selftests/vDSO: Explicit unsigned char conversion for elf_hash
Subsystems (2)
PATCH v3 1/1 selftests/vDSO: Explicit unsigned char conversion for elf_hash
Fixes clang compilation warnings by adding explicit unsigned conversion:
parse_vdso.c:206:22: warning: passing 'const char *' to parameter of
type 'const unsigned char *' converts between pointers to integer types where one is of the unique plain 'char' type and the other is not [-Wpointer-sign]
ver_hash = elf_hash(version);
^~~~~~~
parse_vdso.c:59:52: note: passing argument to parameter 'name' here
static unsigned long elf_hash(const unsigned char *name)
^parse_vdso.c:207:46: warning: passing 'const char *' to parameter of
type 'const unsigned char *' converts between pointers to integer types where one is of the unique plain 'char' type and the other is not [-Wpointer-sign]
ELF(Word) chain = vdso_info.bucket[elf_hash(name) % vdso_info.nbucket];
^~~~parse_vdso.c:59:52: note: passing argument to parameter 'name' here static unsigned long elf_hash(const unsigned char *name)
---
v2: update commit message with correct compiler warning
v3: fix checkpatch errors and indentation
--- tools/testing/selftests/vDSO/parse_vdso.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)
tools/testing/selftests/vDSO/parse_vdso.c +3 -2
@@ -203,8 +203,9 @@ void *vdso_sym(const char *version, const char *name)if (!vdso_info.valid)return 0 ; - ver_hash =elf_hash (version); -ELF (Word) chain = vdso_info.bucket[elf_hash (name) % vdso_info.nbucket]; + ver_hash =elf_hash ((const unsigned char *)version); +ELF (Word) chain = vdso_info.bucket[ +elf_hash ((const unsigned char *)name) % vdso_info.nbucket];for (; chain !=STN_UNDEF ; chain = vdso_info.chain[chain]) {ELF (Sym) *sym = &vdso_info.symtab[chain];
--
2.45.0.rc0.197.gbae5840b3b-goog
Hi,
tools/testing/selftests/vDSO/parse_vdso.c void *vdso_sym(const char *version, const char *name)
Quoted 35 lines · Edward Liaw
Fixes clang compilation warnings by adding explicit unsigned conversion:
parse_vdso.c:206:22: warning: passing 'const char *' to parameter of
type 'const unsigned char *' converts between pointers to integer types where one is of the unique plain 'char' type and the other is not [-Wpointer-sign]ver_hash = elf_hash(version); ^~~~~~~ parse_vdso.c:59:52: note: passing argument to parameter 'name' here static unsigned long elf_hash(const unsigned char *name) ^parse_vdso.c:207:46: warning: passing 'const char *' to parameter of
type 'const unsigned char *' converts between pointers to integer types where one is of the unique plain 'char' type and the other is not [-Wpointer-sign]ELF(Word) chain = vdso_info.bucket[elf_hash(name) % vdso_info.nbucket]; ^~~~parse_vdso.c:59:52: note: passing argument to parameter 'name' here static unsigned long
elf_hash(const unsigned char *name)Fixes 98eedc3a9dbf ("Document the vDSO and add a reference parser")Signed-off-by Edward Liaw <[email protected]>---
v2: update commit message with correct compiler warning
v3: fix checkpatch errors and indentation--- tools/testing/selftests/vDSO/parse_vdso.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)tools/testing/selftests/vDSO/parse_vdso.c +0
-0@@ -203,8 +203,9 @@ void *vdso_sym(const char *version, const char *name)
Is it possible to just change the types of the parameters of vdso_sym() or does that trigger even more warnings on the callsites of vdso_sym()?
Either way, this looks ok to me.
