From c05780ef3c190c2dafbf0be8e65d4f01103ad577 Mon Sep 17 00:00:00 2001 From: Palmer Dabbelt Date: Fri, 7 Jul 2023 09:00:51 -0700 Subject: [PATCH 01/16] module: Ignore RISC-V mapping symbols too RISC-V has an extended form of mapping symbols that we use to encode the ISA when it changes in the middle of an ELF. This trips up modpost as a build failure, I haven't yet verified it yet but I believe the kallsyms difference should result in stacks looking sane again. Reported-by: Randy Dunlap Closes: https://lore.kernel.org/all/9d9e2902-5489-4bf0-d9cb-556c8e5d71c2@infradead.org/ Signed-off-by: Palmer Dabbelt Reviewed-by: Randy Dunlap Tested-by: Randy Dunlap # build-tested Signed-off-by: Luis Chamberlain --- include/linux/module_symbol.h | 12 +++++++++++- kernel/module/kallsyms.c | 2 +- scripts/mod/modpost.c | 2 +- 3 files changed, 13 insertions(+), 3 deletions(-) diff --git a/include/linux/module_symbol.h b/include/linux/module_symbol.h index 7ace7ba30203..5b799942b243 100644 --- a/include/linux/module_symbol.h +++ b/include/linux/module_symbol.h @@ -3,12 +3,22 @@ #define _LINUX_MODULE_SYMBOL_H /* This ignores the intensely annoying "mapping symbols" found in ELF files. */ -static inline int is_mapping_symbol(const char *str) +static inline int is_mapping_symbol(const char *str, int is_riscv) { if (str[0] == '.' && str[1] == 'L') return true; if (str[0] == 'L' && str[1] == '0') return true; + /* + * RISC-V defines various special symbols that start with "$".  The + * mapping symbols, which exist to differentiate between incompatible + * instruction encodings when disassembling, show up all over the place + * and are generally not meant to be treated like other symbols.  So + * just ignore any of the special symbols. + */ + if (is_riscv) + return str[0] == '$'; + return str[0] == '$' && (str[1] == 'a' || str[1] == 'd' || str[1] == 't' || str[1] == 'x') && (str[2] == '\0' || str[2] == '.'); diff --git a/kernel/module/kallsyms.c b/kernel/module/kallsyms.c index ef73ae7c8909..78a1ffc399d9 100644 --- a/kernel/module/kallsyms.c +++ b/kernel/module/kallsyms.c @@ -289,7 +289,7 @@ static const char *find_kallsyms_symbol(struct module *mod, * and inserted at a whim. */ if (*kallsyms_symbol_name(kallsyms, i) == '\0' || - is_mapping_symbol(kallsyms_symbol_name(kallsyms, i))) + is_mapping_symbol(kallsyms_symbol_name(kallsyms, i), IS_ENABLED(CONFIG_RISCV))) continue; if (thisval <= addr && thisval > bestval) { diff --git a/scripts/mod/modpost.c b/scripts/mod/modpost.c index b29b29707f10..7c71429d6502 100644 --- a/scripts/mod/modpost.c +++ b/scripts/mod/modpost.c @@ -1052,7 +1052,7 @@ static inline int is_valid_name(struct elf_info *elf, Elf_Sym *sym) if (!name || !strlen(name)) return 0; - return !is_mapping_symbol(name); + return !is_mapping_symbol(name, elf->hdr->e_machine == EM_RISCV); } /* Look up the nearest symbol based on the section and the address */ From 9ce170cef66916526dcaa868a67cfcc0c2f0c3d6 Mon Sep 17 00:00:00 2001 From: Li zeming Date: Wed, 12 Jul 2023 03:43:31 +0800 Subject: [PATCH 02/16] =?UTF-8?q?kernel:=20params:=20Remove=20unnecessary?= =?UTF-8?q?=20=E2=80=980=E2=80=99=20values=20from=20err?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit err is assigned first, so it does not need to initialize the assignment. Signed-off-by: Li zeming Reviewed-by: Randy Dunlap Signed-off-by: Luis Chamberlain --- kernel/params.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/params.c b/kernel/params.c index 07d01f6ce9a2..2d4a0564697e 100644 --- a/kernel/params.c +++ b/kernel/params.c @@ -331,7 +331,7 @@ EXPORT_SYMBOL(param_ops_bool); int param_set_bool_enable_only(const char *val, const struct kernel_param *kp) { - int err = 0; + int err; bool new_value; bool orig_value = *(bool *)kp->arg; struct kernel_param dummy_kp = *kp; From ff09f6fd297293175eaa0ed492495e36b3eb1a8e Mon Sep 17 00:00:00 2001 From: Palmer Dabbelt Date: Fri, 21 Jul 2023 08:01:48 -0700 Subject: [PATCH 03/16] modpost, kallsyms: Treat add '$'-prefixed symbols as mapping symbols Trying to restrict the '$'-prefix change to RISC-V caused some fallout, so let's just treat all those symbols as special. Fixes: c05780ef3c190 ("module: Ignore RISC-V mapping symbols too") Link: https://lore.kernel.org/all/20230712015747.77263-1-wangkefeng.wang@huawei.com/ Signed-off-by: Palmer Dabbelt Reviewed-by: Masahiro Yamada Signed-off-by: Luis Chamberlain --- include/linux/module_symbol.h | 16 ++-------------- kernel/module/kallsyms.c | 2 +- scripts/mod/modpost.c | 2 +- 3 files changed, 4 insertions(+), 16 deletions(-) diff --git a/include/linux/module_symbol.h b/include/linux/module_symbol.h index 5b799942b243..1269543d0634 100644 --- a/include/linux/module_symbol.h +++ b/include/linux/module_symbol.h @@ -3,25 +3,13 @@ #define _LINUX_MODULE_SYMBOL_H /* This ignores the intensely annoying "mapping symbols" found in ELF files. */ -static inline int is_mapping_symbol(const char *str, int is_riscv) +static inline int is_mapping_symbol(const char *str) { if (str[0] == '.' && str[1] == 'L') return true; if (str[0] == 'L' && str[1] == '0') return true; - /* - * RISC-V defines various special symbols that start with "$".  