x94/0x1b0 [ 370.451035] [<ffffffff801ca8e0>] flush_smp_call_function_queue+0xf8/0x230 [ 370.457991] [<ffffffff8011610c>] ipi_call_interrupt+0xc/0x20 [ 370.463814] [<ffffffff801a5084>] __handle_irq_event_percpu+0xc4/0x1a8 [ 370.470404] [<ffffffff801a5188>] handle_irq_event_percpu+0x20/0x68 [ 370.476734] [<ffffffff801a521c>] handle_irq_event+0x4c/0x88 [ 370.482486] [<ffffffff801a8f84>] handle_edge_irq+0x12c/0x210 [ 370.488316] [<ffffffff801a47a0>] generic_handle_irq+0x38/0x48 [ 370.494280] [<ffffffff804a2dbc>] gic_handle_shared_int+0x194/0x268 [ 370.500616] [<ffffffff801a47a0>] generic_handle_irq+0x38/0x48 [ 370.506529] [<ffffffff80107e60>] do_IRQ+0x18/0x28 [ 370.511445] [<ffffffff804a1524>] plat_irq_dispatch+0xc4/0x140 [ 370.517339] [<ffffffff80106230>] ret_from_irq+0x0/0x4 [ 370.522583] [<ffffffff8010fad4>] do_ri+0x4fc/0x7e8 [ 370.527546] [<ffffffff80106220>] ret_from_exception+0x0/0x10 2) The IPI may arrive during kernel use of the FPU, since we generally only disable preemption around use of the FPU & leave interrupts enabled. This can lead to us unexpectedly losing access to the FPU in places where it previously had not been possible. For example: do_cpu invoked from kernel context![#2]: CPU: 2 PID: 7338 Comm: fp-prctl Tainted: G D 4.7.0-00424-g49b0c82 #2 task: 838e4000 ti: 88d38000 task.ti: 88d38000 $ 0 : 00000000 00000001 ffffffff 88d3fef8 $ 4 : 838e4000 88d38004 00000000 00000001 $ 8 : 3400fc01 801f8020 808e9100 24000000 $12 : dbffffff 807b69d8 807b0000 00000000 $16 : 00000000 80786150 00400fc4 809c0398 $20 : 809c0338 0040273c 88d3ff28 808e9d30 $24 : 808e9d30 00400fb4 $28 : 88d38000 88d3fe88 00000000 8011a2ac Hi : 0040273c Lo : 88d3ff28 epc : 80114178 _restore_fp+0x10/0xa0 ra : 8011a2ac mipsr2_decoder+0xd5c/0x1660 Status: 1400fc03 KERNEL EXL IE Cause : 1080002c (ExcCode 0b) PrId : 0001a920 (MIPS I6400) Modules linked in: Process fp-prctl (pid: 7338, threadinfo=88d38000, task=838e4000, tls=766527d0) Stack : 00000000 00000000 00000000 88d3fe98 00000000 00000000 809c0398 809c0338 808e9100 00000000 88d3ff28 00400fc4 00400fc4 0040273c 7fb69e18 004a0000 004a0000 004a0000 7664add0 8010de18 00000000 00000000 88d3fef8 88d3ff28 808e9100 00000000 766527d0 8010e534 000c0000 85755000 8181d580 00000000 00000000 00000000 004a0000 00000000 766527d0 7fb69e18 004a0000 80105c20 ... Call Trace: [<80114178>] _restore_fp+0x10/0xa0 [<8011a2ac>] mipsr2_decoder+0xd5c/0x1660 [<8010de18>] do_ri+0x90/0x6b8 [<80105c20>] ret_from_exception+0x0/0x10 At first glance a simple fix may seem to be to disable interrupts around kernel use of the FPU rather than merely preemption, however this would introduce further overhead outside of the mode switch path & doesn't solve the third problem: 3) The IPI may arrive whilst the kernel is running code that will lead to a preempt_disable() call & FPU usage soon. If this happens then the IPI will be serviced & we'll proceed to enable an FPU whilst the mode switch is in progress, leading to strange & inconsistent behaviour. Further to all of this is a separate but related problem: 4) There are various paths through which we may enable the FPU without the user having triggered a coprocessor 1 disabled exception. These paths are those in which we emulate instructions & then enable the FPU with the expectation that the user might execute an FP instruction shortly afterwards. However these paths have not previously checked whether an FP mode switch is underway for the task, and therefore could enable the FPU whilst such a mode switch is in progress leading to strange & inconsistent behaviour for user code. This patch fixes all of the above by taking a step back & re-examining our approach to FP mode switches. Up until now we have taken these basic steps: a) Prevent any threads that are part of the affected process from being able to obtain ownership of the FPU. b) Cause any threads that are part of the affected process and already have ownership of an FPU to lose it. c) Set the thread flags for each thread that is part of the affected process to reflect the new FP mode. d) Allow threads to obtain ownership of the FPU again. This approach is however more complex than necessary. All that we really require is that the mode switch has occurred for all threads that are part of the affected process before mips_set_process_fp_mode(), and thus the PR_SET_FP_MODE prctl() syscall, returns. This doesn't require that we stop threads from owning or using an FPU whilst a mode switch occurs, only that we force them to relinquish it after the mode switch has occurred such that they next own an FPU with the correct mode configured. Our basic steps therefore simplify to: A) Set the thread flags for each thread that is part of the affected process to reflect the new FP mode. B) Cause any threads that are part of the affected process and already have ownership of an FPU to lose it. We implement B) by forcing each CPU