shinkos6145: Support job combining (4x6 *2 -> 8x6-div2)

This code could be genericized, but as the other 6" sinfonia models
automatically combine prints it's not terribly useful.

The 6" printers don't appear to support a 5x7-div2 cut size, so
this remains only useful for 4x6.  Similarly, we can't do a 6x8-div4.

But as all 6" models can rewind, this is only a print speed optimization
rather than a media use optimization!
This commit is contained in:
Solomon Peachy 2020-01-31 11:34:08 -05:00
parent bff5d1df5c
commit 3d01e35e5e
3 changed files with 111 additions and 8 deletions

View File

@ -342,7 +342,7 @@ static void *dnp_combine_jobs(const void *vjob1,
if (gap_bytes > 0) {
/* Insert gap/padding after first image */
memset(newjob->databuf + newjob->datalen + i, 0, gap_bytes);
memset(newjob->databuf + newjob->datalen + i, 0xff, gap_bytes);
newjob->datalen += gap_bytes;
} else {
// uint8_t *ptrA = newjob->databuf + newjob->datalen + 1088;

View File

@ -1167,6 +1167,7 @@ static int shinkos6145_read_parse(void *vctx, const void **vjob, int data_fd, in
return CUPS_BACKEND_RETRY_CURRENT;
}
memset(job, 0, sizeof(*job));
job->jobsize = sizeof(*job);
/* Common read/parse code */
if (ctx->dev.type == P_KODAK_6900) {
@ -1185,6 +1186,16 @@ static int shinkos6145_read_parse(void *vctx, const void **vjob, int data_fd, in
else
job->copies = copies;
/* S6145 can only combine 2* 4x6 -> 8x6.
2x6 strips and 3.5x5 -> 5x7 can't. */
if (job->jp.columns == 1844 &&
job->jp.rows == 1240 &&
job->jp.method == PRINT_METHOD_STD &&
(ctx->media.ribbon_code == RIBBON_6x8 ||
ctx->media.ribbon_code == RIBBON_6x9)) {
job->can_combine = 1;
}
/* Extended spool format to re-purpose an unused header field.
When bit 0 is set, this tells the backend that the data is
already in planar YMC format (vs packed RGB) so we don't need
@ -1215,16 +1226,104 @@ static int shinkos6145_read_parse(void *vctx, const void **vjob, int data_fd, in
job->databuf = databuf3;
}
// if (job->copies > 1 && hdr->media == 0 && hdr->method == 0)
// and if printer_media == 6x8 or 6x9
// combine 4x6 + 4x6 -> 8x6
// 1844x2492 = 1844x1240.. delta = 12.
*vjob = job;
return CUPS_BACKEND_OK;
}
#define JOB_EQUIV(__x) if (job1->__x != job2->__x) goto done
// XXX this code could be "generic sinfonia?"
// but only the S6145 and S2245 can't automatically combine
// or rewind. So it's really only useful here.
static void *shinkos6145_combine_jobs(const void *vjob1,
const void *vjob2)
{
const struct sinfonia_printjob *job1 = vjob1;
const struct sinfonia_printjob *job2 = vjob2;
struct sinfonia_printjob *newjob = NULL;
uint16_t newrows;
uint16_t newpad;
if (!job1 || !job2)
goto done;
/* Make sure we're okay to proceed */
JOB_EQUIV(jp.columns);
JOB_EQUIV(jp.rows);
JOB_EQUIV(jp.oc_mode);
JOB_EQUIV(jp.method);
JOB_EQUIV(jp.media);
JOB_EQUIV(jp.quality);
switch (job1->jp.rows) {
case 1240: /* 4x6 */
if (job1->jp.method != PRINT_METHOD_STD)
goto done;
newrows = 2492;
newpad = 12;
break;
default:
/* All other sizes, we can't combine */
goto done;
}
newjob = malloc(sizeof(*newjob));
if (!newjob) {
ERROR("Memory allocation failure!\n");
goto done;
}
memcpy(newjob, job1, sizeof(*newjob));
newjob->jp.rows = newrows;
newjob->jp.media = CODE_6x8;
newjob->jp.method = PRINT_METHOD_SPLIT;
/* Allocate new buffer */
newjob->databuf = malloc(newjob->jp.rows * newjob->jp.columns * 3);
newjob->datalen = 0;
if (!newjob->databuf) {
sinfonia_cleanup_job(newjob);
newjob = NULL;
goto done;
}
/* Copy Planar YMC payload into new buffer! */
memcpy(newjob->databuf + newjob->datalen,
job1->databuf,
job1->jp.rows * job1->jp.columns);
newjob->datalen += job1->jp.rows * job1->jp.columns;
memset(newjob->databuf + newjob->datalen, 0xff, newpad * newjob->jp.columns);
memcpy(newjob->databuf + newjob->datalen,
job2->databuf,
job2->jp.rows * job2->jp.columns);
newjob->datalen += job2->jp.rows * job2->jp.columns;
memcpy(newjob->databuf + newjob->datalen,
job1->databuf + (job1->jp.rows * job1->jp.columns),
job1->jp.rows * job1->jp.columns);
newjob->datalen += job1->jp.rows * job1->jp.columns;
memset(newjob->databuf + newjob->datalen, 0xff, newpad * newjob->jp.columns);
memcpy(newjob->databuf + newjob->datalen,
job2->databuf + (job2->jp.rows * job2->jp.columns),
job2->jp.rows * job2->jp.columns);
newjob->datalen += job2->jp.rows * job2->jp.columns;
memcpy(newjob->databuf + newjob->datalen,
job1->databuf + 2*(job1->jp.rows * job1->jp.columns),
job1->jp.rows * job1->jp.columns);
newjob->datalen += job1->jp.rows * job1->jp.columns;
memset(newjob->databuf + newjob->datalen, 0xff, newpad * newjob->jp.columns);
memcpy(newjob->databuf + newjob->datalen,
job2->databuf + 2*(job2->jp.rows * job2->jp.columns),
job2->jp.rows * job2->jp.columns);
newjob->datalen += job2->jp.rows * job2->jp.columns;
done:
return newjob;
}
static int shinkos6145_main_loop(void *vctx, const void *vjob) {
struct shinkos6145_ctx *ctx = vctx;
@ -1644,7 +1743,7 @@ static const char *shinkos6145_prefixes[] = {
struct dyesub_backend shinkos6145_backend = {
.name = "Shinko/Sinfonia CHC-S6145/CS2/S2245/S3",
.version = "0.47" " (lib " LIBSINFONIA_VER ")",
.version = "0.48" " (lib " LIBSINFONIA_VER ")",
.uri_prefixes = shinkos6145_prefixes,
.cmdline_usage = shinkos6145_cmdline,
.cmdline_arg = shinkos6145_cmdline_arg,
@ -1657,6 +1756,7 @@ struct dyesub_backend shinkos6145_backend = {
.query_serno = sinfonia_query_serno,
.query_markers = shinkos6145_query_markers,
.query_stats = shinkos6145_query_stats,
.combine_jobs = shinkos6145_combine_jobs,
.devices = {
{ USB_VID_SHINKO, USB_PID_SHINKO_S6145, P_SHINKO_S6145, NULL, "sinfonia-chcs6145"},
{ USB_VID_SHINKO, USB_PID_SHINKO_S6145, P_SHINKO_S6145, NULL, "shinko-chcs6145"}, /* Duplicate */

View File

@ -52,11 +52,14 @@ struct sinfonia_job_param {
#define EXT_FLAG_DOUBLESLUG 0x04
struct sinfonia_printjob {
size_t jobsize;
int copies;
int can_combine;
struct sinfonia_job_param jp;
uint8_t *databuf;
int datalen;
int copies;
};
int sinfonia_read_parse(int data_fd, uint32_t model,