The - * mapping symbols, which exist to differentiate between incompatible - * instruction encodings when disassembling, show up all over the place - * and are generally not meant to be treated like other symbols.  So - * just ignore any of the special symbols. - */ - if (is_riscv) - return str[0] == '$'; - - return str[0] == '$' && - (str[1] == 'a' || str[1] == 'd' || str[1] == 't' || str[1] == 'x') - && (str[2] == '\0' || str[2] == '.'); + return str[0] == '$'; } #endif /* _LINUX_MODULE_SYMBOL_H */ diff --git a/kernel/module/kallsyms.c b/kernel/module/kallsyms.c index 78a1ffc399d9..ef73ae7c8909 100644 --- a/kernel/module/kallsyms.c +++ b/kernel/module/kallsyms.c @@ -289,7 +289,7 @@ static const char *find_kallsyms_symbol(struct module *mod, * and inserted at a whim. */ if (*kallsyms_symbol_name(kallsyms, i) == '\0' || - is_mapping_symbol(kallsyms_symbol_name(kallsyms, i), IS_ENABLED(CONFIG_RISCV))) + is_mapping_symbol(kallsyms_symbol_name(kallsyms, i))) continue; if (thisval <= addr && thisval > bestval) { diff --git a/scripts/mod/modpost.c b/scripts/mod/modpost.c index 7c71429d6502..b29b29707f10 100644 --- a/scripts/mod/modpost.c +++ b/scripts/mod/modpost.c @@ -1052,7 +1052,7 @@ static inline int is_valid_name(struct elf_info *elf, Elf_Sym *sym) if (!name || !strlen(name)) return 0; - return !is_mapping_symbol(name, elf->hdr->e_machine == EM_RISCV); + return !is_mapping_symbol(name); } /* Look up the nearest symbol based on the section and the address */ From 910e230d5f1bb72c54532e94fbb1705095c7bab6 Mon Sep 17 00:00:00 2001 From: Rong Tao Date: Sun, 16 Apr 2023 23:05:17 +0800 Subject: [PATCH 04/16] samples/hw_breakpoint: Fix kernel BUG 'invalid opcode: 0000' Macro symbol_put() is defined as __symbol_put(__stringify(x)) ksym_name = "jiffies" symbol_put(ksym_name) will be resolved as __symbol_put("ksym_name") which is clearly wrong. So symbol_put must be replaced with __symbol_put. When we uninstall hw_breakpoint.ko (rmmod), a kernel bug occurs with the following error: [11381.854152] kernel BUG at kernel/module/main.c:779! [11381.854159] invalid opcode: 0000 [#2] PREEMPT SMP PTI [11381.854163] CPU: 8 PID: 59623 Comm: rmmod Tainted: G D OE 6.2.9-200.fc37.x86_64 #1 [11381.854167] Hardware name: To Be Filled By O.E.M. To Be Filled By O.E.M./B360M-HDV, BIOS P3.20 10/23/2018 [11381.854169] RIP: 0010:__symbol_put+0xa2/0xb0 [11381.854175] Code: 00 e8 92 d2 f7 ff 65 8b 05 c3 2f e6 78 85 c0 74 1b 48 8b 44 24 30 65 48 2b 04 25 28 00 00 00 75 12 48 83 c4 38 c3 cc cc cc cc <0f> 0b 0f 1f 44 00 00 eb de e8 c0 df d8 00 90 90 90 90 90 90 90 90 [11381.854178] RSP: 0018:ffffad8ec6ae7dd0 EFLAGS: 00010246 [11381.854181] RAX: 0000000000000000 RBX: ffffffffc1fd1240 RCX: 000000000000000c [11381.854184] RDX: 000000000000006b RSI: ffffffffc02bf7c7 RDI: ffffffffc1fd001c [11381.854186] RBP: 000055a38b76e7c8 R08: ffffffff871ccfe0 R09: 0000000000000000 [11381.854188] R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000000000 [11381.854190] R13: 0000000000000000 R14: 0000000000000000 R15: 0000000000000000 [11381.854192] FS: 00007fbf7c62c740(0000) GS:ffff8c5badc00000(0000) knlGS:0000000000000000 [11381.854195] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [11381.854197] CR2: 000055a38b7793f8 CR3: 0000000363e1e001 CR4: 00000000003726e0 [11381.854200] DR0: ffffffffb3407980 DR1: 0000000000000000 DR2: 0000000000000000 [11381.854202] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 [11381.854204] Call Trace: [11381.854207] [11381.854212] s_module_exit+0xc/0xff0 [symbol_getput] [11381.854219] __do_sys_delete_module.constprop.0+0x198/0x2f0 [11381.854225] do_syscall_64+0x58/0x80 [11381.854231] ? exit_to_user_mode_prepare+0x180/0x1f0 [11381.854237] ? syscall_exit_to_user_mode+0x17/0x40 [11381.854241] ? do_syscall_64+0x67/0x80 [11381.854245] ? syscall_exit_to_user_mode+0x17/0x40 [11381.854248] ? do_syscall_64+0x67/0x80 [11381.854252] ? exc_page_fault+0x70/0x170 [11381.854256] entry_SYSCALL_64_after_hwframe+0x72/0xdc Signed-off-by: Rong Tao Reviewed-by: Petr Mladek Signed-off-by: Luis Chamberlain --- samples/hw_breakpoint/data_breakpoint.