Message ID | 20191018143643.29676-1-miquel.raynal@bootlin.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
Series | mtd: spear_smi: Fix nonalignment not handled in memcpy_toio | expand |
Hello, +Russell who might have thoughts about the issue Miquel Raynal <miquel.raynal@bootlin.com> wrote on Fri, 18 Oct 2019 16:36:43 +0200: > Any write with either dd or flashcp to a device driven by the > spear_smi.c driver will pass through the spear_smi_cpy_toio() > function. This function will get called for chunks of up to 256 bytes. > If the amount of data is smaller, we may have a problem if the data > length is not 4-byte aligned. In this situation, the kernel panics > during the memcpy: > > # dd if=/dev/urandom bs=1001 count=1 of=/dev/mtd6 > spear_smi_cpy_toio [620] dest c9070000, src c7be8800, len 256 > spear_smi_cpy_toio [620] dest c9070100, src c7be8900, len 256 > spear_smi_cpy_toio [620] dest c9070200, src c7be8a00, len 256 > spear_smi_cpy_toio [620] dest c9070300, src c7be8b00, len 233 > Unhandled fault: external abort on non-linefetch (0x808) at 0xc90703e8 > [...] > PC is at memcpy+0xcc/0x330 > > Workaround this issue by using the alternate _memcpy_toio() method > which at least does not present the same problem. > > Fixes: f18dbbb1bfe0 ("mtd: ST SPEAr: Add SMI driver for serial NOR flash") > Cc: stable@vger.kernel.org > Suggested-by: Boris Brezillon <boris.brezillon@collabora.com> > Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com> > --- > > Hello, > > This patch could not be tested with a mainline kernel (only compiled) > but was tested with a stable 4.14.x kernel. I have really no idea why > memcpy fails in this situation that's why I propose this workaround > but I bet there is something deeper not working. I did not mention that the opposite direction using memcpy_fromio() does not present any issue. > > Thanks, > Miquèl > > drivers/mtd/devices/spear_smi.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/drivers/mtd/devices/spear_smi.c b/drivers/mtd/devices/spear_smi.c > index 986f81d2f93e..d888625a3244 100644 > --- a/drivers/mtd/devices/spear_smi.c > +++ b/drivers/mtd/devices/spear_smi.c > @@ -614,7 +614,7 @@ static inline int spear_smi_cpy_toio(struct spear_smi *dev, u32 bank, > ctrlreg1 = readl(dev->io_base + SMI_CR1); > writel((ctrlreg1 | WB_MODE) & ~SW_MODE, dev->io_base + SMI_CR1); > > - memcpy_toio(dest, src, len); > + _memcpy_toio(dest, src, len); > > writel(ctrlreg1, dev->io_base + SMI_CR1); > Thanks, Miquèl
On Fri, 18 Oct 2019 16:36:43 +0200 Miquel Raynal <miquel.raynal@bootlin.com> wrote: > Any write with either dd or flashcp to a device driven by the > spear_smi.c driver will pass through the spear_smi_cpy_toio() > function. This function will get called for chunks of up to 256 bytes. > If the amount of data is smaller, we may have a problem if the data > length is not 4-byte aligned. In this situation, the kernel panics > during the memcpy: > > # dd if=/dev/urandom bs=1001 count=1 of=/dev/mtd6 > spear_smi_cpy_toio [620] dest c9070000, src c7be8800, len 256 > spear_smi_cpy_toio [620] dest c9070100, src c7be8900, len 256 > spear_smi_cpy_toio [620] dest c9070200, src c7be8a00, len 256 > spear_smi_cpy_toio [620] dest c9070300, src c7be8b00, len 233 > Unhandled fault: external abort on non-linefetch (0x808) at 0xc90703e8 > [...] > PC is at memcpy+0xcc/0x330 Can you find out which instruction is at memcpy+0xcc/0x330? For the record, the assembly is here [1]. > > Workaround this issue by using the alternate _memcpy_toio() method > which at least does not present the same problem. > > Fixes: f18dbbb1bfe0 ("mtd: ST SPEAr: Add SMI driver for serial NOR flash") > Cc: stable@vger.kernel.org > Suggested-by: Boris