which might be running a thread which is part of the affected process to schedule a no-op function, which causes the affected thread to lose its FPU ownership when it is descheduled. The end result is simpler FP mode switching with less overhead in the FPU enable path (ie. enable_restore_fp_context()) and fewer moving parts. Signed-off-by: Paul Burton <paul.burton(a)mips.com> Fixes: 9791554b45a2 ("MIPS,prctl: add PR_[GS]ET_FP_MODE prctl options for MIPS") Fixes: 6b8322576e9d ("MIPS: Force CPUs to lose FP context during mode switches") Cc: James Hogan <jhogan(a)kernel.org> Cc: Ralf Baechle <ralf(a)linux-mips.org> Cc: linux-mips(a)linux-mips.org Cc: stable <stable(a)vger.kernel.org> # v4.0+ diff --git a/arch/mips/include/asm/mmu_context.h b/arch/mips/include/asm/mmu_context.h index da2004cef2d5..b509371a6b0c 100644 --- a/arch/mips/include/asm/mmu_context.h +++ b/arch/mips/include/asm/mmu_context.h @@ -126,8 +126,6 @@ init_new_context(struct task_struct *tsk, struct mm_struct *mm) for_each_possible_cpu(i) cpu_context(i, mm) = 0; - atomic_set(&mm->context.fp_mode_switching, 0); - mm->context.bd_emupage_allocmap = NULL; spin_lock_init(&mm->context.bd_emupage_lock); init_waitqueue_head(&mm->context.bd_emupage_queue); diff --git a/arch/mips/kernel/process.c b/arch/mips/kernel/process.c index 8d85046adcc8..fe6001d748cf 100644 --- a/arch/mips/kernel/process.c +++ b/arch/mips/kernel/process.c @@ -29,6 +29,7 @@ #include <linux/kallsyms.h> #include <linux/random.h> #include <linux/prctl.h> +#include <linux/cpu.h> #include <asm/asm.h> #include <asm/bootinfo.h> @@ -691,19 +692,25 @@ int mips_get_process_fp_mode(struct task_struct *task) return value; } -static void prepare_for_fp_mode_switch(void *info) +static long prepare_for_fp_mode_switch(void *unused) { - struct mm_struct *mm = info; - - if (current->mm == mm) - lose_fpu(1); + /* + * This is icky, but we use this to simply ensure that all CPUs have + * context switched, regardless of whether they were previously running + * kernel or user code. This ensures that no CPU currently has its FPU + * enabled, or is about to attempt to enable it through any path other + * than enable_restore_fp_context() which will wait appropriately for + * fp_mode_switching to be zero. + */ + return 0; } int mips_set_process_fp_mode(struct task_struct *task, unsigned int value) { const unsigned int known_bits = PR_FP_MODE_FR | PR_FP_MODE_FRE; struct task_struct *t; - int max_users; + struct cpumask process_cpus; + int cpu; /* If nothing to change, return right away, successfully. */ if (value == mips_get_process_fp_mode(task)) @@ -736,35 +743,7 @@ int mips_set_process_fp_mode(struct task_struct *task, unsigned int value) if (!(value & PR_FP_MODE_FR) && raw_cpu_has_fpu && cpu_has_mips_r6) return -EOPNOTSUPP; - /* Proceed with the mode switch */ - preempt_disable(); - - /* Save FP & vector context, then disable FPU & MSA */ - if (task->signal == current->signal) - lose_fpu(1); - - /* Prevent any threads from obtaining live FP context */ - atomic_set(&task->mm->context.fp_mode_switching, 1); - smp_mb__after_atomic(); - - /* - * If there are multiple online CPUs then force any which are running - * threads in this process to lose their FPU context, which they can't - * regain until fp_mode_switching is cleared later. - */ - if (num_online_cpus() > 1) { - /* No need to send an IPI for the local CPU */ - max_users = (task->mm == current->mm) ? 1 : 0; - - if (atomic_read(&current->mm->mm_users) > max_users) - smp_call_function(prepare_for_fp_mode_switch, - (void *)current->mm, 1); - } - - /* - * There are now no threads of the process with live FP context, so it - * is safe to proceed with the FP mode switch. - */ + /* Indicate the new FP mode in each thread */ for_each_thread(task, t) { /* Update desired FP register width */ if (value & PR_FP_MODE_FR) { @@ -781,9 +760,34 @@ int mips_set_process_fp_mode(struct task_struct *task, unsigned int value) clear_tsk_thread_flag(t, TIF_HYBRID_FPREGS); } - /* Allow threads to use FP again */ - atomic_set(&task->mm->context.fp_mode_switching, 0); - preempt_enable(); + /* + * We need to ensure that all threads in the process have switched mode + * before returning, in order to allow userland to not worry about + * races. We can do this by forcing all CPUs that any thread in the + * process may be running on to schedule something else - in this case + * prepare_for_fp_mode_switch(). + * + * We begin by generating a mask of all CPUs that any thread in the + * process may be running on. + */ + cpumask_clear(&process_cpus); + for_each_thread(task, t) + cpumask_set_cpu(task_cpu(t), &process_cpus); + + /* + * Now we schedule prepare_for_fp_mode_switch() on each of those CPUs. + * + * The CPUs may have rescheduled already since we switched mode or + * generated the cpumask, but that doesn't matter. If the task in this + * process is scheduled out then our