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/samples/hw_breakpoint/data_breakpoint.c b/samples/hw_breakpoint/data_breakpoint.c index 418c46fe5ffc..9debd128b2ab 100644 --- a/samples/hw_breakpoint/data_breakpoint.c +++ b/samples/hw_breakpoint/data_breakpoint.c @@ -70,7 +70,7 @@ fail: static void __exit hw_break_module_exit(void) { unregister_wide_hw_breakpoint(sample_hbp); - symbol_put(ksym_name); + __symbol_put(ksym_name); printk(KERN_INFO "HW Breakpoint for %s write uninstalled\n", ksym_name); } From b9080468caeddc58a91edd1c3a7d212ea82b0d1d Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Tue, 25 Jul 2023 10:25:36 +0200 Subject: [PATCH 05/16] samples/hw_breakpoint: fix building without module unloading __symbol_put() is really meant as an internal helper and is not available when module unloading is disabled, unlike the previously used symbol_put(): samples/hw_breakpoint/data_breakpoint.c: In function 'hw_break_module_exit': samples/hw_breakpoint/data_breakpoint.c:73:9: error: implicit declaration of function '__symbol_put'; did you mean '__symbol_get'? [-Werror=implicit-function-declaration] The hw_break_module_exit() function is not actually used when module unloading is disabled, but it still causes the build failure for an undefined identifier. Enclose this one call in an appropriate #ifdef to clarify what the requirement is. Leaving out the entire exit function would also work but feels less clar in this case. Fixes: 910e230d5f1bb ("samples/hw_breakpoint: Fix kernel BUG 'invalid opcode: 0000'") Fixes: d8a84d33a4954 ("samples/hw_breakpoint: drop use of kallsyms_lookup_name()") Signed-off-by: Arnd Bergmann Reviewed-by: Petr Mladek Signed-off-by: Luis Chamberlain --- samples/hw_breakpoint/data_breakpoint.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/samples/hw_breakpoint/data_breakpoint.c b/samples/hw_breakpoint/data_breakpoint.c index 9debd128b2ab..b99322f188e5 100644 --- a/samples/hw_breakpoint/data_breakpoint.c +++ b/samples/hw_breakpoint/data_breakpoint.c @@ -70,7 +70,9 @@ fail: static void __exit hw_break_module_exit(void) { unregister_wide_hw_breakpoint(sample_hbp); +#ifdef CONFIG_MODULE_UNLOAD __symbol_put(ksym_name); +#endif printk(KERN_INFO "HW Breakpoint for %s write uninstalled\n", ksym_name); } From 4b49df654cd0d1c3c359fc0f52182943a650b746 Mon Sep 17 00:00:00 2001 From: Chen Jiahao Date: Wed, 26 Jul 2023 10:17:59 +0000 Subject: [PATCH 06/16] samples/hw_breakpoint: mark sample_hbp as static There is a sparse warning shown as below: samples/hw_breakpoint/data_breakpoint.c:24:19: warning: symbol 'sample_hbp' was not declared. Should it be static? Since 'sample_hbp' is only called within data_breakpoint.c, mark it as static to fix the warning. Fixes: 44ee63587dce ("percpu: Add __percpu sparse annotations to hw_breakpoint") Signed-off-by: Chen Jiahao Signed-off-by: Luis Chamberlain --- samples/hw_breakpoint/data_breakpoint.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/samples/hw_breakpoint/data_breakpoint.c b/samples/hw_breakpoint/data_breakpoint.c index b99322f188e5..a2c831e89ce0 100644 --- a/samples/hw_breakpoint/data_breakpoint.c +++ b/samples/hw_breakpoint/data_breakpoint.c @@ -21,7 +21,7 @@ #include #include -struct perf_event * __percpu *sample_hbp; +static struct perf_event * __percpu *sample_hbp; static char ksym_name[KSYM_NAME_LEN] = "jiffies"; module_param_string(ksym, ksym_name, KSYM_NAME_LEN, S_IRUGO); From 0faa29c4207e6e29cfc81b427df60e326c37083a Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Tue, 1 Aug 2023 19:35:40 +0200 Subject: [PATCH 07/16] ARM: pxa: remove use of symbol_get() The spitz board file uses the obscure symbol_get() function to optionally call a function from sharpsl_pm.c if that is built. However, the two files are always built together these days, and have been for a long time, so this can be changed to a normal function call. Link: https://lore.kernel.org/lkml/20230731162639.GA9441@lst.de/ Signed-off-by: Arnd Bergmann Signed-off-by: Christoph Hellwig Signed-off-by: Luis Chamberlain --- arch/arm/mach-pxa/sharpsl_pm.c | 2 -- arch/arm/mach-pxa/spitz.c | 14 +------------- 2 files changed, 1 insertion(+), 15 deletions(-) diff --git a/arch/arm/mach-pxa/sharpsl_pm.c b/arch/arm/mach-pxa/sharpsl_pm.c