Brezillon <boris.brezillon@collabora.com> I don't remember suggesting that as a final solution. I probably suggested to test with _memcpy_toio() to see if using a byte accessor was fixing the problem, but it's definitely not the right solution (using byte access with a memory barrier for 256 bytes buffers is likely to cause a huge perf penalty). > Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com> > --- > > Hello, > > This patch could not be tested with a mainline kernel (only compiled) > but was tested with a stable 4.14.x kernel. I have really no idea why > memcpy fails in this situation that's why I propose this workaround > but I bet there is something deeper not working. > > Thanks, > Miquèl > > drivers/mtd/devices/spear_smi.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/drivers/mtd/devices/spear_smi.c b/drivers/mtd/devices/spear_smi.c > index 986f81d2f93e..d888625a3244 100644 > --- a/drivers/mtd/devices/spear_smi.c > +++ b/drivers/mtd/devices/spear_smi.c > @@ -614,7 +614,7 @@ static inline int spear_smi_cpy_toio(struct spear_smi *dev, u32 bank, > ctrlreg1 = readl(dev->io_base + SMI_CR1); > writel((ctrlreg1 | WB_MODE) & ~SW_MODE, dev->io_base + SMI_CR1); > > - memcpy_toio(dest, src, len); > + _memcpy_toio(dest, src, len); > > writel(ctrlreg1, dev->io_base + SMI_CR1); > [1]https://elixir.bootlin.com/linux/v5.4-rc2/source/arch/arm/lib/memcpy.S
Hello, Boris Brezillon <boris.brezillon@collabora.com> wrote on Mon, 21 Oct 2019 10:01:05 +0200: > On Fri, 18 Oct 2019 16:36:43 +0200 > Miquel Raynal <miquel.raynal@bootlin.com> wrote: > > > Any write with either dd or flashcp to a device driven by the > > spear_smi.c driver will pass through the spear_smi_cpy_toio() > > function. This function will get called for chunks of up to 256 bytes. > > If the amount of data is smaller, we may have a problem if the data > > length is not 4-byte aligned. In this situation, the kernel panics > > during the memcpy: > > > > # dd if=/dev/urandom bs=1001 count=1 of=/dev/mtd6 > > spear_smi_cpy_toio [620] dest c9070000, src c7be8800, len 256 > > spear_smi_cpy_toio [620] dest c9070100, src c7be8900, len 256 > > spear_smi_cpy_toio [620] dest c9070200, src c7be8a00, len 256 > > spear_smi_cpy_toio [620] dest c9070300, src c7be8b00, len 233 > > Unhandled fault: external abort on non-linefetch (0x808) at 0xc90703e8 > > [...] > > PC is at memcpy+0xcc/0x330 > > Can you find out which instruction is at memcpy+0xcc/0x330? For the > record, the assembly is here [1]. The full disassembled file is attached, here is the failing part: 7: ldmfd sp!, {r5 - r8} b8: e8bd01e0 pop {r5, r6, r7, r8} UNWIND( .fnend ) @ end of second stmfd block UNWIND( .fnstart ) usave r4, lr @ still in first stmdb block 8: movs r2, r2, lsl #31 bc: e1b02f82 lsls r2, r2, #31 ldr1b r1, r3, ne, abort=21f c0: 14d13001 ldrbne r3, [r1], #1 ldr1b r1, r4, cs, abort=21f c4: 24d14001 ldrbcs r4, [r1], #1 ldr1b r1, ip, cs, abort=21f c8: 24d1c001 ldrbcs ip, [r1], #1 str1b r0, r3, ne, abort=21f cc: 14c03001 strbne r3, [r0], #1 str1b r0, r4, cs, abort=21f d0: 24c04001 strbcs r4, [r0], #1 str1b r0, ip, cs, abort=21f d4: 24c0c001 strbcs ip, [r0], #1 exit r4, pc d8: e8bd8011 pop {r0, r4, pc} So the fault is triggered on a strbne instruction. > > > > Workaround this issue by using the alternate _memcpy_toio() method > > which at least does not present the same problem. > > > > Fixes: f18dbbb1bfe0 ("mtd: ST SPEAr: Add SMI driver for serial NOR flash") > > Cc: stable@vger.kernel.org > > Suggested-by: Boris Brezillon <boris.brezillon@collabora.com> > > I don't remember suggesting that as a final solution. I probably > suggested to test with _memcpy_toio() to see if using a byte accessor > was fixing the problem, but