scheduling + * prepare_for_fp_mode_switch() will simply be redundant. If it's + * scheduled in then it will already have picked up the new FP mode + * whilst doing so. + */ + get_online_cpus(); + for_each_cpu_and(cpu, &process_cpus, cpu_online_mask) + work_on_cpu(cpu, prepare_for_fp_mode_switch, NULL); + put_online_cpus(); wake_up_var(&task->mm->context.fp_mode_switching); diff --git a/arch/mips/kernel/traps.c b/arch/mips/kernel/traps.c index d67fa74622ee..4d9ca9b465ae 100644 --- a/arch/mips/kernel/traps.c +++ b/arch/mips/kernel/traps.c @@ -1220,13 +1220,6 @@ static int enable_restore_fp_context(int msa) { int err, was_fpu_owner, prior_msa; - /* - * If an FP mode switch is currently underway, wait for it to - * complete before proceeding. - */ - wait_var_event(&current->mm->context.fp_mode_switching, - !atomic_read(&current->mm->context.fp_mode_switching)); - if (!used_math()) { /* First time FP context user. */ preempt_disable();

6 years, 11 months

1
0
0 0

FAILED: patch "[PATCH] MIPS: Schedule on CPUs we need to lose FPU for a mode switch" failed to apply to 4.9-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 4.9-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. thanks, greg k-h ------------------ original commit in Linus's tree ------------------ >From 8c8d953c28000045e5e823f3398319f04d49a7f1 Mon Sep 17 00:00:00 2001 From: Paul Burton <paul.burton(a)mips.com> Date: Tue, 19 Dec 2017 15:11:08 -0800 Subject: [PATCH] MIPS: Schedule on CPUs we need to lose FPU for a mode switch Commit 6b8322576e9d ("MIPS: Force CPUs to lose FP context during mode switches") ensures that we react to PR_SET_FP_MODE prctl syscalls quickly by broadcasting an IPI in order to cause CPUs to lose FPU access when necessary. Whilst it achieves that, unfortunately it causes all sorts of strange race conditions because: 1) The IPI may arrive at a point where the FPU is in the process of being enabled, but that process is not yet complete leading to a state we aren't prepared to handle. For example: [ 370.215903] do_cpu invoked from kernel context![#1]: [ 370.221064] CPU: 0 PID: 963 Comm: fp-prctl Not tainted 4.9.0-rc5-00323-g210db32-dirty #226 [ 370.229420] task: a8000000fd672e00 task.stack: a8000000fd630000 [ 370.235399] $ 0 : 0000000000000000 0000000000000001 0000000000000001 a8000000fd630000 [ 370.243882] $ 4 : a8000000fd672e00 0000000000000000 0000000000000453 0000000000000000 [ 370.252317] $ 8 : 0000000000000000 a8000000fd637c28 1000000000000000 0000000000000010 [ 370.260753] $12 : 00000000140084e0 ffffffff80109c00 0000000000000000 0000000000000002 [ 370.269179] $16 : ffffffff8092f080 a8000000fd672e00 ffffffff80107fe8 a8000000fd485000 [ 370.277612] $20 : ffffffff8084d328 ffffffff80940000 0000000000000009 ffffffff80930000 [ 370.286038] $24 : 0000000000000000 900000001612048c [ 370.294476] $28 : a8000000fd630000 a8000000fd637ac0 ffffffff80937300 ffffffff8010807c [ 370.302909] Hi : 0000000000000000 [ 370.306595] Lo : 0000000000000200 [ 370.310376] epc : ffffffff80115d38 _save_fp+0x10/0xa0 [ 370.315784] ra : ffffffff8010807c prepare_for_fp_mode_switch+0x94/0x1b0 [ 370.322707] Status: 140084e2 KX SX UX KERNEL EXL [ 370.327980] Cause : 1080002c (ExcCode 0b) [ 370.332091] PrId : 0001a428 (MIPS P6600) [ 370.336179] Modules linked in: [ 370.339486] Process fp-prctl (pid: 963, threadinfo=a8000000fd630000, task=a8000000fd672e00, tls=00000000756e67d0) [ 370.349724] Stack : 0000000000000000 a8000000fd557dc0 0000000000000000 ffffffff801ca8e0 [ 370.358161] 0000000000000000 a8000000fd637b9c 0000000000000009 ffffffff80923780 [ 370.366575] ffffffff80850000 ffffffff8011610c 00000000000000b8 ffffffff801a5084 [ 370.374989] ffffffff8084a370 ffffffff8084a388 ffffffff80923780 ffffffff80923828 [ 370.383395] 0000000000010000 ffffffff809237a8 0000000000020000 ffffffff80a40000 [ 370.391817] 000000000000007c 00000000004a0000 00000000756dedd0 ffffffff801a5188 [ 370.400230] a800000002014900 0000000000000001 ffffffff80923780 0000000080923828 [ 370.408644] ffffffff80923780 ffffffff80923780 ffffffff80923828 ffffffff801a521c [ 370.417066] ffffffff80923780 ffffffff80923828 0000000000010000 ffffffff801a8f84 [ 370.425472] ffffffff80a40000 a8000000fd637c20 ffffffff80a39240 0000000000000001 [ 370.433885] ... [ 370.436562] Call Trace: [ 370.439222] [<ffffffff80115d38>] _save_fp+0x10/0xa0 [ 370.444305] [<ffffffff8010807c>] prepare_for_fp_mode_switch+0x94/0x1b0 [ 370.451035] [<ffffffff801ca8e0>] flush_smp_call_function_queue+0xf8/0x230 [ 370.457991] [<ffffffff8011610c>] ipi_call_interrupt+0xc/0x20 [ 370.463814] [<ffffffff801a5084>] __handle_irq_event_percpu+0xc4/0x1a8 [ 370.470404] [<ffffffff801a5188>] handle_irq_event_percpu+0x20/0x68 [ 370.476734] [<ffffffff801a521c>] handle_irq_event+0x4c/0x88 [ 370.482486] [<ffffffff801a8f84>] handle_edge_irq+0x12c/0x210 [ 370.488316] [<ffffffff801a47a0>] generic_handle_irq+0x38/0x48 [ 370.494280] [<ffffffff804a2dbc>] gic_handle_shared_int+0x194/0x268 [ 370.500616] [<ffffffff801a47a0>] generic_handle_irq+0x38/0x48 [ 370.506529] [<ffffffff80107e60>] do_IRQ+0x18/0x28 [ 370.511445] [<ffffffff804a1524>] plat_irq_dispatch+0xc4/0x140 [ 370.517339] [<ffffffff80106230>] ret_from_irq+0x0/0x4 [ 370.522583] [<ffffffff8010fad4>] do_ri+0x4fc/0x7e8 [ 370.527546] [<ffffffff80106220>] ret_from_exception+0x0/0x10 2) The IPI may arrive during kernel use of the FPU, since we generally only disable preemption around use of the FPU & leave interrupts enabled. This can lead to us unexpectedly losing access to the FPU in places where it previously had not been possible. For example: do_cpu invoked from kernel context![#2]: CPU: 2 PID: 7338 Comm: fp-prctl Tainted: G D 4.7.0-00424-g49b0c82 #2 task: 838e4000 ti: 88d38000 task.ti: 88d38000 $ 0 : 00000000 00000001 ffffffff 88d3fef8 $ 4 : 838e4000 88d38004 00000000 00000001 $ 8 : 3400fc01 801f8020 808e9100 24000000 $12 : dbffffff 807b69d8 807b0000 00000000 $16 : 00000000 80786150 00400fc4 809c0398 $20 : 809c0338 0040273c 88d3ff28 808e9d30 $24 : 808e9d30 00400fb4 $28 : 88d38000 88d3fe88 00000000 8011a2ac Hi : 0040273c Lo : 88d3ff28 epc : 80114178 _restore_fp+0x10/0xa0 ra : 8011a2ac mipsr2_decoder+0xd5c/0x1660 Status: 1400fc03 KERNEL EXL IE Cause : 1080002c (ExcCode 0b) PrId : 0001a920 (MIPS I6400) Modules linked in: Process fp-prctl (pid: 7338, threadinfo=88d38000, task=838e4000, tls=766527d0) Stack : 00000000 00000000 00000000 88d3fe98 00000000 00000000 809c0398 809c0338 808e9100 00000000 88d3ff28 00400fc4 00400fc4 0040273c 7fb69e18 004a0000 004a0000 004a0000 7664add0 8010de18 00000000 00000000 88d3fef8 88d3ff28 808e9100 00000000 766527d0 8010e534 000c0000 85755000 8181d580 00000000 00000000 00000000 004a0000 00000000 766527d0 7fb69e18 004a0000 80105c20 ... Call Trace: [<80114178>] _restore_fp+0x10/0xa0 [<8011a2ac>] mipsr2_decoder+0xd5c/0x1660 [<8010de18>] do_ri+0x90/0x6b8 [<80105c20>] ret_from_exception+0x0/0x10 At first glance a simple fix may seem to be to disable interrupts around kernel use of the FPU rather than merely preemption, however this would introduce further overhead outside of the mode switch path & doesn't solve the third problem: 3) The IPI may arrive whilst the kernel is running code that will lead to a preempt_disable() call & FPU usage soon. If this happens then the IPI will be serviced & we'll proceed to enable an FPU whilst the mode switch is in progress, leading to strange & inconsistent behaviour. Further to all of this is a separate but related problem: 4) There are various paths through which we may enable the FPU without the user having triggered a coprocessor 1 disabled exception. These paths are those in which we emulate instructions & then enable the FPU with the expectation that the user might execute an FP instruction shortly afterwards. However these paths have not previously checked whether an FP mode switch is underway for the task, and therefore could enable the FPU whilst such a mode switch is in progress leading to strange & inconsistent behaviour for user code. This patch fixes all of the above by taking a step back & re-examining our approach to FP mode switches. Up until now we have taken these basic steps: a) Prevent any threads that are part of the affected process from being able to obtain ownership of the FPU. b) Cause any threads that are part of the affected process and already have ownership of an FPU to lose it. c) Set the thread flags for each thread that is part of the affected process to reflect the new FP mode. d) Allow threads to obtain ownership of the FPU again. This approach is however more complex than necessary. All that we really require is that the mode switch has occurred for all threads that are part of the affected process before mips_set_process_fp_mode(), and thus the PR_SET_FP_MODE prctl() syscall, returns. This doesn't require that we stop threads from owning or using an FPU whilst a mode switch occurs, only that we force them to relinquish it after the mode switch has occurred such that they next own an FPU with the correct mode configured. Our basic steps therefore simplify to: A) Set the thread flags for each thread that is part of the affected process to reflect the new FP mode. B) Cause any threads that are part of the affected process and already have ownership of an FPU to lose it. We implement B) by forcing each CPU which might be running a thread which is part of the affected process to schedule a no-op function, which causes the affected thread to lose its FPU ownership when it is descheduled. The end result is simpler FP mode switching with less overhead in the FPU enable path (ie. enable_restore_fp_context()) and fewer moving parts. Signed-off-by: Paul Burton <paul.burton(a)mips.com> Fixes: 9791554b45a2 ("MIPS,prctl: add PR_[GS]ET_FP_MODE prctl options for MIPS") Fixes: 6b8322576e9d ("MIPS: Force CPUs to lose FP context during