index d29bdcd5270e..72fa2e3fd353 100644 --- a/arch/arm/mach-pxa/sharpsl_pm.c +++ b/arch/arm/mach-pxa/sharpsl_pm.c @@ -216,8 +216,6 @@ void sharpsl_battery_kick(void) { schedule_delayed_work(&sharpsl_bat, msecs_to_jiffies(125)); } -EXPORT_SYMBOL(sharpsl_battery_kick); - static void sharpsl_battery_thread(struct work_struct *private_) { diff --git a/arch/arm/mach-pxa/spitz.c b/arch/arm/mach-pxa/spitz.c index d01ea54b0b78..cc691b199429 100644 --- a/arch/arm/mach-pxa/spitz.c +++ b/arch/arm/mach-pxa/spitz.c @@ -9,7 +9,6 @@ */ #include -#include /* symbol_get ; symbol_put */ #include #include #include @@ -518,17 +517,6 @@ static struct gpiod_lookup_table spitz_ads7846_gpio_table = { }, }; -static void spitz_bl_kick_battery(void) -{ - void (*kick_batt)(void); - - kick_batt = symbol_get(sharpsl_battery_kick); - if (kick_batt) { - kick_batt(); - symbol_put(sharpsl_battery_kick); - } -} - static struct gpiod_lookup_table spitz_lcdcon_gpio_table = { .dev_id = "spi2.1", .table = { @@ -556,7 +544,7 @@ static struct corgi_lcd_platform_data spitz_lcdcon_info = { .max_intensity = 0x2f, .default_intensity = 0x1f, .limit_mask = 0x0b, - .kick_battery = spitz_bl_kick_battery, + .kick_battery = sharpsl_battery_kick, }; static struct spi_board_info spitz_spi_devices[] = { From d4a5c59a955bba96b273ec1a5885bada24c56979 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Tue, 1 Aug 2023 19:35:41 +0200 Subject: [PATCH 08/16] mmc: au1xmmc: force non-modular build and remove symbol_get usage au1xmmc is split somewhat awkwardly into the main mmc subsystem driver, and callbacks in platform_data that sit under arch/mips/ and are always built in. The latter than call mmc_detect_change through symbol_get. Remove the use of symbol_get by requiring the driver to be built in. In the future the interrupt handlers for card insert/eject detection should probably be moved into the main driver, and which point it can be built modular again. Signed-off-by: Christoph Hellwig Acked-by: Manuel Lauss Reviewed-by: Arnd Bergmann [mcgrof: squashed in depends on MMC=y suggested by Arnd] Signed-off-by: Luis Chamberlain --- arch/mips/alchemy/devboards/db1000.c | 8 +------- arch/mips/alchemy/devboards/db1200.c | 19 ++----------------- arch/mips/alchemy/devboards/db1300.c | 10 +--------- drivers/mmc/host/Kconfig | 5 +++-- 4 files changed, 7 insertions(+), 35 deletions(-) diff --git a/arch/mips/alchemy/devboards/db1000.c b/arch/mips/alchemy/devboards/db1000.c index 79d66faa8482..012da042d0a4 100644 --- a/arch/mips/alchemy/devboards/db1000.c +++ b/arch/mips/alchemy/devboards/db1000.c @@ -14,7 +14,6 @@ #include #include #include -#include #include #include #include @@ -167,12 +166,7 @@ static struct platform_device db1x00_audio_dev = { static irqreturn_t db1100_mmc_cd(int irq, void *ptr) { - void (*mmc_cd)(struct mmc_host *, unsigned long); - /* link against CONFIG_MMC=m */ - mmc_cd = symbol_get(mmc_detect_change); - mmc_cd(ptr, msecs_to_jiffies(500)); - symbol_put(mmc_detect_change); - + mmc_detect_change(ptr, msecs_to_jiffies(500)); return IRQ_HANDLED; } diff --git a/arch/mips/alchemy/devboards/db1200.c b/arch/mips/alchemy/devboards/db1200.c index 1864eb935ca5..76080c71a2a7 100644 --- a/arch/mips/alchemy/devboards/db1200.c +++ b/arch/mips/alchemy/devboards/db1200.c @@ -10,7 +10,6 @@ #include #include #include -#include #include #include #include @@ -340,14 +339,7 @@ static irqreturn_t db1200_mmc_cd(int irq, void *ptr) static irqreturn_t db1200_mmc_cdfn(int irq, void *ptr) { - void (*mmc_cd)(struct mmc_host *, unsigned long); - - /* link against CONFIG_MMC=m */ - mmc_cd = symbol_get(mmc_detect_change); - if (mmc_cd) { - mmc_cd(ptr, msecs_to_jiffies(200)); - symbol_put(mmc_detect_change); - } + mmc_detect_change(ptr, msecs_to_jiffies(200)); msleep(100); /* debounce */ if (irq == DB1200_SD0_INSERT_INT) @@ -431,14 +423,7 @@ static irqreturn_t pb1200_mmc1_cd(int irq, void *ptr) static irqreturn_t pb1200_mmc1_cdfn(int irq, void *ptr) { - void (*mmc_cd)(struct mmc_host *, unsigned long); - - /* link against CONFIG_MMC=m */ - mmc_cd = symbol_get(mmc_detect_change); - if (mmc_cd) { - mmc_cd(ptr, msecs_to_jiffies(200)); - symbol_put(mmc_detect_change); - } + mmc_detect_change(ptr, msecs_to_jiffies(200)); msleep(100); /* debounce */ if (irq == PB1200_SD1_INSERT_INT) diff --git a/arch/mips/alchemy/devboards/db1300.c b/arch/mips/alchemy/devboards/db1300.c index e70e529ddd91..ff61901329c6 100644 --- a/arch/mips/alchemy/devboards/db1300.c +++ b/arch/mips/alchemy/devboards/db1300.c @@ -17,7 +17,6 @@ #include #include #include -#include #include #include #include @@ -459,14 +458,7 @@ static irqreturn_t db1300_mmc_cd(int irq, void *ptr) static irqreturn_t db1300_mmc_cdfn(int irq, void *ptr) { - void (*mmc_cd)(struct mmc_host *, unsigned long); - - /* link against CONFIG_MMC=m. We can only be called once MMC core has - * initialized the controller, so symbol_get() should always succeed. - */ - mmc_cd = symbol_get(mmc_detect_change); - mmc_cd(ptr, msecs_to_jiffies(200)); - symbol_put(mmc_detect_change); + mmc_detect_change(ptr, msecs_to_jiffies(200)); msleep(100); /* debounce */ if (irq == DB1300_SD1_INSERT_INT) diff --git a/drivers/mmc/host/Kconfig b/drivers/mmc/host/Kconfig index 159a3e9490ae..554e67103c1a 100644 --- a/drivers/mmc/host/Kconfig +++ b/drivers/mmc/host/Kconfig @@ -526,11 +526,12 @@ config MMC_ALCOR of Alcor Micro PCI-E card reader config MMC_AU1X - tristate "Alchemy AU1XX0 MMC Card Interface support" + bool "Alchemy AU1XX0 MMC Card Interface support" depends on MIPS_ALCHEMY + depends on MMC=y help This selects the AMD Alchemy(R) Multimedia card interface. - If you have a Alchemy platform with a MMC slot, say Y or M here. + If you have a Alchemy platform with a MMC slot, say Y here. If unsure, say N. From 569820befb16ffc755ab7af71f4f08cc5f68f0fe Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Tue, 1 Aug 2023 19:35:42 +0200 Subject: [PATCH 09/16] net: enetc: use EXPORT_SYMBOL_GPL for enetc_phc_index enetc_phc_index is only used via symbol_get, which was only ever intended for very internal symbols like this one. Use EXPORT_SYMBOL_GPL for it so that symbol_get can enforce only being used on EXPORT_SYMBOL_GPL symbols. Signed-off-by: Christoph Hellwig Reviewed-by: Jakub Kicinski Reviewed-by: Greg Kroah-Hartman Signed-off-by: Luis Chamberlain --- drivers/net/ethernet/freescale/enetc/enetc_ptp.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/freescale/enetc/enetc_ptp.c b/drivers/net/ethernet/freescale/enetc/enetc_ptp.c index 17c097cef7d4..5243fc031058 100644 --- a/drivers/net/ethernet/freescale/enetc/enetc_ptp.c +++ b/drivers/net/ethernet/freescale/enetc/enetc_ptp.c @@ -8,7 +8,7 @@ #include "enetc.h" int enetc_phc_index = -1; -EXPORT_SYMBOL(enetc_phc_index); +EXPORT_SYMBOL_GPL(enetc_phc_index); static struct ptp_clock_info enetc_ptp_caps = { .owner = THIS_MODULE, From 95e7ebc6823170256a8ce19fad87912805bfa001 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Tue, 1 Aug 2023 19:35:43 +0200 Subject: [PATCH 10/16] rtc: ds1685: use EXPORT_SYMBOL_GPL for ds1685_rtc_poweroff ds1685_rtc_poweroff is only used externally via symbol_get, which was only ever intended for very internal symbols like this one. Use EXPORT_SYMBOL_GPL for it so that symbol_get can enforce only being used on EXPORT_SYMBOL_GPL symbols. Signed-off-by: Christoph Hellwig Acked-by: Joshua Kinard Reviewed-by: Greg Kroah-Hartman Signed-off-by: Luis Chamberlain --- drivers/rtc/rtc-ds1685.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/rtc/rtc-ds1685.c b/drivers/rtc/rtc-ds1685.c index 0f707be0eb87..04dbf35cf3b7 100644 --- a/drivers/rtc/rtc-ds1685.c +++ b/drivers/rtc/rtc-ds1685.c @@ -1432,7 +1432,7 @@ ds1685_rtc_poweroff(struct platform_device *pdev) unreachable(); } } -EXPORT_SYMBOL(ds1685_rtc_poweroff); +EXPORT_SYMBOL_GPL(ds1685_rtc_poweroff); /* ----------------------------------------------------------------------- */ From 9011e49d54dcc7653ebb8a1e05b5badb5ecfa9f9 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Tue, 1 Aug 2023 19:35:44 +0200 Subject: [PATCH 11/16] modules: only allow symbol_get of EXPORT_SYMBOL_GPL modules It has recently come to my attention that nvidia is circumventing the protection added in 262e6ae7081d ("modules: inherit TAINT_PROPRIETARY_MODULE") by importing exports from their proprietary modules into an allegedly GPL licensed module and then rexporting them. Given that symbol_get was only ever intended for tightly cooperating modules using very internal symbols it is logical to restrict it to being used on EXPORT_SYMBOL_GPL and prevent nvidia from costly DMCA Circumvention of Access Controls law suites. All symbols except for four used through symbol_get were already exported as EXPORT_SYMBOL_GPL, and the remaining four ones were switched over in the preparation patches. Fixes: 262e6ae7081d ("modules: inherit TAINT_PROPRIETARY_MODULE") Signed-off-by: Christoph Hellwig Reviewed-by: Greg Kroah-Hartman Signed-off-by: Luis Chamberlain --- kernel/module/main.c | 14 +++++++++++--- 1 file changed, 11 insertions(+), 3 deletions(-) diff --git a/kernel/module/main.c b/kernel/module/main.c index 59b1d067e528..c395af9eced1 100644 --- a/kernel/module/main.c +++ b/kernel/module/main.c @@ -1295,12 +1295,20 @@ void *__symbol_get(const char *symbol) }; preempt_disable(); - if (!find_symbol(&fsa) || strong_try_module_get(fsa.owner)) { - preempt_enable(); - return NULL; + if (!find_symbol(&fsa)) + goto fail; + if (fsa.license != GPL_ONLY) { + pr_warn("failing symbol_get of non-GPLONLY symbol %s.