Quoted 12 lines
tools/testing/selftests/vDSO/parse_vdso.c +3
-3if (!vdso_info.valid)return 0 ; - ver_hash =elf_hash (version); -ELF (Word) chain = vdso_info.bucket[elf_hash (name) % vdso_info.nbucket]; + ver_hash =elf_hash ((const unsigned char *)version); +ELF (Word) chain = vdso_info.bucket[ +elf_hash ((const unsigned char *)name) % vdso_info.nbucket];for (; chain !=STN_UNDEF ; chain = vdso_info.chain[chain]) {ELF (Sym) *sym = &vdso_info.symtab[chain]; --
Quoted 2 lines
2.45.0.rc0.197.gbae5840b3b-goog
Thanks
Justin
Quoted 34 lines · Justin Stitt
Hi,
Quoted 30 lines · Edward Liaw
Fixes clang compilation warnings by adding explicit unsigned conversion:
parse_vdso.c:206:22: warning: passing 'const char *' to parameter of
type 'const unsigned char *' converts between pointers to integer types where one is of the unique plain 'char' type and the other is not [-Wpointer-sign]ver_hash = elf_hash(version); ^~~~~~~ parse_vdso.c:59:52: note: passing argument to parameter 'name' here static unsigned long elf_hash(const unsigned char *name) ^parse_vdso.c:207:46: warning: passing 'const char *' to parameter of
type 'const unsigned char *' converts between pointers to integer types where one is of the unique plain 'char' type and the other is not [-Wpointer-sign]ELF(Word) chain = vdso_info.bucket[elf_hash(name) % vdso_info.nbucket]; ^~~~parse_vdso.c:59:52: note: passing argument to parameter 'name' here static unsigned long
elf_hash(const unsigned char *name)Fixes 98eedc3a9dbf ("Document the vDSO and add a reference parser")Signed-off-by Edward Liaw <[email protected]>---
v2: update commit message with correct compiler warning
v3: fix checkpatch errors and indentation--- tools/testing/selftests/vDSO/parse_vdso.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)
Quoted 8 lines
tools/testing/selftests/vDSO/parse_vdso.c void *vdso_sym(const char *version, const char *name)
Quoted 5 lines
tools/testing/selftests/vDSO/parse_vdso.c +0
-0@@ -203,8 +203,9 @@ void *vdso_sym(const char *version, const char *name)Is it possible to just change the types of the parameters of
vdso_sym()or does that trigger even more warnings on the callsites ofvdso_sym()?
It does trigger more warnings on the callsites unfortunately.
Quoted 22 lines
Either way, this looks ok to me.
Acked-by Justin Stitt <[email protected]>Quoted 12 lines
tools/testing/selftests/vDSO/parse_vdso.c +3
-3if (!vdso_info.valid)return 0 ; - ver_hash =elf_hash (version); -ELF (Word) chain = vdso_info.bucket[elf_hash (name) % vdso_info.nbucket]; + ver_hash =elf_hash ((const unsigned char *)version); +ELF (Word) chain = vdso_info.bucket[ +elf_hash ((const unsigned char *)name) % vdso_info.nbucket];for (; chain !=STN_UNDEF ; chain = vdso_info.chain[chain]) {ELF (Sym) *sym = &vdso_info.symtab[chain]; --Quoted 2 lines
2.45.0.rc0.197.gbae5840b3b-goog
Thanks
Justin
From: Justin Stitt
Quoted 1 line
Sent: 01 May 2024 20:55
...
Quoted 1 line
Quoted 1 line
static unsigned long
elf_hash(const unsigned char *name)
...
Quoted 2 lines
Is it possible to just change the types of the parameters of
vdso_sym()or does that trigger even more warnings on the callsites ofvdso_sym()?
Isn't the problem the definition of elf_hash()?
A '\0' terminated string really ought to be 'char *' not 'unsigned char *'.
David
-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK Registration No: 1397386 (Wales)
Quoted 11 lines · David Laight
From: Justin Stitt
Quoted 1 line
Sent: 01 May 2024 20:55
...
Quoted 1 line
Quoted 1 line
static unsigned long
elf_hash(const unsigned char *name)...
Quoted 2 lines
Is it possible to just change the types of the parameters of
vdso_sym()or does that trigger even more warnings on the callsites ofvdso_sym()?Isn't the problem the definition of
elf_hash()?
A '\0' terminated string really ought to be 'char *' not 'unsigned char *'.
Right, although note this comment just about its definition:
/* Straight from the ELF specification. */ static unsigned long elf_hash(const unsigned char *name) {
which indeed matches [1]
[1]: https://man.freebsd.org/cgi/man.cgi?query=elf_hash&sektion=3&apropos=0&manpath=FreeBSD+7.1-RELEASE
Quoted 7 lines
David
-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK Registration No: 1397386 (Wales)
Quoted 16 lines · Justin Stitt
Quoted 11 lines · David Laight
From: Justin Stitt
Quoted 1 line
Sent: 01 May 2024 20:55
...