mode switches") Cc: James Hogan <jhogan(a)kernel.org> Cc: Ralf Baechle <ralf(a)linux-mips.org> Cc: linux-mips(a)linux-mips.org Cc: stable <stable(a)vger.kernel.org> # v4.0+ diff --git a/arch/mips/include/asm/mmu_context.h b/arch/mips/include/asm/mmu_context.h index da2004cef2d5..b509371a6b0c 100644 --- a/arch/mips/include/asm/mmu_context.h +++ b/arch/mips/include/asm/mmu_context.h @@ -126,8 +126,6 @@ init_new_context(struct task_struct *tsk, struct mm_struct *mm) for_each_possible_cpu(i) cpu_context(i, mm) = 0; - atomic_set(&mm->context.fp_mode_switching, 0); - mm->context.bd_emupage_allocmap = NULL; spin_lock_init(&mm->context.bd_emupage_lock); init_waitqueue_head(&mm->context.bd_emupage_queue); diff --git a/arch/mips/kernel/process.c b/arch/mips/kernel/process.c index 8d85046adcc8..fe6001d748cf 100644 --- a/arch/mips/kernel/process.c +++ b/arch/mips/kernel/process.c @@ -29,6 +29,7 @@ #include <linux/kallsyms.h> #include <linux/random.h> #include <linux/prctl.h> +#include <linux/cpu.h> #include <asm/asm.h> #include <asm/bootinfo.h> @@ -691,19 +692,25 @@ int mips_get_process_fp_mode(struct task_struct *task) return value; } -static void prepare_for_fp_mode_switch(void *info) +static long prepare_for_fp_mode_switch(void *unused) { - struct mm_struct *mm = info; - - if (current->mm == mm) - lose_fpu(1); + /* + * This is icky, but we use this to simply ensure that all CPUs have + * context switched, regardless of whether they were previously running + * kernel or user code. This ensures that no CPU currently has its FPU + * enabled, or is about to attempt to enable it through any path other + * than enable_restore_fp_context() which will wait appropriately for + * fp_mode_switching to be zero. + */ + return 0; } int mips_set_process_fp_mode(struct task_struct *task, unsigned int value) { const unsigned int known_bits = PR_FP_MODE_FR | PR_FP_MODE_FRE; struct task_struct *t; - int max_users; + struct cpumask process_cpus; + int cpu; /* If nothing to change, return right away, successfully. */ if (value == mips_get_process_fp_mode(task)) @@ -736,35 +743,7 @@ int mips_set_process_fp_mode(struct task_struct *task, unsigned int value) if (!(value & PR_FP_MODE_FR) && raw_cpu_has_fpu && cpu_has_mips_r6) return -EOPNOTSUPP; - /* Proceed with the mode switch */ - preempt_disable(); - - /* Save FP & vector context, then disable FPU & MSA */ - if (task->signal == current->signal) - lose_fpu(1); - - /* Prevent any threads from obtaining live FP context */ - atomic_set(&task->mm->context.fp_mode_switching, 1); - smp_mb__after_atomic(); - - /* - * If there are multiple online CPUs then force any which are running - * threads in this process to lose their FPU context, which they can't - * regain until fp_mode_switching is cleared later. - */ - if (num_online_cpus() > 1) { - /* No need to send an IPI for the local CPU */ - max_users = (task->mm == current->mm) ? 1 : 0; - - if (atomic_read(&current->mm->mm_users) > max_users) - smp_call_function(prepare_for_fp_mode_switch, - (void *)current->mm, 1); - } - - /* - * There are now no threads of the process with live FP context, so it - * is safe to proceed with the FP mode switch. - */ + /* Indicate the new FP mode in each thread */ for_each_thread(task, t) { /* Update desired FP register width */ if (value & PR_FP_MODE_FR) { @@ -781,9 +760,34 @@ int mips_set_process_fp_mode(struct task_struct *task, unsigned int value) clear_tsk_thread_flag(t, TIF_HYBRID_FPREGS); } - /* Allow threads to use FP again */ - atomic_set(&task->mm->context.fp_mode_switching, 0); - preempt_enable(); + /* + * We need to ensure that all threads in the process have switched mode + * before returning, in order to allow userland to not worry about + * races. We can do this by forcing all CPUs that any thread in the + * process may be running on to schedule something else - in this case + * prepare_for_fp_mode_switch(). + * + * We begin by generating a mask of all CPUs that any thread in the + * process may be running on. + */ + cpumask_clear(&process_cpus); + for_each_thread(task, t) + cpumask_set_cpu(task_cpu(t), &process_cpus); + + /* + * Now we schedule prepare_for_fp_mode_switch() on each of those CPUs. + * + * The CPUs may have rescheduled already since we switched mode or + * generated the cpumask, but that doesn't matter. If the task in this + * process is scheduled out then our scheduling + * prepare_for_fp_mode_switch() will simply be redundant. If it's + * scheduled in then it will already have picked up the new FP mode + * whilst doing so. + */ + get_online_cpus(); + for_each_cpu_and(cpu, &process_cpus, cpu_online_mask) + work_on_cpu(cpu, prepare_for_fp_mode_switch, NULL); + put_online_cpus(); wake_up_var(&task->mm->context.fp_mode_switching); diff --git a/arch/mips/kernel/traps.c b/arch/mips/kernel/traps.c index d67fa74622ee..4d9ca9b465ae 100644 --- a/arch/mips/kernel/traps.c +++ b/arch/mips/kernel/traps.c @@ -1220,13 +1220,6 @@ static int enable_restore_fp_context(int msa) { int err, was_fpu_owner, prior_msa; - /* - * If an FP mode switch is currently underway, wait for it to - * complete before proceeding. - */ - wait_var_event(&current->mm->context.fp_mode_switching, - !atomic_read(&current->mm->context.fp_mode_switching)); - if (!used_math()) { /* First time FP context user. */ preempt_disable();