\n", + symbol); + goto fail; } + if (strong_try_module_get(fsa.owner)) + goto fail; preempt_enable(); return (void *)kernel_symbol_value(fsa.sym); +fail: + preempt_enable(); + return NULL; } EXPORT_SYMBOL_GPL(__symbol_get); From 2abcc4b5a64a65a2d2287ba0be5c2871c1552416 Mon Sep 17 00:00:00 2001 From: James Morse Date: Tue, 1 Aug 2023 14:54:07 +0000 Subject: [PATCH 12/16] module: Expose module_init_layout_section() module_init_layout_section() choses whether the core module loader considers a section as init or not. This affects the placement of the exit section when module unloading is disabled. This code will never run, so it can be free()d once the module has been initialised. arm and arm64 need to count the number of PLTs they need before applying relocations based on the section name. The init PLTs are stored separately so they can be free()d. arm and arm64 both use within_module_init() to decide which list of PLTs to use when applying the relocation. Because within_module_init()'s behaviour changes when module unloading is disabled, both architecture would need to take this into account when counting the PLTs. Today neither architecture does this, meaning when module unloading is disabled there are insufficient PLTs in the init section to load some modules, resulting in warnings: | WARNING: CPU: 2 PID: 51 at arch/arm64/kernel/module-plts.c:99 module_emit_plt_entry+0x184/0x1cc | Modules linked in: crct10dif_common | CPU: 2 PID: 51 Comm: modprobe Not tainted 6.5.0-rc4-yocto-standard-dirty #15208 | Hardware name: QEMU KVM Virtual Machine, BIOS 0.0.0 02/06/2015 | pstate: 20400005 (nzCv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--) | pc : module_emit_plt_entry+0x184/0x1cc | lr : module_emit_plt_entry+0x94/0x1cc | sp : ffffffc0803bba60 [...] | Call trace: | module_emit_plt_entry+0x184/0x1cc | apply_relocate_add+0x2bc/0x8e4 | load_module+0xe34/0x1bd4 | init_module_from_file+0x84/0xc0 | __arm64_sys_finit_module+0x1b8/0x27c | invoke_syscall.constprop.0+0x5c/0x104 | do_el0_svc+0x58/0x160 | el0_svc+0x38/0x110 | el0t_64_sync_handler+0xc0/0xc4 | el0t_64_sync+0x190/0x194 Instead of duplicating module_init_layout_section()s logic, expose it. Reported-by: Adam Johnston Fixes: 055f23b74b20 ("module: check for exit sections in layout_sections() instead of module_init_section()") Cc: stable@vger.kernel.org Signed-off-by: James Morse Signed-off-by: Luis Chamberlain --- include/linux/moduleloader.h | 5 +++++ kernel/module/main.c | 2 +- 2 files changed, 6 insertions(+), 1 deletion(-) diff --git a/include/linux/moduleloader.h b/include/linux/moduleloader.h index 03be088fb439..001b2ce83832 100644 --- a/include/linux/moduleloader.h +++ b/include/linux/moduleloader.h @@ -42,6 +42,11 @@ bool module_init_section(const char *name); */ bool module_exit_section(const char *name); +/* Describes whether within_module_init() will consider this an init section + * or not. This behaviour changes with CONFIG_MODULE_UNLOAD. + */ +bool module_init_layout_section(const char *sname); + /* * Apply the given relocation to the (simplified) ELF. Return -error * or 0. diff --git a/kernel/module/main.c b/kernel/module/main.c index c395af9eced1..98fedfdb8db5 100644 --- a/kernel/module/main.c +++ b/kernel/module/main.c @@ -1492,7 +1492,7 @@ long module_get_offset_and_type(struct module *mod, enum mod_mem_type type, return offset | mask; } -static bool module_init_layout_section(const char *sname) +bool module_init_layout_section(const char *sname) { #ifndef CONFIG_MODULE_UNLOAD if (module_exit_section(sname)) From f928f8b1a2496e7af95b860f9acf553f20f68f16 Mon Sep 17 00:00:00 2001 From: James Morse Date: Tue, 1 Aug 2023 14:54:08 +0000 Subject: [PATCH 13/16] arm64: module: Use module_init_layout_section() to spot init sections Today module_frob_arch_sections() spots init sections from their 'init' prefix, and uses this to keep the init PLTs separate from the rest. module_emit_plt_entry() uses within_module_init() to determine if a