[v2] scsi: ipr: work around fortify-string warning
Commit Message
From: Arnd Bergmann <arnd@arndb.de>
The ipr_log_vpd_compact() function triggers a fortified memcpy() warning
about a potential string overflow with all versions of clang:
In file included from drivers/scsi/ipr.c:43:
In file included from include/linux/string.h:254:
include/linux/fortify-string.h:520:4: error: call to '__write_overflow_field' declared with 'warning' attribute: detected write beyond size of field (1st parameter); maybe use struct_group()? [-Werror,-Wattribute-warning]
__write_overflow_field(p_size_field, size);
^
include/linux/fortify-string.h:520:4: error: call to '__write_overflow_field' declared with 'warning' attribute: detected write beyond size of field (1st parameter); maybe use struct_group()? [-Werror,-Wattribute-warning]
2 errors generated.
I don't see anything actually wrong with the function, but this is the
only instance I can reproduce of the fortification going wrong in the
kernel at the moment, so the easiest solution may be to rewrite the
function into something that does not trigger the warning.
Instead of having a combined buffer for vendor/device/serial strings,
use three separate local variables and just truncate the whitespace
individually.
Fixes: 8cf093e275d0 ("[SCSI] ipr: Improved dual adapter errors")
Cc: Kees Cook <keescook@chromium.org>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
---
I did not try to bisect which commit introduced this behavior into
the fortified memcpy(), the Fixes: commit is the one that introduced
the ipr_log_vpd_compact() function but this predates the fortified
string helpers.
v2 changes:
- fix off-by-one error
---
drivers/scsi/ipr.c | 41 +++++++++++++++++++++--------------------
1 file changed, 21 insertions(+), 20 deletions(-)
Comments
On Tue, Feb 14, 2023 at 02:28:08PM +0100, Arnd Bergmann wrote:
> From: Arnd Bergmann <arnd@arndb.de>
>
> The ipr_log_vpd_compact() function triggers a fortified memcpy() warning
> about a potential string overflow with all versions of clang:
Perhaps this is arch or config specific? I haven't been able to reproduce
this for some reason.
On Tue, Feb 14, 2023, at 22:49, Kees Cook wrote:
> On Tue, Feb 14, 2023 at 02:28:08PM +0100, Arnd Bergmann wrote:
>> From: Arnd Bergmann <arnd@arndb.de>
>>
>> The ipr_log_vpd_compact() function triggers a fortified memcpy() warning
>> about a potential string overflow with all versions of clang:
>
> Perhaps this is arch or config specific? I haven't been able to reproduce
> this for some reason.
I saw it on three arm32 randconfig builds out
of a few thousand, with clang-15, clang-16 and
clang-17. See [1] for one config that triggered
it with recent linux-next based tree.
Arnd
[1] https://pastebin.com/c9FEjzKe
On 2/14/23 22:28, Arnd Bergmann wrote:
> From: Arnd Bergmann <arnd@arndb.de>
>
> The ipr_log_vpd_compact() function triggers a fortified memcpy() warning
> about a potential string overflow with all versions of clang:
>
> In file included from drivers/scsi/ipr.c:43:
> In file included from include/linux/string.h:254:
> include/linux/fortify-string.h:520:4: error: call to '__write_overflow_field' declared with 'warning' attribute: detected write beyond size of field (1st parameter); maybe use struct_group()? [-Werror,-Wattribute-warning]
> __write_overflow_field(p_size_field, size);
> ^
> include/linux/fortify-string.h:520:4: error: call to '__write_overflow_field' declared with 'warning' attribute: detected write beyond size of field (1st parameter); maybe use struct_group()? [-Werror,-Wattribute-warning]
> 2 errors generated.
>
> I don't see anything actually wrong with the function, but this is the
> only instance I can reproduce of the fortification going wrong in the
> kernel at the moment, so the easiest solution may be to rewrite the
> function into something that does not trigger the warning.
>
> Instead of having a combined buffer for vendor/device/serial strings,
> use three separate local variables and just truncate the whitespace
> individually.
>
> Fixes: 8cf093e275d0 ("[SCSI] ipr: Improved dual adapter errors")
> Cc: Kees Cook <keescook@chromium.org>
> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Looks good to me.
Reviewed-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
On Tue, Feb 14, 2023 at 02:28:08PM +0100, Arnd Bergmann wrote:
> From: Arnd Bergmann <arnd@arndb.de>
>
> The ipr_log_vpd_compact() function triggers a fortified memcpy() warning
> about a potential string overflow with all versions of clang:
>
> In file included from drivers/scsi/ipr.c:43:
> In file included from include/linux/string.h:254:
> include/linux/fortify-string.h:520:4: error: call to '__write_overflow_field' declared with 'warning' attribute: detected write beyond size of field (1st parameter); maybe use struct_group()? [-Werror,-Wattribute-warning]
> __write_overflow_field(p_size_field, size);
> ^
> include/linux/fortify-string.h:520:4: error: call to '__write_overflow_field' declared with 'warning' attribute: detected write beyond size of field (1st parameter); maybe use struct_group()? [-Werror,-Wattribute-warning]
> 2 errors generated.
>
> I don't see anything actually wrong with the function, but this is the
> only instance I can reproduce of the fortification going wrong in the
> kernel at the moment, so the easiest solution may be to rewrite the
> function into something that does not trigger the warning.