6 years, 11 months

1
0
0 0

FAILED: patch "[PATCH] MIPS: Schedule on CPUs we need to lose FPU for a mode switch" failed to apply to 4.14-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 4.14-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. thanks, greg k-h ------------------ original commit in Linus's tree ------------------ >From 8c8d953c28000045e5e823f3398319f04d49a7f1 Mon Sep 17 00:00:00 2001 From: Paul Burton <paul.burton(a)mips.com> Date: Tue, 19 Dec 2017 15:11:08 -0800 Subject: [PATCH] MIPS: Schedule on CPUs we need to lose FPU for a mode switch Commit 6b8322576e9d ("MIPS: Force CPUs to lose FP context during mode switches") ensures that we react to PR_SET_FP_MODE prctl syscalls quickly by broadcasting an IPI in order to cause CPUs to lose FPU access when necessary. Whilst it achieves that, unfortunately it causes all sorts of strange race conditions because: 1) The IPI may arrive at a point where the FPU is in the process of being enabled, but that process is not yet complete leading to a state we aren't prepared to handle. For example: [ 370.215903] do_cpu invoked from kernel context![#1]: [ 370.221064] CPU: 0 PID: 963 Comm: fp-prctl Not tainted 4.9.0-rc5-00323-g210db32-dirty #226 [ 370.229420] task: a8000000fd672e00 task.stack: a8000000fd630000 [ 370.235399] $ 0 : 0000000000000000 0000000000000001 0000000000000001 a8000000fd630000 [ 370.243882] $ 4 : a8000000fd672e00 0000000000000000 0000000000000453 0000000000000000 [ 370.252317] $ 8 : 0000000000000000 a8000000fd637c28 1000000000000000 0000000000000010 [ 370.260753] $12 : 00000000140084e0 ffffffff80109c00 0000000000000000 0000000000000002 [ 370.269179] $16 : ffffffff8092f080 a8000000fd672e00 ffffffff80107fe8 a8000000fd485000 [ 370.277612] $20 : ffffffff8084d328 ffffffff80940000 0000000000000009 ffffffff80930000 [ 370.286038] $24 : 0000000000000000 900000001612048c [ 370.294476] $28 : a8000000fd630000 a8000000fd637ac0 ffffffff80937300 ffffffff8010807c [ 370.302909] Hi : 0000000000000000 [ 370.306595] Lo : 0000000000000200 [ 370.310376] epc : ffffffff80115d38 _save_fp+0x10/0xa0 [ 370.315784] ra : ffffffff8010807c prepare_for_fp_mode_switch+0x94/0x1b0 [ 370.322707] Status: 140084e2 KX SX UX KERNEL EXL [ 370.327980] Cause : 1080002c (ExcCode 0b) [ 370.332091] PrId : 0001a428 (MIPS P6600) [ 370.336179] Modules linked in: [ 370.339486] Process fp-prctl (pid: 963, threadinfo=a8000000fd630000, task=a8000000fd672e00, tls=00000000756e67d0) [ 370.349724] Stack : 0000000000000000 a8000000fd557dc0 0000000000000000 ffffffff801ca8e0 [ 370.358161] 0000000000000000 a8000000fd637b9c 0000000000000009 ffffffff80923780 [ 370.366575] ffffffff80850000 ffffffff8011610c 00000000000000b8 ffffffff801a5084 [ 370.374989] ffffffff8084a370 ffffffff8084a388 ffffffff80923780 ffffffff80923828 [ 370.383395] 0000000000010000 ffffffff809237a8 0000000000020000 ffffffff80a40000 [ 370.391817] 000000000000007c 00000000004a0000 00000000756dedd0 ffffffff801a5188 [ 370.400230] a800000002014900 0000000000000001 ffffffff80923780 0000000080923828 [ 370.408644] ffffffff80923780 ffffffff80923780 ffffffff80923828 ffffffff801a521c [ 370.417066] ffffffff80923780 ffffffff80923828 0000000000010000 ffffffff801a8f84 [ 370.425472] ffffffff80a40000 a8000000fd637c20 ffffffff80a39240 0000000000000001 [ 370.433885] ... [ 370.436562] Call Trace: [ 370.439222] [<ffffffff80115d38>] _save_fp+0x10/0xa0 [ 370.444305] [<ffffffff8010807c>] prepare_for_fp_mode_switch+0x94/0x1b0 [ 370.451035] [<ffffffff801ca8e0>] flush_smp_call_function_queue+0xf8/0x230 [ 370.457991] [<ffffffff8011610c>] ipi_call_interrupt+0xc/0x20 [ 370.463814] [<ffffffff801a5084>] __handle_irq_event_percpu+0xc4/0x1a8 [ 370.470404] [<ffffffff801a5188>] handle_irq_event_percpu+0x20/0x68 [ 370.476734] [<ffffffff801a521c>] handle_irq_event+0x4c/0x88 [ 370.482486] [<ffffffff801a8f84>] handle_edge_irq+0x12c/0x210 [ 370.488316] [<ffffffff801a47a0>] generic_handle_irq+0x38/0x48 [ 370.494280] [<ffffffff804a2dbc>] gic_handle_shared_int+0x194/0x268 [ 370.500616] [<ffffffff801a47a0>] generic_handle_irq+0x38/0x48 [ 370.506529] [<ffffffff80107e60>] do_IRQ+0x18/0x28 [ 370.511445] [<ffffffff804a1524>] plat_irq_dispatch+0xc4/0x140 [ 370.517339] [<ffffffff80106230>] ret_from_irq+0x0/0x4 [ 370.522583] [<ffffffff8010fad4>] do_ri+0x4fc/0x7e8 [ 370.527546] [<ffffffff80106220>] ret_from_exception+0x0/0x10 2) The IPI may arrive during kernel use of the FPU, since we generally only disable preemption around use of the FPU & leave interrupts