it's definitely not the right solution > (using byte access with a memory barrier for 256 bytes buffers is likely > to cause a huge perf penalty). > > > Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com> > > --- > > > > Hello, > > > > This patch could not be tested with a mainline kernel (only compiled) > > but was tested with a stable 4.14.x kernel. I have really no idea why > > memcpy fails in this situation that's why I propose this workaround > > but I bet there is something deeper not working. > > > > Thanks, > > Miquèl > > > > drivers/mtd/devices/spear_smi.c | 2 +- > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > diff --git a/drivers/mtd/devices/spear_smi.c b/drivers/mtd/devices/spear_smi.c > > index 986f81d2f93e..d888625a3244 100644 > > --- a/drivers/mtd/devices/spear_smi.c > > +++ b/drivers/mtd/devices/spear_smi.c > > @@ -614,7 +614,7 @@ static inline int spear_smi_cpy_toio(struct spear_smi *dev, u32 bank, > > ctrlreg1 = readl(dev->io_base + SMI_CR1); > > writel((ctrlreg1 | WB_MODE) & ~SW_MODE, dev->io_base + SMI_CR1); > > > > - memcpy_toio(dest, src, len); > > + _memcpy_toio(dest, src, len); > > > > writel(ctrlreg1, dev->io_base + SMI_CR1); > > > > [1]https://elixir.bootlin.com/linux/v5.4-rc2/source/arch/arm/lib/memcpy.S Thanks, Miquèl
Russel was out of the loop, re-adding him as he may have interesting thoughts about this. Sorry for the double e-mail. Miquel Raynal <miquel.raynal@bootlin.com> wrote on Tue, 22 Oct 2019 09:44:51 +0200: > Hello, > > Boris Brezillon <boris.brezillon@collabora.com> wrote on Mon, 21 Oct > 2019 10:01:05 +0200: > > > On Fri, 18 Oct 2019 16:36:43 +0200 > > Miquel Raynal <miquel.raynal@bootlin.com> wrote: > > > > > Any write with either dd or flashcp to a device driven by the > > > spear_smi.c driver will pass through the spear_smi_cpy_toio() > > > function. This function will get called for chunks of up to 256 bytes. > > > If the amount of data is smaller, we may have a problem if the data > > > length is not 4-byte aligned. In this situation, the kernel panics > > > during the memcpy: > > > > > > # dd if=/dev/urandom bs=1001 count=1 of=/dev/mtd6 > > > spear_smi_cpy_toio [620] dest c9070000, src c7be8800, len 256 > > > spear_smi_cpy_toio [620] dest c9070100, src c7be8900, len 256 > > > spear_smi_cpy_toio [620] dest c9070200, src c7be8a00, len 256 > > > spear_smi_cpy_toio [620] dest c9070300, src c7be8b00, len 233 > > > Unhandled fault: external abort on non-linefetch (0x808) at 0xc90703e8 > > > [...] > > > PC is at memcpy+0xcc/0x330 > > > > Can you find out which instruction is at memcpy+0xcc/0x330? For the > > record, the assembly is here [1]. > > The full disassembled file is attached, here is the failing part: > > 7: ldmfd sp!, {r5 - r8} > b8: e8bd01e0 pop {r5, r6, r7, r8} > UNWIND( .fnend ) @ end of second stmfd block > > UNWIND( .fnstart ) > usave r4, lr @ still in first stmdb block > 8: movs r2, r2, lsl #31 > bc: e1b02f82 lsls r2, r2, #31 > ldr1b r1, r3, ne, abort=21f > c0: 14d13001 ldrbne r3, [r1], #1 > ldr1b r1, r4, cs, abort=21f > c4: 24d14001 ldrbcs r4, [r1], #1 > ldr1b r1, ip, cs, abort=21f > c8: 24d1c001 ldrbcs ip, [r1], #1 > str1b r0, r3, ne, abort=21f > cc: 14c03001 strbne r3, [r0], #1 > str1b r0, r4, cs, abort=21f > d0: 24c04001 strbcs r4, [r0], #1 > str1b r0, ip, cs, abort=21f > d4: 24c0c001 strbcs ip, [r0], #1 > > exit r4, pc > d8: e8bd8011 pop {r0, r4, pc} > > > So the fault is triggered on a strbne instruction. > > > > > > > Workaround this issue by using the alternate _memcpy_toio() method > > > which at least does not present the same problem. > > > > > > Fixes: f18dbbb1bfe0 ("mtd: ST SPEAr: Add SMI driver for serial NOR flash") > > > Cc: stable@vger.kernel.org > > > Suggested-by: Boris Brezillon <boris.brezillon@collabora.com> > > > > I don't remember suggesting that as a final solution. I probably > > suggested to test with _memcpy_toio() to see if using a byte accessor > > was fixing the problem, but it's definitely not the right solution > > (using byte access with a memory barrier for 256 bytes buffers is likely > > to cause a huge perf penalty). > > > > > Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com> > > > --- > > > > > > Hello, > > > > > > This patch could not be tested with a mainline kernel (only compiled) > > > but was tested with a stable 4.14.x kernel. I have really no idea why > > > memcpy fails in this situation that's why I propose this workaround > > > but I bet there is something deeper not working. > > > > > > Thanks, > > > Miquèl > > > > > > drivers/mtd/devices/spear_smi.c | 2 +- > > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > > > diff --git a/drivers/mtd/devices/spear_smi.c b/drivers/mtd/devices/spear_smi.c > > > index 986f81d2f93e..d888625a3244 100644 > > > --- a/drivers/mtd/devices/spear_smi.c > > > +++ b/drivers/mtd/devices/spear_smi.c > > > @@ -614,7 +614,7 @@ static inline int spear_smi_cpy_toio(struct spear_smi *dev, u32 bank, > > > ctrlreg1 = readl(dev->io_base + SMI_CR1); > > > writel((ctrlreg1 | WB_MODE) & ~SW_MODE, dev->io_base + SMI_CR1); > > > > > > - memcpy_toio(dest, src, len); > > > + _memcpy_toio(dest, src, len); > > > > > > writel(ctrlreg1, dev->io_base + SMI_CR1); > > > > > > > [1]https://elixir.bootlin.com/linux/v5.4-rc2/source/arch/arm/lib/memcpy.S > Thanks, Miquèl
Hello, +Russell in Cc. On Tue, 22 Oct 2019 09:44:51 +0200 Miquel Raynal <miquel.raynal@bootlin.com> wrote: > Boris Brezillon <boris.brezillon@collabora.com> wrote on Mon, 21 Oct > 2019 10:01:05 +0200: > > > On Fri, 18 Oct 2019 16:36:43 +0200 > > Miquel Raynal <miquel.raynal@bootlin.com> wrote: > > > > > Any write with either dd or flashcp to a device driven by the > > > spear_smi.c driver will pass through the spear_smi_cpy_toio() > > > function. This function will get called for chunks of up to 256 bytes. > > > If the amount of data is smaller, we may have a problem if the data > > > length is not 4-byte aligned. In this situation, the kernel panics > > > during the memcpy: > > > > > > # dd if=/dev/urandom bs=1001 count=1 of=/dev/mtd6 > > > spear_smi_cpy_toio [620] dest c9070000, src c7be8800, len 256 > > > spear_smi_cpy_toio [620] dest c9070100, src c7be8900, len 256 > > > spear_smi_cpy_toio [620] dest c9070200, src c7be8a00, len 256 > > > spear_smi_cpy_toio [620] dest c9070300, src c7be8b00, len 233 > > > Unhandled fault: external abort on non-linefetch (0x808) at 0xc90703e8 > > > [...] > > > PC is at memcpy+0xcc/0x330 > > > > Can you find out which instruction is at memcpy+0xcc/0x330? For the > > record, the assembly is here [1]. > > The full disassembled file is attached, here is the failing part: > > 7: ldmfd sp!, {r5 - r8} > b8: e8bd01e0 pop {r5, r6, r7, r8} > UNWIND( .fnend ) @ end of second stmfd block > > UNWIND( .fnstart ) > usave r4, lr @ still in first stmdb block > 8: movs r2, r2, lsl #31 > bc: e1b02f82 lsls r2, r2, #31 > ldr1b r1, r3, ne, abort=21f > c0: 14d13001 ldrbne r3, [r1], #1 > ldr1b r1, r4, cs, abort=21f > c4: 24d14001 ldrbcs r4, [r1], #1 > ldr1b r1, ip, cs, abort=21f > c8: 24d1c001 ldrbcs ip, [r1], #1 > str1b r0, r3, ne, abort=21f > cc: 14c03001 strbne r3, [r0], #1 > str1b r0, r4, cs, abort=21f > d0: 24c04001 strbcs r4, [r0], #1 > str1b r0, ip, cs, abort=21f > d4: 24c0c001 strbcs ip, [r0], #1 > > exit r4, pc > d8: e8bd8011 pop {r0, r4, pc} > > > So the fault is triggered on a strbne instruction. What I find odd is: (1) Failing on a 1-byte store