Quoted 1 line
Quoted 1 line
static unsigned long
elf_hash(const unsigned char *name)...
Quoted 2 lines
Is it possible to just change the types of the parameters of
vdso_sym()or does that trigger even more warnings on the callsites ofvdso_sym()?Isn't the problem the definition of
elf_hash()?
A '\0' terminated string really ought to be 'char *' not 'unsigned char *'.Right, although note this comment just about its definition:
Quoted 7 lines
/* Straight from the ELF specification. */ static unsigned long
elf_hash(const unsigned char *name){which indeed matches [1]
[1]: https://man.freebsd.org/cgi/man.cgi?query=elf_hash&sektion=3&apropos=0&manpath=FreeBSD+7.1-RELEASE
How about I move the type cast into elf_hash, like libelf does
https://sourceforge.net/p/elftoolchain/code/HEAD/tree/trunk/libelf/elf_hash.c#l47
(quoted diff) +0 -0
b/tools/testing/selftests/vDSO/parse_vdso.c
index 413f75620a35..33db8abd7d59 100644
tools/testing/selftests/vDSO/parse_vdso.c +6 -3
@@ -56,12 +56,15 @@ static struct vdso_info } vdso_info;/* Straight from the ELF specification. */ -static unsigned long elf_hash (const unsigned char *name) +static unsigned long elf_hash (const char *name) {unsigned long h =0 , g; -while (*name) +const unsigned char *s; + + s = (const unsigned char *) name; +while (*s) { - h = (h <<4 ) + *name++; + h = (h <<4 ) + *s++;if (g = h &0xf0000000 ) h ^= g >>24 ; h &= ~g;
Quoted 8 lines
Quoted 8 lines
David
-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK Registration No: 1397386 (Wales)
From: Edward Liaw
Quoted 45 lines
Sent: 06 May 2024 18:34
Quoted 16 lines · Justin Stitt
Quoted 11 lines · David Laight
From: Justin Stitt
Quoted 1 line
Sent: 01 May 2024 20:55
...
Quoted 1 line
Quoted 1 line
static unsigned long
elf_hash(const unsigned char *name)...
Quoted 2 lines
Is it possible to just change the types of the parameters of
vdso_sym()or does that trigger even more warnings on the callsites ofvdso_sym()?Isn't the problem the definition of
elf_hash()?
A '\0' terminated string really ought to be 'char *' not 'unsigned char *'.Right, although note this comment just about its definition:
Quoted 7 lines
/* Straight from the ELF specification. */ static unsigned long
elf_hash(const unsigned char *name){which indeed matches [1]
[1]: https://man.freebsd.org/cgi/man.cgi?query=elf_hash&sektion=3&apropos=0&manpath=FreeBSD+7.1-
RELEASE
How about I move the type cast into
elf_hash, like libelf does
https://sourceforge.net/p/elftoolchain/code/HEAD/tree/trunk/libelf/elf_hash.c#l47tools/testing/selftests/vDSO/parse_vdso.c +1
-2@@ -56,12 +56,15 @@ static struct vdso_info } vdso_info;/* Straight from the ELF specification. */ -static unsigned long elf_hash (const unsigned char *name) +static unsigned long elf_hash (const char *name) {unsigned long h =0 , g; -while (*name)
tools/testing/selftests/vDSO/parse_vdso.c static struct vdso_info · line 70
Quoted 10 lines
tools/testing/selftests/vDSO/parse_vdso.c +5
-1+const unsigned char *s; + + s = (const unsigned char *) name; +while (*s) { - h = (h <<4 ) + *name++; + h = (h <<4 ) + *s++;if (g = h &0xf0000000 ) h ^= g >>24 ; h &= ~g;
That makes a lot more sense.
David
-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK Registration No: 1397386 (Wales)
Quoted 3 lines · David Laight
That makes a lot more sense.
I sent v4 https://lore.kernel.org/all/[email protected]/.
I updated the commit log to reflect the change but I'm not sure if I did it right for the mailing list. Should I have added --in-reply-to this thread?
Edward