enabled. This can lead to us unexpectedly losing access to the FPU in places where it previously had not been possible. For example: do_cpu invoked from kernel context![#2]: CPU: 2 PID: 7338 Comm: fp-prctl Tainted: G D 4.7.0-00424-g49b0c82 #2 task: 838e4000 ti: 88d38000 task.ti: 88d38000 $ 0 : 00000000 00000001 ffffffff 88d3fef8 $ 4 : 838e4000 88d38004 00000000 00000001 $ 8 : 3400fc01 801f8020 808e9100 24000000 $12 : dbffffff 807b69d8 807b0000 00000000 $16 : 00000000 80786150 00400fc4 809c0398 $20 : 809c0338 0040273c 88d3ff28 808e9d30 $24 : 808e9d30 00400fb4 $28 : 88d38000 88d3fe88 00000000 8011a2ac Hi : 0040273c Lo : 88d3ff28 epc : 80114178 _restore_fp+0x10/0xa0 ra : 8011a2ac mipsr2_decoder+0xd5c/0x1660 Status: 1400fc03 KERNEL EXL IE Cause : 1080002c (ExcCode 0b) PrId : 0001a920 (MIPS I6400) Modules linked in: Process fp-prctl (pid: 7338, threadinfo=88d38000, task=838e4000, tls=766527d0) Stack : 00000000 00000000 00000000 88d3fe98 00000000 00000000 809c0398 809c0338 808e9100 00000000 88d3ff28 00400fc4 00400fc4 0040273c 7fb69e18 004a0000 004a0000 004a0000 7664add0 8010de18 00000000 00000000 88d3fef8 88d3ff28 808e9100 00000000 766527d0 8010e534 000c0000 85755000 8181d580 00000000 00000000 00000000 004a0000 00000000 766527d0 7fb69e18 004a0000 80105c20 ... Call Trace: [<80114178>] _restore_fp+0x10/0xa0 [<8011a2ac>] mipsr2_decoder+0xd5c/0x1660 [<8010de18>] do_ri+0x90/0x6b8 [<80105c20>] ret_from_exception+0x0/0x10 At first glance a simple fix may seem to be to disable interrupts around kernel use of the FPU rather than merely preemption, however this would introduce further overhead outside of the mode switch path & doesn't solve the third problem: 3) The IPI may arrive whilst the kernel is running code that will lead to a preempt_disable() call & FPU usage soon. If this happens then the IPI will be serviced & we'll proceed to enable an FPU whilst the mode switch is in progress, leading to strange & inconsistent behaviour. Further to all of this is a separate but related problem: 4) There are various paths through which we may enable the FPU without the user having triggered a coprocessor 1 disabled exception. These paths are those in which we emulate instructions & then enable the FPU with the expectation that the user might execute an FP instruction shortly afterwards. However these paths have not previously checked whether an FP mode switch is underway for the task, and therefore could enable the FPU whilst such a mode switch is in progress leading to strange & inconsistent behaviour for user code. This patch fixes all of the above by taking a step back & re-examining our approach to FP mode switches. Up until now we have taken these basic steps: a) Prevent any threads that are part of the affected process from being able to obtain ownership of the FPU. b) Cause any threads that are part of the affected process and already have ownership of an FPU to lose it. c) Set the thread flags for each thread that is part of the affected process to reflect the new FP mode. d) Allow threads to obtain ownership of the FPU again. This approach is however more complex than necessary. All that we really require is that the mode switch has occurred for all threads that are part of the affected process before mips_set_process_fp_mode(), and thus the PR_SET_FP_MODE prctl() syscall, returns. This doesn't require that we stop threads from owning or using an FPU whilst a mode switch occurs, only that we force them to relinquish it after the mode switch has occurred such that they next own an FPU with the correct mode configured. Our basic steps therefore simplify to: A) Set the thread flags for each thread that is part of the affected process to reflect the new FP mode. B) Cause any threads that are part of the affected process and already have ownership of an FPU to lose it. We implement B) by forcing each CPU which might be running a thread which is part of the affected process to schedule a no-op function, which causes the affected thread to lose its FPU ownership when it is descheduled. The end result is simpler FP mode switching with less overhead in the FPU enable path (ie. enable_restore_fp_context()) and fewer moving parts. Signed-off-by: Paul Burton <paul.burton(a)mips.com> Fixes: 9791554b45a2 ("MIPS,prctl: add PR_[GS]ET_FP_MODE prctl options for MIPS") Fixes: 6b8322576e9d ("MIPS: Force CPUs to lose FP context during mode switches") Cc: James Hogan <jhogan(a)kernel.org> Cc: Ralf Baechle <ralf(a)linux-mips.org> Cc: linux-mips(a)linux-mips.org Cc: stable <stable(a)vger.kernel.org> # v4.0+ diff --git a/arch/mips/include/asm/mmu_context.h b/arch/mips/include/asm/mmu_context.h index da2004cef2d5..b509371a6b0c 100644 --- a/arch/mips/include/asm/mmu_context.h +++ b/arch/mips/include/asm/mmu_context.h @@ -126,8 +126,6 @@ init_new_context(struct task_struct *tsk, struct mm_struct *mm) for_each_possible_cpu(i) cpu_context(i, mm) = 