Fixup vpd page length calculation
PATCH 1/1 Fixup vpd page length calculation
The VPD page length parameter is not only used for the length, but also to hold any error code. So it needs to be moved to an integer.
And we should be resetting the default VPD page length after each iteration so that we do not allocate overly large buffers.
Signed-off-by Hannes Reinecke <[email protected]>
--- drivers/scsi/scsi.c | 4 +++- include/scsi/scsi_device.h | 7 +++++-- 2 files changed, 8 insertions(+), 3 deletions(-)
drivers/scsi/scsi.c +3 -1
@@ -1041,7 +1041,7 @@ EXPORT_SYMBOL_GPL(scsi_get_vpd_page);void scsi_attach_vpd (struct scsi_device *sdev) {int result, i; -int vpd_len =255 ; +int vpd_len =SCSI_VPD_PG_LEN ;int pg80_supported =0 ;int pg83_supported =0 ;unsigned char *vpd_buf, *tmp_pg;
@@ -1073,6 +1073,7 @@ retry_pg0: pg83_supported =1 ; }kfree (vpd_buf); + vpd_len =SCSI_VPD_PG_LEN ; retry_pg80:if (pg80_supported) {
@@ -1106,6 +1107,7 @@ retry_pg80: sdev->vpd_pg80_len = result;spin_unlock (&sdev->reconfig_lock); } + vpd_len =SCSI_VPD_PG_LEN ; }else {spin_lock (&sdev->reconfig_lock); tmp_pg = sdev->vpd_pg80;
include/scsi/scsi_device.h +5 -2
@@ -113,10 +113,13 @@ struct scsi_device {const char * vendor;/* [back_compat] point into 'inquiry' ... */ const char * model;/* ... after scan; point to static string */ const char * rev;/* ... "nullnullnullnull" before scan */ -unsigned char vpd_pg83_len; + +#define SCSI_VPD_PG_LEN 255 +int vpd_pg83_len;unsigned char *vpd_pg83; -unsigned char vpd_pg80_len; +int vpd_pg80_len;unsigned char *vpd_pg80; +spinlock_t reconfig_lock;unsigned char current_tag;/* current tag */ struct scsi_target *sdev_target;/* used only for single_lun */
Signature · 1 line
1.7.12.4
Quoted 5 lines · Hannes Reinecke
The VPD page length parameter is not only used for the length, but also to hold any error code. So it needs to be moved to an integer.
And we should be resetting the default VPD page length after each iteration so that we do not allocate overly large buffers.
This is pretty difficult to apply, since three patches are affected. I think I got it right (please check), but next time, I'd prefer the fixed patches.
Thanks,
James