instruction, which means it should have no alignment constraints. (2) Failing on a 1-byte store instruction, while switching to _memcpy_toio(), which does *only* 1-byte stores, works around the problem. _memcpy_toio() looks like this: void _memcpy_toio(volatile void __iomem *to, const void *from, size_t count) { const unsigned char *f = from; while (count) { 6c: e3520000 cmp r2, #0 70: 012fff1e bxeq lr 74: e0802002 add r2, r0, r2 count--; writeb(*f, to); 78: e4d13001 ldrb r3, [r1], #1 asm volatile("strb %1, %0" 7c: e5c03000 strb r3, [r0] f++; to++; 80: e2800001 add r0, r0, #1 while (count) { 84: e1500002 cmp r0, r2 88: 1afffffa bne 78 <_memcpy_toio+0xc> 8c: e12fff1e bx lr So it's also doing a strb, nothing different. Best regards, Thomas
On Fri, Oct 18, 2019 at 04:36:43PM +0200, Miquel Raynal wrote: > Any write with either dd or flashcp to a device driven by the > spear_smi.c driver will pass through the spear_smi_cpy_toio() > function. This function will get called for chunks of up to 256 bytes. > If the amount of data is smaller, we may have a problem if the data > length is not 4-byte aligned. In this situation, the kernel panics > during the memcpy: > > # dd if=/dev/urandom bs=1001 count=1 of=/dev/mtd6 > spear_smi_cpy_toio [620] dest c9070000, src c7be8800, len 256 > spear_smi_cpy_toio [620] dest c9070100, src c7be8900, len 256 > spear_smi_cpy_toio [620] dest c9070200, src c7be8a00, len 256 > spear_smi_cpy_toio [620] dest c9070300, src c7be8b00, len 233 > Unhandled fault: external abort on non-linefetch (0x808) at 0xc90703e8 > [...] > PC is at memcpy+0xcc/0x330 I need the full oops if you want me to comment on this. > > Workaround this issue by using the alternate _memcpy_toio() method > which at least does not present the same problem. > > Fixes: f18dbbb1bfe0 ("mtd: ST SPEAr: Add SMI driver for serial NOR flash") > Cc: stable@vger.kernel.org > Suggested-by: Boris Brezillon <boris.brezillon@collabora.com> > Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com> > --- > > Hello, > > This patch could not be tested with a mainline kernel (only compiled) > but was tested with a stable 4.14.x kernel. I have really no idea why > memcpy fails in this situation that's why I propose this workaround > but I bet there is something deeper not working. > > Thanks, > Miquèl > > drivers/mtd/devices/spear_smi.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/drivers/mtd/devices/spear_smi.c b/drivers/mtd/devices/spear_smi.c > index 986f81d2f93e..d888625a3244 100644 > --- a/drivers/mtd/devices/spear_smi.c > +++ b/drivers/mtd/devices/spear_smi.c > @@ -614,7 +614,7 @@ static inline int spear_smi_cpy_toio(struct spear_smi *dev, u32 bank, > ctrlreg1 = readl(dev->io_base + SMI_CR1); > writel((ctrlreg1 | WB_MODE) & ~SW_MODE, dev->io_base + SMI_CR1); > > - memcpy_toio(dest, src, len); > + _memcpy_toio(dest, src, len); > > writel(ctrlreg1, dev->io_base + SMI_CR1); > > -- > 2.20.1 > > > _______________________________________________ > linux-arm-kernel mailing list > linux-arm-kernel@lists.infradead.org > http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
Hi Russell, Russell King - ARM Linux admin <linux@armlinux.org.uk> wrote on Tue, 22 Oct 2019 09:26:43 +0100: > On Fri, Oct 18, 2019 at 04:36:43PM +0200, Miquel Raynal wrote: > > Any write with either dd or flashcp to a device driven by the > > spear_smi.c driver will pass through the spear_smi_cpy_toio() > > function. This function will get called for chunks of up to 256 bytes. > > If the amount of data is smaller, we may have a problem if the data > > length is not 4-byte aligned. In this situation, the kernel panics > > during the memcpy: > > > > # dd if=/dev/urandom bs=1001 count=1 of=/dev/mtd6 > > spear_smi_cpy_toio [620] dest c9070000, src c7be8800, len 256 > > spear_smi_cpy_toio [620] dest c9070100, src c7be8900, len 256 > > spear_smi_cpy_toio [620] dest c9070200, src c7be8a00, len 256 > > spear_smi_cpy_toio [620] dest c9070300, src c7be8b00, len 233 > > Unhandled fault: external abort on non-linefetch (0x808) at 0xc90703e8 > > [...] > > PC is at memcpy+0xcc/0x330 > > I need the full oops if you want me to comment on this. FYI, I ran the dd command within a for loop, incrementing the block size (bs) by one byte, if failed with bs=6. Disabling WB_MODE (burst mode) does not change anything. Adding a wmb() right after the memcpy_toio() prevents the fault. Here is the full trace when writing 1001 bytes: # dd if=/dev/urandom bs=1001 count=1 of=/dev/mtd6 Unhandled fault: external abort on non-linefetch (0x808) at 0xc90703e8 pgd = c7be8000 [c90703e8] *pgd=f8000452(bad) Internal error: : 808 [#1] ARM Modules linked in: CPU: 0 PID: 660 Comm: dd Not tainted 4.14.0-00045-gf5d08192704f-dirty #6 Hardware name: ST SPEAr600 (Flattened Device Tree) task: c7a05080 task.stack: c7bd2000 PC is at memcpy+0xcc/0x330 LR is at 0x13f0ec28 pc : [<c044344c>] lr : [<13f0ec28>] psr: 80000013 sp : c7bd3e44 ip : 00000018 fp : 000003e9 r10: 00000000 r9 : c7a9959c r8 : c7bd3eac r7 : c7a99590 r6 : c7afb438 r5 : 00000300 r4 : 5171436c r3 : 00000058 r2 : 80000000 r1 : c7be4be9 r0 : c90703e8 Flags: Nzcv IRQs on FIQs on Mode SVC_32 ISA ARM Segment none Control: 0005317f Table: 07be8000 DAC: 00000051 Process dd (pid: 660, stack limit = 0xc7bd2190) Stack: (0xc7bd3e44 to 0xc7bd4000) 3e40: c9070300 000000e9 c0290d14 c9070300 c7be4b00 0001046f c9070000 3e60: c7afb418 00000000 c7bd3e98 000003e9 c7bd3f88 000003e9 00000000 000c0008 3e80: 00000051 c7bd2000 c7be4800 c028e57c 000003e9 c7bd3eac c7be4800 00000000 3ea0: c7bf73c0 c7addc00 000003e9 00000300 00000000 00000000 00000000 00000000 3ec0: 00000000 00000000 00000000 00000000 00000000 000003e9 c028e4bc c7962a80 3ee0: c7bd3f88 00000000 c7bd2000 00000000 000bf990 c00bdb0c 000bf990 c00bd878 3f00: 00000000 00000000 00000000 c7bd3f10 c7a688c0 c009eedc c7becb58 000c0000 3f20: 00000003 c7962460 c7962484 00000000 00000000 c045c2f8 00000003 c00d9e58 3f40: 000003e9 000c0008 c7962a80 c7bd3f88 00000000 c7bd2000 00000000 c00bddb4 3f60: 000bf990 c00bda6c 00000000 c7962a80 c7962a80 000c0008 000003e9 c000a804 3f80: c7bd2000 c00bdfa4 00000000 00000000 00000000 000bfd94 00000001 000c0008 3fa0: 00000004 c000a640 000bfd94 00000001 00000001 000c0008 000003e9 be8e8f53 3fc0: 000bfd94 00000001 000c0008 00000004 000c0008 000c0008 000003e9 000bf990 3fe0: 00000000 be8e8ba4 0000ea3c b6eba7ec 60000010 00000001 00000000 00000000 [<c044344c>] (memcpy) from [<c0290d14>] (spear_mtd_write+0x240/0x294) [<c0290d14>] (spear_mtd_write) from [<c028e57c>] (mtdchar_write+0xc0/0x230) [<c028e57c>] (mtdchar_write) from [<c00bdb0c>] (__vfs_write+0x1c/0x128) [<c00bdb0c>] (__vfs_write) from [<c00bddb4>] (vfs_write+0xa0/0x168) [<c00bddb4>] (vfs_write) from [<c00bdfa4>] (SyS_write+0x3c/0x90) [<c00bdfa4>] (SyS_write) from [<c000a640>] (ret_fast_syscall+0x0/0x44) Code: e1b02f82 14d13001 24d14001 24d1c001 (14c03001) ---[ end trace f9a736cc2841cf14 ]--- Segmentation fault Thanks, Miquèl