>
> Instead of having a combined buffer for vendor/device/serial strings,
> use three separate local variables and just truncate the whitespace
> individually.
>
> Fixes: 8cf093e275d0 ("[SCSI] ipr: Improved dual adapter errors")
> Cc: Kees Cook <keescook@chromium.org>
> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Kees Cook <keescook@chromium.org>
Reproduced this locally -- I agree your fix looks like the best
approach. I think Clang was seeing the old "i + 2" return as potentially
overflowing in the case where there was no space-padding on any strings.
On 2/14/23 7:28 AM, Arnd Bergmann wrote:
> From: Arnd Bergmann <arnd@arndb.de>
>
> The ipr_log_vpd_compact() function triggers a fortified memcpy() warning
> about a potential string overflow with all versions of clang:
>
> In file included from drivers/scsi/ipr.c:43:
> In file included from include/linux/string.h:254:
> include/linux/fortify-string.h:520:4: error: call to '__write_overflow_field' declared with 'warning' attribute: detected write beyond size of field (1st parameter); maybe use struct_group()? [-Werror,-Wattribute-warning]
> __write_overflow_field(p_size_field, size);
> ^
> include/linux/fortify-string.h:520:4: error: call to '__write_overflow_field' declared with 'warning' attribute: detected write beyond size of field (1st parameter); maybe use struct_group()? [-Werror,-Wattribute-warning]
> 2 errors generated.
>
> I don't see anything actually wrong with the function, but this is the
> only instance I can reproduce of the fortification going wrong in the
> kernel at the moment, so the easiest solution may be to rewrite the
> function into something that does not trigger the warning.
>
> Instead of having a combined buffer for vendor/device/serial strings,
> use three separate local variables and just truncate the whitespace
> individually.
>
> Fixes: 8cf093e275d0 ("[SCSI] ipr: Improved dual adapter errors")
> Cc: Kees Cook <keescook@chromium.org>
> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Brian King <brking@linux.vnet.ibm.com>
Arnd,
> The ipr_log_vpd_compact() function triggers a fortified memcpy()
> warning about a potential string overflow with all versions of clang:
Applied to 6.3/scsi-staging, thanks!
@@ -1516,23 +1516,22 @@ static void ipr_process_ccn(struct ipr_cmnd *ipr_cmd)
}
/**
- * strip_and_pad_whitespace - Strip and pad trailing whitespace.
- * @i: index into buffer
- * @buf: string to modify
+ * strip_whitespace - Strip and pad trailing whitespace.
+ * @i: size of buffer
+ * @buf: string to modify
*
- * This function will strip all trailing whitespace, pad the end
- * of the string with a single space, and NULL terminate the string.
+ * This function will strip all trailing whitespace and
+ * NUL terminate the string.
*
- * Return value:
- * new length of string
**/
-static int strip_and_pad_whitespace(int i, char *buf)
+static void strip_whitespace(int i, char *buf)
{
+ if (i < 1)
+ return;
+ i--;
while (i && buf[i] == ' ')
i--;
- buf[i+1] = ' ';
- buf[i+2] = '\0';
- return i + 2;
+ buf[i+1] = '\0';
}
/**
@@ -1547,19 +1546,21 @@ static int strip_and_pad_whitespace(int i, char *buf)
static void ipr_log_vpd_compact(char *prefix, struct ipr_hostrcb *hostrcb,
struct ipr_vpd *vpd)
{
- char buffer[IPR_VENDOR_ID_LEN + IPR_PROD_ID_LEN + IPR_SERIAL_NUM_LEN + 3];
- int i = 0;
+ char vendor_id[IPR_VENDOR_ID_LEN + 1];
+ char product_id[IPR_PROD_ID_LEN + 1];
+ char sn[IPR_SERIAL_NUM_LEN + 1];
- memcpy(buffer, vpd->vpids.vendor_id, IPR_VENDOR_ID_LEN);
- i = strip_and_pad_whitespace(IPR_VENDOR_ID_LEN - 1, buffer);
+ memcpy(vendor_id, vpd->vpids.vendor_id, IPR_VENDOR_ID_LEN);
+ strip_whitespace(IPR_VENDOR_ID_LEN, vendor_id);
- memcpy(&buffer[i], vpd->vpids.product_id, IPR_PROD_ID_LEN);
- i = strip_and_pad_whitespace(i + IPR_PROD_ID_LEN - 1, buffer);
+ memcpy(product_id, vpd->vpids.product_id, IPR_PROD_ID_LEN);
+ strip_whitespace(IPR_PROD_ID_LEN, product_id);
- memcpy(&buffer[i], vpd->sn, IPR_SERIAL_NUM_LEN);
- buffer[IPR_SERIAL_NUM_LEN + i] = '\0';
+ memcpy(sn, vpd->sn, IPR_SERIAL_NUM_LEN);
+ strip_whitespace(IPR_SERIAL_NUM_LEN, sn);
- ipr_hcam_err(hostrcb, "%s VPID/SN: %s\n", prefix, buffer);
+ ipr_hcam_err(hostrcb, "%s VPID/SN: %s %s %s\n", prefix,
+ vendor_id, product_id, sn);
}
/**