0; - atomic_set(&mm->context.fp_mode_switching, 0); - mm->context.bd_emupage_allocmap = NULL; spin_lock_init(&mm->context.bd_emupage_lock); init_waitqueue_head(&mm->context.bd_emupage_queue); diff --git a/arch/mips/kernel/process.c b/arch/mips/kernel/process.c index 8d85046adcc8..fe6001d748cf 100644 --- a/arch/mips/kernel/process.c +++ b/arch/mips/kernel/process.c @@ -29,6 +29,7 @@ #include <linux/kallsyms.h> #include <linux/random.h> #include <linux/prctl.h> +#include <linux/cpu.h> #include <asm/asm.h> #include <asm/bootinfo.h> @@ -691,19 +692,25 @@ int mips_get_process_fp_mode(struct task_struct *task) return value; } -static void prepare_for_fp_mode_switch(void *info) +static long prepare_for_fp_mode_switch(void *unused) { - struct mm_struct *mm = info; - - if (current->mm == mm) - lose_fpu(1); + /* + * This is icky, but we use this to simply ensure that all CPUs have + * context switched, regardless of whether they were previously running + * kernel or user code. This ensures that no CPU currently has its FPU + * enabled, or is about to attempt to enable it through any path other + * than enable_restore_fp_context() which will wait appropriately for + * fp_mode_switching to be zero. + */ + return 0; } int mips_set_process_fp_mode(struct task_struct *task, unsigned int value) { const unsigned int known_bits = PR_FP_MODE_FR | PR_FP_MODE_FRE; struct task_struct *t; - int max_users; + struct cpumask process_cpus; + int cpu; /* If nothing to change, return right away, successfully. */ if (value == mips_get_process_fp_mode(task)) @@ -736,35 +743,7 @@ int mips_set_process_fp_mode(struct task_struct *task, unsigned int value) if (!(value & PR_FP_MODE_FR) && raw_cpu_has_fpu && cpu_has_mips_r6) return -EOPNOTSUPP; - /* Proceed with the mode switch */ - preempt_disable(); - - /* Save FP & vector context, then disable FPU & MSA */ - if (task->signal == current->signal) - lose_fpu(1); - - /* Prevent any threads from obtaining live FP context */ - atomic_set(&task->mm->context.fp_mode_switching, 1); - smp_mb__after_atomic(); - - /* - * If there are multiple online CPUs then force any which are running - * threads in this process to lose their FPU context, which they can't - * regain until fp_mode_switching is cleared later. - */ - if (num_online_cpus() > 1) { - /* No need to send an IPI for the local CPU */ - max_users = (task->mm == current->mm) ? 1 : 0; - - if (atomic_read(&current->mm->mm_users) > max_users) - smp_call_function(prepare_for_fp_mode_switch, - (void *)current->mm, 1); - } - - /* - * There are now no threads of the process with live FP context, so it - * is safe to proceed with the FP mode switch. - */ + /* Indicate the new FP mode in each thread */ for_each_thread(task, t) { /* Update desired FP register width */ if (value & PR_FP_MODE_FR) { @@ -781,9 +760,34 @@ int mips_set_process_fp_mode(struct task_struct *task, unsigned int value) clear_tsk_thread_flag(t, TIF_HYBRID_FPREGS); } - /* Allow threads to use FP again */ - atomic_set(&task->mm->context.fp_mode_switching, 0); - preempt_enable(); + /* + * We need to ensure that all threads in the process have switched mode + * before returning, in order to allow userland to not worry about + * races. We can do this by forcing all CPUs that any thread in the + * process may be running on to schedule something else - in this case + * prepare_for_fp_mode_switch(). + * + * We begin by generating a mask of all CPUs that any thread in the + * process may be running on. + */ + cpumask_clear(&process_cpus); + for_each_thread(task, t) + cpumask_set_cpu(task_cpu(t), &process_cpus); + + /* + * Now we schedule prepare_for_fp_mode_switch() on each of those CPUs. + * + * The CPUs may have rescheduled already since we switched mode or + * generated the cpumask, but that doesn't matter. If the task in this + * process is scheduled out then our scheduling + * prepare_for_fp_mode_switch() will simply be redundant. If it's + * scheduled in then it will already have picked up the new FP mode + * whilst doing so. + */ + get_online_cpus(); + for_each_cpu_and(cpu, &process_cpus, cpu_online_mask) + work_on_cpu(cpu, prepare_for_fp_mode_switch, NULL); + put_online_cpus(); wake_up_var(&task->mm->context.fp_mode_switching); diff --git a/arch/mips/kernel/traps.c b/arch/mips/kernel/traps.c index d67fa74622ee..4d9ca9b465ae 100644 --- a/arch/mips/kernel/traps.c +++ b/arch/mips/kernel/traps.c @@ -1220,13 +1220,6 @@ static int enable_restore_fp_context(int msa) { int err, was_fpu_owner, prior_msa; - /* - * If an FP mode switch is currently underway, wait for it to - * complete before proceeding. - */ - wait_var_event(&current->mm->context.fp_mode_switching, - !atomic_read(&current->mm->context.fp_mode_switching)); - if (!used_math()) { /* First time FP context user. */ preempt_disable();

6 years, 11 months

1
0
0 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

Linux-stable-mirror