On Tue, Oct 22, 2019 at 11:17:07AM +0200, Miquel Raynal wrote: > Hi Russell, > > Russell King - ARM Linux admin <linux@armlinux.org.uk> wrote on Tue, 22 > Oct 2019 09:26:43 +0100: > > > On Fri, Oct 18, 2019 at 04:36:43PM +0200, Miquel Raynal wrote: > > > Any write with either dd or flashcp to a device driven by the > > > spear_smi.c driver will pass through the spear_smi_cpy_toio() > > > function. This function will get called for chunks of up to 256 bytes. > > > If the amount of data is smaller, we may have a problem if the data > > > length is not 4-byte aligned. In this situation, the kernel panics > > > during the memcpy: > > > > > > # dd if=/dev/urandom bs=1001 count=1 of=/dev/mtd6 > > > spear_smi_cpy_toio [620] dest c9070000, src c7be8800, len 256 > > > spear_smi_cpy_toio [620] dest c9070100, src c7be8900, len 256 > > > spear_smi_cpy_toio [620] dest c9070200, src c7be8a00, len 256 > > > spear_smi_cpy_toio [620] dest c9070300, src c7be8b00, len 233 > > > Unhandled fault: external abort on non-linefetch (0x808) at 0xc90703e8 > > > [...] > > > PC is at memcpy+0xcc/0x330 > > > > I need the full oops if you want me to comment on this. > > FYI, I ran the dd command within a for loop, incrementing the block size > (bs) by one byte, if failed with bs=6. > > Disabling WB_MODE (burst mode) does not change anything. > > Adding a wmb() right after the memcpy_toio() prevents the fault. Thanks. Can you check what the result of the write buffer test earlier in the kernel boot is? CPU: Testing write buffer coherency: ... ? Thanks. > > Here is the full trace when writing 1001 bytes: > > # dd if=/dev/urandom bs=1001 count=1 of=/dev/mtd6 > Unhandled fault: external abort on non-linefetch (0x808) at 0xc90703e8 > pgd = c7be8000 > [c90703e8] *pgd=f8000452(bad) > Internal error: : 808 [#1] ARM > Modules linked in: > CPU: 0 PID: 660 Comm: dd Not tainted 4.14.0-00045-gf5d08192704f-dirty #6 > Hardware name: ST SPEAr600 (Flattened Device Tree) > task: c7a05080 task.stack: c7bd2000 > PC is at memcpy+0xcc/0x330 > LR is at 0x13f0ec28 > pc : [<c044344c>] lr : [<13f0ec28>] psr: 80000013 > sp : c7bd3e44 ip : 00000018 fp : 000003e9 > r10: 00000000 r9 : c7a9959c r8 : c7bd3eac > r7 : c7a99590 r6 : c7afb438 r5 : 00000300 r4 : 5171436c > r3 : 00000058 r2 : 80000000 r1 : c7be4be9 r0 : c90703e8 > Flags: Nzcv IRQs on FIQs on Mode SVC_32 ISA ARM Segment none > Control: 0005317f Table: 07be8000 DAC: 00000051 > Process dd (pid: 660, stack limit = 0xc7bd2190) > Stack: (0xc7bd3e44 to 0xc7bd4000) > 3e40: c9070300 000000e9 c0290d14 c9070300 c7be4b00 0001046f c9070000 > 3e60: c7afb418 00000000 c7bd3e98 000003e9 c7bd3f88 000003e9 00000000 000c0008 > 3e80: 00000051 c7bd2000 c7be4800 c028e57c 000003e9 c7bd3eac c7be4800 00000000 > 3ea0: c7bf73c0 c7addc00 000003e9 00000300 00000000 00000000 00000000 00000000 > 3ec0: 00000000 00000000 00000000 00000000 00000000 000003e9 c028e4bc c7962a80 > 3ee0: c7bd3f88 00000000 c7bd2000 00000000 000bf990 c00bdb0c 000bf990 c00bd878 > 3f00: 00000000 00000000 00000000 c7bd3f10 c7a688c0 c009eedc c7becb58 000c0000 > 3f20: 00000003 c7962460 c7962484 00000000 00000000 c045c2f8 00000003 c00d9e58 > 3f40: 000003e9 000c0008 c7962a80 c7bd3f88 00000000 c7bd2000 00000000 c00bddb4 > 3f60: 000bf990 c00bda6c 00000000 c7962a80 c7962a80 000c0008 000003e9 c000a804 > 3f80: c7bd2000 c00bdfa4 00000000 00000000 00000000 000bfd94 00000001 000c0008 > 3fa0: 00000004 c000a640 000bfd94 00000001 00000001 000c0008 000003e9 be8e8f53 > 3fc0: 000bfd94 00000001 000c0008 00000004 000c0008 000c0008 000003e9 000bf990 > 3fe0: 00000000 be8e8ba4 0000ea3c b6eba7ec 60000010 00000001 00000000 00000000 > [<c044344c>] (memcpy) from [<c0290d14>] (spear_mtd_write+0x240/0x294) > [<c0290d14>] (spear_mtd_write) from [<c028e57c>] (mtdchar_write+0xc0/0x230) > [<c028e57c>] (mtdchar_write) from [<c00bdb0c>] (__vfs_write+0x1c/0x128) > [<c00bdb0c>] (__vfs_write) from [<c00bddb4>] (vfs_write+0xa0/0x168) > [<c00bddb4>] (vfs_write) from [<c00bdfa4>] (SyS_write+0x3c/0x90) > [<c00bdfa4>] (SyS_write) from [<c000a640>] (ret_fast_syscall+0x0/0x44) > Code: e1b02f82 14d13001 24d14001 24d1c001 (14c03001) > ---[ end trace f9a736cc2841cf14 ]--- > Segmentation fault > > > Thanks, > Miquèl >
