Skip to contentSkip to conversation

block/nvme: Remove QEMU_PACKED from naturally aligned NVMeRegs struct

PATCH Applied Thomas Huth 3 participants 5 messages Feb 25, 2019 Block layer

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 block/nvme: Remove QEMU_PACKED from naturally aligned NVMeRegs struct

The QEMU_PACKED is causing a compiler warning/error with GCC 9:

CC block/nvme.o
block/nvme.c: In function ‘nvme_create_queue_pair’:
block/nvme.c:209:22: error: taking address of packed member of
‘struct <anonymous>’ may result in an unaligned pointer value [-Werror=address-of-packed-member]
209 | q->sq.doorbell = &s->regs->doorbells[idx * 2 * s->doorbell_scale];

All members of the struct are naturally aligned, so there should not be the need for QEMU_PACKED here, and the following QEMU_BUILD_BUG_ON also ensures that there is no padding. Thus simply remove the QEMU_PACKED here.

---
 block/nvme.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
block/nvme.c +1 -1
index b5952c9..6c2ce7d 100644
@@ -82,7 +82,7 @@ typedef volatile struct {
     uint8_t  reserved1[0xec0];
     uint8_t  cmd_set_specfic[0x100];
     uint32_t doorbells[];
-} QEMU_PACKED NVMeRegs;
+} NVMeRegs;
 
 QEMU_BUILD_BUG_ON(offsetof(NVMeRegs, doorbells) != 0x1000);
 
Signature · 1 line
1.8.3.1
Reviewed-by on v1 1/1
#1 lore raw
Quoted 27 lines · Thomas Huth

The QEMU_PACKED is causing a compiler warning/error with GCC 9:

CC block/nvme.o
block/nvme.c: In function ‘nvme_create_queue_pair’:
block/nvme.c:209:22: error: taking address of packed member of
‘struct <anonymous>’ may result in an unaligned pointer value [-Werror=address-of-packed-member]
209 | q->sq.doorbell = &s->regs->doorbells[idx * 2 * s->doorbell_scale];

All members of the struct are naturally aligned, so there should not be the need for QEMU_PACKED here, and the following QEMU_BUILD_BUG_ON also ensures that there is no padding. Thus simply remove the QEMU_PACKED here.

---
 block/nvme.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
block/nvme.c +0 -0
index b5952c9..6c2ce7d 100644

block/nvme.c typedef volatile struct { · line 88

Quoted 8 lines
block/nvme.c +1 -1
@@ -82,7 +82,7 @@ typedef volatile struct {
     uint8_t  reserved1[0xec0];
     uint8_t  cmd_set_specfic[0x100];
     uint32_t doorbells[];
-} QEMU_PACKED NVMeRegs;
+} NVMeRegs;
 
 QEMU_BUILD_BUG_ON(offsetof(NVMeRegs, doorbells) != 0x1000);
 

thanks
-- PMM

Applied
#2 lore raw
In reply to Thomas Huth

Am 25.02.2019 um 12:59 hat Thomas Huth geschrieben:

Quoted 9 lines

The QEMU_PACKED is causing a compiler warning/error with GCC 9:

CC block/nvme.o
block/nvme.c: In function ‘nvme_create_queue_pair’:
block/nvme.c:209:22: error: taking address of packed member of
‘struct <anonymous>’ may result in an unaligned pointer value [-Werror=address-of-packed-member]
209 | q->sq.doorbell = &s->regs->doorbells[idx * 2 * s->doorbell_scale];

Quoted 8 lines

All members of the struct are naturally aligned, so there should not be the need for QEMU_PACKED here, and the following QEMU_BUILD_BUG_ON also ensures that there is no padding. Thus simply remove the QEMU_PACKED here.

Thanks, applied to the block branch.

Though I'm not sure why a compiler should warn if packed and non-packed result in the exact same layout anyway...

Kevin

#3 lore raw
Quoted 2 lines · Kevin Wolf

Though I'm not sure why a compiler should warn if packed and non-packed result in the exact same layout anyway...

Packed implies "and any time you see a pointer to this struct it might not be aligned". Non-packed implies "you can assume that the base of the struct is aligned".

thanks
-- PMM

#4 lore raw
Quoted 7 lines · Peter Maydell
Quoted 2 lines · Kevin Wolf

Though I'm not sure why a compiler should warn if packed and non-packed result in the exact same layout anyway...

Packed implies "and any time you see a pointer to this struct it might not be aligned". Non-packed implies "you can assume that the base of the struct is aligned".

Right. FWIW, consider this small program:

#include <stdio.h>
#include <stdint.h>

typedef struct {
    uint64_t a;
    uint64_t b;
    uint64_t c;
    uint64_t d;
} __attribute__((packed)) s1_t;
struct {
    char x;
    s1_t s1;
} s2;

int main()

{
    printf("address of s2.s1.d = %p\n", &s2.s1.d);
    return 0;
}

Though all members of s1_t look like they are naturally aligned, I get an odd pointer for one of its members in this case.

So passing along that pointer to other functions will certainly cause crashes on architectures like Sparc, thus the compiler warning is indeed justified here.

Thomas