location is in the init text or not, but this depends on whether core code thought this was an init section. Naturally the logic is different. module_init_layout_section() groups the init and exit text together if module unloading is disabled, as the exit code will never run. The result is kernels with this configuration can't load all their modules because there are not enough PLTs for the combined init+exit section. This results in the following: | WARNING: CPU: 2 PID: 51 at arch/arm64/kernel/module-plts.c:99 module_emit_plt_entry+0x184/0x1cc | Modules linked in: crct10dif_common | CPU: 2 PID: 51 Comm: modprobe Not tainted 6.5.0-rc4-yocto-standard-dirty #15208 | Hardware name: QEMU KVM Virtual Machine, BIOS 0.0.0 02/06/2015 | pstate: 20400005 (nzCv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--) | pc : module_emit_plt_entry+0x184/0x1cc | lr : module_emit_plt_entry+0x94/0x1cc | sp : ffffffc0803bba60 [...] | Call trace: | module_emit_plt_entry+0x184/0x1cc | apply_relocate_add+0x2bc/0x8e4 | load_module+0xe34/0x1bd4 | init_module_from_file+0x84/0xc0 | __arm64_sys_finit_module+0x1b8/0x27c | invoke_syscall.constprop.0+0x5c/0x104 | do_el0_svc+0x58/0x160 | el0_svc+0x38/0x110 | el0t_64_sync_handler+0xc0/0xc4 | el0t_64_sync+0x190/0x194 A previous patch exposed module_init_layout_section(), use that so the logic is the same. Reported-by: Adam Johnston Tested-by: Adam Johnston Fixes: 055f23b74b20 ("module: check for exit sections in layout_sections() instead of module_init_section()") Cc: # 5.15.x: 60a0aab7463ee69 arm64: module-plts: inline linux/moduleloader.h Cc: # 5.15.x Signed-off-by: James Morse Acked-by: Catalin Marinas Signed-off-by: Luis Chamberlain --- arch/arm64/kernel/module-plts.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/arch/arm64/kernel/module-plts.c b/arch/arm64/kernel/module-plts.c index ad02058756b5..bd69a4e7cd60 100644 --- a/arch/arm64/kernel/module-plts.c +++ b/arch/arm64/kernel/module-plts.c @@ -339,7 +339,7 @@ int module_frob_arch_sections(Elf_Ehdr *ehdr, Elf_Shdr *sechdrs, if (nents) sort(rels, nents, sizeof(Elf64_Rela), cmp_rela, NULL); - if (!str_has_prefix(secstrings + dstsec->sh_name, ".init")) + if (!module_init_layout_section(secstrings + dstsec->sh_name)) core_plts += count_plts(syms, rels, numrels, sechdrs[i].sh_info, dstsec); else From a6846234f45801441f0e31a8b37f901ef0abd2df Mon Sep 17 00:00:00 2001 From: James Morse Date: Tue, 1 Aug 2023 14:54:09 +0000 Subject: [PATCH 14/16] ARM: module: Use module_init_layout_section() to spot init sections Today module_frob_arch_sections() spots init sections from their 'init' prefix, and uses this to keep the init PLTs separate from the rest. get_module_plt() uses within_module_init() to determine if a location is in the init text or not, but this depends on whether core code thought this was an init section. Naturally the logic is different. module_init_layout_section() groups the init and exit text together if module unloading is disabled, as the exit code will never run. The result is kernels with this configuration can't load all their modules because there are not enough PLTs for the combined init+exit section. A previous patch exposed module_init_layout_section(), use that so the logic is the same. Fixes: 055f23b74b20 ("module: check for exit sections in layout_sections() instead of module_init_section()") Cc: stable@vger.kernel.org Signed-off-by: James Morse Signed-off-by: Luis Chamberlain --- arch/arm/kernel/module-plts.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/arch/arm/kernel/module-plts.c b/arch/arm/kernel/module-plts.c index f5a43fd8c163..da2ee8d6ef1a 100644 --- a/arch/arm/kernel/module-plts.c +++ b/arch/arm/kernel/module-plts.c @@ -251,7 +251,7 @@ int module_frob_arch_sections(Elf_Ehdr *ehdr, Elf_Shdr *sechdrs, /* sort by type and symbol index */ sort(rels, numrels, sizeof(Elf32_Rel), cmp_rel, NULL); - if (strncmp(secstrings + dstsec->sh_name, ".init", 5) != 0) + if (!module_init_layout_section(secstrings + dstsec->sh_name)) core_plts += count_plts(syms, dstsec->sh_addr, rels, numrels, s->sh_info); else From 33c24bee4b787ef58d8c9a36316ba85363b93617 Mon Sep 17 00:00:00 2001 From: Kees Cook Date: Thu, 24 Aug 2023 13:58:00 -0700 Subject: [PATCH 15/16] kallsyms: Add more debug output for selftest While debugging a recent kallsyms_selftest failure[1], I needed more details on what specifically was failing. This adds those details for each failure state that is checked. [1] https://lore.kernel.org/all/202308232200.1c932a90-oliver.sang@intel.com/ Cc: Luis Chamberlain Cc: Yonghong Song Cc: "Erhard F." Cc: Zhen Lei Cc: kernel test robot Cc: Petr Mladek Cc: Nicholas Piggin Cc: Yang Li Signed-off-by: Kees Cook Signed-off-by: Luis Chamberlain --- kernel/kallsyms_selftest.c | 22 ++++++++++++++++++---- 1 file changed, 18 insertions(+), 4 deletions(-) diff --git a/kernel/kallsyms_selftest.c b/kernel/kallsyms_selftest.c index a2e3745d15c4..232e8faefb46 100644 --- a/kernel/kallsyms_selftest.c +++ b/kernel/kallsyms_selftest.c @@ -341,6 +341,7 @@ static int test_kallsyms_basic_function(void) ret = lookup_symbol_name(addr, namebuf); if (unlikely(ret)) { namebuf[0] = 0; + pr_info("%d: lookup_symbol_name(%lx) failed\n", i, addr); goto failed; } @@ -388,8 +389,11 @@ static int test_kallsyms_basic_function(void) if (stat->addr != stat2->addr || stat->real_cnt != stat2->real_cnt || memcmp(stat->addrs, stat2->addrs, - stat->save_cnt * sizeof(stat->addrs[0]))) + stat->save_cnt * sizeof(stat->addrs[0]))) { + pr_info("%s: mismatch between kallsyms_on_each_symbol() and kallsyms_on_each_match_symbol()\n", + namebuf); goto failed; + } /* * The average of random increments is 128, that is, one of @@ -400,15 +404,23 @@ static int test_kallsyms_basic_function(void) } /* Need to be found at least once */ - if (!stat->real_cnt) + if (!stat->real_cnt) { + pr_info("%s: Never found\n", namebuf); goto failed; + } /* * kallsyms_lookup_name() returns the address of the first * symbol found and cannot be NULL. */ - if (!lookup_addr || lookup_addr != stat->addrs[0]) + if (!lookup_addr) { + pr_info("%s: NULL lookup_addr?!\n", namebuf); goto failed; + } + if (lookup_addr != stat->addrs[0]) { + pr_info("%s: lookup_addr != stat->addrs[0]\n", namebuf); + goto failed; + } /* * If the addresses of all matching symbols are recorded, the @@ -420,8 +432,10 @@ static int test_kallsyms_basic_function(void) break; } - if (j == stat->save_cnt) + if (j == stat->save_cnt) { + pr_info("%s: j == save_cnt?!\n", namebuf); goto failed; + } } } From a419beac4a070aff63c520f36ebf7cb8a76a8ae5 Mon Sep 17 00:00:00 2001 From: Andrea Righi Date: Tue, 29 Aug 2023 14:05:08 +0200 Subject: [PATCH 16/16] module/decompress: use vmalloc() for zstd decompression workspace Using kmalloc() to allocate the decompression workspace for zstd may trigger the following warning when large modules are loaded (i.e., xfs): [ 2.961884] WARNING: CPU: 1 PID: 254 at mm/page_alloc.c:4453 __alloc_pages+0x2c3/0x350 ... [ 2.989033] Call Trace: [ 2.989841] [ 2.990614] ? show_regs+0x6d/0x80 [ 2.991573] ? __warn+0x89/0x160 [ 2.992485] ? __alloc_pages+0x2c3/0x350 [ 2.993520] ? report_bug+0x17e/0x1b0 [ 2.994506] ? handle_bug+0x51/0xa0 [ 2.995474] ? exc_invalid_op+0x18/0x80 [ 2.996469] ? asm_exc_invalid_op+0x1b/0x20 [ 2.997530] ? module_zstd_decompress+0xdc/0x2a0 [ 2.998665] ? __alloc_pages+0x2c3/0x350 [ 2.999695] ? module_zstd_decompress+0xdc/0x2a0 [ 3.000821] __kmalloc_large_node+0x7a/0x150 [ 3.001920] __kmalloc+0xdb/0x170 [ 3.002824] module_zstd_decompress+0xdc/0x2a0 [ 3.003857] module_decompress+0x37/0xc0 [ 3.004688] init_module_from_file+0xd0/0x100 [ 3.005668] idempotent_init_module+0x11c/0x2b0 [ 3.006632] __x64_sys_finit_module+0x64/0xd0 [ 3.007568] do_syscall_64+0x59/0x90 [ 3.008373] ? ksys_read+0x73/0x100 [ 3.009395] ? exit_to_user_mode_prepare+0x30/0xb0 [ 3.010531] ? syscall_exit_to_user_mode+0x37/0x60 [ 3.011662] ? do_syscall_64+0x68/0x90 [ 3.012511] ? do_syscall_64+0x68/0x90 [ 3.013364] entry_SYSCALL_64_after_hwframe+0x6e/0xd8 However, continuous physical memory does not seem to be required in module_zstd_decompress(), so use vmalloc() instead, to prevent the warning and avoid potential failures at loading compressed modules. Fixes: 169a58ad824d ("module/decompress: Support zstd in-kernel decompression") Signed-off-by: Andrea Righi Signed-off-by: Luis Chamberlain --- kernel/module/decompress.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/kernel/module/decompress.c b/kernel/module/decompress.c index 8a5d6d63b06c..87440f714c0c 100644 --- a/kernel/module/decompress.c +++ b/kernel/module/decompress.c @@ -241,7 +241,7 @@ static ssize_t module_zstd_decompress(struct load_info *info, } wksp_size = zstd_dstream_workspace_bound(header.windowSize); - wksp = kmalloc(wksp_size, GFP_KERNEL); + wksp = vmalloc(wksp_size); if (!wksp) { retval = -ENOMEM; goto out; @@ -284,7 +284,7 @@ static ssize_t module_zstd_decompress(struct load_info *info, retval = new_size; out: - kfree(wksp); + vfree(wksp); return retval; } #else