Hi Russell, Russell King - ARM Linux admin <linux@armlinux.org.uk> wrote on Tue, 22 Oct 2019 10:26:19 +0100: > On Tue, Oct 22, 2019 at 11:17:07AM +0200, Miquel Raynal wrote: > > Hi Russell, > > > > Russell King - ARM Linux admin <linux@armlinux.org.uk> wrote on Tue, 22 > > Oct 2019 09:26:43 +0100: > > > > > On Fri, Oct 18, 2019 at 04:36:43PM +0200, Miquel Raynal wrote: > > > > Any write with either dd or flashcp to a device driven by the > > > > spear_smi.c driver will pass through the spear_smi_cpy_toio() > > > > function. This function will get called for chunks of up to 256 bytes. > > > > If the amount of data is smaller, we may have a problem if the data > > > > length is not 4-byte aligned. In this situation, the kernel panics > > > > during the memcpy: > > > > > > > > # dd if=/dev/urandom bs=1001 count=1 of=/dev/mtd6 > > > > spear_smi_cpy_toio [620] dest c9070000, src c7be8800, len 256 > > > > spear_smi_cpy_toio [620] dest c9070100, src c7be8900, len 256 > > > > spear_smi_cpy_toio [620] dest c9070200, src c7be8a00, len 256 > > > > spear_smi_cpy_toio [620] dest c9070300, src c7be8b00, len 233 > > > > Unhandled fault: external abort on non-linefetch (0x808) at 0xc90703e8 > > > > [...] > > > > PC is at memcpy+0xcc/0x330 > > > > > > I need the full oops if you want me to comment on this. > > > > FYI, I ran the dd command within a for loop, incrementing the block size > > (bs) by one byte, if failed with bs=6. > > > > Disabling WB_MODE (burst mode) does not change anything. > > > > Adding a wmb() right after the memcpy_toio() prevents the fault. > > Thanks. Can you check what the result of the write buffer test earlier > in the kernel boot is? > > CPU: Testing write buffer coherency: ... > > ? CPU: Testing write buffer coherency: ok
diff --git a/drivers/mtd/devices/spear_smi.c b/drivers/mtd/devices/spear_smi.c index 986f81d2f93e..d888625a3244 100644 --- a/drivers/mtd/devices/spear_smi.c +++ b/drivers/mtd/devices/spear_smi.c @@ -614,7 +614,7 @@ static inline int spear_smi_cpy_toio(struct spear_smi *dev, u32 bank, ctrlreg1 = readl(dev->io_base + SMI_CR1); writel((ctrlreg1 | WB_MODE) & ~SW_MODE, dev->io_base + SMI_CR1); - memcpy_toio(dest, src, len); + _memcpy_toio(dest, src, len); writel(ctrlreg1, dev->io_base + SMI_CR1);
Any write with either dd or flashcp to a device driven by the spear_smi.c driver will pass through the spear_smi_cpy_toio() function. This function will get called for chunks of up to 256 bytes. If the amount of data is smaller, we may have a problem if the data length is not 4-byte aligned. In this situation, the kernel panics during the memcpy: # dd if=/dev/urandom bs=1001 count=1 of=/dev/mtd6 spear_smi_cpy_toio [620] dest c9070000, src c7be8800, len 256 spear_smi_cpy_toio [620] dest c9070100, src c7be8900, len 256 spear_smi_cpy_toio [620] dest c9070200, src c7be8a00, len 256 spear_smi_cpy_toio [620] dest c9070300, src c7be8b00, len 233 Unhandled fault: external abort on non-linefetch (0x808) at 0xc90703e8 [...] PC is at memcpy+0xcc/0x330 Workaround this issue by using the alternate _memcpy_toio() method which at least does not present the same problem. Fixes: f18dbbb1bfe0 ("mtd: ST SPEAr: Add SMI driver for serial NOR flash") Cc: stable@vger.kernel.org Suggested-by: Boris Brezillon <boris.brezillon@collabora.com> Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com> --- Hello, This patch could not be tested with a mainline kernel (only compiled) but was tested with a stable 4.14.x kernel. I have really no idea why memcpy fails in this situation that's why I propose this workaround but I bet there is something deeper not working. Thanks, Miquèl drivers/mtd/devices/spear_smi.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)